mirror of
https://github.com/truewhile/MeBox.git
synced 2026-11-15 20:34:23 +08:00
Compare commits
34 Commits
| Author | SHA1 | Date | |
|---|---|---|---|
| 1a3248f4bb | |||
| 0be7f8b97e | |||
| 6984158f24 | |||
| d545ff1a7b | |||
| 227e0a9197 | |||
| e85df5fb6b | |||
| 6656f9c398 | |||
| dacfec7492 | |||
| b0f1ab39b9 | |||
| 5242d54a06 | |||
| e961969207 | |||
| b8b02cb3a7 | |||
| 60cfad47e5 | |||
| 47c8bfdd6c | |||
| ce3baee8f1 | |||
| 2494f0c0cd | |||
| c4e9926352 | |||
| f3b88fad7c | |||
| 2c6646beea | |||
| 83ec494a36 | |||
| b8648f0049 | |||
| 5acfa6292a | |||
| 42343212c5 | |||
| 6584519bd8 | |||
| 565a55b194 | |||
| 9036ad04e7 | |||
| ce4319ec25 | |||
| c1ff55ffbf | |||
| fa41936095 | |||
| 9cc21a47c5 | |||
| 5976a8d310 | |||
| d0b4255aa8 | |||
| b8e6418326 | |||
| c84f2604ed |
@@ -187,7 +187,7 @@ jobs:
|
||||
- uses: actions/checkout@v4
|
||||
- uses: actions/setup-go@v5
|
||||
with:
|
||||
go-version: '1.25'
|
||||
go-version: '1.26'
|
||||
cache: true
|
||||
- name: Download web/dist
|
||||
uses: actions/download-artifact@v4
|
||||
|
||||
@@ -4,10 +4,12 @@
|
||||
# 产出:
|
||||
# 1. 前端 + 后端编译验证(go vet / go test / go build)
|
||||
# 2. 多平台可执行二进制 artifact(linux/amd64、linux/arm64、windows/amd64)
|
||||
# 3. ghcr.io/{owner}/mebox:beta 多架构 Docker 镜像(linux/amd64 + linux/arm64)
|
||||
# 3. ghcr.io/{owner}/mebox 多架构 Docker 镜像(linux/amd64 + linux/arm64),
|
||||
# 标签:beta(滚动)+ {基版本}-beta.{短 SHA}(本次构建,只增不改)
|
||||
#
|
||||
# 与 main 分支的发布流(Auto-docker-publish.yml)隔离:beta 不做版本递增、
|
||||
# 不打 release tag,只构建带 -beta 标识的产物供测试。
|
||||
# 与 main 分支发布流(Auto-docker-publish.yml)的关系:镜像的构建方式、标签策略
|
||||
# 和基版本号来源(version 分支)与发布流保持一致,区别是 beta 不做版本递增、
|
||||
# 不打 release tag、不发布 GitHub Release、不部署服务器。
|
||||
|
||||
name: Beta Build
|
||||
|
||||
@@ -22,9 +24,6 @@ permissions:
|
||||
contents: read
|
||||
packages: write
|
||||
|
||||
env:
|
||||
BETA_VERSION_PREFIX: beta
|
||||
|
||||
jobs:
|
||||
# ─────────────────────────────────────────────────────────────────────────────
|
||||
# 1) 编译验证 + 多平台二进制产物
|
||||
@@ -32,17 +31,24 @@ jobs:
|
||||
test-and-build:
|
||||
name: Test & build artifacts
|
||||
runs-on: ubuntu-latest
|
||||
outputs:
|
||||
version: ${{ steps.version.outputs.version }}
|
||||
binary_version: ${{ steps.version.outputs.binary_version }}
|
||||
steps:
|
||||
- uses: actions/checkout@v4
|
||||
with:
|
||||
fetch-depth: 0
|
||||
|
||||
# 基版本号与发布流同源:读 version 分支上由 main 发布流水线维护的 VERSION。
|
||||
# beta 只读不自增,避免与 main 的发布流程争抢版本号。
|
||||
- name: Resolve beta version
|
||||
id: version
|
||||
run: |
|
||||
BASE_VERSION=$(cat VERSION 2>/dev/null || echo "0.0.0")
|
||||
git fetch origin version || true
|
||||
BASE_VERSION=$(git show FETCH_HEAD:VERSION 2>/dev/null || echo "0.0.0")
|
||||
SHA_SHORT=${GITHUB_SHA:0:7}
|
||||
echo "full_version=${BASE_VERSION}-beta.${SHA_SHORT}" >> "$GITHUB_OUTPUT"
|
||||
echo "version=${BASE_VERSION}-beta.${SHA_SHORT}" >> "$GITHUB_OUTPUT"
|
||||
echo "binary_version=mebox-v${BASE_VERSION}-beta.${SHA_SHORT}" >> "$GITHUB_OUTPUT"
|
||||
|
||||
# The binary embeds the SPA (web/dist) via go:embed, so dist must exist
|
||||
# before the Go toolchain touches the web package.
|
||||
@@ -59,7 +65,7 @@ jobs:
|
||||
|
||||
- uses: actions/setup-go@v5
|
||||
with:
|
||||
go-version: '1.25'
|
||||
go-version: '1.26'
|
||||
cache: true
|
||||
|
||||
- name: go vet
|
||||
@@ -71,13 +77,13 @@ jobs:
|
||||
- name: go build (host)
|
||||
run: go build ./...
|
||||
|
||||
# 多平台可执行文件(嵌入刚构建的 web/dist)
|
||||
# 多平台可执行文件(嵌入刚构建的 web/dist),版本号写入口径与发布流一致
|
||||
- name: Build linux/amd64
|
||||
run: CGO_ENABLED=0 GOOS=linux GOARCH=amd64 go build -trimpath -ldflags="-s -w -X main.version=${{ steps.version.outputs.full_version }}" -o dist/mebox-beta-linux-amd64 ./cmd/server
|
||||
run: CGO_ENABLED=0 GOOS=linux GOARCH=amd64 go build -trimpath -ldflags="-s -w -X main.version=${{ steps.version.outputs.binary_version }}" -o dist/mebox-beta-linux-amd64 ./cmd/server
|
||||
- name: Build linux/arm64
|
||||
run: CGO_ENABLED=0 GOOS=linux GOARCH=arm64 go build -trimpath -ldflags="-s -w -X main.version=${{ steps.version.outputs.full_version }}" -o dist/mebox-beta-linux-arm64 ./cmd/server
|
||||
run: CGO_ENABLED=0 GOOS=linux GOARCH=arm64 go build -trimpath -ldflags="-s -w -X main.version=${{ steps.version.outputs.binary_version }}" -o dist/mebox-beta-linux-arm64 ./cmd/server
|
||||
- name: Build windows/amd64
|
||||
run: CGO_ENABLED=0 GOOS=windows GOARCH=amd64 go build -trimpath -ldflags="-s -w -H=windowsgui -X main.version=${{ steps.version.outputs.full_version }}" -o dist/mebox-beta-windows-amd64.exe ./cmd/server
|
||||
run: CGO_ENABLED=0 GOOS=windows GOARCH=amd64 go build -trimpath -ldflags="-s -w -H=windowsgui -X main.version=${{ steps.version.outputs.binary_version }}" -o dist/mebox-beta-windows-amd64.exe ./cmd/server
|
||||
|
||||
- name: Upload artifacts
|
||||
uses: actions/upload-artifact@v4
|
||||
@@ -87,7 +93,8 @@ jobs:
|
||||
if-no-files-found: error
|
||||
|
||||
# ─────────────────────────────────────────────────────────────────────────────
|
||||
# 2) Beta Docker 镜像(ghcr.io/{owner}/mebox:beta)
|
||||
# 2) Beta Docker 镜像(ghcr.io/{owner}/mebox:beta + :{基版本}-beta.{短 SHA})
|
||||
# 构建步骤与 Auto-docker-publish.yml 的 build-image 一致,只是不带 deploy。
|
||||
# ─────────────────────────────────────────────────────────────────────────────
|
||||
docker-beta:
|
||||
name: Build & push beta Docker image
|
||||
@@ -98,13 +105,6 @@ jobs:
|
||||
steps:
|
||||
- uses: actions/checkout@v4
|
||||
|
||||
- name: Resolve beta version
|
||||
id: version
|
||||
run: |
|
||||
BASE_VERSION=$(cat VERSION 2>/dev/null || echo "0.0.0")
|
||||
SHA_SHORT=${GITHUB_SHA:0:7}
|
||||
echo "full_version=${BASE_VERSION}-beta.${SHA_SHORT}" >> "$GITHUB_OUTPUT"
|
||||
|
||||
- uses: docker/setup-qemu-action@v3
|
||||
- uses: docker/setup-buildx-action@v3
|
||||
|
||||
@@ -115,6 +115,15 @@ jobs:
|
||||
username: ${{ github.actor }}
|
||||
password: ${{ secrets.GITHUB_TOKEN }}
|
||||
|
||||
- name: Extract image metadata
|
||||
id: meta
|
||||
uses: docker/metadata-action@v5
|
||||
with:
|
||||
images: ghcr.io/${{ github.repository_owner }}/mebox
|
||||
tags: |
|
||||
type=raw,value=beta
|
||||
type=raw,value=${{ needs.test-and-build.outputs.version }}
|
||||
|
||||
- name: Build & push
|
||||
uses: docker/build-push-action@v6
|
||||
with:
|
||||
@@ -123,11 +132,9 @@ jobs:
|
||||
push: true
|
||||
provenance: false
|
||||
sbom: false
|
||||
tags: ghcr.io/${{ github.repository_owner }}/mebox:beta
|
||||
labels: |
|
||||
org.opencontainers.image.revision=${{ github.sha }}
|
||||
org.opencontainers.image.source=${{ github.repository }}
|
||||
tags: ${{ steps.meta.outputs.tags }}
|
||||
labels: ${{ steps.meta.outputs.labels }}
|
||||
build-args: |
|
||||
VERSION=${{ steps.version.outputs.full_version }}
|
||||
VERSION=${{ needs.test-and-build.outputs.binary_version }}
|
||||
cache-from: type=gha
|
||||
cache-to: type=gha,mode=max
|
||||
|
||||
@@ -15,7 +15,7 @@ jobs:
|
||||
|
||||
- uses: actions/setup-go@v5
|
||||
with:
|
||||
go-version: '1.25'
|
||||
go-version: '1.26'
|
||||
cache: true
|
||||
|
||||
# The binary embeds the SPA (web/dist) via go:embed, so the dist must exist
|
||||
|
||||
+2
-1
@@ -80,4 +80,5 @@ media/
|
||||
tools/
|
||||
verify-cache/
|
||||
verify-data/
|
||||
.zcode/
|
||||
.zcode/
|
||||
.tmp-src
|
||||
|
||||
+114
@@ -0,0 +1,114 @@
|
||||
# Binaries
|
||||
bin/
|
||||
*.exe
|
||||
*.dll
|
||||
*.so
|
||||
*.dylib
|
||||
|
||||
# Test binary, built with `go test -c`
|
||||
*.test
|
||||
*.out
|
||||
|
||||
# Go workspace
|
||||
go.work
|
||||
|
||||
# Dependency directories
|
||||
node_modules/
|
||||
|
||||
# Build artifacts
|
||||
web/dist/
|
||||
web/.vite/
|
||||
web/coverage/
|
||||
web/tsconfig.tsbuildinfo
|
||||
dist-release/
|
||||
|
||||
# Data / runtime
|
||||
data/
|
||||
cache/
|
||||
logs/
|
||||
.tmp-deploy-data/
|
||||
.tmp-deploy-smoke-data/
|
||||
.tmp-deploy-smoke-cache/
|
||||
.tmp-deploy-cache/
|
||||
.tmp-deploy-server.*
|
||||
.tmp-live-server.*
|
||||
.mebox.pid
|
||||
*.log
|
||||
*.db
|
||||
*.db-journal
|
||||
*.db-shm
|
||||
*.db-wal
|
||||
|
||||
# Editor / OS
|
||||
.idea/
|
||||
.vscode/
|
||||
.DS_Store
|
||||
Thumbs.db
|
||||
|
||||
# Env files
|
||||
.env
|
||||
.env.local
|
||||
.env.*.local
|
||||
|
||||
# Local configs (keep examples)
|
||||
config/secrets.yaml
|
||||
config.yaml
|
||||
|
||||
# WorkBuddy workspace (local AI assistant memory)
|
||||
.workbuddy/
|
||||
|
||||
# Editor backups
|
||||
*~
|
||||
.tmp_*
|
||||
|
||||
# Runtime / local-only artifacts (清理补充)
|
||||
.tmp/
|
||||
.tmp-live-backups/
|
||||
.tmp-*
|
||||
.codex-*
|
||||
.codex/
|
||||
downloads/
|
||||
media/
|
||||
*.pid
|
||||
|
||||
# 本地开发运行产物
|
||||
.agents/
|
||||
.claude/
|
||||
.dev-cache/
|
||||
.dev-data/
|
||||
.dev-logs/
|
||||
tools/
|
||||
verify-cache/
|
||||
verify-data/
|
||||
.zcode/
|
||||
|
||||
# ===== ↑ 以上同步自 .gitignore(「从 .gitignore 同步」只重写以上部分)=====
|
||||
.git/
|
||||
.hg/
|
||||
.svn/
|
||||
bower_components/
|
||||
jspm_packages/
|
||||
__pycache__/
|
||||
site-packages/
|
||||
venv/
|
||||
coverage/
|
||||
htmlcov/
|
||||
lcov-report/
|
||||
cmakefiles/
|
||||
cmake-build-*/
|
||||
bazel-*/
|
||||
pods/
|
||||
deriveddata/
|
||||
storybook-static/
|
||||
playwright-report/
|
||||
test-results/
|
||||
allure-results/
|
||||
allure-report/
|
||||
cdk.out/
|
||||
*.egg-info/
|
||||
*.dist-info/
|
||||
eggs/
|
||||
pip-wheel-metadata/
|
||||
wheels/
|
||||
# ----- ↑ 以上为 ZCode 默认排除规则(自定义规则请写在本行下方,不会被同步/恢复改动)-----
|
||||
# 自定义规则写在下方(本行提示可删除)
|
||||
+2
-2
@@ -3,7 +3,7 @@
|
||||
# Multi-architecture build for MeBox.
|
||||
#
|
||||
# Stage 1 (frontend) : Node 20.19+ -> static SPA bundle
|
||||
# Stage 2 (backend) : Go 1.25 -> single static binary (CGO_ENABLED=0)
|
||||
# Stage 2 (backend) : Go 1.26 -> single static binary (CGO_ENABLED=0)
|
||||
# Stage 3 (runtime) : Alpine 3.23 -> ffmpeg + tzdata + non-root user
|
||||
#
|
||||
# Build:
|
||||
@@ -25,7 +25,7 @@ COPY web/ .
|
||||
RUN npm run build
|
||||
|
||||
# ---- Stage 2: backend (cross-compiled to TARGETPLATFORM) -------------------
|
||||
FROM --platform=$BUILDPLATFORM golang:1.25-alpine AS backend
|
||||
FROM --platform=$BUILDPLATFORM golang:1.26-alpine AS backend
|
||||
ARG TARGETOS
|
||||
ARG TARGETARCH
|
||||
ARG GOPROXY=https://proxy.golang.org,direct
|
||||
|
||||
@@ -22,7 +22,7 @@
|
||||
</p>
|
||||
|
||||
<p align="center">
|
||||
<img alt="Go" src="https://img.shields.io/badge/Go-1.25+-00ADD8?style=flat-square&logo=go&logoColor=white" />
|
||||
<img alt="Go" src="https://img.shields.io/badge/Go-1.26+-00ADD8?style=flat-square&logo=go&logoColor=white" />
|
||||
<img alt="React" src="https://img.shields.io/badge/React-18-61DAFB?style=flat-square&logo=react&logoColor=111827" />
|
||||
<img alt="Docker" src="https://img.shields.io/badge/Docker-ready-2496ED?style=flat-square&logo=docker&logoColor=white" />
|
||||
<img alt="License" src="https://img.shields.io/badge/License-GPL--3.0-blue?style=flat-square" />
|
||||
|
||||
+1
-1
@@ -21,7 +21,7 @@
|
||||
</p>
|
||||
|
||||
<p align="center">
|
||||
<img alt="Go" src="https://img.shields.io/badge/Go-1.25+-00ADD8?style=flat-square&logo=go&logoColor=white" />
|
||||
<img alt="Go" src="https://img.shields.io/badge/Go-1.26+-00ADD8?style=flat-square&logo=go&logoColor=white" />
|
||||
<img alt="React" src="https://img.shields.io/badge/React-18-61DAFB?style=flat-square&logo=react&logoColor=111827" />
|
||||
<img alt="Docker" src="https://img.shields.io/badge/Docker-ready-2496ED?style=flat-square&logo=docker&logoColor=white" />
|
||||
<img alt="License" src="https://img.shields.io/badge/License-GPL--3.0-blue?style=flat-square" />
|
||||
|
||||
@@ -0,0 +1,197 @@
|
||||
// reader-smoke 是阅读书源兼容性冒烟工具:
|
||||
// 对批量书源逐个跑「搜索 → 详情 → 目录 → 正文」全链路,输出兼容率报告。
|
||||
//
|
||||
// 用法:
|
||||
//
|
||||
// go run ./cmd/reader-smoke -file sources.json -key 斗破苍穹 -c 8
|
||||
// go run ./cmd/reader-smoke -url https://example.com/sources.json -json > report.json
|
||||
package main
|
||||
|
||||
import (
|
||||
"context"
|
||||
"encoding/json"
|
||||
"flag"
|
||||
"fmt"
|
||||
"net/http"
|
||||
"os"
|
||||
"strings"
|
||||
"sync"
|
||||
"time"
|
||||
|
||||
"go.uber.org/zap"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/config"
|
||||
"github.com/truewhile/MeBox/internal/helper"
|
||||
"github.com/truewhile/MeBox/internal/repository"
|
||||
"github.com/truewhile/MeBox/internal/service/reader"
|
||||
)
|
||||
|
||||
func main() {
|
||||
file := flag.String("file", "", "书源文件路径(JSON 数组/对象/Base64/每行一个)")
|
||||
urlFlag := flag.String("url", "", "书源网络地址(与 -file 二选一)")
|
||||
key := flag.String("key", "斗破苍穹", "搜索关键词")
|
||||
concurrency := flag.Int("c", 4, "并发数")
|
||||
timeout := flag.Int("timeout", 90, "单源全链路超时(秒)")
|
||||
jsonOut := flag.Bool("json", false, "输出完整 JSON 报告(追加在汇总后)")
|
||||
flag.Parse()
|
||||
|
||||
payload := ""
|
||||
switch {
|
||||
case *file != "":
|
||||
data, err := os.ReadFile(*file)
|
||||
if err != nil {
|
||||
fatal("读取文件失败: %v", err)
|
||||
}
|
||||
payload = string(data)
|
||||
case *urlFlag != "":
|
||||
client := helper.NewSiteHTTPClient(30, true)
|
||||
req, err := http.NewRequest("GET", *urlFlag, nil)
|
||||
if err != nil {
|
||||
fatal("构造请求失败: %v", err)
|
||||
}
|
||||
for k, v := range helper.HTTPHeaderPresets() {
|
||||
req.Header.Set(k, v)
|
||||
}
|
||||
resp, err := client.Do(req)
|
||||
if err != nil {
|
||||
fatal("拉取书源失败: %v", err)
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
var sb strings.Builder
|
||||
buf := make([]byte, 32*1024)
|
||||
for {
|
||||
n, err := resp.Body.Read(buf)
|
||||
sb.Write(buf[:n])
|
||||
if err != nil {
|
||||
break
|
||||
}
|
||||
}
|
||||
payload = sb.String()
|
||||
default:
|
||||
fatal("需要 -file 或 -url 指定书源来源")
|
||||
}
|
||||
|
||||
sources := reader.ParseSourcePayload(payload)
|
||||
if len(sources) == 0 {
|
||||
fatal("未从输入中识别到书源")
|
||||
}
|
||||
|
||||
svc := reader.NewReaderService(&config.Config{}, zap.NewNop(), &repository.Container{})
|
||||
ctx := context.Background()
|
||||
|
||||
results := make([]*reader.SmokeChainResult, len(sources))
|
||||
sem := make(chan struct{}, max(1, *concurrency))
|
||||
var wg sync.WaitGroup
|
||||
for i, raw := range sources {
|
||||
wg.Add(1)
|
||||
sem <- struct{}{}
|
||||
go func(i int, raw string) {
|
||||
defer wg.Done()
|
||||
defer func() { <-sem }()
|
||||
ctxSrc, cancel := context.WithTimeout(ctx, time.Duration(*timeout)*time.Second)
|
||||
defer cancel()
|
||||
res := svc.SmokeSource(ctxSrc, raw, *key)
|
||||
results[i] = res
|
||||
status := "✓"
|
||||
if !res.OK {
|
||||
status = "✗"
|
||||
}
|
||||
fmt.Fprintf(os.Stderr, "%s %-24s [%s] hits=%d chapters=%d content=%d %s\n",
|
||||
status, truncate(res.SourceName, 24), stageCN(res), res.SearchHits, res.Chapters, res.ContentLen, res.Error)
|
||||
}(i, raw)
|
||||
}
|
||||
wg.Wait()
|
||||
|
||||
// 汇总
|
||||
var searchOK, infoOK, tocOK, contentOK, allOK int
|
||||
failedAt := map[string]int{}
|
||||
for _, r := range results {
|
||||
if r == nil {
|
||||
continue
|
||||
}
|
||||
switch r.FailedAt {
|
||||
case "":
|
||||
allOK++
|
||||
searchOK++
|
||||
infoOK++
|
||||
tocOK++
|
||||
contentOK++
|
||||
case "search":
|
||||
failedAt["search"]++
|
||||
case "info":
|
||||
searchOK++
|
||||
failedAt["info"]++
|
||||
case "toc":
|
||||
searchOK++
|
||||
infoOK++
|
||||
failedAt["toc"]++
|
||||
case "content":
|
||||
searchOK++
|
||||
infoOK++
|
||||
tocOK++
|
||||
failedAt["content"]++
|
||||
}
|
||||
}
|
||||
n := len(results)
|
||||
pct := func(v int) string {
|
||||
if n == 0 {
|
||||
return "0%"
|
||||
}
|
||||
return fmt.Sprintf("%.1f%%", float64(v)/float64(n)*100)
|
||||
}
|
||||
fmt.Printf("\n==== 冒烟报告 ====\n")
|
||||
fmt.Printf("书源总数: %d 关键词: %s\n", n, *key)
|
||||
fmt.Printf("搜索通过: %d (%s)\n", searchOK, pct(searchOK))
|
||||
fmt.Printf("详情通过: %d (%s)\n", infoOK, pct(infoOK))
|
||||
fmt.Printf("目录通过: %d (%s)\n", tocOK, pct(tocOK))
|
||||
fmt.Printf("正文通过: %d (%s)\n", contentOK, pct(contentOK))
|
||||
fmt.Printf("全链路通过: %d (%s)\n", allOK, pct(allOK))
|
||||
for _, stage := range []string{"search", "info", "toc", "content"} {
|
||||
if failedAt[stage] > 0 {
|
||||
fmt.Printf(" 失败于 %s: %d\n", stageCN(&reader.SmokeChainResult{FailedAt: stage}), failedAt[stage])
|
||||
}
|
||||
}
|
||||
|
||||
if *jsonOut {
|
||||
out, err := json.MarshalIndent(results, "", " ")
|
||||
if err != nil {
|
||||
fatal("序列化报告失败: %v", err)
|
||||
}
|
||||
fmt.Println(string(out))
|
||||
}
|
||||
}
|
||||
|
||||
func stageCN(r *reader.SmokeChainResult) string {
|
||||
switch r.FailedAt {
|
||||
case "":
|
||||
return "完成"
|
||||
case "search":
|
||||
return "搜索"
|
||||
case "info":
|
||||
return "详情"
|
||||
case "toc":
|
||||
return "目录"
|
||||
case "content":
|
||||
return "正文"
|
||||
case "parse":
|
||||
return "解析"
|
||||
default:
|
||||
return r.FailedAt
|
||||
}
|
||||
}
|
||||
|
||||
func truncate(s string, n int) string {
|
||||
rs := []rune(strings.TrimSpace(s))
|
||||
if len(rs) <= n {
|
||||
if len(rs) == 0 {
|
||||
return "(未命名)"
|
||||
}
|
||||
return string(rs)
|
||||
}
|
||||
return string(rs[:n]) + "…"
|
||||
}
|
||||
|
||||
func fatal(format string, args ...any) {
|
||||
fmt.Fprintf(os.Stderr, "reader-smoke: "+format+"\n", args...)
|
||||
os.Exit(1)
|
||||
}
|
||||
Vendored
+26
@@ -0,0 +1,26 @@
|
||||
<!doctype html>
|
||||
<html lang="zh-CN">
|
||||
<head>
|
||||
<meta charset="UTF-8" />
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<meta name="theme-color" content="#faf8f5" />
|
||||
<meta name="description" content="MeBox — 个人家庭媒体中心" />
|
||||
<link rel="icon" type="image/png" href="/brand/logo-64.png" />
|
||||
<link rel="icon" type="image/x-icon" href="/favicon.ico" />
|
||||
<link rel="apple-touch-icon" href="/brand/logo-192.png" />
|
||||
<!-- 字体已自托管(见 src/index.css 的 @font-face),不再使用 Google Fonts 外链:
|
||||
自托管环境常无法访问 Google,渲染阻塞外链会拖慢首屏。 -->
|
||||
<title>MeBox</title>
|
||||
<script type="module" crossorigin src="/assets/index-CcRvbC_F.js"></script>
|
||||
<link rel="modulepreload" crossorigin href="/assets/rolldown-runtime-B0Z9INg1.js">
|
||||
<link rel="modulepreload" crossorigin href="/assets/vendor-data-CD8zvwOM.js">
|
||||
<link rel="modulepreload" crossorigin href="/assets/vendor-react-Cz-iIQjW.js">
|
||||
<link rel="modulepreload" crossorigin href="/assets/vendor-motion-FEuwvkVZ.js">
|
||||
<link rel="modulepreload" crossorigin href="/assets/auth-D3Hf5Dnr.js">
|
||||
<link rel="modulepreload" crossorigin href="/assets/createLucideIcon-nIJoh1kU.js">
|
||||
<link rel="stylesheet" crossorigin href="/assets/index-BPz3c5Oc.css">
|
||||
</head>
|
||||
<body class="font-body antialiased">
|
||||
<div id="root"></div>
|
||||
|
||||
</body>
|
||||
@@ -0,0 +1,144 @@
|
||||
# MeBox 阅读功能实施计划(legado 书源兼容)
|
||||
|
||||
> 分支:`feature/reading`
|
||||
> 目标:在首页增加「影视 / 阅读」模式切换,阅读模式完整兼容阅读 3.0(legado)书源体系,覆盖 **文本(bookSourceType=0)、音频(=1)、漫画/图片(=2)** 三类源。
|
||||
> 实现方式(用户明确要求):**样式与逻辑全部仿造 refgd/legado 本体**,不参考其他重实现项目;相当于用 Go + React 18 + TypeScript 5 重写该项目。
|
||||
> 界面与交互的唯一规格:`docs/reader-ui-spec.md`(从 legado 源码逐屏调研产出)。
|
||||
> 规则引擎的唯一语义基准:legado `app/src/main/java/io/legado/app/model/analyzeRule/` 源码,Go 侧逐方法移植对拍(源码克隆在 `C:\MyProject\_ref\legado`,仅作对照,不进入构建)。
|
||||
|
||||
## 1. 范围
|
||||
|
||||
**做:**
|
||||
- 首页「影视/阅读」切换,阅读模式下有独立首页(书架/搜索/发现/最近阅读)
|
||||
- legado 书源导入与管理(URL 导入、文本/JSON 粘贴导入、启停、分组、排序)
|
||||
- 三类书源的完整链路:搜索 → 详情 → 目录 → 正文/播放列表/图片列表
|
||||
- Go 侧规则引擎:CSS(jsoup 风格) / JSONPath / XPath / 正则 / 内嵌 JS 五种语法及其组合
|
||||
- 文本阅读器、音频播放器、漫画阅读器三套前端 UI
|
||||
- 追更、阅读进度同步(服务端存储,多端一致)、换源、替换净化规则
|
||||
|
||||
**不做(本期明确排除):**
|
||||
- `webView` 类规则(需要无头浏览器,识别后标记该源为不兼容并提示)
|
||||
- `webView` 真人校验类登录(loginUrl 走 WebView 人工过验证码的场景;
|
||||
纯 JS / 表单类登录已支持,见 P4.5)
|
||||
- RSS/订阅源、TTS 朗读、文件类型源(bookSourceType=3)
|
||||
- 本地 TXT/EPUB 导入(列为后续可选)
|
||||
|
||||
## 2. 总体架构
|
||||
|
||||
沿用 MeBox 现有分层,全部新增代码集中在:
|
||||
|
||||
```
|
||||
internal/
|
||||
model/ # 新增 5 张表,注册进 AllModels() 自动迁移
|
||||
repository/ # reader 相关 GORM 封装
|
||||
service/reader/ # 规则引擎 + 书源业务(核心新增,预计占全部后端代码 70%)
|
||||
handler/ # /api/reader/* 路由组
|
||||
web/src/
|
||||
pages/reader/ # 阅读端独立页面群(懒加载路由)
|
||||
components/reader/
|
||||
```
|
||||
|
||||
**基建复用**:`internal/helper/http.go`(浏览器 UA + 代理回退 HTTP 客户端)、`internal/service/runtime_cache.go`(正文/目录/搜索缓存,内存+Redis)、`internal/service/image_proxy*`(封面与漫画图片代理)、`internal/handler/ws.go` 的 WSHub(搜索进度、追更任务推送,新增 `reader:*` topic)。
|
||||
|
||||
**数据流**:书源 JSON 存库 → 搜索/发现时按启用的源并发抓取(errgroup + 信号量限流,超时熔断)→ 结果聚合 → 前端。正文、播放地址、图片列表由服务端组装(含 `nextContentUrl` 翻页合并)后带 TTL 缓存下发;音频流与漫画图片按需经服务端代理补 UA/Referer 头。
|
||||
|
||||
## 3. 规则引擎(核心工作)
|
||||
|
||||
语义基准:gedoor/legado `app/src/main/java/io/legado/app/model/analyzeRule/` 下的 AnalyzeRule / AnalyzeByJSoup / AnalyzeByJSonPath / AnalyzeByRegex / AnalyzeUrl,逐项对拍测试。
|
||||
|
||||
### 3.1 组件与选型
|
||||
|
||||
| 组件 | 选型 | 说明 |
|
||||
|---|---|---|
|
||||
| HTML/CSS | `PuerkitoBio/goquery` | jsoup 等价物;jsoup 特有语法(class.x / id.x / tag.x / text.x / children / @text / @textNodes / @html / 属性选择)自己包一层 |
|
||||
| XPath | `antchfx/htmlquery` | 对齐 JsoupXpath 语义 |
|
||||
| JSONPath | `PaesslerAG/jsonpath`(备选 ohler55/ojg) | Jayway 语义 + 自实现 `\|\|`/`&&` 合并层,选型阶段需验证 |
|
||||
| 正则 | Go regexp(RE2) | legado 部分源用 Java 正则语法,回退换 `dlclark/regexp2` |
|
||||
| JS | `dop251/goja` | ES2017+,跑书源内嵌 JS |
|
||||
| 字符集 | `golang.org/x/text` | GBK/GB18030 解码 |
|
||||
|
||||
### 3.2 引擎能力清单
|
||||
|
||||
- 规则模式识别:`@css:` / `$.`(JSONPath) / `@XPath:`或`//` / `<js>`与`@js:` / `##` 正则替换段
|
||||
- 列表组合符 `&&` / `||` / `%%`,变量存取 `@put:{}` / `@get:{}`,内嵌 JS `{{ }}`
|
||||
- jsoup 分析器、JSONPath 分析器、XPath 分析器、正则分析器,以及混合规则的链式解析(AnalyzeRule 的分段执行语义)
|
||||
- AnalyzeUrl:`{{key}}`/`{{page}}` 变量、`<js>` 生成 URL、URL 后 `,{...}` 选项(method/body/charset/headers/retry/timeout/type/proxy/js/webView)
|
||||
- JS 沙箱:goja 运行时 + 执行超时中断 + 禁止直接 IO;上下文注入 `java`、`source`、`book`、`baseUrl`、`result` 等对象
|
||||
- `java.*` 桥接函数(按书源实际使用频率分批实现):
|
||||
- 网络:ajax / ajaxAll / connect / get / post / head
|
||||
- 编解码:base64Decode/Encode(含 URL-safe)、hexDecode、encodeURI/decodeURI、htmlDecode
|
||||
- 加解密:md5(16/32)、sha1/sha256、AES/DES/3DES/RSA(CBC/ECB + 常见 padding/key 语义,legado 源里最常见的坑)
|
||||
- 字符串与时间:replaceAll/substring/正则族、timeFormat 等
|
||||
- 规则回调:`java.getString/getElement` 等,桥回 Go 规则引擎(JS 与规则互相嵌套的关键)
|
||||
|
||||
### 3.3 兼容策略
|
||||
|
||||
- 引擎按能力分层实现,每个能力配真实书源样本的单测(fixtures 放 `internal/service/reader/testdata/`)
|
||||
- 提供 CLI 冒烟工具(如 `cmd/reader-smoke`):对批量导入的公开书源集跑 搜索/详情/目录/正文 全链路,输出成功率报告,作为每个阶段验收依据
|
||||
- 含 `webView` 选项的源直接判定不兼容并在书源管理页标注
|
||||
|
||||
## 4. 数据模型(新增表)
|
||||
|
||||
| 表 | 关键字段 |
|
||||
|---|---|
|
||||
| book_sources | name, group, type(0/1/2), source_url, json(原文), enabled, custom_order, last_check_at, comment |
|
||||
| books(书架) | source_url, book_url, name, author, cover_url, intro, kind, type(文本/音频/图片), latest_chapter, total_chapters, last_read_chapter_index, last_read_at |
|
||||
| book_chapters | book_id, index, title, url, is_volume, update_time |
|
||||
| read_progress | book_id(唯一), chapter_index, position(滚动/秒/图片序), updated_at |
|
||||
| replace_rules | name, find, replace, scope, is_regex, enabled, order |
|
||||
|
||||
阅读器显示设置(主题/字体/翻页方式)存前端 localStorage,不上服务端。
|
||||
|
||||
## 5. API 设计(/api/reader/*)
|
||||
|
||||
- 书源:`GET/POST/DELETE /sources`、`POST /sources/import`(URL 或 JSON/base64 文本,自动识别格式与类型)、`PATCH /sources/:id`(启停/排序)
|
||||
- 搜索:`POST /search {keyword}` → 后台聚合任务,结果经 WS `reader:search` 增量推送;结果可一键加入书架
|
||||
- 发现:`GET /explore?source=&group=`(解析 exploreUrl 的 `分组名::url` 结构)
|
||||
- 书架:`GET/POST/DELETE /books`、`GET /books/:id/info`、`GET /books/:id/toc`、`POST /books/:id/refresh`(追更)
|
||||
- 本地书籍:`POST /local/books`(multipart 上传 TXT/EPUB,导入即入书架)
|
||||
- 内容:`GET /books/:id/chapters/:idx/content` —— 按书籍类型返回:
|
||||
- 文本:`{type:"text", content:"..."}`(服务端已合并 nextContentUrl 翻页、已应用替换规则)
|
||||
- 音频:`{type:"audio", tracks:[{url,title}]}`(含代理路径与所需请求头)
|
||||
- 图片:`{type:"image", images:[{url, style}]}`(同样经代理)
|
||||
- 进度:`PUT /books/:id/progress`
|
||||
- 替换规则:`/replace-rules` CRUD
|
||||
- 调试:`POST /debug {source_id, rule, url}`(书源调试器后端)
|
||||
- 图片/流代理:复用现有 image_proxy / stream_proxy 模式,按源配置注入 UA/Referer
|
||||
|
||||
## 6. 前端设计
|
||||
|
||||
- **首页切换**:`HomePage.tsx` 顶部加分段控件(仿 `LibraryTagBar` tab 模式),「阅读」切到阅读首页;选择持久化(zustand + localStorage)
|
||||
- **阅读首页**:继续阅读横排 + 书架封面网格 + 搜索入口 + 追更提示
|
||||
- **页面群**(懒加载,仿 `appRoutes.tsx`):`/reader`(首页)、`/reader/search`(多源并发搜索 + 实时进度)、`/reader/explore`、`/reader/book/:id`(详情 + 目录 + 换源)、`/reader/sources`(书源管理 + 调试器)
|
||||
- **三套阅读器**:
|
||||
- 文本:滚动 + 分页双模式(CSS 分栏测量分页)、主题(含夜间)、字体/行距/边距、点击翻页区、章节预加载、进度上报
|
||||
- 音频:`hls.js`(已是依赖)+ `<audio>` 兜底,播放列表、倍速、记忆进度、锁屏/息屏策略
|
||||
- 漫画:上下滚动 / 左右翻页双模式、相邻图片预加载、`imageStyle` 支持
|
||||
|
||||
## 7. 阶段划分
|
||||
|
||||
| 阶段 | 内容 | 交付物 |
|
||||
|---|---|---|
|
||||
| P0 引擎地基 ✅ | 规则引擎核心(四分析器 + 规则拆分/组合/变量)+ AnalyzeUrl v1(GET/POST/charset/headers/变量/页码模式)+ 表结构 + 书源导入/管理 API + 搜索/详情/目录/正文/书架/进度/调试 API | 已完成:`internal/service/reader/rule/`(规则引擎,~2800 行,对齐 AnalyzeRule/AnalyzeByJSoup/AnalyzeByJSonPath/AnalyzeByXPath/AnalyzeByRegex/AnalyzeUrl/RuleAnalyzer)+ 服务层 + `/api/reader/*` 路由 + 单测/端到端测试全绿 |
|
||||
| P1 文本源全链路 + 首页切换 | 搜索聚合(WS 进度)/详情/目录/正文(nextContentUrl 合并、缓存)+ 前端首页切换、书架、搜索、详情、文本阅读器 v1(阅读器样式仿 legado:9 宫格点击、主题、翻页动画) | 用纯规则型文本源完成「搜书→加入→阅读」全流程 |
|
||||
| P2 JS 与兼容率爬坡 ✅ | goja 接入 + `java.*` 桥(网络/编解码/摘要/对称加密全家桶/规则回调,函数名对齐 JsExtensions)+ URL 规则 JS(analyzeJs/{{}}/js/bodyJs)+ cookie jar + 用户替换净化规则(含正则超时保护)+ 替换净化页 + 结构化冒烟链路(SmokeChain)+ `cmd/reader-smoke` 冒烟 CLI | JS 源可用;冒烟 CLI 跑公开书源集出各阶段通过率报告 |
|
||||
| P3 音频源 ✅ | 正文按音频类型返回播放列表(绝对化)+ `/api/reader/media` 签名媒体代理(HMAC 防滥用、Range 透传支持拖动、m3u8 分片/密钥地址重写)+ 听书面板(hls.js 播 m3u8、直链 `<audio>`、上一章/-15s/播放暂停/+15s/下一章、章节选择抽屉、定时关闭 0–180 分钟、倍速 0.5–3.0、跳过片头片尾按书持久化 `PUT /api/reader/books/:id/audio-config`、进度按秒记忆、播完/片尾到点自动下一章) | 音频源可听 |
|
||||
| P4 漫画/图片源 ✅ | 图片列表绝对化 + 经签名代理(带书源 Referer 防盗链头)+ 漫画阅读器(上下滚动/左右单页双模式、图片懒加载与加载失败占位、点击分区翻页/呼菜单、菜单进度条按图片序号、进度按图片序号记忆、下一章预取)+ imageStyle 透传 | 漫画源可看 |
|
||||
| P4.5 书源登录 ✅ | `jsLib` 一次装载(对应 legado SharedJsScope)+ `source.*` 会话方法(getVariable/setVariable/getLoginInfo/putLoginInfo/getLoginHeader/putLoginHeader/get/put)+ `cookie.*`(getCookie/setCookie/replaceCookie/removeCookie/getKey,按 eTLD+1 隔离)+ 请求自动携带 Cookie 与 loginHeader + `loginUrl`/`loginUi`(解析表单 → 按钮 action 拼在 loginUrl 后执行 → result 为表单值)+ `loginCheckJs`(会话失效自动重登/重取)+ 源变量落库 + 登录信息/Cookie 加密存储 + 服务端 toast/startBrowser 回传前端 + 登录面板与变量编辑器 | 登录类书源可登录、可留存登录态 |
|
||||
| P4.6 本地书籍 ✅ | `POST /api/reader/local/books` 上传 TXT / EPUB(上限 64MB)→ 落盘 `data/reader/local/<bookID>.<txt\|epub>` + 解析目录入 `reader_chapters` + 落库为 `origin=""`、`is_local=true` 的书架条目;TXT 自动识别 BOM/UTF-8/GBK/Big5/UTF-16 并按 legado 默认 TXT 目录规则切章(RE2 无 lookbehind,改行首锚定 + 句子启发式过滤),章定位信息存字节区间 `start:end`,读章只读该区间;EPUB 走 container.xml → OPF spine,标题优先取 NCX/NAV,正文去标签与实体;移出书架同步删落盘文件;同名重复导入覆盖更新并尽量保留进度 | 本地书可上传、可读、可删 |
|
||||
| P5 体验完善 | 换源(ChangeBookSourceDialog 四档排序)、追更(定时刷新目录 + 缓存清理)、发现页(exploreUrl 标签条)、阅读器高级设置(页眉页脚提示、点击区域自定义)、书源编辑器六 Tab、备份导出 | 完整体验 |
|
||||
|
||||
P0–P2 是主体(约全部工作量 60–70%),P3/P4 相对独立可并行。
|
||||
|
||||
## 8. 风险与对策
|
||||
|
||||
- **书源质量参差**:单源解析全程超时与重试上限;正文翻页循环设页数上限防死循环
|
||||
- **goja 性能**:复杂 JS 源解析慢 → 正文/目录强缓存、单源并发限 1、全局执行预算
|
||||
- **Java 正则/JSoup 细节差异**:以对拍测试驱动修正,regex 备选 regexp2
|
||||
- **热源防盗链**:图片与音频一律走带请求头的代理
|
||||
- **合规**:不内置任何书源,全部由用户自行导入,管理页附免责说明
|
||||
|
||||
## 9. 参考
|
||||
|
||||
- **唯一语义与样式基准**:https://github.com/refgd/legado (规则引擎 analyzeRule 源码 + UI 布局/交互,见 docs/reader-ui-spec.md)
|
||||
- 书源规则教程:https://mgz0227.github.io/The-tutorial-of-Legado/
|
||||
@@ -0,0 +1,253 @@
|
||||
# legado UI/交互仿制规格(React 18 + Tailwind Web 版实现依据)
|
||||
|
||||
> 来源:对 `refgd/legado`(克隆于 `C:\MyProject\_ref\legado`)源码的调研。
|
||||
> 路径缩写:`<src>` = `app/src/main/java/io/legado/app/`,布局在 `app/src/main/res/layout/`。
|
||||
|
||||
## 1. 首页/书架
|
||||
|
||||
文件:`<src>ui/main/bookshelf/`(BookshelfFragment、BooksAdapterGrid/List)、`fragment_bookshelf1/2.xml`、`item_bookshelf_grid*.xml`、`item_bookshelf_list*.xml`。主界面底部 BottomNavigation(书架/发现/我的)+ 右上悬浮搜索按钮。
|
||||
|
||||
**布局结构(从上到下)**
|
||||
- style1(标签页风格):标题行(书架标题 + 下拉箭头,点开分组切换菜单)→ 搜索按钮 + 更多按钮 → 一级分组 tab 条(tab 长按可删分组)→ 次级分组扩展标签 → ViewPager(每 tab 一个书籍列表/网格)。
|
||||
- style2(单 RecyclerView 混排):title_bar → SwipeRefresh → 书籍列表(根分组时「分组网格项 + 书籍项」混排,进入分组只显示书籍)→ 空态文案。
|
||||
- 网格项:封面 + 本地角标 + 未读数角标 + 加载中 + 书名 + 长按遮罩。
|
||||
- 列表项:封面 + 本地图标 + 有更新标记 + 未读数 + 书名 + 作者 + 最后更新时间 + 当前读到(章节名)+ 最新章节(章节名)+ 长按遮罩。
|
||||
|
||||
**交互逻辑**
|
||||
- 内置分组 id:全部(IdRoot=-100)、本地(IdLocal=-2)、未分组(IdUngrouped=-4),其余用户分组。
|
||||
- 单击书 → 直接进入阅读;单击分组 → 进入该分组。
|
||||
- 长按书 → 跳转书籍详情页;长按分组 → 重命名/删除对话框;style1 长按 tab 删除分组。
|
||||
- 右上菜单:搜索、Wi-Fi 传书、刷新目录、书架布局切换、分组管理、导出/导入书架、下载离线、本地导入、网址添加、日志。
|
||||
|
||||
**MeBox 的「影视 / 阅读」首页模式**(`stores/readerSettings.homeMode` + 顶栏 `LayoutReaderModeToggle`)
|
||||
- 切换入口在**顶栏**,位于搜索框与账号菜单之间,**只显示图标不显示文字**:影视模式显示场记板图标,阅读模式显示书图标(并带品牌色高亮),点一下切到另一个模块、图标随之变化。悬停提示写明「当前是 X 模式,点击切换到 Y」,不在首页时点击会先跳回 `/`。
|
||||
- 两种模式共用一个首页路由 `/`:影视模式渲染媒体首页,阅读模式渲染书架(`ReaderHomeContent embedded`)。首页内容区不再放分段式「影视 / 阅读」开关(原 `ReaderModeSwitch` 仅留给独立布局的 `/reader` 首页兜底)。
|
||||
- 切到阅读模式时,顶部搜索框换成**书搜索**(`LayoutHeaderBookSearch`,占位「搜索书籍…」):聚焦即下拉书架(本地即时过滤书名/作者,点条目进阅读器),回车优先打开书架首条;下拉底部固定一行「在书源中搜索「xxx」」,带 `?key=` 跳到 `/reader/search` 由多源聚合搜索页自动开搜。账号/主题菜单照常显示。
|
||||
- 阅读模式下移动端底部导航(首页/媒体库/收藏/列表/更多)隐藏,避免影视导航混进书架。
|
||||
- `/reader/*` 本来就是独立全屏布局(不套影视 Layout),不受影响。
|
||||
|
||||
**MeBox 书架已实现的布局与信息(`ReaderHomeContent.tsx` + `BookshelfCard.tsx` + `BookshelfSettingsDialog.tsx`)**
|
||||
- 三套布局,设置在顶栏「书架设置」弹窗里(对应 legado dialog_bookshelf_config),存 `stores/readerSettings.ts`:
|
||||
- 网格:封面宫格,列数 2–6 或「自适应」;窄屏自动降到 2–3 列兜底。卡片含封面 + 本地角标 + 未读章数徽标 + 双行书名 + 详情/移出按钮。
|
||||
- 列表:66×90 封面 + 书名 + 作者 + 更新时间 + 读到(当前章)+ 最新章节;对应 legado `item_bookshelf_list.xml` 的四行信息。
|
||||
- 紧凑列表:48×64 小封面 + 书名 +「作者 · 读到」,一屏放更多书(对应 legado `layout_list_compact`)。
|
||||
- 排序六档,对齐 legado `AppConfig.getBookSortByGroupId`:最近阅读(dur_chapter_time,默认)/ 最近更新(latest_chapter_time)/ 综合 / 按书名 / 按作者 / 手动顺序(order)。顶栏有快捷排序下拉,弹窗里也可选。
|
||||
- 显示项开关:未读章数徽标、更新时间。
|
||||
- 「更新目录」(对应 legado `menu_update_toc`):`POST /reader/shelf/refresh-toc` 并发重抓书架内网络书籍的目录,覆盖章节缓存;末章标题变化时刷新 `latest_chapter_time`,用于「最近更新」排序与更新时间展示。本地书籍与无书源信息的书跳过。返回 `total/updated/unchanged/failed` 四档——抓到但章数没变(完结书)记 `unchanged`「已是最新」,只有抓取或写入真的失败才记 `failed`,避免把正常结果报成「更新失败」;失败会打 Warn 日志(含书名与书源)。
|
||||
- 未实现(与 legado 的差距):书籍二级分组网格(进入分组后的封面网格)、导出/导入书架、离线下载。
|
||||
|
||||
**书架分组(`/reader/book-groups`,仿影视模块的媒体库标签)**
|
||||
- 按用户存服务端:`reader_book_groups` 表每个用户一条,`groups` 列是 `[{name, book_ids}]` 的 JSON。结构、语义与影视的媒体库标签(`User.LibraryTags`)同构——组名 → 成员 ID、整份替换、一个成员只归一个组、组内顺序即展示顺序。
|
||||
- 接口:`GET /reader/book-groups`(没有分组返回空数组)、`PUT /reader/book-groups`(整份替换)。服务端落库前做两条需要「知道书是否存在」的收敛:只保留该用户书架上的书(书移出书架后分组里不留死 ID)、一本书只归一个组(越靠前的分组优先)。空分组保留,方便先建组再放书。
|
||||
- 与影视的差异:书架有两个内置页签「全部」「未分组」(未分组是「没有被任何分组认领」的视图,不持久化,只有存在分组时才出现,且固定在分组栏最后,避免挤开用户自己拖拽排好的分组顺序);书籍不进管理弹窗逐个分配,而是在卡片上用「分组」按钮指定(书架可能有上千本,列出来既慢又难找)。
|
||||
- 前端:`utils/readerBookGroups.ts`(纯逻辑)+ `hooks/useBookGroups.ts`(乐观更新、串行提交、失败回滚,接口不可用时退回 localStorage,照搬 `useLibraryTags`)+ `components/BookGroupBar.tsx`(分组栏)+ `ManageBookGroupsDialogView.tsx`(新建/重命名/删除/拖拽排序)+ `BookGroupPickerDialog.tsx`(移到分组)。
|
||||
|
||||
**阅读器偏好按账号同步(跨设备)**
|
||||
- 主题 / 夜间模式 / 排版(字号、行距、段距)/ 翻页或滚动模式 / 听书倍速与定时 / 书架布局·排序·显示项,原先只存浏览器 localStorage(设备级,换设备就丢),现在按用户落库到 `reader_profiles`(`model.ReaderProfile`,`user_id` 唯一索引),接口 `GET|PUT /reader/profile`。
|
||||
- 服务端是权威来源,localStorage(zustand persist 的 `mebox-reader-settings`)退化成首屏缓存:登录后 `hydrateReaderSettings()` 拉一次覆盖本地,之后本地改动防抖 600ms 回写(`utils/readerSettingsSync.ts`)。
|
||||
- 账号还没有偏好记录时 `GET` 返回 `null`,前端把本地现值推上去「播种」,升级前在本机调好的设置不会被重置。
|
||||
- 服务端只做存储 + 范围收敛(数值夹区间、枚举未知回落默认、主题标识限长),不做语义解释;前端另有本地 clamp,两处范围需同步(见 `service/reader/profile.go` 顶部注释)。
|
||||
- 首页的「影视 / 阅读」模式(homeMode)是设备级偏好,故意不参与同步。
|
||||
|
||||
## 2. 阅读界面(重点)
|
||||
|
||||
文件:`<src>ui/book/read/`(ReadBookActivity、ReadMenu、SearchMenu、MangaMenu、config/*Dialog)、`ui/book/read/page/`(ReadView、PageView、ContentTextView、ChapterProvider)。
|
||||
|
||||
**布局结构(单 FrameLayout 栈)**:read_view 正文 → 文字选择光标 → read_menu 主菜单 → search_menu 全文搜索 → 朗读回原文浮条。
|
||||
|
||||
**正文渲染**
|
||||
- ReadView 内含 3 个 PageView(prev/cur/next 缓存),ContentTextView 纯 Canvas 自绘:按字号/行距/段距/缩进/两端对齐排版,Web 版用 CSS multi-column 或 JS 分页等效实现。
|
||||
- 翻页动画(0–5):0 覆盖、1 平移、2 仿真、3 上下滚动、4 无、5 平移覆盖;速度可调(默认 300ms)。
|
||||
- 滚动模式支持背景跟随、自动翻页(定时滚动/定时翻页)。
|
||||
- 图片/漫画书切换独立配置集 + MangaMenu:顶栏 + 底部胶囊(上一页/页码进度条/下一页),支持横滚、缩放。
|
||||
|
||||
**点击区域(9 宫格,边缘留 pageTouchClick px)**
|
||||
- 区块:左上/中上/右上/左中/中/右中/左下/中下/右下。
|
||||
- 默认:左列与上中=上一页;右列与下中=下一页;正中=呼出菜单。
|
||||
- 全部可配动作:0 呼出菜单、1 下一页、2 上一页、3 下一章、4 上一章、7 书签、10 目录、11 搜索、12 同步进度等。
|
||||
|
||||
**主菜单 ReadMenu(自上而下)**
|
||||
- 半透明遮罩(点击收起)。
|
||||
- 顶部 title_bar:返回 + 书名(点击进详情)+ 章节名/章节链接(点击打开原页面)+ 书源按钮(弹菜单:编辑书源/禁用书源等)。
|
||||
- 底部 bottom_menu(圆角面板):上一章 | 进度条 | 下一章 → 亮度条 + 自动亮度 → 快捷按钮行(搜索、自动翻页、替换规则、夜间/日间切换)→ 动作面板行(目录、朗读、界面、设置)。
|
||||
- 菜单配色可跟随页面背景或用主题底色,透明度 35–100。
|
||||
|
||||
**设置面板**
|
||||
- 「界面」三 Tab:文本(字体文件、字重、字号 5–50、字距、行距 0–20、段距、缩进、下划线、阴影);样式(白天/夜间/E-Ink 三态主题:文字色、背景色/图、强调色、菜单底色、菜单透明度、背景透明度,可恢复/管理/分享);页面(6 种翻页动画、状态栏深浅、滚动背景跟随、四边距、页眉页脚提示、简繁转换)。
|
||||
- 「设置」:屏幕方向、保持亮屏、隐藏状态栏、两端/底部对齐、竖排、双页、进度条行为(按页/按章)、音量键/滚轮翻页、点击区域配置、动画速度、自动换源等。
|
||||
- 页眉页脚提示项:无/章节名/时间/电量/电量百分比/页码/总进度/页码+总页/时间+电量/书名;页眉默认 左=时间/右=电量,页脚默认 左=章节名/右=页码/总页。
|
||||
|
||||
**内置主题与配色(assets/defaultData/readConfig.json,6 套 + 自定义)**
|
||||
|
||||
| 名称 | 背景(日/夜) | 文字(日/夜) | 强调(日/夜) | 字号 |
|
||||
|---|---|---|---|---|
|
||||
| 微信读书 | #FFC0EDC6 / #000000 | #FF0B0B0B / #ADADAD | #E53935 / #FE4D55 | 24 |
|
||||
| 预设1 | #FFFFFF / #000000 | #000000 / #FFFFFF | #E53935 / #FE4D55 | 20 |
|
||||
| 预设2(羊皮纸) | #DDC090 / #3C3F43 | #3E3422 / #DCDFE1 | #834E00 / #FE4D55 | 20 |
|
||||
| 预设3(护眼绿) | #C2D8AA / #3C3F43 | #596C44 / #88C16F | #E53935 / #FE4D55 | 20 |
|
||||
| 预设4(粉紫) | #DBB8E2 / #3C3F43 | #68516C / #F6AEAE | #801314 / #90BFF5 | 20 |
|
||||
| 预设5(淡蓝) | #ABCEE0 / #3C3F43 | #3D4C54 / #90BFF5 | #E53935 / #FE4D55 | 20 |
|
||||
|
||||
默认排版:textSize 20、letterSpacing 0.1、lineSpacingExtra 12、paragraphSpacing 2、缩进「 」、padding 上下6/左右16、页脚线 true。未选样式时默认:bg #EEEEEE / 夜 #000000 / E-Ink #FFFFFF,文字 #3E3D3B / 夜 #ADADAD,强调 #E53935 / 夜 #FE4D55。
|
||||
|
||||
**MeBox 文本阅读器已实现的排版/菜单细节**(`ReaderViewPage.tsx`)
|
||||
- 正文留边:左右 16px、上下 8px(对齐 legado 默认左右16/上下6),分页列宽按留边后的视口宽计算,正文不贴屏幕边。
|
||||
- 菜单打开时正文整体下移一个顶栏高度(用 `transform`,不改视口高度、不触发重新分页),顶栏不再压住开头 1–2 行。
|
||||
- 界面面板的字号/行距/段距三个调节组用 `flex-wrap`,窄屏自动折行,不会把「段距」挤出屏幕。
|
||||
- 点击分区:翻页模式渲染一层 `absolute inset-0` 的三分区按钮(左 30% 上一页 / 中 40% 呼出菜单 / 右 30% 下一页);**滚动模式不能渲染这层覆盖层**——它不是滚动容器的子节点,手机上手指落在覆盖层按钮上时浏览器找不到可滚动的祖先,纵向滑动完全失效、只剩点左右能翻屏。滚动模式改为在滚动容器自身的 `onClick` 上按 x 坐标分区(`handleZoneTap`),手指拖动不会产生 click,原生纵向滚动照常。
|
||||
- 滚动模式的滚轮(`useSmoothWheelScroll`):把滚轮格数累加成一个目标位置,再逐帧向它逼近,滚动连续、松手后自己滑行一段,仿手机上下滑动,而不是浏览器整格跳变。只在文本/漫画的上下滚动模式接管;翻页模式仍走下面的翻页监听器。按真实帧间隔换算逼近比例,120Hz 屏上速度不会翻倍;`ctrl/cmd+滚轮` 保留缩放;手指按下(滑动、拖滚动条、点分区)立刻停掉动画把控制权交还用户。
|
||||
- 漫画双页铺开(`hooks/useComicSpreads.ts` + `utils/comicSpread.ts`,legado 无对应项,是桌面端补充):窗口宽度 ≥900px 且漫画处于翻页模式时,一屏并排显示两页(`settings.comicDoublePage` 为设备级偏好,默认开,可在菜单里关;窗口不够宽时自动退回单页)。分组规则同 Tachiyomi 的 dual page——宽度明显大于高度(比值 ≥1.1)的图视为「跨页宽图」独占一屏,其余连续的竖版页两两配对;列表接口只给 URL,尺寸靠 `new Image()` 探测(`<img>` 渲染时顺带回填),只探测当前页往后 8 张的窗口,避免整章预载。进度仍按图片序号记,翻页时换算成「第几屏」,所以关掉再开双页进度不会错位。菜单动作行改为 `flex` 等分,容纳「双页/单页」入口(`grid-cols-N` 得写死列数字面量,多一个按钮无法复用)。
|
||||
- 漫画图片显示尺寸(`utils/comicImageFit.ts`,仅上下滚动模式):legado 的漫画阅读有「缩放」,Web 版原先只有一种写死的宽度(正文列封顶 900px),桌面端放不大也缩不小,所以补上一组档位,在界面面板的「图片尺寸」一行切换。`default` 就是老样子(900px 居中),`width` 适应宽度(正文列不再封顶、铺满窗口),`height` 适应高度(一屏一页),`long` 适应长边(整页完整可见),`original` 原图(原始像素 1:1,超出部分横向滚动)。`default`/`width` 由外层列宽决定,其余三个由 `<img>` 自身约束,**所以后三种必须放开 900px 列宽上限**,否则会被列再压回去。`height`/`long` 要按「一屏多高」换算,所以量的是滚动容器本身而不是 `window.innerHeight`(正文区是 root 里的 `flex-1`,另挂上下条之后就不等于窗口高度了),并用 `ResizeObserver` 跟随;只在启用了自尺寸档位时才挂观察器,默认档位不白白重渲染整章。`settings.comicImageFit` 是设备级偏好(不同屏幕和源分辨率合适的尺寸不同),和 `comicDoublePage` 一样只存本机、不参与账号同步。翻页模式是整页缩放进视口,本来就不存在「太大/太小」,档位对它不生效、也不显示入口。`ctrl/cmd+滚轮` 依然留给浏览器缩放,本功能只用档位做粗调。
|
||||
- **文本型漫画源按图片下发**(`internal/service/reader` 的 `imageMarkersOnly`):拷贝漫画这类书源 `bookSourceType=0`(文本),正文规则直接给 `<img>`,服务端把整行标签转成 `[img]<签名地址>` 标记行后类型仍是 `text`,前端就会走文本阅读器——分页把每张图当成一列,一屏只看得到一张,桌面端也谈不上双页铺开。现在 `GetContentForBook` / `LocalChapterContent` 在标记化之后判断「非空行全部是 `[img]` 标记」,是则把 `type` 改成 `image` 并给出 `images`,交由漫画阅读器渲染(整页缩放、双页铺开、进度按图片序号)。判定只认「全是标记」:混了正文的章节一律保持 `text`,孤立的 HTML 标签行(`<div>`/`</div>`)与空行当作无内容忽略,绝不吞掉可读文字。
|
||||
- 鼠标滚轮翻页(仅翻页模式):向上滚=上一页,向下滚=下一页;菜单打开时也不翻页。鼠标滚轮一格一页(间隔至少 220ms,与翻页动画对齐),触控板小步长累计到阈值翻一页且一次手势只翻一页(避免惯性连翻)。`ctrl/cmd+滚轮` 保留浏览器缩放。
|
||||
- 目录:整屏面板(顶部返回 + 书名 + 章数,Virtuoso 虚拟列表,定位并高亮当前章,点章跳转)。**必须渲染在底部菜单之外**:菜单带 `backdrop-blur`,会成为 `fixed` 后代的包含块,放里面 `h-full` 只能拿到菜单高度;历史上它写的是 `top-0 + bottom-full`,两者同时存在时高度被算成 0,整块目录完全看不见。
|
||||
|
||||
**本地书籍(对应 legado 本地 TXT / EPUB)**
|
||||
- 入口:书架页右上「本地导入」按钮(书架为空时另有「上传本地书籍」),支持 TXT / EPUB,单文件上限 64MB,上传后自动入库并直接进入阅读页。
|
||||
- 书架卡片:本地书打「本地」角标,右上角有删除按钮(二次确认),删除会连服务器上的文件一起清掉。
|
||||
- 存储:正文落盘 `data/reader/local/<bookID>.<txt|epub>`;目录信息与网络书共用 `reader_chapters`,用 `Tag` 记定位:TXT 存 UTF-8 规范化后文件内的字节区间 `start:end`,EPUB 存 zip 内的 XHTML 条目路径。读章只取所需区间/条目,不整本载入内存。
|
||||
- TXT:自动识别 BOM(UTF-8/UTF-16) 与 UTF-8 / GBK / Big5,统一转 UTF-8 落盘;目录按 legado 默认 TXT 规则切章(`第X章/节/卷/集/部/篇`、序章、楔子、番外等),并对「第一章的正文内容」这类正文行做启发式过滤,切不出章名时整本当一章「全文」,章前内容(书名/简介)并入第一章。
|
||||
- EPUB:`META-INF/container.xml` → OPF → `spine` 顺序出章。目录标题优先取 NCX / EPUB3 NAV,**逐 token 走并用栈收任意层级的 navPoint**——Epubor 等工具导出的 EPUB 常漏 `</navPoint>`,标题会整棵嵌进上一个节点,按固定层级解会丢掉一大半标题;取不到时依次退回正文首行 → 封面页(文件名含 cover 且该页只有图)标「封面」→ `<title>`(过滤 Cover/Table of Contents 这类无信息量的)→ `第 N 章`。
|
||||
- EPUB 图片:正文里的 `<img>` 在转纯文本时就地换成 `[img]<zip 条目>` 标记行(相对路径按该 XHTML 所在目录解析),下发前把标记换成签名地址 `/api/reader/local/asset?b=&p=&s=`(HMAC,`<img>` 带不了 JWT),前端把 `[img]` 开头的行渲染成居中图片(`max-width:100%` + `max-height:70vh`,保证不撑破分栏)。网络书正文不受影响。
|
||||
- 进度:与网络书同一套 `durChapter*` 字段,跨端一致;同名文件重复上传按覆盖更新处理(章数不变则保留进度)。
|
||||
|
||||
**音频书播放条(AudioPlayActivity + AudioPlayService,即「听书」)**
|
||||
- 背景与封面:书籍封面强模糊(blur 32px)铺满做底,叠一层很淡的主题底色(opacity 0.3)保住日/夜对比度;正中圆形显示封面原图(对应 legado `upCover` 的 ivBg 模糊图 + ivCover 圆图)。封面缺失或加载失败时退回主题色圆点,不留破图。
|
||||
- transport 行:上一章 | -15s | 播放暂停 | +15s | 下一章(SEEK_STEP=15s,进度按秒)。
|
||||
- 动作行(常驻底部,抽屉打开时仍可点):章节(目录选择,见下)、定时关闭、倍速、片头片尾。
|
||||
- 章节选择:底部抽屉列全部章节,定位到当前章、当前章高亮、点章即跳;卷名行不可点。
|
||||
- 定时关闭:0/5/10/15/30/60/90/180 分钟;暂停期间不倒计时;归零自动暂停播放;选定值持久化为下次默认(对应 AppConfig.ttsTimer 在服务启动时 setTimer)。
|
||||
- 倍速:滑杆 0.5–3.0(步进 0.1)+ 0.5/0.75/1/1.25/1.5/1.75/2/2.5/3 快捷档;持久化(对应 AudioPlay.playSpeed,Android 6 以下不支持调速)。
|
||||
- 跳过片头片尾:抽屉内两条滑杆(片头/片尾,秒,0 不跳过,上限 300),按书持久化(Book.openCredits/closeCredits,落库 books.open_credits / close_credits)。语义:全新开播(该章进度为 0)时 seek 到片头秒数;播放到 duration-片尾秒数即等同播完,有下一章则自动续播,末章则停在片尾处。**单位是秒,不是章数。**
|
||||
- 播放进度按秒记忆(节流 5s 上报),跨端一致;播完自动下一章。
|
||||
- legado 的「播放模式」(顺序/单章循环/随机/列表循环)与「音频服务唤醒锁」是客户端能力,Web 端未实现。
|
||||
|
||||
## 3. 搜索
|
||||
|
||||
文件:`<src>ui/book/search/` + `<src>model/webBook/SearchModel.kt`。
|
||||
|
||||
- 布局:顶部 SearchView + 转圈进度条 → 结果流 → 书架命中提示卡 → 搜索历史 chips + 清空 → 右下 开始/停止 悬浮按钮。
|
||||
- 结果项:封面、在书架角标、书源数角标(可换源数)、书名、作者、分类标签、最新章节、简介。
|
||||
- 多源并发:所有启用书源(按搜索范围过滤),固定线程池并发(全局设置 threadCount),单源 30s 超时。
|
||||
- 实时结果流:每源返回即刷新列表;结束后 onSearchFinish(isEmpty, hasMore)(任一源有结果即可翻页,searchPage++)。
|
||||
- 聚合去重 mergeItems 四档:书名或作者等于关键词 > kind 含关键词 > 书名/作者含关键词 > 其他;同名同作者合并为一项并 addOrigin(角标显示可换源数);档内按书源数降序。
|
||||
- 顶部菜单:精准搜索开关、搜索范围(分组/全选)、书源管理、日志。点结果 → 换源对话框/直接打开。
|
||||
|
||||
## 4. 发现
|
||||
|
||||
`<src>ui/main/explore/` + `ui/book/explore/`。布局:标题栏 → SwipeRefresh → 源选择行(当前源名 + 下拉、源内搜索、标签筛选、更多)→ 二级筛选条 → 分类标签条(横滑 chip,可展开)→ 书籍网格(封面+书名)→ 兜底列表。数据来自启用书源的 exploreUrl + ruleExplore;切源刷新标签;点标签拼 URL 加载、分页加载更多。
|
||||
|
||||
## 5. 书籍详情
|
||||
|
||||
`<src>ui/book/info/BookInfoActivity.kt`。布局:模糊封面背景 → 标题栏(返回/刷新/菜单)→ 下拉刷新 → 滚动区:封面 + 书名 + 作者 + 最新章节(点击刷新目录)+ 阅读时长 + 分类标签 + 分组行 + 书源行(换源)+ 目录入口行 → 双 Tab(简介 / 目录预览,可全屏、点章节跳读)→ 底部动作区:tv_shelf 加入/移出书架(描边按钮)+ tv_read 开始阅读/继续阅读(实心主色按钮)。菜单含置顶、去书源网页、登录、删除(勾选删缓存)、分享。
|
||||
|
||||
## 6. 目录页
|
||||
|
||||
`<src>ui/book/toc/`。布局:标题栏(返回 + 书名 + SearchView 搜章节)→ Tab(章节/书签)→ 章节 RecyclerView + 底部信息条(当前位置 + 跳顶/跳底)。菜单:倒序开关、使用净化替换、加载字数、长章分卷合并、TXT 目录规则。倒序即列表反转;卷名行可折叠;已读章节变色;点击章节回传 index 跳读。
|
||||
|
||||
## 7. 书源管理
|
||||
|
||||
- **主列表**:标题栏(SearchView 过滤 + 菜单)→ RecyclerView(item:域名、复选、启停 Switch、编辑、更多菜单、发现、快速调试、响应耗时条)→ 多选操作栏。菜单:排序(手动/自动/名称/URL/更新时间/响应时间/启用)、分组筛选(启用/禁用/需登录/无分组/启用发现/禁用发现)、添加书源(本地/网络/二维码)。
|
||||
- **编辑**:标题栏(保存/调试入口)→ 类型下拉 + 六个开关(启用/启用发现/CookieJar/段评/事件监听/自定义按钮)→ 六个 Tab:基本/搜索/发现/详情/目录/正文,每 Tab 为「标签 + 输入框 + 帮助弹层」表单。
|
||||
- **调试**:顶部 SearchView 输入关键词 → 串行执行 搜索→详情→目录→正文,日志逐条流式显示;可查看各阶段原始 HTML、切换探索分类调试。
|
||||
|
||||
## 8. 替换规则
|
||||
|
||||
- 列表页:标题栏(搜索 + 菜单:分组管理、启停筛选、添加、导入)→ RecyclerView(名称、分组、内容摘要、启停 Switch)→ 多选操作栏。
|
||||
- 编辑页字段:规则名 → 分组 → 替换规则(+ 正则开关 + 帮助)→ 替换为 → 作用于标题 / 作用于正文 → 作用范围(书名) → 排除范围 → 超时毫秒(order 隐含)。
|
||||
|
||||
## 9. 数据实体字段全集
|
||||
|
||||
**BookSource.kt**(rule* 为内嵌对象)
|
||||
|
||||
| 字段 | 类型 | 默认值 |
|
||||
|---|---|---|
|
||||
| bookSourceUrl | String | "" |
|
||||
| bookSourceName | String | "" |
|
||||
| bookSourceGroup | String? | null |
|
||||
| bookSourceType | Int (0文本/1音频/2图片/3文件/4视频) | 0 |
|
||||
| bookUrlPattern | String? | null |
|
||||
| customOrder | Int | 0 |
|
||||
| enabled | Boolean | true |
|
||||
| jsLib | String? | null |
|
||||
| enabledExplore | Boolean | true |
|
||||
| enabledCookieJar | Boolean? | true |
|
||||
| concurrentRate | String? | null |
|
||||
| header | String? | null |
|
||||
| loginUrl / loginUi / loginCheckJs / coverDecodeJs | String? | null |
|
||||
| bookSourceComment | String? | null |
|
||||
| variableComment | String? | null |
|
||||
| lastUpdateTime | Long | 0 |
|
||||
| respondTime | Long | 180000 |
|
||||
| weight | Int | 0 |
|
||||
| exploreUrl / exploreScreen | String? | null |
|
||||
| ruleExplore / ruleSearch / ruleBookInfo / ruleToc / ruleContent / ruleReview | 对象 | null |
|
||||
| eventListener / customButton | Boolean | false |
|
||||
|
||||
**Book.kt**
|
||||
|
||||
| 字段 | 类型 | 默认值 |
|
||||
|---|---|---|
|
||||
| bookUrl (PK) / tocUrl | String | "" / "" |
|
||||
| origin / originName | String | localTag / "" |
|
||||
| name / author | String | "" / "" |
|
||||
| kind / customTag / coverUrl / customCoverUrl / intro / customIntro / charset | String? | null |
|
||||
| type | Int (BookType) | text |
|
||||
| group | Long | 0 |
|
||||
| latestChapterTitle | String? | null |
|
||||
| latestChapterTime / lastCheckTime | Long | now |
|
||||
| lastCheckCount / totalChapterNum | Int | 0 |
|
||||
| durChapterTitle | String? | null |
|
||||
| durChapterIndex / durVolumeIndex / chapterInVolumeIndex / durChapterPos | Int | 0 |
|
||||
| durChapterTime | Long | now |
|
||||
| wordCount | String? | null |
|
||||
| canUpdate | Boolean | true |
|
||||
| order / originOrder | Int | 0 |
|
||||
| variable | String? | null |
|
||||
| readConfig(ReadConfig 内嵌) | reverseToc、pageAnim、reSegment、imageStyle、useReplaceRule、delTag、ttsEngine、splitLongChapter=true、readSimulating、startDate、startChapter、dailyChapters=3、openCredits=0、closeCredits=0、playMode=0、playSpeed=1.0、manga* 系列、mangaPageAnim | |
|
||||
|
||||
**BookChapter.kt**
|
||||
|
||||
| 字段 | 类型 | 默认值 |
|
||||
|---|---|---|
|
||||
| url / title | String | "" / "" |
|
||||
| isVolume | Boolean | false |
|
||||
| baseUrl / bookUrl | String | "" / "" |
|
||||
| index | Int | 0 |
|
||||
| isVip / isPay | Boolean | false |
|
||||
| resourceUrl / tag / wordCount | String? | null |
|
||||
| start / end | Long? | null |
|
||||
| startFragmentId / endFragmentId / variable / imgUrl | String? | null |
|
||||
|
||||
**ReplaceRule.kt**
|
||||
|
||||
| 字段 | 类型 | 默认值 |
|
||||
|---|---|---|
|
||||
| id | Long | now |
|
||||
| name | String | "" |
|
||||
| group | String? | null |
|
||||
| pattern | String | "" |
|
||||
| replacement | String | "" |
|
||||
| scope | String? | null |
|
||||
| scopeTitle | Boolean | false |
|
||||
| scopeContent | Boolean | true |
|
||||
| excludeScope | String? | null |
|
||||
| isEnabled | Boolean | true |
|
||||
| isRegex | Boolean | true |
|
||||
| timeoutMillisecond | Long | 3000 |
|
||||
| order | Int | Int.MIN_VALUE |
|
||||
|
||||
## 10. 主题与视觉
|
||||
|
||||
- 全局:主背景 `#F7F7FA`、主文字 `#DE000000`、品牌色 `#FFF6FBF8`(浅绿白调)。
|
||||
- 界面主题由 ThemeConfig(themes.json)控制 6 套;阅读菜单/底栏用 bottomBackground(跟随所选界面主题),沉浸模式(readBarStyleFollowPage)直接取页面背景/文字色。
|
||||
- 强调色 accentColor 用于选中态、光标、Tab 选中;E-Ink 模式菜单改描边。
|
||||
- 视觉风格:大圆角面板、1dp 描边、玻璃拟态半透明条、图标+文字标签的动作面板。
|
||||
|
||||
## Web 版实现要点
|
||||
|
||||
- 正文分页仿 ReadView 三页缓存 + 9 宫格点击映射。
|
||||
- 主题模型照抄 ReadBookConfig.Config(含日/夜/E-Ink 三态字段),内置 6 套主题色值直接使用上表。
|
||||
- 搜索并发用 Promise 池 + 流式合并,合并逻辑照抄 mergeItems 四档排序。
|
||||
- 搜索分页:滚到底自动取下一页(`POST /reader/search` 的 `page`,从 1 开始,对应书源 `searchUrl` 的 `{{page}}`),逐页增量合并(跨页按「书名+作者」去重、并书源,见 `utils/searchBooks.ts`)。收口比 legado 严一档:**新一页没带来新书就停止翻页**——不支持分页的源每页都重复返回首页结果,照 legado「任一源有结果即可翻页」会一直转圈加载;单次搜索最多 20 页。
|
||||
- 书源/替换编辑器按第 7/8 节 Tab/字段一一对应。
|
||||
@@ -1,6 +1,6 @@
|
||||
module github.com/truewhile/MeBox
|
||||
|
||||
go 1.25.0
|
||||
go 1.26.0
|
||||
|
||||
require (
|
||||
fyne.io/systray v1.12.2
|
||||
@@ -19,9 +19,9 @@ require (
|
||||
github.com/stretchr/testify v1.11.1
|
||||
github.com/ulikunitz/xz v0.5.15
|
||||
go.uber.org/zap v1.27.0
|
||||
golang.org/x/crypto v0.49.0
|
||||
golang.org/x/crypto v0.55.0
|
||||
golang.org/x/image v0.45.0
|
||||
golang.org/x/sync v0.22.0
|
||||
golang.org/x/sync v0.23.0
|
||||
golang.org/x/sys v0.47.0
|
||||
golang.org/x/time v0.15.0
|
||||
gopkg.in/yaml.v3 v3.0.1
|
||||
@@ -30,6 +30,12 @@ require (
|
||||
)
|
||||
|
||||
require (
|
||||
github.com/PaesslerAG/gval v1.0.0 // indirect
|
||||
github.com/PaesslerAG/jsonpath v0.1.1 // indirect
|
||||
github.com/PuerkitoBio/goquery v1.13.0 // indirect
|
||||
github.com/andybalholm/cascadia v1.3.4 // indirect
|
||||
github.com/antchfx/htmlquery v1.3.6 // indirect
|
||||
github.com/antchfx/xpath v1.3.6 // indirect
|
||||
github.com/bytedance/gopkg v0.1.4 // indirect
|
||||
github.com/bytedance/sonic v1.15.0 // indirect
|
||||
github.com/bytedance/sonic/loader v0.5.0 // indirect
|
||||
@@ -38,6 +44,9 @@ require (
|
||||
github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc // indirect
|
||||
github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f // indirect
|
||||
github.com/disintegration/imaging v1.6.3-0.20201218193011-d40f48ce0f09 // indirect
|
||||
github.com/dlclark/regexp2 v1.12.0 // indirect
|
||||
github.com/dlclark/regexp2/v2 v2.5.2 // indirect
|
||||
github.com/dop251/goja v0.0.0-20260926152631-39ec2650adc9 // indirect
|
||||
github.com/dustin/go-humanize v1.0.1 // indirect
|
||||
github.com/esimov/pigo v1.4.7-0.20240801095032-7465ed14de47 // indirect
|
||||
github.com/gabriel-vasile/mimetype v1.4.13 // indirect
|
||||
@@ -47,9 +56,12 @@ require (
|
||||
github.com/go-playground/locales v0.14.1 // indirect
|
||||
github.com/go-playground/universal-translator v0.18.1 // indirect
|
||||
github.com/go-playground/validator/v10 v10.30.1 // indirect
|
||||
github.com/go-sourcemap/sourcemap v2.1.3+incompatible // indirect
|
||||
github.com/goccy/go-json v0.10.6 // indirect
|
||||
github.com/goccy/go-yaml v1.19.2 // indirect
|
||||
github.com/godbus/dbus/v5 v5.1.0 // indirect
|
||||
github.com/golang/groupcache v0.0.0-20241129210726-2c02b8208cf8 // indirect
|
||||
github.com/google/pprof v0.0.0-20250317173921-a4b03ec1a45e // indirect
|
||||
github.com/hashicorp/hcl v1.0.0 // indirect
|
||||
github.com/jackc/pgpassfile v1.0.0 // indirect
|
||||
github.com/jackc/pgservicefile v0.0.0-20240606120523-5a60cdf6a761 // indirect
|
||||
@@ -90,8 +102,8 @@ require (
|
||||
go.uber.org/multierr v1.10.0 // indirect
|
||||
golang.org/x/arch v0.25.0 // indirect
|
||||
golang.org/x/exp v0.0.0-20251023183803-a4bb9ffd2546 // indirect
|
||||
golang.org/x/net v0.52.0 // indirect
|
||||
golang.org/x/text v0.41.0 // indirect
|
||||
golang.org/x/net v0.58.0 // indirect
|
||||
golang.org/x/text v0.42.0 // indirect
|
||||
google.golang.org/protobuf v1.36.11 // indirect
|
||||
gopkg.in/ini.v1 v1.67.0 // indirect
|
||||
modernc.org/libc v1.70.0 // indirect
|
||||
|
||||
@@ -1,7 +1,20 @@
|
||||
fyne.io/systray v1.12.2 h1:Y8DZxgLHsVQt6rY9Zrkkg+j67S7vv/1F2viOWKPpVeA=
|
||||
fyne.io/systray v1.12.2/go.mod h1:RVwqP9nYMo7h5zViCBHri2FgjXF7H2cub7MAq4NSoLs=
|
||||
github.com/PaesslerAG/gval v1.0.0 h1:GEKnRwkWDdf9dOmKcNrar9EA1bz1z9DqPIO1+iLzhd8=
|
||||
github.com/PaesslerAG/gval v1.0.0/go.mod h1:y/nm5yEyTeX6av0OfKJNp9rBNj2XrGhAf5+v24IBN1I=
|
||||
github.com/PaesslerAG/jsonpath v0.1.0/go.mod h1:4BzmtoM/PI8fPO4aQGIusjGxGir2BzcV0grWtFzq1Y8=
|
||||
github.com/PaesslerAG/jsonpath v0.1.1 h1:c1/AToHQMVsduPAa4Vh6xp2U0evy4t8SWp8imEsylIk=
|
||||
github.com/PaesslerAG/jsonpath v0.1.1/go.mod h1:lVboNxFGal/VwW6d9JzIy56bUsYAP6tH/x80vjnCseY=
|
||||
github.com/PuerkitoBio/goquery v1.13.0 h1:mqHbjD7Jmnul4DTR24LKTjo1uUmHUh072kteGV+xpFM=
|
||||
github.com/PuerkitoBio/goquery v1.13.0/go.mod h1:Hip5mdBL8K2wEGKJdr27sRaNwIdDajmCwB/ExUPwW+g=
|
||||
github.com/aliyun/alibabacloud-oss-go-sdk-v2 v1.5.2 h1:40yUSXwdkWN851BHCq6uiDhleh7A4+0yIBS+IUAqZVY=
|
||||
github.com/aliyun/alibabacloud-oss-go-sdk-v2 v1.5.2/go.mod h1:FTzydeQVmR24FI0D6XWUOMKckjXehM/jgMn1xC+DA9M=
|
||||
github.com/andybalholm/cascadia v1.3.4 h1:vM2lgh0Vru9Vwyfm4cQqWP2HHMW0u0+2PAW7Q38Qufg=
|
||||
github.com/andybalholm/cascadia v1.3.4/go.mod h1:BLRmbRjpEtNKieZOCCvYj4RqN+KRA41GBe/5O+G93kM=
|
||||
github.com/antchfx/htmlquery v1.3.6 h1:RNHHL7YehO5XdO8IM8CynwLKONwRHWkrghbYhQIk9ag=
|
||||
github.com/antchfx/htmlquery v1.3.6/go.mod h1:kcVUqancxPygm26X2rceEcagZFFVkLEE7xgLkGSDl/4=
|
||||
github.com/antchfx/xpath v1.3.6 h1:s0y+ElRRtTQdfHP609qFu0+c6bglDv20pqOViQjjdPI=
|
||||
github.com/antchfx/xpath v1.3.6/go.mod h1:i54GszH55fYfBmoZXapTHN8T8tkcHfRgLyVwwqzXNcs=
|
||||
github.com/bsm/ginkgo/v2 v2.12.0 h1:Ny8MWAHyOepLGlLKYmXG4IEkioBysk6GpaRTLC8zwWs=
|
||||
github.com/bsm/ginkgo/v2 v2.12.0/go.mod h1:SwYbGRRDovPVboqFv0tPTcG1sN61LM1Z4ARdbAV9g4c=
|
||||
github.com/bsm/gomega v1.27.10 h1:yeMWxP2pV2fG3FgAODIY8EiRE3dy0aeFYt4l7wh6yKA=
|
||||
@@ -24,6 +37,12 @@ github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f h1:lO4WD4F/r
|
||||
github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f/go.mod h1:cuUVRXasLTGF7a8hSLbxyZXjz+1KgoB3wDUb6vlszIc=
|
||||
github.com/disintegration/imaging v1.6.3-0.20201218193011-d40f48ce0f09 h1:MJFqtdxTq94XqUgg7DcGCaOIXrDTJE/tPHK66Jshguc=
|
||||
github.com/disintegration/imaging v1.6.3-0.20201218193011-d40f48ce0f09/go.mod h1:44/5580QXChDfwIclfc/PCwrr44amcmDAg8hxG0Ewe4=
|
||||
github.com/dlclark/regexp2 v1.12.0 h1:0j4c5qQmnC6XOWNjP3PIXURXN2gWx76rd3KvgdPkCz8=
|
||||
github.com/dlclark/regexp2 v1.12.0/go.mod h1:DHkYz0B9wPfa6wondMfaivmHpzrQ3v9q8cnmRbL6yW8=
|
||||
github.com/dlclark/regexp2/v2 v2.5.2 h1:HAsucWRhsqcDzl6Ua9aR8JwYOTzrZyPrF0/FNxJVAI0=
|
||||
github.com/dlclark/regexp2/v2 v2.5.2/go.mod h1:avUrQvPaLz2DrFNHJF0taWAFFX2C1GMSSoeiqFjcBmU=
|
||||
github.com/dop251/goja v0.0.0-20260926152631-39ec2650adc9 h1:Ksd3Vil12b/ykT5NjGzd3qNc688cJB9OUXJZZChEktI=
|
||||
github.com/dop251/goja v0.0.0-20260926152631-39ec2650adc9/go.mod h1:u8yZRUavu+N4EnFFy6J5fVtjE7lEcZ2YyV2GcBXY9c8=
|
||||
github.com/dustin/go-humanize v1.0.1 h1:GzkhY7T5VNhEkwH0PVJgjz+fX1rhBrR7pRT3mDkpeCY=
|
||||
github.com/dustin/go-humanize v1.0.1/go.mod h1:Mu1zIs6XwVuF/gI1OepvI0qD18qycQx+mFykh5fBlto=
|
||||
github.com/esimov/pigo v1.4.7-0.20240801095032-7465ed14de47 h1:48iGRx9HamDuG4pCbPG5IXt4bKHhgn33KGynzHUgeIA=
|
||||
@@ -54,6 +73,8 @@ github.com/go-playground/universal-translator v0.18.1 h1:Bcnm0ZwsGyWbCzImXv+pAJn
|
||||
github.com/go-playground/universal-translator v0.18.1/go.mod h1:xekY+UJKNuX9WP91TpwSH2VMlDf28Uj24BCp08ZFTUY=
|
||||
github.com/go-playground/validator/v10 v10.30.1 h1:f3zDSN/zOma+w6+1Wswgd9fLkdwy06ntQJp0BBvFG0w=
|
||||
github.com/go-playground/validator/v10 v10.30.1/go.mod h1:oSuBIQzuJxL//3MelwSLD5hc2Tu889bF0Idm9Dg26cM=
|
||||
github.com/go-sourcemap/sourcemap v2.1.3+incompatible h1:W1iEw64niKVGogNgBN3ePyLFfuisuzeidWPMPWmECqU=
|
||||
github.com/go-sourcemap/sourcemap v2.1.3+incompatible/go.mod h1:F8jJfvm2KbVjc5NqelyYJmf/v5J0dwNLS2mL4sNA1Jg=
|
||||
github.com/goccy/go-json v0.10.6 h1:p8HrPJzOakx/mn/bQtjgNjdTcN+/S6FcG2CTtQOrHVU=
|
||||
github.com/goccy/go-json v0.10.6/go.mod h1:oq7eo15ShAhp70Anwd5lgX2pLfOS3QCiwU/PULtXL6M=
|
||||
github.com/goccy/go-yaml v1.19.2 h1:PmFC1S6h8ljIz6gMRBopkjP1TVT7xuwrButHID66PoM=
|
||||
@@ -62,7 +83,11 @@ github.com/godbus/dbus/v5 v5.1.0 h1:4KLkAxT3aOY8Li4FRJe/KvhoNFFxo0m6fNuFUO8QJUk=
|
||||
github.com/godbus/dbus/v5 v5.1.0/go.mod h1:xhWf0FNVPg57R7Z0UbKHbJfkEywrmjJnf7w5xrFpKfA=
|
||||
github.com/golang-jwt/jwt/v5 v5.2.2 h1:Rl4B7itRWVtYIHFrSNd7vhTiz9UpLdi6gZhZ3wEeDy8=
|
||||
github.com/golang-jwt/jwt/v5 v5.2.2/go.mod h1:pqrtFR0X4osieyHYxtmOUWsAWrfe1Q5UVIyoH402zdk=
|
||||
github.com/golang/groupcache v0.0.0-20210331224755-41bb18bfe9da/go.mod h1:cIg4eruTrX1D+g88fzRXU5OdNfaM+9IcxsU14FzY7Hc=
|
||||
github.com/golang/groupcache v0.0.0-20241129210726-2c02b8208cf8 h1:f+oWsMOmNPc8JmEHVZIycC7hBoQxHH9pNKQORJNozsQ=
|
||||
github.com/golang/groupcache v0.0.0-20241129210726-2c02b8208cf8/go.mod h1:wcDNUvekVysuuOpQKo3191zZyTpiI6se1N1ULghS0sw=
|
||||
github.com/google/go-cmp v0.5.6/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
|
||||
github.com/google/go-cmp v0.6.0/go.mod h1:17dUlkBOakJ0+DkrSSNjCkIjxS6bF9zb3elmeNGIjoY=
|
||||
github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8=
|
||||
github.com/google/go-cmp v0.7.0/go.mod h1:pXiqmnSA92OHEEa9HXL2W4E7lf9JzCmGVUdgjX3N/iU=
|
||||
github.com/google/gofuzz v1.0.0/go.mod h1:dBl0BpW6vV/+mYPU4Po3pmUjxk6FQPldtuIdl/M65Eg=
|
||||
@@ -177,6 +202,7 @@ github.com/ugorji/go/codec v1.3.1 h1:waO7eEiFDwidsBN6agj1vJQ4AG7lh2yqXyOXqhgQuyY
|
||||
github.com/ugorji/go/codec v1.3.1/go.mod h1:pRBVtBSKl77K30Bv8R2P+cLSGaTtex6fsA2Wjqmfxj4=
|
||||
github.com/ulikunitz/xz v0.5.15 h1:9DNdB5s+SgV3bQ2ApL10xRc35ck0DuIX/isZvIk+ubY=
|
||||
github.com/ulikunitz/xz v0.5.15/go.mod h1:nbz6k7qbPmH4IRqmfOplQw/tblSgqTqBwxkY0oWt/14=
|
||||
github.com/yuin/goldmark v1.4.13/go.mod h1:6yULJ656Px+3vBD8DxQVa3kxgyrAnzto9xy5taEt/CY=
|
||||
github.com/yusufpapurcu/wmi v1.2.4 h1:zFUKzehAFReQwLys1b/iSMl+JQGSCSjtVqQn9bBrPo0=
|
||||
github.com/yusufpapurcu/wmi v1.2.4/go.mod h1:SBZ9tNy3G9/m5Oi98Zks0QjeHVDvuK0qfxQmPyzfmi0=
|
||||
go.mongodb.org/mongo-driver/v2 v2.5.0 h1:yXUhImUjjAInNcpTcAlPHiT7bIXhshCTL3jVBkF3xaE=
|
||||
@@ -191,33 +217,104 @@ go.uber.org/zap v1.27.0 h1:aJMhYGrd5QSmlpLMr2MftRKl7t8J8PTZPA732ud/XR8=
|
||||
go.uber.org/zap v1.27.0/go.mod h1:GB2qFLM7cTU87MWRP2mPIjqfIDnGu+VIO4V/SdhGo2E=
|
||||
golang.org/x/arch v0.25.0 h1:qnk6Ksugpi5Bz32947rkUgDt9/s5qvqDPl/gBKdMJLE=
|
||||
golang.org/x/arch v0.25.0/go.mod h1:0X+GdSIP+kL5wPmpK7sdkEVTt2XoYP0cSjQSbZBwOi8=
|
||||
golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
|
||||
golang.org/x/crypto v0.0.0-20210921155107-089bfa567519/go.mod h1:GvvjBRRGRdwPK5ydBHafDWAxML/pGHZbMvKqRZ5+Abc=
|
||||
golang.org/x/crypto v0.13.0/go.mod h1:y6Z2r+Rw4iayiXXAIxJIDAJ1zMW4yaTpebo8fPOliYc=
|
||||
golang.org/x/crypto v0.19.0/go.mod h1:Iy9bg/ha4yyC70EfRS8jz+B6ybOBKMaSxLj6P6oBDfU=
|
||||
golang.org/x/crypto v0.23.0/go.mod h1:CKFgDieR+mRhux2Lsu27y0fO304Db0wZe70UKqHu0v8=
|
||||
golang.org/x/crypto v0.31.0/go.mod h1:kDsLvtWBEx7MV9tJOj9bnXsPbxwJQ6csT/x4KIN4Ssk=
|
||||
golang.org/x/crypto v0.49.0 h1:+Ng2ULVvLHnJ/ZFEq4KdcDd/cfjrrjjNSXNzxg0Y4U4=
|
||||
golang.org/x/crypto v0.49.0/go.mod h1:ErX4dUh2UM+CFYiXZRTcMpEcN8b/1gxEuv3nODoYtCA=
|
||||
golang.org/x/crypto v0.55.0 h1:+KWHjbgOaAQ66dh/YlkZKHlz9ZUlq61AFirAR9ntP8M=
|
||||
golang.org/x/crypto v0.55.0/go.mod h1:uq0V9dE/fzQuJtbnL+2EhWOE63vo164FY8xqEnV9xis=
|
||||
golang.org/x/exp v0.0.0-20251023183803-a4bb9ffd2546 h1:mgKeJMpvi0yx/sU5GsxQ7p6s2wtOnGAHZWCHUM4KGzY=
|
||||
golang.org/x/exp v0.0.0-20251023183803-a4bb9ffd2546/go.mod h1:j/pmGrbnkbPtQfxEe5D0VQhZC6qKbfKifgD0oM7sR70=
|
||||
golang.org/x/image v0.0.0-20191009234506-e7c1f5e7dbb8/go.mod h1:FeLwcggjj3mMvU+oOTbSwawSJRM1uh48EjtB4UJZlP0=
|
||||
golang.org/x/image v0.45.0 h1:FMb1nTbH5H9vF55SriQHgFw5GnNL9Jg6L25BwXKzhB0=
|
||||
golang.org/x/image v0.45.0/go.mod h1:n62x/7RqlwXDvGsSU4u6IUTUf6KghUZ9Bt7cG/T9Fx4=
|
||||
golang.org/x/mod v0.6.0-dev.0.20220419223038-86c51ed26bb4/go.mod h1:jJ57K6gSWd91VN4djpZkiMVwK6gcyfeH4XE8wZrZaV4=
|
||||
golang.org/x/mod v0.8.0/go.mod h1:iBbtSCu2XBx23ZKBPSOrRkjjQPZFPuis4dIYUhu/chs=
|
||||
golang.org/x/mod v0.12.0/go.mod h1:iBbtSCu2XBx23ZKBPSOrRkjjQPZFPuis4dIYUhu/chs=
|
||||
golang.org/x/mod v0.15.0/go.mod h1:hTbmBsO62+eylJbnUtE2MGJUyE7QWk4xUqPFrRgJ+7c=
|
||||
golang.org/x/mod v0.17.0/go.mod h1:hTbmBsO62+eylJbnUtE2MGJUyE7QWk4xUqPFrRgJ+7c=
|
||||
golang.org/x/mod v0.38.0 h1:MECBjubtXD7yj4HrhIUcywNaGeNVUdfVnxmPajOk4yk=
|
||||
golang.org/x/mod v0.38.0/go.mod h1:V6Xz0pq8TQ3dGqVQ1FVHuelZpAL0uNhSkk9ogYP3c40=
|
||||
golang.org/x/mod v0.41.0 h1:qJmnOUb4YB+FsEuM3HcWucdZASCPGhsX6uljO6pog0c=
|
||||
golang.org/x/net v0.0.0-20190620200207-3b0461eec859/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
|
||||
golang.org/x/net v0.0.0-20210226172049-e18ecbb05110/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
|
||||
golang.org/x/net v0.0.0-20220722155237-a158d28d115b/go.mod h1:XRhObCWvk6IyKnWLug+ECip1KBveYUHfp+8e9klMJ9c=
|
||||
golang.org/x/net v0.6.0/go.mod h1:2Tu9+aMcznHK/AK1HMvgo6xiTLG5rD5rZLDS+rp2Bjs=
|
||||
golang.org/x/net v0.10.0/go.mod h1:0qNGK6F8kojg2nk9dLZ2mShWaEBan6FAoqfSigmmuDg=
|
||||
golang.org/x/net v0.15.0/go.mod h1:idbUs1IY1+zTqbi8yxTbhexhEEk5ur9LInksu6HrEpk=
|
||||
golang.org/x/net v0.21.0/go.mod h1:bIjVDfnllIU7BJ2DNgfnXvpSvtn8VRwhlsaeUTyUS44=
|
||||
golang.org/x/net v0.25.0/go.mod h1:JkAGAh7GEvH74S6FOH42FLoXpXbE/aqXSrIQjXgsiwM=
|
||||
golang.org/x/net v0.33.0/go.mod h1:HXLR5J+9DxmrqMwG9qjGCxZ+zKXxBru04zlTvWlWuN4=
|
||||
golang.org/x/net v0.52.0 h1:He/TN1l0e4mmR3QqHMT2Xab3Aj3L9qjbhRm78/6jrW0=
|
||||
golang.org/x/net v0.52.0/go.mod h1:R1MAz7uMZxVMualyPXb+VaqGSa3LIaUqk0eEt3w36Sw=
|
||||
golang.org/x/net v0.58.0 h1:ynWG7rqYi4ccpTEuPZ2QGWHktVEM9DMCj9yzDE0Q7To=
|
||||
golang.org/x/net v0.58.0/go.mod h1:YwCddHnFlT7eLQqVprV19OnhLGtc5xOKgE0RyqgfWAU=
|
||||
golang.org/x/sync v0.0.0-20190423024810-112230192c58/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
|
||||
golang.org/x/sync v0.0.0-20220722155255-886fb9371eb4/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
|
||||
golang.org/x/sync v0.1.0/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
|
||||
golang.org/x/sync v0.3.0/go.mod h1:FU7BRWz2tNW+3quACPkgCx/L+uEAv1htQ0V83Z9Rj+Y=
|
||||
golang.org/x/sync v0.6.0/go.mod h1:Czt+wKu1gCyEFDUtn0jG5QVvpJ6rzVqr5aXyt9drQfk=
|
||||
golang.org/x/sync v0.7.0/go.mod h1:Czt+wKu1gCyEFDUtn0jG5QVvpJ6rzVqr5aXyt9drQfk=
|
||||
golang.org/x/sync v0.10.0/go.mod h1:Czt+wKu1gCyEFDUtn0jG5QVvpJ6rzVqr5aXyt9drQfk=
|
||||
golang.org/x/sync v0.22.0 h1:SZjpbeLmrCk4xhRSZFNZW5gFUeCeFgjekvI/+gfScek=
|
||||
golang.org/x/sync v0.22.0/go.mod h1:9xrNwdLfx4jkKbNva9FpL6vEN7evnE43NNNJQ2LF3+0=
|
||||
golang.org/x/sync v0.23.0 h1:KameEIfc1IkluZyXWLn39Wd4tURc6GbCiISGiZm2bQk=
|
||||
golang.org/x/sync v0.23.0/go.mod h1:sUUOizhqBxiL6pEWpqNLUiaJn1ShEbZ6BBqskPbjZm0=
|
||||
golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
|
||||
golang.org/x/sys v0.0.0-20190916202348-b4ddaad3f8a3/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
|
||||
golang.org/x/sys v0.0.0-20201119102817-f84b799fce68/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
|
||||
golang.org/x/sys v0.0.0-20201204225414-ed752295db88/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
|
||||
golang.org/x/sys v0.0.0-20210615035016-665e8c7367d1/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
|
||||
golang.org/x/sys v0.0.0-20220520151302-bc2c85ada10a/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
|
||||
golang.org/x/sys v0.0.0-20220722155257-8c9f86f7a55f/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
|
||||
golang.org/x/sys v0.5.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
|
||||
golang.org/x/sys v0.6.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
|
||||
golang.org/x/sys v0.8.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
|
||||
golang.org/x/sys v0.11.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
|
||||
golang.org/x/sys v0.12.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
|
||||
golang.org/x/sys v0.17.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
|
||||
golang.org/x/sys v0.20.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
|
||||
golang.org/x/sys v0.28.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
|
||||
golang.org/x/sys v0.47.0 h1:o7XGOvZQCADBQQ4Y7VNq2dRWQR7JmOUW8Kxx4ZsNgWs=
|
||||
golang.org/x/sys v0.47.0/go.mod h1:4GL1E5IUh+htKOUEOaiffhrAeqysfVGipDYzABqnCmw=
|
||||
golang.org/x/telemetry v0.0.0-20240228155512-f48c80bd79b2/go.mod h1:TeRTkGYfJXctD9OcfyVLyj2J3IxLnKwHJR8f4D8a3YE=
|
||||
golang.org/x/term v0.0.0-20201126162022-7de9c90e9dd1/go.mod h1:bj7SfCRtBDWHUb9snDiAeCFNEtKQo2Wmx5Cou7ajbmo=
|
||||
golang.org/x/term v0.0.0-20210927222741-03fcf44c2211/go.mod h1:jbD1KX2456YbFQfuXm/mYQcufACuNUgVhRMnK/tPxf8=
|
||||
golang.org/x/term v0.5.0/go.mod h1:jMB1sMXY+tzblOD4FWmEbocvup2/aLOaQEp7JmGp78k=
|
||||
golang.org/x/term v0.8.0/go.mod h1:xPskH00ivmX89bAKVGSKKtLOWNx2+17Eiy94tnKShWo=
|
||||
golang.org/x/term v0.12.0/go.mod h1:owVbMEjm3cBLCHdkQu9b1opXd4ETQWc3BhuQGKgXgvU=
|
||||
golang.org/x/term v0.17.0/go.mod h1:lLRBjIVuehSbZlaOtGMbcMncT+aqLLLmKrsjNrUguwk=
|
||||
golang.org/x/term v0.20.0/go.mod h1:8UkIAJTvZgivsXaD6/pH6U9ecQzZ45awqEOzuCvwpFY=
|
||||
golang.org/x/term v0.27.0/go.mod h1:iMsnZpn0cago0GOrHO2+Y7u7JPn5AylBrcoWkElMTSM=
|
||||
golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
|
||||
golang.org/x/text v0.3.3/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
|
||||
golang.org/x/text v0.3.7/go.mod h1:u+2+/6zg+i71rQMx5EYifcz6MCKuco9NR6JIITiCfzQ=
|
||||
golang.org/x/text v0.7.0/go.mod h1:mrYo+phRRbMaCq/xk9113O4dZlRixOauAjOtrjsXDZ8=
|
||||
golang.org/x/text v0.9.0/go.mod h1:e1OnstbJyHTd6l/uOt8jFFHp6TRDWZR/bV3emEE/zU8=
|
||||
golang.org/x/text v0.13.0/go.mod h1:TvPlkZtksWOMsz7fbANvkp4WM8x/WCo/om8BMLbz+aE=
|
||||
golang.org/x/text v0.14.0/go.mod h1:18ZOQIKpY8NJVqYksKHtTdi31H5itFRjB5/qKTNYzSU=
|
||||
golang.org/x/text v0.15.0/go.mod h1:18ZOQIKpY8NJVqYksKHtTdi31H5itFRjB5/qKTNYzSU=
|
||||
golang.org/x/text v0.21.0/go.mod h1:4IBbMaMmOPCJ8SecivzSH54+73PCFmPWxNTLm+vZkEQ=
|
||||
golang.org/x/text v0.41.0 h1:vz/seA0lnX87Othu2f/0L24RcgrXD9/YFTSuGjj3rH8=
|
||||
golang.org/x/text v0.41.0/go.mod h1:jvf1O8ajNzZqhSrQBPbutR/EB83Cc0CFrezNQIwbb5M=
|
||||
golang.org/x/text v0.42.0 h1:JbOZXgfeCPU9gacVtYliJqOhD+zhrEqK4LfdpmlUZqI=
|
||||
golang.org/x/text v0.42.0/go.mod h1:ojzP1Z+2QtioaF8DTtO8K5q7JWVVYwZKenzujK0Zd0E=
|
||||
golang.org/x/time v0.15.0 h1:bbrp8t3bGUeFOx08pvsMYRTCVSMk89u4tKbNOZbp88U=
|
||||
golang.org/x/time v0.15.0/go.mod h1:Y4YMaQmXwGQZoFaVFk4YpCt4FLQMYKZe9oeV/f4MSno=
|
||||
golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
|
||||
golang.org/x/tools v0.0.0-20191119224855-298f0cb1881e/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
|
||||
golang.org/x/tools v0.1.12/go.mod h1:hNGJHUnrk76NpqgfD5Aqm5Crs+Hm0VOH/i9J2+nxYbc=
|
||||
golang.org/x/tools v0.6.0/go.mod h1:Xwgl3UAJ/d3gWutnCtw505GrjyAbvKui8lOU390QaIU=
|
||||
golang.org/x/tools v0.13.0/go.mod h1:HvlwmtVNQAhOuCjW7xxvovg8wbNq7LwfXh/k7wXUl58=
|
||||
golang.org/x/tools v0.21.1-0.20240508182429-e35e4ccd0d2d/go.mod h1:aiJjzUbINMkxbQROHiO6hDPo2LHcIPhhQsa9DLh0yGk=
|
||||
golang.org/x/tools v0.48.0 h1:3+hClM1aLL5mjMKm5ovokw9epgRXPuu2tILgismM6RE=
|
||||
golang.org/x/tools v0.48.0/go.mod h1:08xX0orndb/F7jJxGDicx061tyd5pcMto75YMAXr6lk=
|
||||
golang.org/x/tools v0.49.0 h1:3NI7VXzL9+1WZD52Dx2ttoPwD5DWrFGpl9mFZDlmisI=
|
||||
golang.org/x/xerrors v0.0.0-20190717185122-a985d3407aa7/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
|
||||
golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
|
||||
google.golang.org/protobuf v1.36.11 h1:fV6ZwhNocDyBLK0dj+fg8ektcVegBBuEolpbTQyBNVE=
|
||||
google.golang.org/protobuf v1.36.11/go.mod h1:HTf+CrKn2C3g5S8VImy6tdcUvCska2kB7j23XfzDpco=
|
||||
|
||||
@@ -26,6 +26,20 @@ func Register(r *gin.Engine, cfg *config.Config, log *zap.Logger, svc *service.C
|
||||
// STRM 播放端点:strm 文件内容指向这里,Emby/Infuse 直接请求(无 JWT)。
|
||||
api.GET("/strm/play/:provider/:file", strmPlayHandler(svc))
|
||||
api.HEAD("/strm/play/:provider/:file", strmPlayHandler(svc))
|
||||
// 阅读媒体代理:音频流/漫画图片,鉴权靠 HMAC 签名(audio/img 元素带不了 JWT)。
|
||||
api.GET("/reader/media", readerMediaProxyHandler(svc))
|
||||
// 本地书籍内嵌资源(EPUB 图片):同样靠 HMAC 签名鉴权。
|
||||
api.GET("/reader/local/asset", readerLocalAssetHandler(svc))
|
||||
// 本地有声书音频流:同样靠 HMAC 签名鉴权,服务端处理 Range。
|
||||
api.GET("/reader/local/audio", readerLocalAudioHandler(svc))
|
||||
api.HEAD("/reader/local/audio", readerLocalAudioHandler(svc))
|
||||
// 需要转码的音轨(WMA 等浏览器解不了的格式):签名鉴权 + 服务端 ffmpeg 转码。
|
||||
api.GET("/reader/audio/transcode", readerAudioTranscodeHandler(svc))
|
||||
api.HEAD("/reader/audio/transcode", readerAudioTranscodeHandler(svc))
|
||||
// 阅读书源浏览器页面与资源代理:java.startBrowserAwait 承载的页面填进
|
||||
// <iframe src>,同样带不了 JWT,鉴权靠绑定待办 ID 的 HMAC 签名。
|
||||
api.GET("/reader/browser/page", readerBrowserPageHandler(svc))
|
||||
api.GET("/reader/browser/asset", readerBrowserAssetHandler(svc))
|
||||
// 115 中继/CloudDrive 授权回跳(authorization_id 会话 + 共享密钥校验)
|
||||
api.POST("/strm/oauth/callback", strm115OAuthCallbackHandler(svc))
|
||||
api.GET("/strm/oauth/callback", strm115OAuthCallbackHandler(svc))
|
||||
|
||||
@@ -0,0 +1,247 @@
|
||||
package handler
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"net/url"
|
||||
"strings"
|
||||
"testing"
|
||||
"time"
|
||||
|
||||
"github.com/gin-gonic/gin"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/service"
|
||||
)
|
||||
|
||||
// 书源宿主浏览器的 HTTP 层测试:轮询待办 → 承载页面 → 回传 DOM。
|
||||
//
|
||||
// 这条链路对应真实书源的「切换线路」「用户后台」:
|
||||
// java.startBrowserAwait 会阻塞在 POST /sources/:id/login 上,
|
||||
// 前端必须能从 /reader/browser/pending 拿到页面,再把 DOM POST 回去。
|
||||
|
||||
// browserHandlerSource 构造一个用 startBrowserAwait 切换线路的书源。
|
||||
func browserHandlerSource(sourceURL string) string {
|
||||
loginJS := `function switchLine() {
|
||||
let html = '<!DOCTYPE html><html><head></head><body><span id="serverValue">线路甲</span></body></html>';
|
||||
let body = java.startBrowserAwait('data:text/html;base64,' + java.base64Encode(html), '线路设置', false).body();
|
||||
let m = body.match(/id="serverValue"\s*>\s*([^<]*?)\s*<\/span>/);
|
||||
source.setVariable(JSON.stringify({线路: m ? m[1] : ''}));
|
||||
}`
|
||||
src := map[string]any{
|
||||
"bookSourceUrl": sourceURL,
|
||||
"bookSourceName": "浏览器面板源",
|
||||
"loginUrl": loginJS,
|
||||
"loginUi": `[{"name":"切换线路","type":"button","action":"switchLine()"}]`,
|
||||
}
|
||||
out, _ := json.Marshal(src)
|
||||
return string(out)
|
||||
}
|
||||
|
||||
// registerBrowserRoutesForTest 挂载阅读路由 + 承载页面/资源代理(跳过鉴权)。
|
||||
// 生产环境里后两个是公开路由(iframe 带不了 JWT),鉴权靠 HMAC 签名。
|
||||
func registerBrowserRoutesForTest(cfg *service.Container) *gin.Engine {
|
||||
r := registerReaderRoutesForTest(cfg)
|
||||
r.GET("/api/reader/browser/page", readerBrowserPageHandler(cfg))
|
||||
r.GET("/api/reader/browser/asset", readerBrowserAssetHandler(cfg))
|
||||
return r
|
||||
}
|
||||
|
||||
func TestReaderBrowserPanelRoutes(t *testing.T) {
|
||||
container := newReaderHandlerContainer(t)
|
||||
router := registerBrowserRoutesForTest(container)
|
||||
|
||||
// 导入书源
|
||||
body, _ := json.Marshal(map[string]string{"text": browserHandlerSource("https://panel.example.com")})
|
||||
w := httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodPost, "/api/reader/sources/import", strings.NewReader(string(body)))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("导入书源 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
w = httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, "/api/reader/sources", nil))
|
||||
var list struct {
|
||||
Sources []struct {
|
||||
ID string `json:"id"`
|
||||
} `json:"sources"`
|
||||
}
|
||||
_ = json.Unmarshal(w.Body.Bytes(), &list)
|
||||
if len(list.Sources) != 1 {
|
||||
t.Fatalf("应导入 1 个书源: %s", w.Body.String())
|
||||
}
|
||||
sourceID := list.Sources[0].ID
|
||||
|
||||
// 执行切换线路:这个请求会阻塞等待用户操作,放到后台跑
|
||||
actionDone := make(chan *httptest.ResponseRecorder, 1)
|
||||
go func() {
|
||||
rec := httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodPost, "/api/reader/sources/"+sourceID+"/login",
|
||||
strings.NewReader(`{"action":"switchLine()","fields":{}}`))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
router.ServeHTTP(rec, req)
|
||||
actionDone <- rec
|
||||
}()
|
||||
|
||||
// 轮询待办页面
|
||||
var page struct {
|
||||
ID string `json:"id"`
|
||||
Title string `json:"title"`
|
||||
Mode string `json:"mode"`
|
||||
PageURL string `json:"page_url"`
|
||||
SourceID string `json:"source_id"`
|
||||
}
|
||||
deadline := time.Now().Add(5 * time.Second)
|
||||
for {
|
||||
if time.Now().After(deadline) {
|
||||
t.Fatal("等待超时:未出现待办页面")
|
||||
}
|
||||
w = httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet,
|
||||
"/api/reader/browser/pending?source_id="+sourceID, nil))
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("轮询 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
var pending struct {
|
||||
Pages []json.RawMessage `json:"pages"`
|
||||
}
|
||||
if err := json.Unmarshal(w.Body.Bytes(), &pending); err != nil {
|
||||
t.Fatalf("待办返回不是 JSON: %s", w.Body.String())
|
||||
}
|
||||
if len(pending.Pages) > 0 {
|
||||
if err := json.Unmarshal(pending.Pages[0], &page); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
break
|
||||
}
|
||||
time.Sleep(20 * time.Millisecond)
|
||||
}
|
||||
if page.Mode != "wait" || page.Title != "线路设置" {
|
||||
t.Fatalf("待办描述异常: %+v", page)
|
||||
}
|
||||
if page.SourceID != sourceID {
|
||||
t.Fatalf("待办应带书源 ID: %+v", page)
|
||||
}
|
||||
|
||||
// 承载页面:签名正确应返回 HTML,签名错误应 403
|
||||
w = httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, page.PageURL, nil))
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("承载页面 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
if !strings.Contains(w.Body.String(), `id="serverValue"`) {
|
||||
t.Fatalf("承载页面内容异常: %s", w.Body.String())
|
||||
}
|
||||
// 服务端注入的 DOM 回传脚本必须在(父窗口靠它取回用户操作后的页面)
|
||||
if !strings.Contains(w.Body.String(), "data-mebox-browser-bridge") {
|
||||
t.Fatalf("承载页面缺少 DOM 回传脚本: %s", w.Body.String())
|
||||
}
|
||||
w = httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, "/api/reader/browser/page?id="+page.ID+"&s=forged", nil))
|
||||
if w.Code != http.StatusForbidden {
|
||||
t.Fatalf("伪造签名应 403,实际 %d", w.Code)
|
||||
}
|
||||
|
||||
// 回传用户操作后的 DOM(模拟点了 √)
|
||||
result, _ := json.Marshal(map[string]any{
|
||||
"id": page.ID,
|
||||
"body": `<html><body><span id="serverValue">线路乙</span></body></html>`,
|
||||
})
|
||||
w = httptest.NewRecorder()
|
||||
req = httptest.NewRequest(http.MethodPost, "/api/reader/browser/result", strings.NewReader(string(result)))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("回传结果 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
|
||||
// 阻塞的登录动作应被放行
|
||||
select {
|
||||
case rec := <-actionDone:
|
||||
if rec.Code != http.StatusOK {
|
||||
t.Fatalf("登录动作 status=%d body=%s", rec.Code, rec.Body.String())
|
||||
}
|
||||
var res struct {
|
||||
OK bool `json:"ok"`
|
||||
}
|
||||
_ = json.Unmarshal(rec.Body.Bytes(), &res)
|
||||
if !res.OK {
|
||||
t.Fatalf("动作未成功: %s", rec.Body.String())
|
||||
}
|
||||
case <-time.After(5 * time.Second):
|
||||
t.Fatal("回传后阻塞未解除")
|
||||
}
|
||||
|
||||
// 线路应已写进源变量
|
||||
w = httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, "/api/reader/sources/"+sourceID+"/login", nil))
|
||||
var info struct {
|
||||
Variable string `json:"variable"`
|
||||
}
|
||||
_ = json.Unmarshal(w.Body.Bytes(), &info)
|
||||
if !strings.Contains(info.Variable, "线路乙") {
|
||||
t.Fatalf("线路未写入源变量: %q", info.Variable)
|
||||
}
|
||||
|
||||
// 完成后待办清空
|
||||
w = httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet,
|
||||
"/api/reader/browser/pending?source_id="+sourceID, nil))
|
||||
var after struct {
|
||||
Pages []json.RawMessage `json:"pages"`
|
||||
}
|
||||
_ = json.Unmarshal(w.Body.Bytes(), &after)
|
||||
if len(after.Pages) != 0 {
|
||||
t.Fatalf("完成后待办未清理: %s", w.Body.String())
|
||||
}
|
||||
}
|
||||
|
||||
func TestReaderBrowserResultRejectsUnknownPage(t *testing.T) {
|
||||
container := newReaderHandlerContainer(t)
|
||||
router := registerBrowserRoutesForTest(container)
|
||||
|
||||
w := httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodPost, "/api/reader/browser/result",
|
||||
strings.NewReader(`{"id":"not-a-real-page","body":"x"}`))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusBadRequest {
|
||||
t.Fatalf("未知待办应 400,实际 %d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
}
|
||||
|
||||
func TestReaderBrowserResultRequiresID(t *testing.T) {
|
||||
container := newReaderHandlerContainer(t)
|
||||
router := registerBrowserRoutesForTest(container)
|
||||
|
||||
w := httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodPost, "/api/reader/browser/result", strings.NewReader(`{}`))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusBadRequest {
|
||||
t.Fatalf("缺少 id 应 400,实际 %d", w.Code)
|
||||
}
|
||||
}
|
||||
|
||||
// TestReaderBrowserPageBadQuery 承载地址参数缺失/非法时不应 500。
|
||||
func TestReaderBrowserPageBadQuery(t *testing.T) {
|
||||
container := newReaderHandlerContainer(t)
|
||||
router := registerBrowserRoutesForTest(container)
|
||||
|
||||
for _, target := range []string{"/api/reader/browser/page", "/api/reader/browser/page?id=x",
|
||||
"/api/reader/browser/asset", "/api/reader/browser/asset?id=x&u=!!!&s=y"} {
|
||||
w := httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, target, nil))
|
||||
if w.Code != http.StatusForbidden {
|
||||
t.Fatalf("%s 应 403,实际 %d", target, w.Code)
|
||||
}
|
||||
}
|
||||
// 合法 base64 但不匹配的签名
|
||||
w := httptest.NewRecorder()
|
||||
q := url.Values{"id": {"x"}, "u": {"aHR0cHM6Ly9leGFtcGxlLmNvbS8"}, "s": {"deadbeef"}}
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, "/api/reader/browser/asset?"+q.Encode(), nil))
|
||||
if w.Code != http.StatusForbidden {
|
||||
t.Fatalf("签名不匹配应 403,实际 %d", w.Code)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,40 @@
|
||||
package handler
|
||||
|
||||
import (
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"strings"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// 段评打开接口的 HTTP 层契约:路由存在、参数校验生效、内网地址被拒。
|
||||
//
|
||||
// 正常路径(真的去抓评论页)依赖外部站点,放在服务层解析单测里覆盖;
|
||||
// 这里只钉住契约与安全边界,避免把测试绑到网络。
|
||||
func TestReaderOpenCommentRejectsBadURL(t *testing.T) {
|
||||
container := newReaderHandlerContainer(t)
|
||||
router := registerReaderRoutesForTest(container)
|
||||
|
||||
post := func(body string) int {
|
||||
w := httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodPost, "/api/reader/comments/open", strings.NewReader(body))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
router.ServeHTTP(w, req)
|
||||
return w.Code
|
||||
}
|
||||
|
||||
cases := []struct {
|
||||
name string
|
||||
body string
|
||||
}{
|
||||
{"缺少 url", `{"book_id":"x"}`},
|
||||
{"非 http 协议", `{"book_id":"x","url":"javascript:alert(1)"}`},
|
||||
{"回环地址", `{"book_id":"x","url":"http://127.0.0.1/c"}`},
|
||||
{"内网地址", `{"book_id":"x","url":"http://192.168.1.1/c"}`},
|
||||
}
|
||||
for _, c := range cases {
|
||||
if code := post(c.body); code != http.StatusBadRequest {
|
||||
t.Fatalf("%s:应 400,得到 %d", c.name, code)
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,240 @@
|
||||
package handler
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"strings"
|
||||
"testing"
|
||||
|
||||
"github.com/gin-gonic/gin"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
"github.com/truewhile/MeBox/internal/service"
|
||||
"github.com/truewhile/MeBox/internal/service/reader"
|
||||
)
|
||||
|
||||
// 书架分组接口(/reader/book-groups)的 HTTP 契约测试。
|
||||
//
|
||||
// 分组仿影视模块的媒体库标签:整份替换、一本书只归一个组、按用户隔离。
|
||||
// 另外两条是书籍特有的:只保留书架上的书(书移出后不留死 ID)、空分组保留。
|
||||
|
||||
func addBookForGroups(t *testing.T, svc *service.Container, userID, name, bookURL string) *model.ReaderBook {
|
||||
t.Helper()
|
||||
book, err := svc.Reader.AddBook(t.Context(), userID, reader.SearchOrigin{
|
||||
Origin: "https://example.com",
|
||||
BookURL: bookURL,
|
||||
}, name, "作者", "")
|
||||
if err != nil {
|
||||
t.Fatalf("加入书架失败: %v", err)
|
||||
}
|
||||
return book
|
||||
}
|
||||
|
||||
func setBookGroups(t *testing.T, router *gin.Engine, user string, groups []model.BookGroupSet) ([]model.BookGroupSet, int) {
|
||||
t.Helper()
|
||||
body, _ := json.Marshal(map[string]any{"groups": groups})
|
||||
w := httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodPut, "/api/reader/book-groups", strings.NewReader(string(body)))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
req.Header.Set("X-Test-User", user)
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusOK {
|
||||
return nil, w.Code
|
||||
}
|
||||
var res struct {
|
||||
Groups []model.BookGroupSet `json:"groups"`
|
||||
}
|
||||
if err := json.Unmarshal(w.Body.Bytes(), &res); err != nil {
|
||||
t.Fatalf("解析响应失败: %v body=%s", err, w.Body.String())
|
||||
}
|
||||
return res.Groups, w.Code
|
||||
}
|
||||
|
||||
func getBookGroups(t *testing.T, router *gin.Engine, user string) ([]model.BookGroupSet, int) {
|
||||
t.Helper()
|
||||
w := httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodGet, "/api/reader/book-groups", nil)
|
||||
req.Header.Set("X-Test-User", user)
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusOK {
|
||||
return nil, w.Code
|
||||
}
|
||||
var res struct {
|
||||
Groups []model.BookGroupSet `json:"groups"`
|
||||
}
|
||||
if err := json.Unmarshal(w.Body.Bytes(), &res); err != nil {
|
||||
t.Fatalf("解析响应失败: %v body=%s", err, w.Body.String())
|
||||
}
|
||||
return res.Groups, w.Code
|
||||
}
|
||||
|
||||
func namesOf(groups []model.BookGroupSet) string {
|
||||
out := make([]string, 0, len(groups))
|
||||
for _, g := range groups {
|
||||
out = append(out, g.Name)
|
||||
}
|
||||
return strings.Join(out, ",")
|
||||
}
|
||||
|
||||
func idsOf(groups []model.BookGroupSet, name string) string {
|
||||
for _, g := range groups {
|
||||
if g.Name == name {
|
||||
return strings.Join(g.BookIDs, ",")
|
||||
}
|
||||
}
|
||||
return "<missing>"
|
||||
}
|
||||
|
||||
// TestBookGroupsEmptyStateIsArray 没有分组时应返回空数组,前端可直接遍历。
|
||||
func TestBookGroupsEmptyStateIsArray(t *testing.T) {
|
||||
svc := newReaderHandlerContainer(t)
|
||||
router := profileRouterForTest(t, svc)
|
||||
|
||||
w := httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodGet, "/api/reader/book-groups", nil)
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
// 必须是 [] 而不是 null
|
||||
if !strings.Contains(w.Body.String(), `"groups":[]`) {
|
||||
t.Fatalf("空态应返回 [],实际 %s", w.Body.String())
|
||||
}
|
||||
}
|
||||
|
||||
func TestBookGroupsNormalizeFilterAndDedupe(t *testing.T) {
|
||||
svc := newReaderHandlerContainer(t)
|
||||
router := profileRouterForTest(t, svc)
|
||||
|
||||
b1 := addBookForGroups(t, svc, "u1", "书一", "https://example.com/1")
|
||||
b2 := addBookForGroups(t, svc, "u1", "书二", "https://example.com/2")
|
||||
b3 := addBookForGroups(t, svc, "u1", "书三", "https://example.com/3")
|
||||
foreign := addBookForGroups(t, svc, "u2", "别人的书", "https://example.com/9")
|
||||
|
||||
got, code := setBookGroups(t, router, "u1", []model.BookGroupSet{
|
||||
{Name: "科幻", BookIDs: []string{b1.ID, b2.ID}},
|
||||
{Name: "在读", BookIDs: []string{b1.ID}}, // b1 已被「科幻」认领 → 应变空
|
||||
{Name: "玄幻", BookIDs: []string{b3.ID}},
|
||||
{Name: "空组", BookIDs: []string{}},
|
||||
{Name: " ", BookIDs: []string{b2.ID}}, // 空名 → 丢弃整个分组
|
||||
{Name: "外来", BookIDs: []string{foreign.ID}}, // 不属于 u1 的书架 → 过滤掉
|
||||
})
|
||||
if code != http.StatusOK {
|
||||
t.Fatalf("保存失败 status=%d", code)
|
||||
}
|
||||
if names := namesOf(got); names != "科幻,在读,玄幻,空组,外来" {
|
||||
t.Fatalf("分组顺序/数量 = %q,期望 科幻,在读,玄幻,空组,外来", names)
|
||||
}
|
||||
if ids := idsOf(got, "科幻"); ids != b1.ID+","+b2.ID {
|
||||
t.Fatalf("科幻 = %q,期望 %s,%s", ids, b1.ID, b2.ID)
|
||||
}
|
||||
if ids := idsOf(got, "在读"); ids != "" {
|
||||
t.Fatalf("在读 应为空(b1 归了科幻),实际 %q", ids)
|
||||
}
|
||||
if ids := idsOf(got, "玄幻"); ids != b3.ID {
|
||||
t.Fatalf("玄幻 = %q,期望 %s", ids, b3.ID)
|
||||
}
|
||||
if ids := idsOf(got, "外来"); ids != "" {
|
||||
t.Fatalf("外来 应被过滤为空,实际 %q", ids)
|
||||
}
|
||||
|
||||
// 读回来应当一致
|
||||
loaded, code := getBookGroups(t, router, "u1")
|
||||
if code != http.StatusOK {
|
||||
t.Fatalf("读回 status=%d", code)
|
||||
}
|
||||
if names := namesOf(loaded); names != namesOf(got) {
|
||||
t.Fatalf("读回分组 = %q,期望 %q", names, namesOf(got))
|
||||
}
|
||||
if ids := idsOf(loaded, "玄幻"); ids != b3.ID {
|
||||
t.Fatalf("读回 玄幻 = %q,期望 %s", ids, b3.ID)
|
||||
}
|
||||
}
|
||||
|
||||
// TestBookGroupsDropsBooksRemovedFromShelf 书被移出书架后,分组里不应留下死 ID。
|
||||
func TestBookGroupsDropsBooksRemovedFromShelf(t *testing.T) {
|
||||
svc := newReaderHandlerContainer(t)
|
||||
router := profileRouterForTest(t, svc)
|
||||
|
||||
b1 := addBookForGroups(t, svc, "u1", "书一", "https://example.com/1")
|
||||
b2 := addBookForGroups(t, svc, "u1", "书二", "https://example.com/2")
|
||||
|
||||
if _, code := setBookGroups(t, router, "u1", []model.BookGroupSet{
|
||||
{Name: "科幻", BookIDs: []string{b1.ID, b2.ID}},
|
||||
}); code != http.StatusOK {
|
||||
t.Fatalf("保存失败 status=%d", code)
|
||||
}
|
||||
|
||||
if err := svc.Reader.RemoveBook(t.Context(), "u1", b2.ID); err != nil {
|
||||
t.Fatalf("移出书架失败: %v", err)
|
||||
}
|
||||
|
||||
loaded, _ := getBookGroups(t, router, "u1")
|
||||
if ids := idsOf(loaded, "科幻"); ids != b1.ID {
|
||||
t.Fatalf("科幻 = %q,期望只剩 %s", ids, b1.ID)
|
||||
}
|
||||
}
|
||||
|
||||
func TestBookGroupsArePerUser(t *testing.T) {
|
||||
svc := newReaderHandlerContainer(t)
|
||||
router := profileRouterForTest(t, svc)
|
||||
|
||||
u1Book := addBookForGroups(t, svc, "u1", "书一", "https://example.com/1")
|
||||
u2Book := addBookForGroups(t, svc, "u2", "书二", "https://example.com/2")
|
||||
|
||||
if _, code := setBookGroups(t, router, "u1", []model.BookGroupSet{
|
||||
{Name: "u1 的分组", BookIDs: []string{u1Book.ID}},
|
||||
}); code != http.StatusOK {
|
||||
t.Fatalf("u1 保存失败 status=%d", code)
|
||||
}
|
||||
|
||||
// u2 看不到 u1 的分组
|
||||
if groups, code := getBookGroups(t, router, "u2"); code != http.StatusOK || len(groups) != 0 {
|
||||
t.Fatalf("u2 分组 = %+v (status=%d),期望空", groups, code)
|
||||
}
|
||||
|
||||
// u2 把 u1 的书塞进自己的分组也不会生效(不属于 u2 的书架)
|
||||
if groups, code := setBookGroups(t, router, "u2", []model.BookGroupSet{
|
||||
{Name: "抢过来", BookIDs: []string{u1Book.ID, u2Book.ID}},
|
||||
}); code != http.StatusOK || idsOf(groups, "抢过来") != u2Book.ID {
|
||||
t.Fatalf("u2 分组 = %+v (status=%d),期望只含自己的书", groups, code)
|
||||
}
|
||||
|
||||
// u1 的分组不受影响
|
||||
back, _ := getBookGroups(t, router, "u1")
|
||||
if names := namesOf(back); names != "u1 的分组" || idsOf(back, "u1 的分组") != u1Book.ID {
|
||||
t.Fatalf("u1 分组被 u2 影响: %+v", back)
|
||||
}
|
||||
}
|
||||
|
||||
// TestBookGroupsSaveTwiceReplacesWholeSet 整份替换语义:第二次保存覆盖第一次。
|
||||
func TestBookGroupsSaveTwiceReplacesWholeSet(t *testing.T) {
|
||||
svc := newReaderHandlerContainer(t)
|
||||
router := profileRouterForTest(t, svc)
|
||||
|
||||
b1 := addBookForGroups(t, svc, "u1", "书一", "https://example.com/1")
|
||||
|
||||
if _, code := setBookGroups(t, router, "u1", []model.BookGroupSet{
|
||||
{Name: "旧组", BookIDs: []string{b1.ID}},
|
||||
}); code != http.StatusOK {
|
||||
t.Fatalf("首次保存 status=%d", code)
|
||||
}
|
||||
got, code := setBookGroups(t, router, "u1", []model.BookGroupSet{
|
||||
{Name: "新组", BookIDs: []string{}},
|
||||
})
|
||||
if code != http.StatusOK {
|
||||
t.Fatalf("二次保存 status=%d", code)
|
||||
}
|
||||
if names := namesOf(got); names != "新组" {
|
||||
t.Fatalf("二次保存后 = %q,期望只有 新组", names)
|
||||
}
|
||||
|
||||
// 清空分组
|
||||
if groups, code := setBookGroups(t, router, "u1", nil); code != http.StatusOK || len(groups) != 0 {
|
||||
t.Fatalf("清空后 = %+v (status=%d),期望空", groups, code)
|
||||
}
|
||||
if groups, _ := getBookGroups(t, router, "u1"); len(groups) != 0 {
|
||||
t.Fatalf("清空后读回 = %+v,期望空", groups)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,174 @@
|
||||
package handler
|
||||
|
||||
import (
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"os"
|
||||
"path/filepath"
|
||||
"strings"
|
||||
"testing"
|
||||
|
||||
"github.com/gin-gonic/gin"
|
||||
)
|
||||
|
||||
// TestReaderLocalAudioStream 本地有声书音频流端点:
|
||||
// 签名鉴权 + Range 透传(播放器拖进度靠它)。
|
||||
func TestReaderLocalAudioStream(t *testing.T) {
|
||||
container := newReaderHandlerContainer(t)
|
||||
gin.SetMode(gin.TestMode)
|
||||
router := gin.New()
|
||||
router.GET("/api/reader/local/audio", readerLocalAudioHandler(container))
|
||||
router.HEAD("/api/reader/local/audio", readerLocalAudioHandler(container))
|
||||
|
||||
payload := []byte("0123456789abcdefghij")
|
||||
path := filepath.Join(t.TempDir(), "track.mp3")
|
||||
if err := os.WriteFile(path, payload, 0o640); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
signed := container.Reader.LocalAudioURL("book-1", path)
|
||||
|
||||
// 全量拉取
|
||||
w := httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, signed, nil))
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("全量拉取 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
if w.Body.String() != string(payload) {
|
||||
t.Fatalf("全量拉取 body=%q", w.Body.String())
|
||||
}
|
||||
|
||||
// Range 请求要能拖进度
|
||||
w = httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodGet, signed, nil)
|
||||
req.Header.Set("Range", "bytes=2-5")
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusPartialContent {
|
||||
t.Fatalf("Range 请求 status=%d", w.Code)
|
||||
}
|
||||
if w.Body.String() != "2345" {
|
||||
t.Fatalf("Range 请求 body=%q", w.Body.String())
|
||||
}
|
||||
|
||||
// 篡改签名应被拒绝
|
||||
w = httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, signed+"x", nil))
|
||||
if w.Code != http.StatusForbidden {
|
||||
t.Fatalf("坏签名 status=%d,期望 403", w.Code)
|
||||
}
|
||||
|
||||
// 文件不存在 → 404(而不是 500)
|
||||
missing := container.Reader.LocalAudioURL("book-1", filepath.Join(t.TempDir(), "nope.mp3"))
|
||||
w = httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, missing, nil))
|
||||
if w.Code != http.StatusNotFound {
|
||||
t.Fatalf("缺失文件 status=%d,期望 404", w.Code)
|
||||
}
|
||||
}
|
||||
|
||||
// TestReaderAudioTranscodeRoute 转码端点:签名鉴权,命中缓存后按 Range 下发。
|
||||
// 用预置的缓存文件代替真实转码,避免测试依赖 ffmpeg。
|
||||
func TestReaderAudioTranscodeRoute(t *testing.T) {
|
||||
container := newReaderHandlerContainer(t)
|
||||
container.Cfg.App.DataDir = t.TempDir() // 转码缓存落在临时目录,别写进仓库
|
||||
gin.SetMode(gin.TestMode)
|
||||
router := gin.New()
|
||||
router.GET("/api/reader/audio/transcode", readerAudioTranscodeHandler(container))
|
||||
router.HEAD("/api/reader/audio/transcode", readerAudioTranscodeHandler(container))
|
||||
|
||||
const bookID = "book-1"
|
||||
source := `D:\media\斗破苍穹\001.wma`
|
||||
payload := []byte("ID3FAKEMP3PAYLOAD0123456789")
|
||||
|
||||
cachePath, err := container.Reader.AudioTranscodeCachePath(bookID, source)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if err := os.MkdirAll(filepath.Dir(cachePath), 0o750); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if err := os.WriteFile(cachePath, payload, 0o640); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
signed := container.Reader.AudioTranscodeURL(bookID, source)
|
||||
|
||||
w := httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, signed, nil))
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("全量拉取 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
if w.Body.String() != string(payload) {
|
||||
t.Fatalf("全量拉取 body=%q", w.Body.String())
|
||||
}
|
||||
|
||||
// 拖进度靠 Range
|
||||
w = httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodGet, signed, nil)
|
||||
req.Header.Set("Range", "bytes=3-7")
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusPartialContent {
|
||||
t.Fatalf("Range 请求 status=%d", w.Code)
|
||||
}
|
||||
if w.Body.String() != "FAKEM" {
|
||||
t.Fatalf("Range 请求 body=%q", w.Body.String())
|
||||
}
|
||||
|
||||
// 篡改签名应被拒绝
|
||||
w = httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, signed+"x", nil))
|
||||
if w.Code != http.StatusForbidden {
|
||||
t.Fatalf("坏签名 status=%d,期望 403", w.Code)
|
||||
}
|
||||
|
||||
// 换一本书(签名不匹配)同样拒绝
|
||||
forged := strings.Replace(signed, "b=book-1", "b=book-2", 1)
|
||||
w = httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, forged, nil))
|
||||
if w.Code != http.StatusForbidden {
|
||||
t.Fatalf("换书后 status=%d,期望 403", w.Code)
|
||||
}
|
||||
}
|
||||
|
||||
// TestReaderAudioTranscodeRouteWithoutFFmpeg 没装 ffmpeg 且无缓存时返回明确错误,
|
||||
// 而不是空响应(前端据此提示用户去装 ffmpeg)。
|
||||
func TestReaderAudioTranscodeRouteWithoutFFmpeg(t *testing.T) {
|
||||
container := newReaderHandlerContainer(t)
|
||||
container.Cfg.App.DataDir = t.TempDir()
|
||||
container.Cfg.App.FFmpegPath = filepath.Join(t.TempDir(), "definitely-missing-ffmpeg")
|
||||
gin.SetMode(gin.TestMode)
|
||||
router := gin.New()
|
||||
router.GET("/api/reader/audio/transcode", readerAudioTranscodeHandler(container))
|
||||
|
||||
signed := container.Reader.AudioTranscodeURL("book-1", `D:\media\001.wma`)
|
||||
w := httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, signed, nil))
|
||||
if w.Code != http.StatusBadGateway {
|
||||
t.Fatalf("status=%d,期望 502", w.Code)
|
||||
}
|
||||
if !strings.Contains(w.Body.String(), "ffmpeg") {
|
||||
t.Fatalf("错误正文应提到 ffmpeg: %q", w.Body.String())
|
||||
}
|
||||
}
|
||||
|
||||
// TestReaderLocalAudioRejectsOtherBook 换一本书的签名不通用:防止拿到别人的音频地址。
|
||||
func TestReaderLocalAudioRejectsOtherBook(t *testing.T) {
|
||||
container := newReaderHandlerContainer(t)
|
||||
gin.SetMode(gin.TestMode)
|
||||
router := gin.New()
|
||||
router.GET("/api/reader/local/audio", readerLocalAudioHandler(container))
|
||||
|
||||
path := filepath.Join(t.TempDir(), "track.mp3")
|
||||
if err := os.WriteFile(path, []byte("x"), 0o640); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
signed := container.Reader.LocalAudioURL("book-1", path)
|
||||
forged := strings.Replace(signed, "b=book-1", "b=book-2", 1)
|
||||
if forged == signed {
|
||||
t.Fatal("测试用例未改写 book 参数")
|
||||
}
|
||||
w := httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, forged, nil))
|
||||
if w.Code != http.StatusForbidden {
|
||||
t.Fatalf("换书后签名应失效,status=%d", w.Code)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,307 @@
|
||||
package handler
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"fmt"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"slices"
|
||||
"strings"
|
||||
"testing"
|
||||
|
||||
"github.com/gin-gonic/gin"
|
||||
"github.com/glebarez/sqlite"
|
||||
"go.uber.org/zap"
|
||||
"gorm.io/gorm"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/config"
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
"github.com/truewhile/MeBox/internal/repository"
|
||||
"github.com/truewhile/MeBox/internal/service"
|
||||
"github.com/truewhile/MeBox/internal/service/reader"
|
||||
)
|
||||
|
||||
// 书源登录相关的 HTTP 层测试:路由注册与 JSON 契约。
|
||||
|
||||
func newReaderHandlerContainer(t *testing.T) *service.Container {
|
||||
t.Helper()
|
||||
// 唯一库名 + cache=shared:阅读链路会并发写会话状态,
|
||||
// 而 :memory: 下每个连接各自一个库,并发写入对后续读取不可见。
|
||||
dsn := fmt.Sprintf("file:%s?mode=memory&cache=shared", strings.ReplaceAll(t.Name(), "/", "_"))
|
||||
db, err := gorm.Open(sqlite.Open(dsn), &gorm.Config{})
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if err := db.AutoMigrate(model.AllModels()...); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
repos := repository.New(db)
|
||||
cfg := &config.Config{}
|
||||
cfg.Secrets.JWTSecret = "test-secret"
|
||||
return &service.Container{
|
||||
Repo: repos,
|
||||
Cfg: cfg,
|
||||
Log: zap.NewNop(),
|
||||
Reader: reader.NewReaderService(cfg, zap.NewNop(), repos),
|
||||
}
|
||||
}
|
||||
|
||||
// loginHandlerSource 返回一个声明了 loginUrl/loginUi 的书源 JSON。
|
||||
func loginHandlerSource(server string) string {
|
||||
loginJS := `function login(flag) {
|
||||
var payload = JSON.stringify({register_email: result['邮箱'], password: result['密码']});
|
||||
var res = java.ajax(baseUrl + '/login_api,{"method":"POST","headers":{"Content-Type":"application/json"},"body":' + JSON.stringify(payload) + '}');
|
||||
var data = JSON.parse(res);
|
||||
if (data.code == 0) { cookie.setCookie(baseUrl, 'qttoken=' + data.key); java.toast('登录成功'); return true; }
|
||||
java.toast(data.msg || '登录失败'); return false;
|
||||
}`
|
||||
ui, _ := json.Marshal([]map[string]any{
|
||||
{"name": "邮箱", "type": "text"},
|
||||
{"name": "密码", "type": "password"},
|
||||
{"name": "登录", "type": "button", "action": "login(true)"},
|
||||
})
|
||||
src := map[string]any{
|
||||
"bookSourceUrl": server,
|
||||
"bookSourceName": "登录源",
|
||||
"bookSourceType": 0,
|
||||
"loginUrl": loginJS,
|
||||
"loginUi": string(ui),
|
||||
"searchUrl": server + "/search",
|
||||
"ruleSearch": map[string]any{
|
||||
"bookList": "class.item",
|
||||
"name": "tag.h3@tag.a@text",
|
||||
"bookUrl": "tag.h3@tag.a@href",
|
||||
},
|
||||
}
|
||||
b, _ := json.Marshal(src)
|
||||
return string(b)
|
||||
}
|
||||
|
||||
// loginHandlerServer 模拟需要登录的站点。
|
||||
func loginHandlerServer(t *testing.T) *httptest.Server {
|
||||
t.Helper()
|
||||
return httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
switch r.URL.Path {
|
||||
case "/login_api":
|
||||
var body map[string]string
|
||||
_ = json.NewDecoder(r.Body).Decode(&body)
|
||||
if body["register_email"] != "u@e.com" || body["password"] != "pw" {
|
||||
_, _ = w.Write([]byte(`{"code":1,"msg":"账号或密码错误"}`))
|
||||
return
|
||||
}
|
||||
http.SetCookie(w, &http.Cookie{Name: "qttoken", Value: "HTOKEN_1234567890", Path: "/"})
|
||||
_, _ = w.Write([]byte(`{"code":0,"key":"HTOKEN_1234567890"}`))
|
||||
case "/search":
|
||||
if !strings.Contains(r.Header.Get("Cookie"), "qttoken=HTOKEN_1234567890") {
|
||||
w.WriteHeader(http.StatusUnauthorized)
|
||||
_, _ = w.Write([]byte(`<html>未登录</html>`))
|
||||
return
|
||||
}
|
||||
_, _ = w.Write([]byte(`<html><body><div class="item"><h3><a href="/book/1">登录后可见</a></h3></div></body></html>`))
|
||||
default:
|
||||
http.NotFound(w, r)
|
||||
}
|
||||
}))
|
||||
}
|
||||
|
||||
// registerReaderRoutesForTest 挂载阅读路由(跳过鉴权中间件)。
|
||||
func registerReaderRoutesForTest(cfg *service.Container) *gin.Engine {
|
||||
gin.SetMode(gin.TestMode)
|
||||
r := gin.New()
|
||||
registerReaderRoutes(r.Group("/api"), cfg)
|
||||
return r
|
||||
}
|
||||
|
||||
// TestReaderLoginRoutesEndToEnd 走完整 HTTP 路由:
|
||||
// 查看登录信息 → 执行登录 → 搜索携带 Cookie 成功。
|
||||
func TestReaderLoginRoutesEndToEnd(t *testing.T) {
|
||||
srv := loginHandlerServer(t)
|
||||
defer srv.Close()
|
||||
cfg := newReaderHandlerContainer(t)
|
||||
router := registerReaderRoutesForTest(cfg)
|
||||
|
||||
// 导入书源
|
||||
body, _ := json.Marshal(map[string]string{"text": loginHandlerSource(srv.URL)})
|
||||
w := httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodPost, "/api/reader/sources/import", strings.NewReader(string(body)))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("导入书源 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
|
||||
// 列表应标注 has_login
|
||||
w = httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, "/api/reader/sources", nil))
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("书源列表 status=%d", w.Code)
|
||||
}
|
||||
var listResp struct {
|
||||
Sources []struct {
|
||||
ID string `json:"id"`
|
||||
HasLogin bool `json:"has_login"`
|
||||
} `json:"sources"`
|
||||
}
|
||||
if err := json.Unmarshal(w.Body.Bytes(), &listResp); err != nil || len(listResp.Sources) != 1 {
|
||||
t.Fatalf("书源列表解析失败: %s", w.Body.String())
|
||||
}
|
||||
if !listResp.Sources[0].HasLogin {
|
||||
t.Fatal("has_login 应为 true")
|
||||
}
|
||||
sourceID := listResp.Sources[0].ID
|
||||
|
||||
// GET 登录信息:应返回表单字段与未登录状态
|
||||
w = httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, "/api/reader/sources/"+sourceID+"/login", nil))
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("登录信息 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
var info struct {
|
||||
HasLoginJS bool `json:"has_login_js"`
|
||||
LoggedIn bool `json:"logged_in"`
|
||||
Fields []struct {
|
||||
Name string `json:"name"`
|
||||
Type string `json:"type"`
|
||||
} `json:"fields"`
|
||||
}
|
||||
if err := json.Unmarshal(w.Body.Bytes(), &info); err != nil {
|
||||
t.Fatalf("登录信息解析失败: %s", w.Body.String())
|
||||
}
|
||||
if !info.HasLoginJS || info.LoggedIn {
|
||||
t.Fatalf("初始状态异常: %+v", info)
|
||||
}
|
||||
var names []string
|
||||
for _, f := range info.Fields {
|
||||
names = append(names, f.Name)
|
||||
}
|
||||
for _, want := range []string{"邮箱", "密码", "登录"} {
|
||||
if !slices.Contains(names, want) {
|
||||
t.Fatalf("缺少表单字段 %q: %v", want, names)
|
||||
}
|
||||
}
|
||||
|
||||
// 密码错误 → ok=false,提示透传
|
||||
loginBody, _ := json.Marshal(map[string]any{
|
||||
"fields": map[string]string{"邮箱": "u@e.com", "密码": "bad"},
|
||||
})
|
||||
w = httptest.NewRecorder()
|
||||
req = httptest.NewRequest(http.MethodPost, "/api/reader/sources/"+sourceID+"/login", strings.NewReader(string(loginBody)))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("登录 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
var failRes struct {
|
||||
OK bool `json:"ok"`
|
||||
LoggedIn bool `json:"logged_in"`
|
||||
Toasts []string `json:"toasts"`
|
||||
}
|
||||
_ = json.Unmarshal(w.Body.Bytes(), &failRes)
|
||||
if failRes.LoggedIn {
|
||||
t.Fatalf("密码错误不应登录成功: %s", w.Body.String())
|
||||
}
|
||||
if !strings.Contains(strings.Join(failRes.Toasts, " "), "密码错误") {
|
||||
t.Fatalf("失败提示未透传: %s", w.Body.String())
|
||||
}
|
||||
|
||||
// 正确密码 → ok=true 且已登录
|
||||
loginBody, _ = json.Marshal(map[string]any{
|
||||
"fields": map[string]string{"邮箱": "u@e.com", "密码": "pw"},
|
||||
})
|
||||
w = httptest.NewRecorder()
|
||||
req = httptest.NewRequest(http.MethodPost, "/api/reader/sources/"+sourceID+"/login", strings.NewReader(string(loginBody)))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("登录 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
var okRes struct {
|
||||
OK bool `json:"ok"`
|
||||
LoggedIn bool `json:"logged_in"`
|
||||
Cookies map[string]string `json:"cookies"`
|
||||
}
|
||||
if err := json.Unmarshal(w.Body.Bytes(), &okRes); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if !okRes.OK || !okRes.LoggedIn || len(okRes.Cookies) == 0 {
|
||||
t.Fatalf("登录应成功并带 Cookie: %s", w.Body.String())
|
||||
}
|
||||
|
||||
// 搜索应携带 Cookie 并拿到结果
|
||||
w = httptest.NewRecorder()
|
||||
req = httptest.NewRequest(http.MethodPost, "/api/reader/search", strings.NewReader(`{"key":"任意"}`))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
router.ServeHTTP(w, req)
|
||||
var searchRes struct {
|
||||
Books []struct{ Name string } `json:"books"`
|
||||
Skipped []struct {
|
||||
Reason string `json:"reason"`
|
||||
} `json:"skipped"`
|
||||
}
|
||||
_ = json.Unmarshal(w.Body.Bytes(), &searchRes)
|
||||
if len(searchRes.Books) != 1 || searchRes.Books[0].Name != "登录后可见" {
|
||||
t.Fatalf("登录后搜索应成功: %s", w.Body.String())
|
||||
}
|
||||
|
||||
// 登出 → 登录态清空
|
||||
w = httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodDelete, "/api/reader/sources/"+sourceID+"/login", nil))
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("登出 status=%d", w.Code)
|
||||
}
|
||||
w = httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, "/api/reader/sources/"+sourceID+"/login", nil))
|
||||
var afterLogout struct {
|
||||
LoggedIn bool `json:"logged_in"`
|
||||
Cookies map[string]string `json:"cookies"`
|
||||
}
|
||||
_ = json.Unmarshal(w.Body.Bytes(), &afterLogout)
|
||||
if afterLogout.LoggedIn || len(afterLogout.Cookies) != 0 {
|
||||
t.Fatalf("登出后应无登录态: %s", w.Body.String())
|
||||
}
|
||||
}
|
||||
|
||||
// TestReaderSetSourceVariableRoute 源变量接口应保存合法 JSON 并拒绝非法 JSON。
|
||||
func TestReaderSetSourceVariableRoute(t *testing.T) {
|
||||
cfg := newReaderHandlerContainer(t)
|
||||
router := registerReaderRoutesForTest(cfg)
|
||||
|
||||
body, _ := json.Marshal(map[string]string{"text": `{"bookSourceUrl":"https://v.example.com","bookSourceName":"变量源"}`})
|
||||
w := httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodPost, "/api/reader/sources/import", strings.NewReader(string(body)))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
router.ServeHTTP(w, req)
|
||||
|
||||
w = httptest.NewRecorder()
|
||||
router.ServeHTTP(w, httptest.NewRequest(http.MethodGet, "/api/reader/sources", nil))
|
||||
var list struct {
|
||||
Sources []struct {
|
||||
ID string `json:"id"`
|
||||
} `json:"sources"`
|
||||
}
|
||||
_ = json.Unmarshal(w.Body.Bytes(), &list)
|
||||
if len(list.Sources) != 1 {
|
||||
t.Fatalf("应导入 1 个书源: %s", w.Body.String())
|
||||
}
|
||||
id := list.Sources[0].ID
|
||||
|
||||
// 合法 JSON
|
||||
w = httptest.NewRecorder()
|
||||
req = httptest.NewRequest(http.MethodPut, "/api/reader/sources/"+id+"/variable",
|
||||
strings.NewReader(`{"variable":"{\"线路\":\"https://v2.example.com\"}"}`))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("保存变量 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
|
||||
// 非法 JSON 应 400
|
||||
w = httptest.NewRecorder()
|
||||
req = httptest.NewRequest(http.MethodPut, "/api/reader/sources/"+id+"/variable",
|
||||
strings.NewReader(`{"variable":"not-json"}`))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusBadRequest {
|
||||
t.Fatalf("非法变量应 400,实际 %d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,248 @@
|
||||
package handler
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"strings"
|
||||
"testing"
|
||||
|
||||
"github.com/gin-gonic/gin"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/middleware"
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
"github.com/truewhile/MeBox/internal/service"
|
||||
"github.com/truewhile/MeBox/internal/service/reader"
|
||||
)
|
||||
|
||||
// 阅读器偏好接口(/reader/profile)的 HTTP 契约测试。
|
||||
//
|
||||
// 这些设置原先只存在浏览器 localStorage(设备级),换设备就丢。现在按用户落库,
|
||||
// 所以要保证:新用户没有记录、写读一致、越界值收敛、以及用户之间互不影响。
|
||||
|
||||
// profileRouterForTest 用请求头 X-Test-User 指定调用者(默认 u1),
|
||||
// 便于在一个测试里切换用户验证隔离。
|
||||
func profileRouterForTest(t *testing.T, cfg *service.Container) *gin.Engine {
|
||||
t.Helper()
|
||||
gin.SetMode(gin.TestMode)
|
||||
r := gin.New()
|
||||
r.Use(func(c *gin.Context) {
|
||||
user := c.GetHeader("X-Test-User")
|
||||
if user == "" {
|
||||
user = "u1"
|
||||
}
|
||||
c.Set(middleware.CtxUserID, user)
|
||||
c.Next()
|
||||
})
|
||||
registerReaderRoutes(r.Group("/api"), cfg)
|
||||
return r
|
||||
}
|
||||
|
||||
func getProfile(t *testing.T, router *gin.Engine, user string) (*reader.ReaderSettings, int) {
|
||||
t.Helper()
|
||||
w := httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodGet, "/api/reader/profile", nil)
|
||||
req.Header.Set("X-Test-User", user)
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusOK {
|
||||
return nil, w.Code
|
||||
}
|
||||
var res struct {
|
||||
Profile *reader.ReaderSettings `json:"profile"`
|
||||
}
|
||||
if err := json.Unmarshal(w.Body.Bytes(), &res); err != nil {
|
||||
t.Fatalf("解析响应失败: %v body=%s", err, w.Body.String())
|
||||
}
|
||||
return res.Profile, w.Code
|
||||
}
|
||||
|
||||
func putProfile(t *testing.T, router *gin.Engine, user, body string) (*reader.ReaderSettings, int) {
|
||||
t.Helper()
|
||||
w := httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodPut, "/api/reader/profile", strings.NewReader(body))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
req.Header.Set("X-Test-User", user)
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusOK {
|
||||
return nil, w.Code
|
||||
}
|
||||
var res struct {
|
||||
Profile *reader.ReaderSettings `json:"profile"`
|
||||
}
|
||||
if err := json.Unmarshal(w.Body.Bytes(), &res); err != nil {
|
||||
t.Fatalf("解析响应失败: %v body=%s", err, w.Body.String())
|
||||
}
|
||||
return res.Profile, w.Code
|
||||
}
|
||||
|
||||
func TestReaderProfileNullBeforeFirstSave(t *testing.T) {
|
||||
svc := newReaderHandlerContainer(t)
|
||||
router := profileRouterForTest(t, svc)
|
||||
|
||||
profile, code := getProfile(t, router, "u1")
|
||||
if code != http.StatusOK {
|
||||
t.Fatalf("status=%d", code)
|
||||
}
|
||||
// 没有记录时返回 null,由前端用它本地的值播种
|
||||
if profile != nil {
|
||||
t.Fatalf("新用户应返回 null,实际 %+v", profile)
|
||||
}
|
||||
}
|
||||
|
||||
func TestReaderProfileRoundTrip(t *testing.T) {
|
||||
svc := newReaderHandlerContainer(t)
|
||||
router := profileRouterForTest(t, svc)
|
||||
|
||||
body := `{
|
||||
"theme_id":"preset2","night":true,"page_mode":"scroll",
|
||||
"font_size":24,"line_height":2.1,"paragraph_spacing":12,
|
||||
"audio_speed":1.5,"audio_timer_minutes":30,
|
||||
"shelf_layout":"compact","shelf_grid_columns":4,"shelf_sort":"update",
|
||||
"shelf_show_unread":false,"shelf_show_update_time":true
|
||||
}`
|
||||
saved, code := putProfile(t, router, "u1", body)
|
||||
if code != http.StatusOK {
|
||||
t.Fatalf("保存失败 status=%d", code)
|
||||
}
|
||||
if saved == nil {
|
||||
t.Fatal("保存应返回偏好")
|
||||
}
|
||||
|
||||
got, code := getProfile(t, router, "u1")
|
||||
if code != http.StatusOK || got == nil {
|
||||
t.Fatalf("读回失败 status=%d profile=%+v", code, got)
|
||||
}
|
||||
if *got != *saved {
|
||||
t.Fatalf("读回与写入不一致:\n写入 %+v\n读回 %+v", *saved, *got)
|
||||
}
|
||||
if got.ThemeID != "preset2" || !got.Night || got.PageMode != "scroll" || got.FontSize != 24 ||
|
||||
got.LineHeight != 2.1 || got.ParagraphSpacing != 12 || got.AudioSpeed != 1.5 ||
|
||||
got.AudioTimerMinutes != 30 || got.ShelfLayout != "compact" || got.ShelfGridColumns != 4 ||
|
||||
got.ShelfSort != "update" || got.ShelfShowUnread || !got.ShelfShowUpdateTime {
|
||||
t.Fatalf("字段未按原值往返: %+v", *got)
|
||||
}
|
||||
}
|
||||
|
||||
func TestReaderProfileClampsAndFallsBack(t *testing.T) {
|
||||
svc := newReaderHandlerContainer(t)
|
||||
router := profileRouterForTest(t, svc)
|
||||
|
||||
// 全部越界 / 枚举非法,服务端应收敛而不是原样入库
|
||||
body := `{
|
||||
"theme_id":" ","night":false,"page_mode":"nope",
|
||||
"font_size":999,"line_height":0.1,"paragraph_spacing":-5,
|
||||
"audio_speed":99,"audio_timer_minutes":9999,
|
||||
"shelf_layout":"diagonal","shelf_grid_columns":42,"shelf_sort":"random",
|
||||
"shelf_show_unread":true,"shelf_show_update_time":false
|
||||
}`
|
||||
got, code := putProfile(t, router, "u1", body)
|
||||
if code != http.StatusOK {
|
||||
t.Fatalf("保存失败 status=%d", code)
|
||||
}
|
||||
if got.ThemeID != "preset1" {
|
||||
t.Fatalf("空主题应回落到 preset1,实际 %q", got.ThemeID)
|
||||
}
|
||||
if got.PageMode != "page" {
|
||||
t.Fatalf("非法 page_mode 应回落到 page,实际 %q", got.PageMode)
|
||||
}
|
||||
if got.FontSize != 32 {
|
||||
t.Fatalf("font_size 应夹到上限 32,实际 %d", got.FontSize)
|
||||
}
|
||||
if got.LineHeight != 1.4 {
|
||||
t.Fatalf("line_height 应夹到下限 1.4,实际 %v", got.LineHeight)
|
||||
}
|
||||
if got.ParagraphSpacing != 0 {
|
||||
t.Fatalf("paragraph_spacing 应夹到下限 0,实际 %d", got.ParagraphSpacing)
|
||||
}
|
||||
if got.AudioSpeed != 3 {
|
||||
t.Fatalf("audio_speed 应夹到上限 3,实际 %v", got.AudioSpeed)
|
||||
}
|
||||
if got.AudioTimerMinutes != 180 {
|
||||
t.Fatalf("audio_timer_minutes 应夹到上限 180,实际 %d", got.AudioTimerMinutes)
|
||||
}
|
||||
if got.ShelfLayout != "grid" {
|
||||
t.Fatalf("非法 shelf_layout 应回落到 grid,实际 %q", got.ShelfLayout)
|
||||
}
|
||||
if got.ShelfGridColumns != 6 {
|
||||
t.Fatalf("shelf_grid_columns 应夹到上限 6,实际 %d", got.ShelfGridColumns)
|
||||
}
|
||||
if got.ShelfSort != "recent" {
|
||||
t.Fatalf("非法 shelf_sort 应回落到 recent,实际 %q", got.ShelfSort)
|
||||
}
|
||||
// 布尔值不参与收敛,必须原样保留
|
||||
if !got.ShelfShowUnread || got.ShelfShowUpdateTime {
|
||||
t.Fatalf("布尔项被改动: unread=%v update=%v", got.ShelfShowUnread, got.ShelfShowUpdateTime)
|
||||
}
|
||||
}
|
||||
|
||||
func TestReaderProfileIsPerUser(t *testing.T) {
|
||||
svc := newReaderHandlerContainer(t)
|
||||
router := profileRouterForTest(t, svc)
|
||||
|
||||
if _, code := putProfile(t, router, "u1", `{
|
||||
"theme_id":"preset3","night":true,"page_mode":"page","font_size":18,
|
||||
"line_height":1.5,"paragraph_spacing":4,"audio_speed":1,
|
||||
"audio_timer_minutes":0,"shelf_layout":"list","shelf_grid_columns":0,
|
||||
"shelf_sort":"name","shelf_show_unread":true,"shelf_show_update_time":true
|
||||
}`); code != http.StatusOK {
|
||||
t.Fatalf("u1 保存失败 status=%d", code)
|
||||
}
|
||||
|
||||
// 另一个用户不应看到 u1 的偏好,也不应在 u1 之外建出记录
|
||||
other, code := getProfile(t, router, "u2")
|
||||
if code != http.StatusOK {
|
||||
t.Fatalf("u2 读取 status=%d", code)
|
||||
}
|
||||
if other != nil {
|
||||
t.Fatalf("u2 应返回 null,实际 %+v", other)
|
||||
}
|
||||
|
||||
if _, code := putProfile(t, router, "u2", `{
|
||||
"theme_id":"preset4","night":false,"page_mode":"scroll","font_size":20,
|
||||
"line_height":1.8,"paragraph_spacing":8,"audio_speed":2,
|
||||
"audio_timer_minutes":60,"shelf_layout":"grid","shelf_grid_columns":3,
|
||||
"shelf_sort":"author","shelf_show_unread":false,"shelf_show_update_time":false
|
||||
}`); code != http.StatusOK {
|
||||
t.Fatalf("u2 保存失败 status=%d", code)
|
||||
}
|
||||
|
||||
// u1 的值不能被 u2 覆盖
|
||||
back, code := getProfile(t, router, "u1")
|
||||
if code != http.StatusOK || back == nil {
|
||||
t.Fatalf("u1 读回失败 status=%d", code)
|
||||
}
|
||||
if back.ThemeID != "preset3" || back.ShelfLayout != "list" || back.ShelfSort != "name" || !back.Night {
|
||||
t.Fatalf("u1 的偏好被 u2 覆盖: %+v", *back)
|
||||
}
|
||||
}
|
||||
|
||||
func TestReaderProfileSaveTwiceKeepsOneRow(t *testing.T) {
|
||||
svc := newReaderHandlerContainer(t)
|
||||
router := profileRouterForTest(t, svc)
|
||||
|
||||
first := `{"theme_id":"preset1","page_mode":"page","font_size":20,"line_height":1.8,
|
||||
"paragraph_spacing":8,"audio_speed":1,"shelf_layout":"grid","shelf_sort":"recent"}`
|
||||
if _, code := putProfile(t, router, "u1", first); code != http.StatusOK {
|
||||
t.Fatalf("首次保存 status=%d", code)
|
||||
}
|
||||
second := `{"theme_id":"preset5","page_mode":"scroll","font_size":22,"line_height":2.0,
|
||||
"paragraph_spacing":10,"audio_speed":1.25,"shelf_layout":"compact","shelf_sort":"mixed"}`
|
||||
if _, code := putProfile(t, router, "u1", second); code != http.StatusOK {
|
||||
t.Fatalf("二次保存 status=%d", code)
|
||||
}
|
||||
|
||||
got, _ := getProfile(t, router, "u1")
|
||||
if got == nil || got.ThemeID != "preset5" || got.ShelfSort != "mixed" {
|
||||
t.Fatalf("二次保存未覆盖旧值: %+v", got)
|
||||
}
|
||||
|
||||
// 一个用户只应有一条记录(唯一索引 + upsert)
|
||||
var count int64
|
||||
if err := svc.Repo.DB.Model(&model.ReaderProfile{}).
|
||||
Where("user_id = ?", "u1").Count(&count).Error; err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if count != 1 {
|
||||
t.Fatalf("u1 的记录数 = %d,期望 1", count)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,170 @@
|
||||
package handler
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"strings"
|
||||
"testing"
|
||||
|
||||
"github.com/gin-gonic/gin"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/middleware"
|
||||
"github.com/truewhile/MeBox/internal/service"
|
||||
"github.com/truewhile/MeBox/internal/service/reader"
|
||||
)
|
||||
|
||||
// 阅读进度接口的 HTTP 契约测试。
|
||||
//
|
||||
// 听书(音频)的进度是 audio.currentTime,天然带小数;漫画是图片序号、文本是页码,
|
||||
// 都是整数。同一个接口必须同时接住这两种形态,否则音频进度会被 400 丢掉
|
||||
// (前端是 fire-and-forget,失败不报错),表现为「听书退出后从头开始」。
|
||||
|
||||
func progressRouterForTest(cfg *service.Container, userID string) *gin.Engine {
|
||||
gin.SetMode(gin.TestMode)
|
||||
r := gin.New()
|
||||
r.Use(func(c *gin.Context) {
|
||||
c.Set(middleware.CtxUserID, userID)
|
||||
c.Next()
|
||||
})
|
||||
registerReaderRoutes(r.Group("/api"), cfg)
|
||||
return r
|
||||
}
|
||||
|
||||
func putProgress(t *testing.T, router *gin.Engine, bookID, body string) *httptest.ResponseRecorder {
|
||||
t.Helper()
|
||||
w := httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodPut, "/api/reader/books/"+bookID+"/progress", strings.NewReader(body))
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
router.ServeHTTP(w, req)
|
||||
return w
|
||||
}
|
||||
|
||||
func TestSaveProgressAcceptsFractionalPos(t *testing.T) {
|
||||
svc := newReaderHandlerContainer(t)
|
||||
ctx := t.Context()
|
||||
|
||||
book, err := svc.Reader.AddBook(ctx, "u1", reader.SearchOrigin{
|
||||
Origin: "https://example.com",
|
||||
BookURL: "https://example.com/book/1",
|
||||
}, "有声书", "作者", "")
|
||||
if err != nil {
|
||||
t.Fatalf("加入书架失败: %v", err)
|
||||
}
|
||||
|
||||
router := progressRouterForTest(svc, "u1")
|
||||
// 37.5 秒:音频进度上报的真实形态
|
||||
w := putProgress(t, router, book.ID, `{"chapter_index":2,"pos":37.5,"chapter_title":"第 3 章"}`)
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("音频进度应被接受,实际 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
|
||||
books, err := svc.Reader.ListBooks(ctx, "u1")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if len(books) != 1 {
|
||||
t.Fatalf("书架应有 1 本,实际 %d", len(books))
|
||||
}
|
||||
// 小数秒截断成整数秒落库(dur_chapter_pos 是 int)
|
||||
if books[0].DurChapterPos != 37 {
|
||||
t.Fatalf("进度 = %d,期望 37(37.5 截断)", books[0].DurChapterPos)
|
||||
}
|
||||
if books[0].DurChapterIndex != 2 || books[0].DurChapterTitle != "第 3 章" {
|
||||
t.Fatalf("章节/标题未保存: index=%d title=%q", books[0].DurChapterIndex, books[0].DurChapterTitle)
|
||||
}
|
||||
}
|
||||
|
||||
func TestSaveProgressAcceptsIntegerPos(t *testing.T) {
|
||||
svc := newReaderHandlerContainer(t)
|
||||
ctx := t.Context()
|
||||
|
||||
book, err := svc.Reader.AddBook(ctx, "u1", reader.SearchOrigin{
|
||||
Origin: "https://example.com",
|
||||
BookURL: "https://example.com/book/2",
|
||||
}, "文本", "", "")
|
||||
if err != nil {
|
||||
t.Fatalf("加入书架失败: %v", err)
|
||||
}
|
||||
|
||||
router := progressRouterForTest(svc, "u1")
|
||||
// 文本/漫画仍是整数页码,不能被破坏
|
||||
w := putProgress(t, router, book.ID, `{"chapter_index":7,"pos":12,"chapter_title":"第 8 章"}`)
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("页码进度应被接受,实际 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
|
||||
books, err := svc.Reader.ListBooks(ctx, "u1")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if books[0].DurChapterPos != 12 || books[0].DurChapterIndex != 7 {
|
||||
t.Fatalf("进度 = pos %d / index %d,期望 12 / 7", books[0].DurChapterPos, books[0].DurChapterIndex)
|
||||
}
|
||||
}
|
||||
|
||||
func TestSaveProgressRejectsNegativePos(t *testing.T) {
|
||||
svc := newReaderHandlerContainer(t)
|
||||
ctx := t.Context()
|
||||
|
||||
book, err := svc.Reader.AddBook(ctx, "u1", reader.SearchOrigin{
|
||||
Origin: "https://example.com",
|
||||
BookURL: "https://example.com/book/3",
|
||||
}, "文本", "", "")
|
||||
if err != nil {
|
||||
t.Fatalf("加入书架失败: %v", err)
|
||||
}
|
||||
|
||||
router := progressRouterForTest(svc, "u1")
|
||||
// 负数不是合法进度,落库前应被夹到 0,而不是写入负值
|
||||
w := putProgress(t, router, book.ID, `{"chapter_index":0,"pos":-5,"chapter_title":"第 1 章"}`)
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("负数进度不应 500,实际 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
|
||||
books, err := svc.Reader.ListBooks(ctx, "u1")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if books[0].DurChapterPos != 0 {
|
||||
t.Fatalf("进度 = %d,期望夹到 0", books[0].DurChapterPos)
|
||||
}
|
||||
}
|
||||
|
||||
// TestSaveProgressRoundTripsThroughJSON 防回归:进度写进去要能原样读回来。
|
||||
func TestSaveProgressRoundTripsThroughJSON(t *testing.T) {
|
||||
svc := newReaderHandlerContainer(t)
|
||||
ctx := t.Context()
|
||||
|
||||
book, err := svc.Reader.AddBook(ctx, "u1", reader.SearchOrigin{
|
||||
Origin: "https://example.com",
|
||||
BookURL: "https://example.com/book/4",
|
||||
}, "有声书", "", "")
|
||||
if err != nil {
|
||||
t.Fatalf("加入书架失败: %v", err)
|
||||
}
|
||||
|
||||
router := progressRouterForTest(svc, "u1")
|
||||
if w := putProgress(t, router, book.ID, `{"chapter_index":0,"pos":123.75,"chapter_title":"第 1 章"}`); w.Code != http.StatusOK {
|
||||
t.Fatalf("保存进度失败 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
|
||||
// 走列表接口(前端书架/阅读器恢复进度读的就是它)
|
||||
w := httptest.NewRecorder()
|
||||
req := httptest.NewRequest(http.MethodGet, "/api/reader/books", nil)
|
||||
router.ServeHTTP(w, req)
|
||||
if w.Code != http.StatusOK {
|
||||
t.Fatalf("书架列表 status=%d body=%s", w.Code, w.Body.String())
|
||||
}
|
||||
var res struct {
|
||||
Books []struct {
|
||||
DurChapterPos int `json:"dur_chapter_pos"`
|
||||
} `json:"books"`
|
||||
}
|
||||
if err := json.Unmarshal(w.Body.Bytes(), &res); err != nil {
|
||||
t.Fatalf("解析书架响应失败: %v", err)
|
||||
}
|
||||
if len(res.Books) != 1 || res.Books[0].DurChapterPos != 123 {
|
||||
t.Fatalf("读回的进度 = %+v,期望 123", res.Books)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,942 @@
|
||||
// Package handler — 阅读(legado 书源兼容)子系统路由。
|
||||
package handler
|
||||
|
||||
import (
|
||||
"io"
|
||||
"net/http"
|
||||
"os"
|
||||
"strconv"
|
||||
"strings"
|
||||
|
||||
"github.com/gin-gonic/gin"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/middleware"
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
"github.com/truewhile/MeBox/internal/service"
|
||||
"github.com/truewhile/MeBox/internal/service/reader"
|
||||
)
|
||||
|
||||
func registerReaderRoutes(authed *gin.RouterGroup, svc *service.Container) {
|
||||
g := authed.Group("/reader")
|
||||
|
||||
// 书源管理
|
||||
g.GET("/sources", readerListSourcesHandler(svc))
|
||||
g.POST("/sources/import", readerImportSourcesHandler(svc))
|
||||
g.PATCH("/sources/:id", readerUpdateSourceHandler(svc))
|
||||
g.DELETE("/sources/:id", readerDeleteSourceHandler(svc))
|
||||
g.POST("/sources/:id/debug", readerDebugSourceHandler(svc))
|
||||
|
||||
// 书源登录与源变量(登录类书源必需)
|
||||
g.GET("/sources/:id/login", readerSourceLoginInfoHandler(svc))
|
||||
g.POST("/sources/:id/login", readerSourceLoginActionHandler(svc))
|
||||
g.DELETE("/sources/:id/login", readerSourceLogoutHandler(svc))
|
||||
g.PUT("/sources/:id/variable", readerSetSourceVariableHandler(svc))
|
||||
g.PUT("/sources/:id/login-info", readerSetSourceLoginInfoHandler(svc))
|
||||
|
||||
// 书源 JS 的宿主浏览器(java.startBrowser / startBrowserAwait):
|
||||
// 前端轮询待办 → 在 iframe 里承载页面 → 用户点 √ 回传 DOM。
|
||||
g.GET("/browser/pending", readerBrowserPendingHandler(svc))
|
||||
g.POST("/browser/result", readerBrowserResultHandler(svc))
|
||||
// 页面内的 fetch/XHR 经此转发(iframe 是不透明源,请求带不上书源 Cookie)
|
||||
g.POST("/browser/xhr", readerBrowserXHRHandler(svc))
|
||||
|
||||
// 段评:用宿主浏览器打开评论地址(带书源 Cookie/登录态,对应书源的 showCmt)
|
||||
g.POST("/comments/open", readerOpenCommentHandler(svc))
|
||||
|
||||
// 搜索(多源聚合)
|
||||
g.POST("/search", readerSearchHandler(svc))
|
||||
|
||||
// 详情 / 目录 / 正文
|
||||
g.GET("/book-info", readerBookInfoHandler(svc))
|
||||
g.GET("/toc", readerTocHandler(svc))
|
||||
g.GET("/content", readerContentHandler(svc))
|
||||
|
||||
// 书架
|
||||
g.GET("/books", readerListBooksHandler(svc))
|
||||
g.POST("/books", readerAddBookHandler(svc))
|
||||
// 本地书籍(TXT / EPUB 上传导入)
|
||||
g.POST("/local/books", readerImportLocalBookHandler(svc))
|
||||
// 服务器已有文件/目录导入:原地引用不复制,仅管理员(会读取允许根目录内的文件)
|
||||
g.POST("/local/books/from-path", middleware.AdminRequired(), readerImportLocalBookFromPathHandler(svc))
|
||||
g.POST("/local/audiobooks", middleware.AdminRequired(), readerImportLocalAudioDirHandler(svc))
|
||||
g.DELETE("/books/:id", readerRemoveBookHandler(svc))
|
||||
g.PUT("/books/:id/progress", readerSaveProgressHandler(svc))
|
||||
// 更新目录:重抓书架里全部网络书籍的目录,刷新章节缓存与「最近更新」时间
|
||||
g.POST("/shelf/refresh-toc", readerRefreshBooksTocHandler(svc))
|
||||
|
||||
// 阅读器偏好(每个用户一条):主题 / 排版 / 听书 / 书架展示设置,跨设备同步
|
||||
g.GET("/profile", readerGetProfileHandler(svc))
|
||||
g.PUT("/profile", readerSaveProfileHandler(svc))
|
||||
|
||||
// 书架分组(每个用户一份):组名 → 书籍 ID,仿影视模块的媒体库标签
|
||||
g.GET("/book-groups", readerGetBookGroupsHandler(svc))
|
||||
g.PUT("/book-groups", readerSetBookGroupsHandler(svc))
|
||||
// 换源:把书架里的书切到另一个书源(保留阅读进度,目录缓存按新源重建)
|
||||
g.POST("/books/:id/origin", readerSwitchOriginHandler(svc))
|
||||
g.PUT("/books/:id/audio-config", readerSaveAudioConfigHandler(svc))
|
||||
g.GET("/books/:id/chapters", readerListChaptersHandler(svc))
|
||||
g.POST("/books/:id/chapters", readerReplaceChaptersHandler(svc))
|
||||
g.GET("/books/:id/content", readerBookContentHandler(svc))
|
||||
|
||||
// 替换净化规则
|
||||
g.GET("/replace-rules", readerListReplaceRulesHandler(svc))
|
||||
g.POST("/replace-rules", readerCreateReplaceRuleHandler(svc))
|
||||
g.PATCH("/replace-rules/:id", readerUpdateReplaceRuleHandler(svc))
|
||||
g.DELETE("/replace-rules/:id", readerDeleteReplaceRuleHandler(svc))
|
||||
}
|
||||
|
||||
func readerListSourcesHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
sources, err := svc.Reader.ListSources(c.Request.Context())
|
||||
if err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"sources": sources})
|
||||
}
|
||||
}
|
||||
|
||||
func readerImportSourcesHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
Text string `json:"text" binding:"required"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
imported, err := svc.Reader.ImportSources(c.Request.Context(), body.Text)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"imported": imported})
|
||||
}
|
||||
}
|
||||
|
||||
func readerUpdateSourceHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
Enabled *bool `json:"enabled"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil || body.Enabled == nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": "enabled 字段必填"})
|
||||
return
|
||||
}
|
||||
if err := svc.Reader.UpdateSourceEnabled(c.Request.Context(), c.Param("id"), *body.Enabled); err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"ok": true})
|
||||
}
|
||||
}
|
||||
|
||||
func readerDeleteSourceHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
if err := svc.Reader.DeleteSource(c.Request.Context(), c.Param("id")); err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"ok": true})
|
||||
}
|
||||
}
|
||||
|
||||
func readerDebugSourceHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
Key string `json:"key" binding:"required"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
logs, err := svc.Reader.Debug(c.Request.Context(), c.Param("id"), body.Key)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"logs": logs})
|
||||
}
|
||||
}
|
||||
|
||||
func readerSourceLoginInfoHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
info, err := svc.Reader.GetSourceLogin(c.Request.Context(), userID, c.Param("id"))
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, info)
|
||||
}
|
||||
}
|
||||
|
||||
func readerSourceLoginActionHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
// Action 为 loginUi 里按钮的 action;留空表示执行 login()(确认登录)。
|
||||
Action string `json:"action"`
|
||||
Fields map[string]string `json:"fields"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
res, err := svc.Reader.RunLoginAction(c.Request.Context(), userID, c.Param("id"), body.Action, body.Fields)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, res)
|
||||
}
|
||||
}
|
||||
|
||||
// readerBrowserPendingHandler 前端轮询:该用户在某书源下待用户完成的页面。
|
||||
// 书源的 java.startBrowserAwait 会阻塞在服务端,前端据此把页面呈现出来。
|
||||
func readerBrowserPendingHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
pages := svc.Reader.PendingBrowserPages(userID, c.Query("source_id"))
|
||||
c.JSON(http.StatusOK, gin.H{"pages": pages})
|
||||
}
|
||||
}
|
||||
|
||||
// readerBrowserResultHandler 用户完成页面后回传 DOM(或取消),解除服务端阻塞。
|
||||
func readerBrowserResultHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
ID string `json:"id" binding:"required"`
|
||||
Body string `json:"body"`
|
||||
URL string `json:"url"`
|
||||
Cancelled bool `json:"cancelled"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
if err := svc.Reader.ResolveBrowser(body.ID, userID, body.Body, body.URL, body.Cancelled); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"ok": true})
|
||||
}
|
||||
}
|
||||
|
||||
// readerBrowserXHRHandler 转发承载页面内的接口请求。
|
||||
//
|
||||
// 页面在 iframe 里是不透明源,自己的 XHR 既带不上书源 Cookie 也会被 CORS 拦,
|
||||
// 所以由父窗口(持 JWT)把请求转交到这里,服务端补上书源凭据再发。
|
||||
func readerBrowserXHRHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
ID string `json:"id" binding:"required"`
|
||||
URL string `json:"url" binding:"required"`
|
||||
Method string `json:"method"`
|
||||
Headers map[string]string `json:"headers"`
|
||||
Body string `json:"body"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
res, err := svc.Reader.ProxyBrowserXHR(
|
||||
c.Request.Context(), body.ID, body.Method, body.URL, body.Headers, body.Body)
|
||||
if err != nil {
|
||||
// 交给页面自己处理失败,别把 4xx 泄成框架错误
|
||||
c.JSON(http.StatusBadGateway, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, res)
|
||||
}
|
||||
}
|
||||
|
||||
// readerOpenCommentHandler 打开一条段评:复用书源登录态在宿主浏览器里承载评论页。
|
||||
// 书源的 showCmt 内部就是「java.ajax 取评论页 → java.showBrowser 展示」,
|
||||
// 这里用承载登录页的同一套机制等价实现(见 reader.OpenContentComment)。
|
||||
func readerOpenCommentHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
BookID string `json:"book_id" binding:"required"`
|
||||
URL string `json:"url" binding:"required"`
|
||||
Title string `json:"title"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
page, err := svc.Reader.OpenContentComment(c.Request.Context(), userID, body.BookID, body.URL, body.Title)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, page)
|
||||
}
|
||||
}
|
||||
|
||||
// readerBrowserPageHandler 承载待办页面本体。//
|
||||
// 鉴权走 HMAC 签名而非 JWT:这个地址要填进 <iframe src>,而 iframe 的请求
|
||||
// 带不上 Authorization 头。签名绑定待办 ID,链接随待办一起过期。
|
||||
func readerBrowserPageHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
snap, err := svc.Reader.VerifyBrowserPage(c.Query("id"), c.Query("s"))
|
||||
if err != nil {
|
||||
c.String(http.StatusForbidden, "%s", err.Error())
|
||||
return
|
||||
}
|
||||
c.Header("Cache-Control", "no-store")
|
||||
c.Data(http.StatusOK, "text/html; charset=utf-8", []byte(snap.HTML))
|
||||
}
|
||||
}
|
||||
|
||||
// readerBrowserAssetHandler 页面资源/表单/站内链接的同源代理。
|
||||
// 服务端补上书源 Cookie 与请求头,使「用户后台」这类页面在 iframe 里保持登录态。
|
||||
func readerBrowserAssetHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
target, _, err := svc.Reader.VerifyBrowserAsset(c.Query("id"), c.Query("u"), c.Query("s"))
|
||||
if err != nil {
|
||||
c.String(http.StatusForbidden, "%s", err.Error())
|
||||
return
|
||||
}
|
||||
contentType, status, data, err := svc.Reader.FetchBrowserAsset(c.Request.Context(), c.Query("id"), target)
|
||||
if err != nil {
|
||||
c.String(http.StatusBadGateway, "资源加载失败: %s", err.Error())
|
||||
return
|
||||
}
|
||||
if strings.Contains(strings.ToLower(contentType), "text/css") {
|
||||
data = []byte(svc.Reader.RewriteBrowserCSS(string(data), target, c.Query("id")))
|
||||
}
|
||||
c.Header("Cache-Control", "no-store")
|
||||
c.Data(status, contentType, data)
|
||||
}
|
||||
}
|
||||
|
||||
func readerSourceLogoutHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
if err := svc.Reader.ClearSourceLogin(c.Request.Context(), c.Param("id")); err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"ok": true})
|
||||
}
|
||||
}
|
||||
|
||||
func readerSetSourceVariableHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
Variable string `json:"variable"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
if err := svc.Reader.SetSourceVariable(c.Request.Context(), c.Param("id"), body.Variable); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"ok": true})
|
||||
}
|
||||
}
|
||||
|
||||
func readerSetSourceLoginInfoHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
Fields map[string]string `json:"fields"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
if err := svc.Reader.SetSourceLoginInfo(c.Request.Context(), c.Param("id"), body.Fields); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"ok": true})
|
||||
}
|
||||
}
|
||||
|
||||
func readerSearchHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
Key string `json:"key" binding:"required"`
|
||||
// SourceIDs 搜索范围:空表示全部启用书源(默认),非空则只搜这些书源。
|
||||
SourceIDs []string `json:"source_ids"`
|
||||
// Page 页码(从 1 开始):书源 searchUrl 支持 {{page}} 时,前端滚到底逐页取。
|
||||
Page int `json:"page"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
books, skipped, err := svc.Reader.Search(c.Request.Context(), body.Key, body.SourceIDs, body.Page)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"books": books, "skipped": skipped})
|
||||
}
|
||||
}
|
||||
|
||||
func readerBookInfoHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
info, err := svc.Reader.GetBookInfo(
|
||||
c.Request.Context(),
|
||||
c.Query("source_id"), c.Query("source_url"), c.Query("book_url"),
|
||||
)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, info)
|
||||
}
|
||||
}
|
||||
|
||||
func readerTocHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
chapters, err := svc.Reader.GetToc(
|
||||
c.Request.Context(),
|
||||
userID, c.Query("source_id"), c.Query("source_url"), c.Query("book_url"), c.Query("toc_url"),
|
||||
)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
// nil 切片会被编码成 null,前端拿到 null 再取 .length 就是一句
|
||||
// 「Cannot read properties of null」。空目录统一给 []。
|
||||
if chapters == nil {
|
||||
chapters = []reader.TocChapter{}
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"chapters": chapters})
|
||||
}
|
||||
}
|
||||
|
||||
func readerContentHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
content, err := svc.Reader.GetContent(
|
||||
c.Request.Context(),
|
||||
c.Query("source_id"), c.Query("source_url"), c.Query("book_url"), c.Query("chapter_url"),
|
||||
)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, content)
|
||||
}
|
||||
}
|
||||
|
||||
func readerListBooksHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
books, err := svc.Reader.ListBooks(c.Request.Context(), userID)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"books": books})
|
||||
}
|
||||
}
|
||||
|
||||
func readerAddBookHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
Origin reader.SearchOrigin `json:"origin" binding:"required"`
|
||||
Name string `json:"name" binding:"required"`
|
||||
Author string `json:"author"`
|
||||
CoverURL string `json:"cover_url"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
book, err := svc.Reader.AddBook(c.Request.Context(), userID, body.Origin, body.Name, body.Author, body.CoverURL)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
// 后台补目录缓存,让书架能显示未读章数;失败不影响加入书架本身。
|
||||
svc.Reader.WarmUpBookChaptersAsync(c.Request.Context(), userID, book)
|
||||
c.JSON(http.StatusOK, book)
|
||||
}
|
||||
}
|
||||
|
||||
// readerSwitchOriginHandler 换源:把书架里的书切到另一个书源。
|
||||
// 阅读进度保留;旧源目录缓存清空后按新源后台重新预热。
|
||||
func readerSwitchOriginHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
Origin reader.SearchOrigin `json:"origin" binding:"required"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
book, err := svc.Reader.SwitchOrigin(c.Request.Context(), userID, c.Param("id"), body.Origin)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
svc.Reader.WarmUpBookChaptersAsync(c.Request.Context(), userID, book)
|
||||
c.JSON(http.StatusOK, book)
|
||||
}
|
||||
}
|
||||
|
||||
// readerImportLocalBookHandler 上传本地书籍(TXT / EPUB)并加入书架。
|
||||
// 正文落盘到 data/reader/local,目录切分后与网络书籍共用阅读器链路。
|
||||
func readerImportLocalBookHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
// 先卡住请求体大小,避免超大文件把内存打满(多给 1MB 放 multipart 头)
|
||||
c.Request.Body = http.MaxBytesReader(c.Writer, c.Request.Body, int64(reader.LocalBookMaxBytes)+(1<<20))
|
||||
header, err := c.FormFile("file")
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": "缺少上传文件(表单字段 file),或文件超过大小上限"})
|
||||
return
|
||||
}
|
||||
f, err := header.Open()
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": "读取上传文件失败: " + err.Error()})
|
||||
return
|
||||
}
|
||||
defer f.Close()
|
||||
data, err := io.ReadAll(io.LimitReader(f, int64(reader.LocalBookMaxBytes)+1))
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": "读取上传文件失败: " + err.Error()})
|
||||
return
|
||||
}
|
||||
if len(data) > reader.LocalBookMaxBytes {
|
||||
c.JSON(http.StatusRequestEntityTooLarge, gin.H{"error": "文件超过大小上限"})
|
||||
return
|
||||
}
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
book, err := svc.Reader.ImportLocalBook(c.Request.Context(), userID, header.Filename, data)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, book)
|
||||
}
|
||||
}
|
||||
|
||||
// readerImportLocalBookFromPathHandler 从服务器已有文件导入书籍(TXT / EPUB),原地引用。
|
||||
// 仅管理员:会读取服务器上允许根目录内的任意文件。
|
||||
func readerImportLocalBookFromPathHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
Path string `json:"path" binding:"required"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
abs, err := svc.FileManager.ResolvePath(body.Path)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusForbidden, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
book, err := svc.Reader.ImportLocalBookFromPath(c.Request.Context(), userID, abs)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, book)
|
||||
}
|
||||
}
|
||||
|
||||
// readerImportLocalAudioDirHandler 把一个服务器目录导入为一本有声书,原地引用。
|
||||
// 仅管理员;目录下的音频文件与 .strm 播放指针按相对路径排序成为章节。
|
||||
func readerImportLocalAudioDirHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
Path string `json:"path" binding:"required"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
abs, err := svc.FileManager.ResolvePath(body.Path)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusForbidden, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
book, err := svc.Reader.ImportLocalAudioDir(c.Request.Context(), userID, abs)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, book)
|
||||
}
|
||||
}
|
||||
|
||||
// readerRefreshBooksTocHandler 更新目录:重抓书架里全部网络书籍的目录。
|
||||
// 本地书籍与没有书源信息的书籍跳过;单本失败只计数,不影响其它书。
|
||||
func readerRefreshBooksTocHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
res, err := svc.Reader.RefreshBooksToc(c.Request.Context(), userID)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, res)
|
||||
}
|
||||
}
|
||||
|
||||
func readerRemoveBookHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
if err := svc.Reader.RemoveBook(c.Request.Context(), userID, c.Param("id")); err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"ok": true})
|
||||
}
|
||||
}
|
||||
|
||||
// readerGetProfileHandler 读当前用户的阅读器偏好。
|
||||
// 没保存过时返回 {"profile": null},前端据此用本地值播种。
|
||||
func readerGetProfileHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
settings, err := svc.Reader.GetReaderSettings(c.Request.Context(), userID)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"profile": settings})
|
||||
}
|
||||
}
|
||||
|
||||
// readerSaveProfileHandler 覆盖保存当前用户的阅读器偏好(服务端做范围收敛)。
|
||||
func readerSaveProfileHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body reader.ReaderSettings
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
saved, err := svc.Reader.SaveReaderSettings(c.Request.Context(), userID, body)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"profile": saved})
|
||||
}
|
||||
}
|
||||
|
||||
// readerGetBookGroupsHandler 读当前用户的书架分组。
|
||||
// 没有分组时返回空数组(不是 null),前端可以直接遍历。
|
||||
func readerGetBookGroupsHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
groups, err := svc.Reader.GetBookGroups(c.Request.Context(), userID)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"groups": groups})
|
||||
}
|
||||
}
|
||||
|
||||
// readerSetBookGroupsHandler 覆盖保存当前用户的书架分组(整份替换)。
|
||||
func readerSetBookGroupsHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
Groups []model.BookGroupSet `json:"groups"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
groups, err := svc.Reader.SetBookGroups(c.Request.Context(), userID, body.Groups)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"groups": groups})
|
||||
}
|
||||
}
|
||||
|
||||
// maxReaderChapterPos 进度上限:秒(听书)/ 页码(文本)/ 图片序号(漫画)都远小于它,
|
||||
// 只用来挡住异常大的浮点数转 int 时溢出。约 115 天,足够覆盖任何单章。
|
||||
const maxReaderChapterPos = 10_000_000
|
||||
|
||||
func readerSaveProgressHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
ChapterIndex int `json:"chapter_index"`
|
||||
// Pos 用 float64 接:听书的进度是 audio.currentTime(秒,天然带小数),
|
||||
// 漫画是图片序号、文本是页码(都是整数)。用 int 接小数会让整个请求
|
||||
// 400,而前端是 fire-and-forget,音频进度就被静默丢掉了。
|
||||
Pos float64 `json:"pos"`
|
||||
ChapterTitle string `json:"chapter_title"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
// dur_chapter_pos 落库是 int,截断成秒;负数没有意义,夹到 0。
|
||||
posFloat := body.Pos
|
||||
if !(posFloat > 0) { // 同时挡住 0 与负数(JSON 不会给出 NaN)
|
||||
posFloat = 0
|
||||
}
|
||||
if posFloat > maxReaderChapterPos {
|
||||
posFloat = maxReaderChapterPos
|
||||
}
|
||||
pos := int(posFloat)
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
if err := svc.Reader.SaveProgress(c.Request.Context(), userID, c.Param("id"), body.ChapterIndex, pos, body.ChapterTitle); err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"ok": true})
|
||||
}
|
||||
}
|
||||
|
||||
// readerSaveAudioConfigHandler 保存听书片头/片尾跳过秒数(0 为不跳过)。
|
||||
func readerSaveAudioConfigHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
OpenCredits int `json:"open_credits"`
|
||||
CloseCredits int `json:"close_credits"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
if err := svc.Reader.SaveAudioConfig(
|
||||
c.Request.Context(), userID, c.Param("id"), body.OpenCredits, body.CloseCredits); err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"ok": true})
|
||||
}
|
||||
}
|
||||
|
||||
func readerListChaptersHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
chapters, err := svc.Reader.ListChapters(c.Request.Context(), c.Param("id"))
|
||||
if err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"chapters": chapters})
|
||||
}
|
||||
}
|
||||
|
||||
func readerReplaceChaptersHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body struct {
|
||||
Chapters []reader.ChapterInput `json:"chapters"`
|
||||
}
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
if err := svc.Reader.SaveChapters(c.Request.Context(), c.Param("id"), body.Chapters); err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"ok": true})
|
||||
}
|
||||
}
|
||||
|
||||
func readerListReplaceRulesHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
rules, err := svc.Reader.ListReplaceRules(c.Request.Context(), userID)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"rules": rules})
|
||||
}
|
||||
}
|
||||
|
||||
func readerCreateReplaceRuleHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body reader.ReplaceRuleInput
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
rule, err := svc.Reader.CreateReplaceRule(c.Request.Context(), userID, body)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, rule)
|
||||
}
|
||||
}
|
||||
|
||||
func readerUpdateReplaceRuleHandler(svc *service.Container) gin.HandlerFunc {
|
||||
var body reader.ReplaceRuleInput
|
||||
return func(c *gin.Context) {
|
||||
if err := c.ShouldBindJSON(&body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
if err := svc.Reader.UpdateReplaceRule(c.Request.Context(), userID, c.Param("id"), body); err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"ok": true})
|
||||
}
|
||||
}
|
||||
|
||||
func readerDeleteReplaceRuleHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
if err := svc.Reader.DeleteReplaceRule(c.Request.Context(), userID, c.Param("id")); err != nil {
|
||||
c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, gin.H{"ok": true})
|
||||
}
|
||||
}
|
||||
|
||||
// readerMediaProxyHandler 音频流/漫画图片签名代理:
|
||||
// 校验 HMAC 签名 → 携书源防盗链头拉取 → Range 透传(音频拖动)/ m3u8 重写。
|
||||
func readerMediaProxyHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
bookID := c.Query("b")
|
||||
rawURL, err := svc.Reader.VerifyProxyURL(bookID, c.Query("u"), c.Query("s"))
|
||||
if err != nil {
|
||||
c.JSON(http.StatusForbidden, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
if !strings.HasPrefix(rawURL, "http://") && !strings.HasPrefix(rawURL, "https://") {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": "仅支持 http(s) 媒体地址"})
|
||||
return
|
||||
}
|
||||
book, err := svc.Reader.GetBook(c.Request.Context(), bookID)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusNotFound, gin.H{"error": "书籍不存在"})
|
||||
return
|
||||
}
|
||||
resp, err := svc.Reader.FetchMedia(c.Request.Context(), book, rawURL, c.GetHeader("Range"))
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadGateway, gin.H{"error": "媒体拉取失败: " + err.Error()})
|
||||
return
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
|
||||
ct := resp.Header.Get("Content-Type")
|
||||
isPlaylist := strings.Contains(ct, "mpegurl") || strings.Contains(ct, "m3u8") ||
|
||||
strings.HasSuffix(strings.ToLower(rawURL), ".m3u8")
|
||||
if isPlaylist {
|
||||
// m3u8:改写分片/密钥地址为签名代理后返回,hls.js 无感续播
|
||||
data, _ := io.ReadAll(io.LimitReader(resp.Body, 4<<20))
|
||||
rewritten := svc.Reader.RewritePlaylist(bookID, rawURL, string(data))
|
||||
c.Data(http.StatusOK, "application/vnd.apple.mpegurl", []byte(rewritten))
|
||||
return
|
||||
}
|
||||
// 流式透传(含 206 Partial Content,支持音频拖动进度)
|
||||
for _, h := range []string{"Content-Type", "Content-Length", "Content-Range", "Accept-Ranges"} {
|
||||
if v := resp.Header.Get(h); v != "" {
|
||||
c.Header(h, v)
|
||||
}
|
||||
}
|
||||
c.Status(resp.StatusCode)
|
||||
_, _ = io.Copy(c.Writer, resp.Body)
|
||||
}
|
||||
}
|
||||
|
||||
// readerLocalAssetHandler 本地书籍内嵌资源(EPUB 图片等):
|
||||
// 鉴权走 HMAC 签名(<img src> 带不上 JWT),与 /reader/media 同一套做法。
|
||||
func readerLocalAssetHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
bookID := c.Query("b")
|
||||
entry, err := svc.Reader.VerifyLocalAssetURL(bookID, c.Query("p"), c.Query("s"))
|
||||
if err != nil {
|
||||
c.JSON(http.StatusForbidden, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
data, contentType, err := svc.Reader.ReadLocalAsset(c.Request.Context(), bookID, entry)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusNotFound, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
// 同一本书的图片不会变,可长缓存
|
||||
c.Header("Cache-Control", "private, max-age=604800")
|
||||
c.Data(http.StatusOK, contentType, data)
|
||||
}
|
||||
}
|
||||
|
||||
// readerAudioTranscodeHandler 需要转码的有声书音轨:签名鉴权(<audio src> 带不上 JWT),
|
||||
// 首次请求跑 ffmpeg 转成 mp3 落缓存,之后按 Range 下发,播放器可以拖动进度。
|
||||
func readerAudioTranscodeHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
bookID := c.Query("b")
|
||||
source, err := svc.Reader.VerifyAudioTranscodeURL(bookID, c.Query("u"), c.Query("s"))
|
||||
if err != nil {
|
||||
c.String(http.StatusForbidden, "%s", err.Error())
|
||||
return
|
||||
}
|
||||
path, err := svc.Reader.EnsureTranscodedAudio(c.Request.Context(), bookID, source)
|
||||
if err != nil {
|
||||
c.String(http.StatusBadGateway, "%s", err.Error())
|
||||
return
|
||||
}
|
||||
f, err := os.Open(path) // #nosec G304 -- 路径由签名校验 + 缓存目录哈希生成
|
||||
if err != nil {
|
||||
c.String(http.StatusNotFound, "转码结果已丢失")
|
||||
return
|
||||
}
|
||||
defer f.Close()
|
||||
info, err := f.Stat()
|
||||
if err != nil {
|
||||
c.String(http.StatusNotFound, "转码结果不可用")
|
||||
return
|
||||
}
|
||||
c.Header("Cache-Control", "private, max-age=604800")
|
||||
http.ServeContent(c.Writer, c.Request, info.Name(), info.ModTime(), f)
|
||||
}
|
||||
}
|
||||
|
||||
// readerLocalAudioHandler 本地有声书音频流:签名鉴权(<audio src> 带不上 JWT),
|
||||
// 交给 http.ServeContent 处理 Range,播放器才能拖动进度。
|
||||
func readerLocalAudioHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
path, err := svc.Reader.VerifyLocalAudioURL(c.Query("b"), c.Query("p"), c.Query("s"))
|
||||
if err != nil {
|
||||
c.String(http.StatusForbidden, "%s", err.Error())
|
||||
return
|
||||
}
|
||||
f, info, err := svc.Reader.OpenLocalAudio(path)
|
||||
if err != nil {
|
||||
c.String(http.StatusNotFound, "%s", err.Error())
|
||||
return
|
||||
}
|
||||
defer f.Close()
|
||||
http.ServeContent(c.Writer, c.Request, info.Name(), info.ModTime(), f)
|
||||
}
|
||||
}
|
||||
|
||||
// readerBookContentHandler 书架维度正文(服务端应用替换净化规则)。
|
||||
func readerBookContentHandler(svc *service.Container) gin.HandlerFunc {
|
||||
return func(c *gin.Context) {
|
||||
chapter, err := strconv.Atoi(c.DefaultQuery("chapter", "0"))
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": "chapter 参数需为整数"})
|
||||
return
|
||||
}
|
||||
userID := c.GetString(middleware.CtxUserID)
|
||||
content, err := svc.Reader.GetContentForBook(c.Request.Context(), userID, c.Param("id"), chapter)
|
||||
if err != nil {
|
||||
c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
|
||||
return
|
||||
}
|
||||
c.JSON(http.StatusOK, content)
|
||||
}
|
||||
}
|
||||
@@ -30,4 +30,5 @@ func registerAuthenticatedRoutes(api *gin.RouterGroup, cfg *config.Config, svc *
|
||||
registerAuthedDLNAControlRoutes(authed, svc)
|
||||
registerAuthedFavoriteAndMediaActionRoutes(authed, svc)
|
||||
registerAuthedPlaybackExtraRoutes(authed, svc)
|
||||
registerReaderRoutes(authed, svc)
|
||||
}
|
||||
|
||||
+63
-1
@@ -3,14 +3,66 @@
|
||||
package helper
|
||||
|
||||
import (
|
||||
"bytes"
|
||||
"compress/flate"
|
||||
"compress/gzip"
|
||||
"compress/zlib"
|
||||
"io"
|
||||
"net/http"
|
||||
"net/url"
|
||||
"strings"
|
||||
"time"
|
||||
)
|
||||
|
||||
// defaultUserAgent 是默认浏览器 User-Agent(用于 HTTP 请求头)。
|
||||
const defaultUserAgent = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.0.0 Safari/537.36"
|
||||
|
||||
// DecompressBody 兜底解压响应体(gzip / deflate)。
|
||||
//
|
||||
// 正常情况下用不到:只要不显式设置 Accept-Encoding,net/http 会自己带上 gzip
|
||||
// 并透明解压。但书源 JSON 的 header 字段、或某些 CDN 的固定策略,都可能让响应
|
||||
// 带着 Content-Encoding 回来,压缩字节一旦进入规则层,书源的 JSON.parse 就会
|
||||
// 报 `invalid character '\x1f'`。这里做一次兜底,保证调用方拿到明文。
|
||||
//
|
||||
// brotli(br)无法在纯 Go 标准库里解,因此上面不再 advertise br;真遇到 br
|
||||
// 响应则原样返回,由调用方按失败处理。
|
||||
func DecompressBody(resp *http.Response, data []byte) []byte {
|
||||
if resp == nil || len(data) == 0 {
|
||||
return data
|
||||
}
|
||||
switch strings.ToLower(strings.TrimSpace(resp.Header.Get("Content-Encoding"))) {
|
||||
case "gzip", "x-gzip":
|
||||
r, err := gzip.NewReader(bytes.NewReader(data))
|
||||
if err != nil {
|
||||
return data
|
||||
}
|
||||
defer r.Close()
|
||||
if out, err := io.ReadAll(r); err == nil {
|
||||
return out
|
||||
}
|
||||
case "deflate":
|
||||
// deflate 有两种实际写法:zlib 包装与裸 DEFLATE,依次尝试。
|
||||
if out, err := io.ReadAll(flate.NewReader(bytes.NewReader(data))); err == nil {
|
||||
return out
|
||||
}
|
||||
if zr, err := zlib.NewReader(bytes.NewReader(data)); err == nil {
|
||||
defer zr.Close()
|
||||
if out, err := io.ReadAll(zr); err == nil {
|
||||
return out
|
||||
}
|
||||
}
|
||||
}
|
||||
return data
|
||||
}
|
||||
|
||||
// StripAcceptEncoding 移除显式设置的 Accept-Encoding,交回 net/http 管理。
|
||||
//
|
||||
// 只有「调用方没设置」时 net/http 才会自动解压,因此任何来源(预设头、书源
|
||||
// header 字段)带来的 Accept-Encoding 都必须清掉。
|
||||
func StripAcceptEncoding(h http.Header) {
|
||||
h.Del("Accept-Encoding")
|
||||
}
|
||||
|
||||
// NewSiteHTTPClient builds an http.Client honoring per-site policies:
|
||||
// - timeout (seconds, defaults to 15)
|
||||
// - proxy via HTTP(S)_PROXY environment variables when site.UseProxy is on
|
||||
@@ -41,12 +93,22 @@ func NewSiteHTTPClient(timeoutSeconds int, useProxy bool) *http.Client {
|
||||
|
||||
// HTTPHeaderPresets returns a map of realistic browser HTTP headers.
|
||||
// These mimic a real Chrome browser to avoid WAF/bot detection.
|
||||
//
|
||||
// 注意:这里刻意不设置 Accept-Encoding。
|
||||
//
|
||||
// net/http 只在「调用方没有显式设置 Accept-Encoding」时才会自己补上 gzip 并
|
||||
// 透明解压;一旦我们显式写了这个头,它就原样把压缩字节交出来。之前这里写了
|
||||
// "gzip, deflate, br",于是所有经历了压缩的响应(静态 JSON、CDN 上的页面等)
|
||||
// 都会以原始压缩字节进入规则层,书源的 JSON.parse 直接报
|
||||
// `SyntaxError: invalid character '\x1f'`,表现为「获取最新配置失败」。
|
||||
//
|
||||
// 交给 net/http 管理后:请求仍会带 `Accept-Encoding: gzip`(浏览器常见取值),
|
||||
// 且响应被自动解压;另外也避免服务端挑选我们无法解码的 br。
|
||||
func HTTPHeaderPresets() map[string]string {
|
||||
return map[string]string{
|
||||
"User-Agent": defaultUserAgent,
|
||||
"Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.7",
|
||||
"Accept-Language": "zh-CN,zh;q=0.9,en;q=0.8",
|
||||
"Accept-Encoding": "gzip, deflate, br",
|
||||
"Connection": "keep-alive",
|
||||
"Upgrade-Insecure-Requests": "1",
|
||||
"Sec-Fetch-Dest": "document",
|
||||
|
||||
@@ -0,0 +1,73 @@
|
||||
package helper
|
||||
|
||||
import (
|
||||
"bytes"
|
||||
"compress/gzip"
|
||||
"net/http"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// 本文件:压缩相关的回归测试。
|
||||
//
|
||||
// 背景:预设头里曾显式写 `Accept-Encoding: gzip, deflate, br`,导致 net/http
|
||||
// 不再自动解压(它只在调用方没设置该头时才解压),压缩字节直接进入规则层。
|
||||
|
||||
// TestHeaderPresetsDoNotSetAcceptEncoding 预设头不得设置 Accept-Encoding。
|
||||
// 一旦设置,net/http 的透明解压就失效,压缩响应会以原始字节交给上层。
|
||||
func TestHeaderPresetsDoNotSetAcceptEncoding(t *testing.T) {
|
||||
for k := range HTTPHeaderPresets() {
|
||||
if http.CanonicalHeaderKey(k) == "Accept-Encoding" {
|
||||
t.Fatal("HTTPHeaderPresets 不应设置 Accept-Encoding(会让 net/http 放弃自动解压)")
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestStripAcceptEncoding 任何来源带来的 Accept-Encoding 都应被清掉。
|
||||
func TestStripAcceptEncoding(t *testing.T) {
|
||||
h := http.Header{}
|
||||
h.Set("Accept-Encoding", "br, gzip")
|
||||
StripAcceptEncoding(h)
|
||||
if got := h.Get("Accept-Encoding"); got != "" {
|
||||
t.Fatalf("Accept-Encoding 未清理: %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
// TestDecompressBody gzip / deflate 响应体应被还原成明文。
|
||||
func TestDecompressBody(t *testing.T) {
|
||||
payload := `{"version":"20260926"}`
|
||||
|
||||
t.Run("gzip", func(t *testing.T) {
|
||||
var buf bytes.Buffer
|
||||
gz := gzip.NewWriter(&buf)
|
||||
_, _ = gz.Write([]byte(payload))
|
||||
_ = gz.Close()
|
||||
|
||||
resp := &http.Response{Header: http.Header{}}
|
||||
resp.Header.Set("Content-Encoding", "gzip")
|
||||
if got := string(DecompressBody(resp, buf.Bytes())); got != payload {
|
||||
t.Fatalf("gzip 解压 = %q", got)
|
||||
}
|
||||
})
|
||||
|
||||
t.Run("未压缩原样返回", func(t *testing.T) {
|
||||
resp := &http.Response{Header: http.Header{}}
|
||||
if got := string(DecompressBody(resp, []byte(payload))); got != payload {
|
||||
t.Fatalf("未压缩响应被改动: %q", got)
|
||||
}
|
||||
})
|
||||
|
||||
t.Run("坏数据不 panic", func(t *testing.T) {
|
||||
resp := &http.Response{Header: http.Header{}}
|
||||
resp.Header.Set("Content-Encoding", "gzip")
|
||||
raw := []byte("not-gzip")
|
||||
if got := string(DecompressBody(resp, raw)); got != "not-gzip" {
|
||||
t.Fatalf("解压失败时应原样返回,实际 %q", got)
|
||||
}
|
||||
})
|
||||
|
||||
t.Run("nil resp", func(t *testing.T) {
|
||||
if got := string(DecompressBody(nil, []byte(payload))); got != payload {
|
||||
t.Fatalf("nil resp 应原样返回,实际 %q", got)
|
||||
}
|
||||
})
|
||||
}
|
||||
@@ -0,0 +1,87 @@
|
||||
// Package helper — AES-256-GCM at-rest secret encryption.
|
||||
//
|
||||
// Sensitive values (third-party API keys, service passwords, book-source
|
||||
// login credentials) are stored in the database. Encrypting them keyed off
|
||||
// a per-install secret means a stolen DB file alone is not enough to
|
||||
// recover the plaintext.
|
||||
//
|
||||
// Format on disk: "enc:v1:" + base64(nonce || ciphertext || tag).
|
||||
// Legacy plaintext rows (no prefix) round-trip unchanged so an upgraded
|
||||
// install needs no migration step.
|
||||
package helper
|
||||
|
||||
import (
|
||||
"crypto/aes"
|
||||
"crypto/cipher"
|
||||
"crypto/rand"
|
||||
"crypto/sha256"
|
||||
"encoding/base64"
|
||||
"strings"
|
||||
)
|
||||
|
||||
// EncPrefix tags ciphertext rows so they can be told apart from plaintext.
|
||||
const EncPrefix = "enc:v1:"
|
||||
|
||||
// SecretCipher encrypts/decrypts strings with AES-256-GCM derived from a
|
||||
// stable per-install secret. A zero-value SecretCipher (empty key) is a
|
||||
// pass-through, which keeps unit tests and keyless dev setups working.
|
||||
type SecretCipher struct {
|
||||
aead cipher.AEAD
|
||||
}
|
||||
|
||||
// NewSecretCipher derives a 256-bit key from secret via SHA-256.
|
||||
func NewSecretCipher(secret string) *SecretCipher {
|
||||
if strings.TrimSpace(secret) == "" {
|
||||
return &SecretCipher{}
|
||||
}
|
||||
sum := sha256.Sum256([]byte(secret))
|
||||
block, err := aes.NewCipher(sum[:])
|
||||
if err != nil {
|
||||
return &SecretCipher{}
|
||||
}
|
||||
aead, err := cipher.NewGCM(block)
|
||||
if err != nil {
|
||||
return &SecretCipher{}
|
||||
}
|
||||
return &SecretCipher{aead: aead}
|
||||
}
|
||||
|
||||
// Ready 表示是否真的能加解密(密钥有效)。
|
||||
func (c *SecretCipher) Ready() bool { return c != nil && c.aead != nil }
|
||||
|
||||
// Encrypt returns the prefixed base64 ciphertext for plain.
|
||||
// Empty inputs and already-encrypted values round-trip unchanged.
|
||||
func (c *SecretCipher) Encrypt(plain string) string {
|
||||
if plain == "" || !c.Ready() || strings.HasPrefix(plain, EncPrefix) {
|
||||
return plain
|
||||
}
|
||||
nonce := make([]byte, c.aead.NonceSize())
|
||||
if _, err := rand.Read(nonce); err != nil {
|
||||
return plain
|
||||
}
|
||||
sealed := c.aead.Seal(nonce, nonce, []byte(plain), nil)
|
||||
return EncPrefix + base64.StdEncoding.EncodeToString(sealed)
|
||||
}
|
||||
|
||||
// Decrypt returns the plaintext for a stored value. Plaintext rows (no
|
||||
// prefix) are returned unchanged, as are values that fail to authenticate.
|
||||
func (c *SecretCipher) Decrypt(value string) string {
|
||||
if value == "" || !c.Ready() || !strings.HasPrefix(value, EncPrefix) {
|
||||
return value
|
||||
}
|
||||
data, err := base64.StdEncoding.DecodeString(strings.TrimPrefix(value, EncPrefix))
|
||||
if err != nil || len(data) < c.aead.NonceSize() {
|
||||
return value
|
||||
}
|
||||
nonce, body := data[:c.aead.NonceSize()], data[c.aead.NonceSize():]
|
||||
plain, err := c.aead.Open(nil, nonce, body, nil)
|
||||
if err != nil {
|
||||
return value
|
||||
}
|
||||
return string(plain)
|
||||
}
|
||||
|
||||
// IsEncrypted reports whether value carries the encrypted prefix.
|
||||
func (c *SecretCipher) IsEncrypted(value string) bool {
|
||||
return strings.HasPrefix(value, EncPrefix)
|
||||
}
|
||||
@@ -61,5 +61,12 @@ func AllModels() []interface{} {
|
||||
&StrmDirCache{},
|
||||
&ScrapeTask{},
|
||||
&EmbyMount{},
|
||||
&ReaderBookSource{},
|
||||
&ReaderSourceState{},
|
||||
&ReaderBook{},
|
||||
&ReaderChapter{},
|
||||
&ReaderReplaceRule{},
|
||||
&ReaderProfile{},
|
||||
&ReaderBookGroups{},
|
||||
}
|
||||
}
|
||||
|
||||
@@ -0,0 +1,145 @@
|
||||
// Package model — 阅读(legado 书源兼容)子系统数据模型。
|
||||
// 字段语义对齐 legado 的 BookSource / Book / BookChapter / ReplaceRule 实体;
|
||||
// 阅读进度沿用 legado 的做法直接挂在书籍上(durChapter* 字段)。
|
||||
package model
|
||||
|
||||
import (
|
||||
"time"
|
||||
)
|
||||
|
||||
// ReaderBookSource 书源:原始 JSON 全量存储 + 常用字段冗余列出便于筛选排序。
|
||||
type ReaderBookSource struct {
|
||||
Base
|
||||
Name string `gorm:"type:varchar(255);index" json:"name"`
|
||||
GroupName string `gorm:"type:varchar(255);index" json:"group"`
|
||||
Type int `gorm:"default:0" json:"type"` // 0文本 1音频 2图片 3文件 4视频
|
||||
SourceURL string `gorm:"type:varchar(512);index" json:"source_url"`
|
||||
RawJSON string `gorm:"type:text" json:"-"`
|
||||
Enabled bool `gorm:"default:true" json:"enabled"`
|
||||
EnabledExplore bool `gorm:"default:true" json:"enabled_explore"`
|
||||
CustomOrder int `json:"custom_order"`
|
||||
Weight int `json:"weight"`
|
||||
ConcurrentRate string `gorm:"type:varchar(64)" json:"concurrent_rate"`
|
||||
Header string `gorm:"type:text" json:"header"` // 书源级请求头 JSON
|
||||
Comment string `gorm:"type:text" json:"comment"`
|
||||
Variables string `gorm:"type:text" json:"variables"` // source 变量 JSON
|
||||
LastUpdateTime int64 `json:"last_update_time"`
|
||||
LastCheckAt *time.Time `json:"last_check_at"`
|
||||
RespondTime int64 `json:"respond_time"` // 最近一次调试响应耗时(ms)
|
||||
// HasLogin 是否声明了登录能力(loginUrl/loginUi),列表接口按需计算,不落库。
|
||||
HasLogin bool `gorm:"-" json:"has_login"`
|
||||
// NeedsBrowser 是否依赖 WebView/无头浏览器(webView 选项或 webjs 规则)。
|
||||
// 服务端没有浏览器,这类源注定不可用,列表接口按需计算让前端能提前提示。
|
||||
NeedsBrowser bool `gorm:"-" json:"needs_browser"`
|
||||
}
|
||||
|
||||
// ReaderSourceState 书源会话状态:对应 legado 中按书源 key 存储的
|
||||
// sourceVariable / userInfo(登录信息)/ loginHeader 与 CookieStore。
|
||||
// 与书源分表存放,避免每次导入更新书源时把用户登录态覆盖掉。
|
||||
type ReaderSourceState struct {
|
||||
Base
|
||||
SourceURL string `gorm:"type:varchar(512);uniqueIndex" json:"source_url"`
|
||||
Variable string `gorm:"type:text" json:"variable"` // source.getVariable/setVariable
|
||||
LoginInfo string `gorm:"type:text" json:"login_info"` // source.getLoginInfo/putLoginInfo(登录表单 JSON)
|
||||
LoginHeader string `gorm:"type:text" json:"login_header"` // source 登录请求头 JSON
|
||||
Cookies string `gorm:"type:text" json:"cookies"` // JSON: domain → "k=v; k=v"
|
||||
}
|
||||
|
||||
// ReaderBook 书架条目(含阅读进度,对应 legado Book)。
|
||||
type ReaderBook struct {
|
||||
Base
|
||||
UserID string `gorm:"type:varchar(36);index" json:"user_id"`
|
||||
Origin string `gorm:"type:varchar(512)" json:"origin"` // 书源 URL
|
||||
OriginName string `gorm:"type:varchar(255)" json:"origin_name"`
|
||||
BookURL string `gorm:"type:varchar(512);index" json:"book_url"`
|
||||
TocURL string `gorm:"type:varchar(512)" json:"toc_url"`
|
||||
Name string `gorm:"type:varchar(255)" json:"name"`
|
||||
Author string `gorm:"type:varchar(255)" json:"author"`
|
||||
Kind string `gorm:"type:varchar(255)" json:"kind"`
|
||||
CoverURL string `gorm:"type:varchar(512)" json:"cover_url"`
|
||||
Intro string `gorm:"type:text" json:"intro"`
|
||||
Charset string `gorm:"type:varchar(32)" json:"charset"`
|
||||
Type int `gorm:"default:0" json:"type"` // 0文本 1音频 2图片
|
||||
LatestChapterTitle string `gorm:"type:varchar(512)" json:"latest_chapter_title"`
|
||||
// LatestChapterTime 最后一次「目录末尾章节发生变化」的时间(毫秒)。
|
||||
// 对应 legado Book.latestChapterTime:书架「最近更新」排序与「更新时间」展示都读它。
|
||||
// 抓目录时发现末章标题与已存值不同才刷新;本地导入书籍不参与。
|
||||
LatestChapterTime int64 `json:"latest_chapter_time"`
|
||||
TotalChapterNum int `json:"total_chapter_num"`
|
||||
DurChapterIndex int `json:"dur_chapter_index"`
|
||||
DurChapterPos int `json:"dur_chapter_pos"`
|
||||
DurChapterTitle string `gorm:"type:varchar(512)" json:"dur_chapter_title"`
|
||||
DurChapterTime int64 `json:"dur_chapter_time"`
|
||||
Order int `json:"order"`
|
||||
Variable string `gorm:"type:text" json:"variable"`
|
||||
// LocalPath 本地导入书籍的位置:默认是 data/reader/local 下的文件名
|
||||
// (如 "<id>.txt");LocalExternal 为真时是服务器上的绝对路径。
|
||||
// 为空表示来自网络书源。不对外暴露路径,前端用 is_local 判断。
|
||||
LocalPath string `gorm:"type:varchar(255)" json:"-"`
|
||||
// LocalExternal 为真表示原地引用服务器上已有的文件/目录(管理员在导入时选定),
|
||||
// 移出书架只解除引用,不删除源文件;为假表示 data/reader/local 下的托管副本。
|
||||
LocalExternal bool `json:"local_external"`
|
||||
// IsLocal 是否本地导入书籍,列表接口按需计算,不落库。
|
||||
IsLocal bool `gorm:"-" json:"is_local"`
|
||||
// 听书(音频源)跳过片头/片尾秒数,对应 legado Book.getOpenCredits/getCloseCredits。
|
||||
// 0 表示不跳过。仅对音频和视频源生效。
|
||||
OpenCredits int `gorm:"default:0" json:"open_credits"`
|
||||
CloseCredits int `gorm:"default:0" json:"close_credits"`
|
||||
}
|
||||
|
||||
// ReaderChapter 章节缓存(对应 legado BookChapter)。
|
||||
type ReaderChapter struct {
|
||||
Base
|
||||
BookID string `gorm:"type:varchar(36);uniqueIndex:idx_reader_book_chapter" json:"book_id"`
|
||||
Index int `gorm:"uniqueIndex:idx_reader_book_chapter" json:"index"`
|
||||
URL string `gorm:"type:varchar(512)" json:"url"`
|
||||
Title string `gorm:"type:varchar(512)" json:"title"`
|
||||
IsVolume bool `json:"is_volume"`
|
||||
Tag string `gorm:"type:varchar(255)" json:"tag"`
|
||||
}
|
||||
|
||||
// ReaderReplaceRule 替换净化规则(对应 legado ReplaceRule)。
|
||||
type ReaderReplaceRule struct {
|
||||
Base
|
||||
UserID string `gorm:"type:varchar(36);index" json:"user_id"`
|
||||
Name string `gorm:"type:varchar(255)" json:"name"`
|
||||
GroupName string `gorm:"type:varchar(255);index" json:"group"`
|
||||
Pattern string `gorm:"type:text" json:"pattern"`
|
||||
Replacement string `gorm:"type:text" json:"replacement"`
|
||||
Scope string `gorm:"type:varchar(255)" json:"scope"`
|
||||
ScopeTitle bool `json:"scope_title"`
|
||||
ScopeContent bool `gorm:"default:true" json:"scope_content"`
|
||||
ExcludeScope string `gorm:"type:varchar(255)" json:"exclude_scope"`
|
||||
IsEnabled bool `gorm:"default:true" json:"is_enabled"`
|
||||
IsRegex bool `gorm:"default:true" json:"is_regex"`
|
||||
TimeoutMillisecond int64 `gorm:"default:3000" json:"timeout_millisecond"`
|
||||
Order int `json:"order"`
|
||||
}
|
||||
|
||||
// ReaderProfile 阅读器偏好(每个用户一条)。
|
||||
//
|
||||
// 主题 / 排版 / 听书 / 书架展示这些设置原先只存在浏览器 localStorage,属于设备级:
|
||||
// 换设备或换浏览器就丢。这里按用户落库作为权威来源,前端 localStorage 退化成首屏缓存。
|
||||
// 首页的「影视 / 阅读」模式(homeMode)是设备级偏好,故意不在这里同步。
|
||||
type ReaderProfile struct {
|
||||
Base
|
||||
UserID string `gorm:"type:varchar(36);uniqueIndex" json:"user_id"`
|
||||
|
||||
ThemeID string `gorm:"type:varchar(32)" json:"theme_id"`
|
||||
Night bool `json:"night"`
|
||||
PageMode string `gorm:"type:varchar(16)" json:"page_mode"` // page 翻页 / scroll 滚动
|
||||
FontSize int `json:"font_size"`
|
||||
LineHeight float64 `json:"line_height"`
|
||||
ParagraphSpacing int `json:"paragraph_spacing"`
|
||||
|
||||
// 听书偏好(对应 legado AudioPlay.playSpeed 与 AppConfig.ttsTimer)
|
||||
AudioSpeed float64 `json:"audio_speed"`
|
||||
AudioTimerMinutes int `json:"audio_timer_minutes"`
|
||||
|
||||
// 书架展示偏好(对应 legado 的书架设置)
|
||||
ShelfLayout string `gorm:"type:varchar(16)" json:"shelf_layout"` // grid / list / compact
|
||||
ShelfGridColumns int `json:"shelf_grid_columns"` // 0 表示自适应
|
||||
ShelfSort string `gorm:"type:varchar(16)" json:"shelf_sort"`
|
||||
ShelfShowUnread bool `json:"shelf_show_unread"`
|
||||
ShelfShowUpdateTime bool `json:"shelf_show_update_time"`
|
||||
}
|
||||
@@ -0,0 +1,125 @@
|
||||
package model
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"strings"
|
||||
)
|
||||
|
||||
// 书架分组(对应 legado 的 BookGroup)。
|
||||
//
|
||||
// 与影视模块的媒体库标签(LibraryTagSet)同构:组名 → 成员 ID 列表,一个成员只归一个组。
|
||||
// 区别只是归属对象从媒体库换成了书籍,因此前端可以照搬那套标签栏 / 管理弹窗的交互。
|
||||
|
||||
const (
|
||||
// MaxBookGroups 是单个用户可创建的分组数量上限,避免恶意写入过大的 JSON。
|
||||
MaxBookGroups = 50
|
||||
// MaxBookGroupNameLen 是单个分组名的最大字符长度(按 rune 计数)。
|
||||
MaxBookGroupNameLen = 24
|
||||
)
|
||||
|
||||
// BookGroupSet 是一个书架分组:组名 + 组内书籍 ID(顺序即组内展示顺序)。
|
||||
type BookGroupSet struct {
|
||||
Name string `json:"name"`
|
||||
BookIDs []string `json:"book_ids"`
|
||||
}
|
||||
|
||||
// ReaderBookGroups 是某个用户的书架分组集合,每个用户一条记录。
|
||||
//
|
||||
// 分组是「整份替换」语义(前端一次 PUT 提交完整分组列表),所以直接存成一份 JSON;
|
||||
// 与阅读器偏好(ReaderProfile)分表存放,两条写入路径互不干扰:
|
||||
// 先建分组、后调偏好时不会因为整行 Save 把对方的列写成零值。
|
||||
type ReaderBookGroups struct {
|
||||
Base
|
||||
UserID string `gorm:"type:varchar(36);uniqueIndex" json:"user_id"`
|
||||
// Groups 是 []BookGroupSet 的 JSON 文本;空字符串表示没有分组。
|
||||
Groups string `gorm:"type:text" json:"-"`
|
||||
}
|
||||
|
||||
// DecodeBookGroups 解析 Groups 字段,忽略损坏的数据。
|
||||
func (r *ReaderBookGroups) DecodeBookGroups() []BookGroupSet {
|
||||
if r == nil || strings.TrimSpace(r.Groups) == "" {
|
||||
return nil
|
||||
}
|
||||
var groups []BookGroupSet
|
||||
if err := json.Unmarshal([]byte(r.Groups), &groups); err != nil {
|
||||
return nil
|
||||
}
|
||||
return NormalizeBookGroups(groups)
|
||||
}
|
||||
|
||||
// EncodeBookGroups 把分组集合序列化为可写入 Groups 字段的 JSON 文本。
|
||||
// 空集合序列化为空字符串,便于用零值表达「没有分组」。
|
||||
func EncodeBookGroups(groups []BookGroupSet) (string, error) {
|
||||
if len(groups) == 0 {
|
||||
return "", nil
|
||||
}
|
||||
raw, err := json.Marshal(groups)
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
return string(raw), nil
|
||||
}
|
||||
|
||||
// NormalizeBookGroups 清洗分组集合:去掉空名分组、合并重名分组、去掉组内重复的
|
||||
// 书籍 ID,并保持传入顺序。不在这里做「一本书只归一个组」的收敛——那需要知道
|
||||
// 书籍是否存在,属于服务层(SetBookGroups)的职责。
|
||||
func NormalizeBookGroups(groups []BookGroupSet) []BookGroupSet {
|
||||
if len(groups) == 0 {
|
||||
return nil
|
||||
}
|
||||
out := make([]BookGroupSet, 0, len(groups))
|
||||
indexByName := make(map[string]int, len(groups))
|
||||
for _, group := range groups {
|
||||
name := TruncateBookGroupName(group.Name)
|
||||
if name == "" {
|
||||
continue
|
||||
}
|
||||
if len(out) >= MaxBookGroups {
|
||||
break
|
||||
}
|
||||
key := strings.ToLower(name)
|
||||
pos, exists := indexByName[key]
|
||||
if !exists {
|
||||
out = append(out, BookGroupSet{Name: name, BookIDs: []string{}})
|
||||
pos = len(out) - 1
|
||||
indexByName[key] = pos
|
||||
}
|
||||
seen := make(map[string]struct{}, len(out[pos].BookIDs))
|
||||
for _, id := range out[pos].BookIDs {
|
||||
seen[id] = struct{}{}
|
||||
}
|
||||
for _, id := range group.BookIDs {
|
||||
trimmed := strings.TrimSpace(id)
|
||||
if trimmed == "" {
|
||||
continue
|
||||
}
|
||||
if _, ok := seen[trimmed]; ok {
|
||||
continue
|
||||
}
|
||||
seen[trimmed] = struct{}{}
|
||||
out[pos].BookIDs = append(out[pos].BookIDs, trimmed)
|
||||
}
|
||||
}
|
||||
if len(out) == 0 {
|
||||
return nil
|
||||
}
|
||||
for i := range out {
|
||||
if out[i].BookIDs == nil {
|
||||
out[i].BookIDs = []string{}
|
||||
}
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// TruncateBookGroupName 去掉首尾空白并按 rune 截断到长度上限。
|
||||
func TruncateBookGroupName(name string) string {
|
||||
name = strings.TrimSpace(name)
|
||||
if name == "" {
|
||||
return ""
|
||||
}
|
||||
runes := []rune(name)
|
||||
if len(runes) > MaxBookGroupNameLen {
|
||||
runes = runes[:MaxBookGroupNameLen]
|
||||
}
|
||||
return strings.TrimSpace(string(runes))
|
||||
}
|
||||
@@ -0,0 +1,84 @@
|
||||
package model
|
||||
|
||||
import (
|
||||
"strings"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// 书架分组的清洗与编解码回归测试(与 library_tag_test.go 同构)。
|
||||
|
||||
func TestNormalizeBookGroupsMergesDuplicatesAndTrims(t *testing.T) {
|
||||
groups := []BookGroupSet{
|
||||
{Name: " 科幻 ", BookIDs: []string{"a", "b"}},
|
||||
{Name: "科幻", BookIDs: []string{"b", "c"}},
|
||||
{Name: " ", BookIDs: []string{"d"}},
|
||||
{Name: "在读", BookIDs: []string{"", "e", "e"}},
|
||||
}
|
||||
got := NormalizeBookGroups(groups)
|
||||
if len(got) != 2 {
|
||||
t.Fatalf("NormalizeBookGroups len = %d, want 2 (%#v)", len(got), got)
|
||||
}
|
||||
if got[0].Name != "科幻" || strings.Join(got[0].BookIDs, ",") != "a,b,c" {
|
||||
t.Fatalf("首个分组 = %#v,期望 科幻[a,b,c]", got[0])
|
||||
}
|
||||
if got[1].Name != "在读" || strings.Join(got[1].BookIDs, ",") != "e" {
|
||||
t.Fatalf("第二个分组 = %#v,期望 在读[e]", got[1])
|
||||
}
|
||||
}
|
||||
|
||||
func TestNormalizeBookGroupsCapsCountAndNameLength(t *testing.T) {
|
||||
long := strings.Repeat("长", MaxBookGroupNameLen+10)
|
||||
got := NormalizeBookGroups([]BookGroupSet{{Name: long}})
|
||||
if len(got) != 1 {
|
||||
t.Fatalf("len = %d, want 1", len(got))
|
||||
}
|
||||
if runes := []rune(got[0].Name); len(runes) != MaxBookGroupNameLen {
|
||||
t.Fatalf("分组名长度 = %d,期望 %d", len(runes), MaxBookGroupNameLen)
|
||||
}
|
||||
|
||||
many := make([]BookGroupSet, 0, MaxBookGroups+5)
|
||||
for i := 0; i < MaxBookGroups+5; i++ {
|
||||
many = append(many, BookGroupSet{Name: string(rune('a' + i))})
|
||||
}
|
||||
if capped := NormalizeBookGroups(many); len(capped) != MaxBookGroups {
|
||||
t.Fatalf("capped len = %d, want %d", len(capped), MaxBookGroups)
|
||||
}
|
||||
}
|
||||
|
||||
func TestNormalizeBookGroupsKeepsEmptyGroups(t *testing.T) {
|
||||
got := NormalizeBookGroups([]BookGroupSet{{Name: "空组"}})
|
||||
if len(got) != 1 {
|
||||
t.Fatalf("len = %d, want 1", len(got))
|
||||
}
|
||||
// 空分组的 BookIDs 应该是空数组而不是 nil,前端直接遍历时不会炸
|
||||
if got[0].BookIDs == nil || len(got[0].BookIDs) != 0 {
|
||||
t.Fatalf("BookIDs = %#v,期望空数组", got[0].BookIDs)
|
||||
}
|
||||
}
|
||||
|
||||
func TestEncodeDecodeBookGroupsRoundTrip(t *testing.T) {
|
||||
if encoded, err := EncodeBookGroups(nil); err != nil || encoded != "" {
|
||||
t.Fatalf("EncodeBookGroups(nil) = %q, %v; want \"\", nil", encoded, err)
|
||||
}
|
||||
|
||||
raw, err := EncodeBookGroups([]BookGroupSet{{Name: "科幻", BookIDs: []string{"b-1"}}})
|
||||
if err != nil {
|
||||
t.Fatalf("EncodeBookGroups: %v", err)
|
||||
}
|
||||
row := &ReaderBookGroups{Groups: raw}
|
||||
decoded := row.DecodeBookGroups()
|
||||
if len(decoded) != 1 || decoded[0].Name != "科幻" || decoded[0].BookIDs[0] != "b-1" {
|
||||
t.Fatalf("DecodeBookGroups = %#v", decoded)
|
||||
}
|
||||
|
||||
row.Groups = "{not json"
|
||||
if decoded := row.DecodeBookGroups(); decoded != nil {
|
||||
t.Fatalf("损坏数据应返回 nil,实际 %#v", decoded)
|
||||
}
|
||||
|
||||
// nil 接收者不应 panic(仓库层查不到记录时会返回 nil)
|
||||
var nilRow *ReaderBookGroups
|
||||
if decoded := nilRow.DecodeBookGroups(); decoded != nil {
|
||||
t.Fatalf("nil 接收者应返回 nil,实际 %#v", decoded)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,283 @@
|
||||
// Package repository — 阅读子系统仓储(书源 / 书架 / 章节 / 替换规则)。
|
||||
package repository
|
||||
|
||||
import (
|
||||
"context"
|
||||
"errors"
|
||||
|
||||
"gorm.io/gorm"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
)
|
||||
|
||||
// ReaderRepository 阅读子系统 GORM 封装。
|
||||
type ReaderRepository struct {
|
||||
db *gorm.DB
|
||||
}
|
||||
|
||||
// ListSources 书源列表(按 customOrder 排序)。
|
||||
func (r *ReaderRepository) ListSources(ctx context.Context) ([]model.ReaderBookSource, error) {
|
||||
var out []model.ReaderBookSource
|
||||
err := r.db.WithContext(ctx).Order("custom_order ASC, updated_at DESC").Find(&out).Error
|
||||
return out, err
|
||||
}
|
||||
|
||||
// GetSource 按 ID 取书源。
|
||||
func (r *ReaderRepository) GetSource(ctx context.Context, id string) (*model.ReaderBookSource, error) {
|
||||
var out model.ReaderBookSource
|
||||
if err := r.db.WithContext(ctx).First(&out, "id = ?", id).Error; err != nil {
|
||||
return nil, err
|
||||
}
|
||||
return &out, nil
|
||||
}
|
||||
|
||||
// GetSourceByURL 按书源 URL 取书源(导入去重用)。
|
||||
func (r *ReaderRepository) GetSourceByURL(ctx context.Context, sourceURL string) (*model.ReaderBookSource, error) {
|
||||
var out model.ReaderBookSource
|
||||
err := r.db.WithContext(ctx).First(&out, "source_url = ?", sourceURL).Error
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
return &out, nil
|
||||
}
|
||||
|
||||
// CreateSource 新增书源。
|
||||
func (r *ReaderRepository) CreateSource(ctx context.Context, src *model.ReaderBookSource) error {
|
||||
return r.db.WithContext(ctx).Create(src).Error
|
||||
}
|
||||
|
||||
// UpdateSource 更新书源字段。
|
||||
func (r *ReaderRepository) UpdateSource(ctx context.Context, src *model.ReaderBookSource) error {
|
||||
return r.db.WithContext(ctx).Save(src).Error
|
||||
}
|
||||
|
||||
// DeleteSource 删除书源(连带清理其会话状态)。
|
||||
func (r *ReaderRepository) DeleteSource(ctx context.Context, id string) error {
|
||||
return r.db.WithContext(ctx).Transaction(func(tx *gorm.DB) error {
|
||||
src := &model.ReaderBookSource{}
|
||||
if err := tx.First(src, "id = ?", id).Error; err == nil && src.SourceURL != "" {
|
||||
// 会话状态是「一源一条」,而 source_url 上有覆盖软删行的唯一索引:
|
||||
// 软删会让这一行继续占着 source_url,之后 SaveSourceState 的
|
||||
// First(默认排除软删行)查不到、Create 就会撞唯一约束,
|
||||
// 表现为「保存书源会话状态失败: UNIQUE constraint failed」,
|
||||
// cookie / 登录态从此再也存不进去。这里必须硬删。
|
||||
if err := tx.Unscoped().Delete(&model.ReaderSourceState{}, "source_url = ?", src.SourceURL).Error; err != nil {
|
||||
return err
|
||||
}
|
||||
}
|
||||
return tx.Delete(&model.ReaderBookSource{}, "id = ?", id).Error
|
||||
})
|
||||
}
|
||||
|
||||
// GetSourceState 取书源会话状态;不存在返回 (nil, nil)。
|
||||
func (r *ReaderRepository) GetSourceState(ctx context.Context, sourceURL string) (*model.ReaderSourceState, error) {
|
||||
var out model.ReaderSourceState
|
||||
err := r.db.WithContext(ctx).First(&out, "source_url = ?", sourceURL).Error
|
||||
if err != nil {
|
||||
if errors.Is(err, gorm.ErrRecordNotFound) {
|
||||
return nil, nil
|
||||
}
|
||||
return nil, err
|
||||
}
|
||||
return &out, nil
|
||||
}
|
||||
|
||||
// SaveSourceState 覆盖保存书源会话状态(不存在则新建)。
|
||||
func (r *ReaderRepository) SaveSourceState(ctx context.Context, st *model.ReaderSourceState) error {
|
||||
// 用 Unscoped 连软删行一起找:source_url 的唯一索引覆盖软删行,
|
||||
// 只按未删行查会漏掉历史行,随后 Create 必然撞唯一约束,
|
||||
// 结果就是该源的会话状态(含 cookie、登录态)永远保存失败。
|
||||
var existing model.ReaderSourceState
|
||||
err := r.db.WithContext(ctx).Unscoped().First(&existing, "source_url = ?", st.SourceURL).Error
|
||||
if errors.Is(err, gorm.ErrRecordNotFound) {
|
||||
return r.db.WithContext(ctx).Create(st).Error
|
||||
}
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
st.ID = existing.ID
|
||||
st.CreatedAt = existing.CreatedAt
|
||||
if existing.DeletedAt.Valid {
|
||||
// 历史行被软删过:连 deleted_at 一起写回,把它复活
|
||||
st.DeletedAt = gorm.DeletedAt{}
|
||||
return r.db.WithContext(ctx).Unscoped().Save(st).Error
|
||||
}
|
||||
return r.db.WithContext(ctx).Save(st).Error
|
||||
}
|
||||
|
||||
// ListBooks 用户书架(按 order 排序)。
|
||||
func (r *ReaderRepository) ListBooks(ctx context.Context, userID string) ([]model.ReaderBook, error) {
|
||||
var out []model.ReaderBook
|
||||
err := r.db.WithContext(ctx).Where("user_id = ?", userID).Order("`order` ASC, updated_at DESC").Find(&out).Error
|
||||
return out, err
|
||||
}
|
||||
|
||||
// GetBook 按 ID 取书。
|
||||
func (r *ReaderRepository) GetBook(ctx context.Context, id string) (*model.ReaderBook, error) {
|
||||
var out model.ReaderBook
|
||||
if err := r.db.WithContext(ctx).First(&out, "id = ?", id).Error; err != nil {
|
||||
return nil, err
|
||||
}
|
||||
return &out, nil
|
||||
}
|
||||
|
||||
// FindBookByURL 按用户 + 书源 + 书 URL 查书(加书架去重)。
|
||||
func (r *ReaderRepository) FindBookByURL(ctx context.Context, userID, origin, bookURL string) (*model.ReaderBook, error) {
|
||||
var out model.ReaderBook
|
||||
err := r.db.WithContext(ctx).First(&out, "user_id = ? AND origin = ? AND book_url = ?", userID, origin, bookURL).Error
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
return &out, nil
|
||||
}
|
||||
|
||||
// CreateBook / UpdateBook / DeleteBook。
|
||||
func (r *ReaderRepository) CreateBook(ctx context.Context, b *model.ReaderBook) error {
|
||||
return r.db.WithContext(ctx).Create(b).Error
|
||||
}
|
||||
|
||||
func (r *ReaderRepository) UpdateBook(ctx context.Context, b *model.ReaderBook) error {
|
||||
return r.db.WithContext(ctx).Save(b).Error
|
||||
}
|
||||
|
||||
func (r *ReaderRepository) DeleteBook(ctx context.Context, userID, id string) error {
|
||||
return r.db.WithContext(ctx).Transaction(func(tx *gorm.DB) error {
|
||||
if err := tx.Delete(&model.ReaderBook{}, "id = ? AND user_id = ?", id, userID).Error; err != nil {
|
||||
return err
|
||||
}
|
||||
return tx.Delete(&model.ReaderChapter{}, "book_id = ?", id).Error
|
||||
})
|
||||
}
|
||||
|
||||
// ReplaceChapters 覆盖式刷新章节列表。
|
||||
func (r *ReaderRepository) ReplaceChapters(ctx context.Context, bookID string, chapters []model.ReaderChapter) error {
|
||||
return r.db.WithContext(ctx).Transaction(func(tx *gorm.DB) error {
|
||||
// 章节是纯缓存(软删会留下行,撞上 (book_id, index) 唯一索引),这里物理删除
|
||||
if err := tx.Unscoped().Delete(&model.ReaderChapter{}, "book_id = ?", bookID).Error; err != nil {
|
||||
return err
|
||||
}
|
||||
if len(chapters) == 0 {
|
||||
return nil
|
||||
}
|
||||
return tx.CreateInBatches(chapters, 500).Error
|
||||
})
|
||||
}
|
||||
|
||||
// ListChapters 按序取章节。
|
||||
func (r *ReaderRepository) ListChapters(ctx context.Context, bookID string) ([]model.ReaderChapter, error) {
|
||||
var out []model.ReaderChapter
|
||||
err := r.db.WithContext(ctx).Where("book_id = ?", bookID).Order("`index` ASC").Find(&out).Error
|
||||
return out, err
|
||||
}
|
||||
|
||||
// CountChaptersByBook 一次统计多本书已缓存的章节数(书架显示未读章数用,避免逐本查询)。
|
||||
// 没有目录缓存的书籍不会出现在返回结果里。
|
||||
func (r *ReaderRepository) CountChaptersByBook(ctx context.Context, bookIDs []string) (map[string]int, error) {
|
||||
out := make(map[string]int, len(bookIDs))
|
||||
if len(bookIDs) == 0 {
|
||||
return out, nil
|
||||
}
|
||||
var rows []struct {
|
||||
BookID string
|
||||
Total int
|
||||
}
|
||||
err := r.db.WithContext(ctx).
|
||||
Model(&model.ReaderChapter{}).
|
||||
Select("book_id, COUNT(*) AS total").
|
||||
Where("book_id IN ?", bookIDs).
|
||||
Group("book_id").
|
||||
Scan(&rows).Error
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
for _, row := range rows {
|
||||
out[row.BookID] = row.Total
|
||||
}
|
||||
return out, nil
|
||||
}
|
||||
|
||||
// GetChapter 取指定章节。
|
||||
func (r *ReaderRepository) GetChapter(ctx context.Context, bookID string, index int) (*model.ReaderChapter, error) {
|
||||
var out model.ReaderChapter
|
||||
if err := r.db.WithContext(ctx).First(&out, "book_id = ? AND `index` = ?", bookID, index).Error; err != nil {
|
||||
return nil, err
|
||||
}
|
||||
return &out, nil
|
||||
}
|
||||
|
||||
// GetReaderProfile 取用户阅读器偏好;没有记录时返回 (nil, nil),
|
||||
// 由调用方决定是否用前端上送的当前值播种。
|
||||
func (r *ReaderRepository) GetReaderProfile(ctx context.Context, userID string) (*model.ReaderProfile, error) {
|
||||
var out model.ReaderProfile
|
||||
err := r.db.WithContext(ctx).First(&out, "user_id = ?", userID).Error
|
||||
if err != nil {
|
||||
if errors.Is(err, gorm.ErrRecordNotFound) {
|
||||
return nil, nil
|
||||
}
|
||||
return nil, err
|
||||
}
|
||||
return &out, nil
|
||||
}
|
||||
|
||||
// SaveReaderProfile 覆盖保存用户阅读器偏好(不存在则新建)。
|
||||
func (r *ReaderRepository) SaveReaderProfile(ctx context.Context, p *model.ReaderProfile) error {
|
||||
var existing model.ReaderProfile
|
||||
err := r.db.WithContext(ctx).First(&existing, "user_id = ?", p.UserID).Error
|
||||
if errors.Is(err, gorm.ErrRecordNotFound) {
|
||||
return r.db.WithContext(ctx).Create(p).Error
|
||||
}
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
p.ID = existing.ID
|
||||
p.CreatedAt = existing.CreatedAt
|
||||
return r.db.WithContext(ctx).Save(p).Error
|
||||
}
|
||||
|
||||
// GetBookGroups 取用户的书架分组;没有记录时返回 (nil, nil)。
|
||||
func (r *ReaderRepository) GetBookGroups(ctx context.Context, userID string) (*model.ReaderBookGroups, error) {
|
||||
var out model.ReaderBookGroups
|
||||
err := r.db.WithContext(ctx).First(&out, "user_id = ?", userID).Error
|
||||
if err != nil {
|
||||
if errors.Is(err, gorm.ErrRecordNotFound) {
|
||||
return nil, nil
|
||||
}
|
||||
return nil, err
|
||||
}
|
||||
return &out, nil
|
||||
}
|
||||
|
||||
// SaveBookGroups 覆盖保存用户的书架分组(不存在则新建)。
|
||||
func (r *ReaderRepository) SaveBookGroups(ctx context.Context, row *model.ReaderBookGroups) error {
|
||||
var existing model.ReaderBookGroups
|
||||
err := r.db.WithContext(ctx).First(&existing, "user_id = ?", row.UserID).Error
|
||||
if errors.Is(err, gorm.ErrRecordNotFound) {
|
||||
return r.db.WithContext(ctx).Create(row).Error
|
||||
}
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
row.ID = existing.ID
|
||||
row.CreatedAt = existing.CreatedAt
|
||||
return r.db.WithContext(ctx).Save(row).Error
|
||||
}
|
||||
|
||||
// ListReplaceRules 用户替换规则(按 order 排序)。
|
||||
func (r *ReaderRepository) ListReplaceRules(ctx context.Context, userID string) ([]model.ReaderReplaceRule, error) {
|
||||
var out []model.ReaderReplaceRule
|
||||
err := r.db.WithContext(ctx).Where("user_id = ?", userID).Order("`order` DESC, created_at ASC").Find(&out).Error
|
||||
return out, err
|
||||
}
|
||||
|
||||
// CreateReplaceRule / UpdateReplaceRule / DeleteReplaceRule。
|
||||
func (r *ReaderRepository) CreateReplaceRule(ctx context.Context, rule *model.ReaderReplaceRule) error {
|
||||
return r.db.WithContext(ctx).Create(rule).Error
|
||||
}
|
||||
|
||||
func (r *ReaderRepository) UpdateReplaceRule(ctx context.Context, rule *model.ReaderReplaceRule) error {
|
||||
return r.db.WithContext(ctx).Save(rule).Error
|
||||
}
|
||||
|
||||
func (r *ReaderRepository) DeleteReplaceRule(ctx context.Context, userID, id string) error {
|
||||
return r.db.WithContext(ctx).Delete(&model.ReaderReplaceRule{}, "id = ? AND user_id = ?", id, userID).Error
|
||||
}
|
||||
@@ -0,0 +1,207 @@
|
||||
package repository
|
||||
|
||||
import (
|
||||
"errors"
|
||||
"testing"
|
||||
"time"
|
||||
|
||||
"github.com/glebarez/sqlite"
|
||||
"gorm.io/gorm"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/database"
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
)
|
||||
|
||||
func newReaderTestRepo(t *testing.T) *ReaderRepository {
|
||||
t.Helper()
|
||||
db, err := gorm.Open(sqlite.Open(":memory:"), &gorm.Config{})
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if err := database.AutoMigrate(db); err != nil {
|
||||
t.Fatalf("migrate: %v", err)
|
||||
}
|
||||
return New(db).Reader
|
||||
}
|
||||
|
||||
// TestSourceStateRoundTrip 验证会话状态按书源 URL 存取。
|
||||
func TestSourceStateRoundTrip(t *testing.T) {
|
||||
repo := newReaderTestRepo(t)
|
||||
ctx := t.Context()
|
||||
|
||||
// 不存在时返回 (nil, nil),调用方据此走默认值
|
||||
got, err := repo.GetSourceState(ctx, "https://a.example.com")
|
||||
if err != nil || got != nil {
|
||||
t.Fatalf("首次读取应为空: got=%v err=%v", got, err)
|
||||
}
|
||||
|
||||
st := &model.ReaderSourceState{
|
||||
SourceURL: "https://a.example.com",
|
||||
Variable: `{"线路":"v1"}`,
|
||||
LoginInfo: `{"邮箱":"u@e.com"}`,
|
||||
LoginHeader: `{"X-Token":"t"}`,
|
||||
Cookies: `{"example.com":"qttoken=abc"}`,
|
||||
}
|
||||
if err := repo.SaveSourceState(ctx, st); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
got, err = repo.GetSourceState(ctx, "https://a.example.com")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got.LoginInfo != st.LoginInfo || got.Variable != st.Variable {
|
||||
t.Fatalf("读回不一致: %+v", got)
|
||||
}
|
||||
|
||||
// 再次保存应为更新而非插入(同 URL 唯一)
|
||||
st.LoginInfo = `{"邮箱":"new@e.com"}`
|
||||
if err := repo.SaveSourceState(ctx, st); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
var count int64
|
||||
if err := repo.db.Model(&model.ReaderSourceState{}).
|
||||
Where("source_url = ?", "https://a.example.com").Count(&count).Error; err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if count != 1 {
|
||||
t.Fatalf("同一书源应只有一行状态,实际 %d", count)
|
||||
}
|
||||
got, _ = repo.GetSourceState(ctx, "https://a.example.com")
|
||||
if got.LoginInfo != `{"邮箱":"new@e.com"}` {
|
||||
t.Fatalf("更新未生效: %+v", got)
|
||||
}
|
||||
}
|
||||
|
||||
// TestDeleteSourceAlsoClearsState 删除书源应连带清理登录态。
|
||||
func TestDeleteSourceAlsoClearsState(t *testing.T) {
|
||||
repo := newReaderTestRepo(t)
|
||||
ctx := t.Context()
|
||||
|
||||
src := &model.ReaderBookSource{Name: "源", SourceURL: "https://b.example.com", Enabled: true}
|
||||
if err := repo.CreateSource(ctx, src); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if err := repo.SaveSourceState(ctx, &model.ReaderSourceState{
|
||||
SourceURL: "https://b.example.com", LoginInfo: `{"a":"b"}`,
|
||||
}); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
if err := repo.DeleteSource(ctx, src.ID); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
st, err := repo.GetSourceState(ctx, "https://b.example.com")
|
||||
if err != nil && !errors.Is(err, gorm.ErrRecordNotFound) {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if st != nil {
|
||||
t.Fatalf("书源删除后登录态应一并清理,实际仍在: %+v", st)
|
||||
}
|
||||
}
|
||||
|
||||
// TestSaveSourceStateAfterSourceDeleted 回归:书源删除后,同一 source_url 的状态必须还能存。
|
||||
//
|
||||
// source_url 上的唯一索引覆盖软删行,而 DeleteSource 曾经只做软删:
|
||||
// 之后 SaveSourceState 的 First(默认排除软删行)查不到 → Create 撞唯一约束,
|
||||
// 表现为「保存书源会话状态失败: UNIQUE constraint failed」,
|
||||
// cookie / 登录态从此永远存不进去(删过或重导入过的书源必现)。
|
||||
func TestSaveSourceStateAfterSourceDeleted(t *testing.T) {
|
||||
repo := newReaderTestRepo(t)
|
||||
ctx := t.Context()
|
||||
const url = "https://d.example.com"
|
||||
|
||||
src := &model.ReaderBookSource{Name: "源", SourceURL: url, Enabled: true}
|
||||
if err := repo.CreateSource(ctx, src); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if err := repo.SaveSourceState(ctx, &model.ReaderSourceState{
|
||||
SourceURL: url, Cookies: `{"example.com":"a=1"}`,
|
||||
}); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if err := repo.DeleteSource(ctx, src.ID); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
// 重新导入同 URL 的书源,再存一次状态
|
||||
again := &model.ReaderBookSource{Name: "源", SourceURL: url, Enabled: true}
|
||||
if err := repo.CreateSource(ctx, again); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if err := repo.SaveSourceState(ctx, &model.ReaderSourceState{
|
||||
SourceURL: url, Cookies: `{"example.com":"a=2"}`,
|
||||
}); err != nil {
|
||||
t.Fatalf("书源删除/重导入后状态必须还能保存: %v", err)
|
||||
}
|
||||
st, err := repo.GetSourceState(ctx, url)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if st == nil || st.Cookies != `{"example.com":"a=2"}` {
|
||||
t.Fatalf("状态未写入: %+v", st)
|
||||
}
|
||||
}
|
||||
|
||||
// TestSaveSourceStateRevivesSoftDeletedRow 回归:老数据里已被软删的状态行要能复活。
|
||||
//
|
||||
// 修复前 DeleteSource 是软删,历史库里可能已经躺着软删行;
|
||||
// 保存时必须把它救活,不能去 INSERT 撞唯一约束。
|
||||
func TestSaveSourceStateRevivesSoftDeletedRow(t *testing.T) {
|
||||
repo := newReaderTestRepo(t)
|
||||
ctx := t.Context()
|
||||
const url = "https://e.example.com"
|
||||
|
||||
if err := repo.SaveSourceState(ctx, &model.ReaderSourceState{
|
||||
SourceURL: url, Cookies: `{"example.com":"old=1"}`,
|
||||
}); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
// 造出老版本 DeleteSource 留下的样子:行还在但被软删
|
||||
if err := repo.db.Exec(
|
||||
"update reader_source_states set deleted_at = ? where source_url = ?",
|
||||
time.Now(), url).Error; err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
if err := repo.SaveSourceState(ctx, &model.ReaderSourceState{
|
||||
SourceURL: url, Cookies: `{"example.com":"new=2"}`,
|
||||
}); err != nil {
|
||||
t.Fatalf("软删过的历史行应被复活,而不是撞唯一约束: %v", err)
|
||||
}
|
||||
st, err := repo.GetSourceState(ctx, url)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if st == nil || st.Cookies != `{"example.com":"new=2"}` {
|
||||
t.Fatalf("复活后状态未写入: %+v", st)
|
||||
}
|
||||
}
|
||||
|
||||
// TestImportUpdateKeepsSourceState 重新导入书源(更新 RawJSON)不应丢失登录态。
|
||||
func TestImportUpdateKeepsSourceState(t *testing.T) {
|
||||
repo := newReaderTestRepo(t)
|
||||
ctx := t.Context()
|
||||
|
||||
src := &model.ReaderBookSource{Name: "源", SourceURL: "https://c.example.com", RawJSON: `{"v":1}`}
|
||||
if err := repo.CreateSource(ctx, src); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if err := repo.SaveSourceState(ctx, &model.ReaderSourceState{
|
||||
SourceURL: "https://c.example.com", LoginInfo: `{"k":"keep"}`,
|
||||
}); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
// 模拟导入更新:只改书源自身字段
|
||||
src.RawJSON = `{"v":2}`
|
||||
if err := repo.UpdateSource(ctx, src); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
st, err := repo.GetSourceState(ctx, "https://c.example.com")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if st == nil || st.LoginInfo != `{"k":"keep"}` {
|
||||
t.Fatalf("更新书源后登录态丢失: %+v", st)
|
||||
}
|
||||
}
|
||||
@@ -36,6 +36,7 @@ type Container struct {
|
||||
StrmDirCache *StrmDirCacheRepository
|
||||
ScrapeTask *ScrapeTaskRepository
|
||||
EmbyMount *EmbyMountRepository
|
||||
Reader *ReaderRepository
|
||||
}
|
||||
|
||||
// New 将每个 repository 连接到单个 *gorm.DB。
|
||||
@@ -68,5 +69,6 @@ func New(db *gorm.DB) *Container {
|
||||
StrmDirCache: &StrmDirCacheRepository{db: db},
|
||||
ScrapeTask: &ScrapeTaskRepository{db: db},
|
||||
EmbyMount: &EmbyMountRepository{db: db},
|
||||
Reader: &ReaderRepository{db: db},
|
||||
}
|
||||
}
|
||||
|
||||
+15
-59
@@ -12,96 +12,52 @@
|
||||
package service
|
||||
|
||||
import (
|
||||
"crypto/aes"
|
||||
"crypto/cipher"
|
||||
"crypto/rand"
|
||||
"crypto/sha256"
|
||||
"encoding/base64"
|
||||
"errors"
|
||||
"strings"
|
||||
|
||||
"go.uber.org/zap"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/helper"
|
||||
)
|
||||
|
||||
// encPrefix tags ciphertext rows so we can tell them apart from legacy
|
||||
// plaintext values.
|
||||
const encPrefix = "enc:v1:"
|
||||
// plaintext values. Kept as an alias of the shared helper's prefix so both
|
||||
// implementations stay wire-compatible.
|
||||
const encPrefix = helper.EncPrefix
|
||||
|
||||
// CryptoService wraps an AES-GCM cipher derived from a stable per-install
|
||||
// secret (the JWT secret).
|
||||
//
|
||||
// The cipher itself lives in helper.SecretCipher so lower layers (e.g. the
|
||||
// reader subsystem, which cannot import this package) can share one
|
||||
// implementation; this type keeps the service-layer logging and API.
|
||||
type CryptoService struct {
|
||||
log *zap.Logger
|
||||
aead cipher.AEAD
|
||||
log *zap.Logger
|
||||
cipher *helper.SecretCipher
|
||||
}
|
||||
|
||||
// NewCryptoService derives a 256-bit key from the given secret via
|
||||
// SHA-256 and constructs an AES-GCM AEAD. Empty secrets yield a service
|
||||
// whose Encrypt/Decrypt methods are pass-throughs (used in unit tests).
|
||||
func NewCryptoService(secret string, log *zap.Logger) *CryptoService {
|
||||
c := &CryptoService{log: log}
|
||||
if strings.TrimSpace(secret) == "" {
|
||||
return c
|
||||
}
|
||||
sum := sha256.Sum256([]byte(secret))
|
||||
block, err := aes.NewCipher(sum[:])
|
||||
if err != nil {
|
||||
log.Error("crypto: aes.NewCipher", zap.Error(err))
|
||||
return c
|
||||
}
|
||||
aead, err := cipher.NewGCM(block)
|
||||
if err != nil {
|
||||
log.Error("crypto: cipher.NewGCM", zap.Error(err))
|
||||
return c
|
||||
}
|
||||
c.aead = aead
|
||||
return c
|
||||
return &CryptoService{log: log, cipher: helper.NewSecretCipher(secret)}
|
||||
}
|
||||
|
||||
// Encrypt returns the base64-encoded ciphertext (with prefix) for plain.
|
||||
// Empty inputs round-trip unchanged.
|
||||
func (c *CryptoService) Encrypt(plain string) string {
|
||||
if plain == "" || c.aead == nil {
|
||||
return plain
|
||||
}
|
||||
if strings.HasPrefix(plain, encPrefix) {
|
||||
return plain
|
||||
}
|
||||
nonce := make([]byte, c.aead.NonceSize())
|
||||
if _, err := rand.Read(nonce); err != nil {
|
||||
return plain
|
||||
}
|
||||
cipherBytes := c.aead.Seal(nonce, nonce, []byte(plain), nil)
|
||||
return encPrefix + base64.StdEncoding.EncodeToString(cipherBytes)
|
||||
return c.cipher.Encrypt(plain)
|
||||
}
|
||||
|
||||
// Decrypt returns the plaintext for an encrypted value. Plaintext rows
|
||||
// (no prefix) are returned unchanged.
|
||||
func (c *CryptoService) Decrypt(value string) string {
|
||||
if c == nil || value == "" || c.aead == nil {
|
||||
return value
|
||||
}
|
||||
if !strings.HasPrefix(value, encPrefix) {
|
||||
return value
|
||||
}
|
||||
raw := strings.TrimPrefix(value, encPrefix)
|
||||
data, err := base64.StdEncoding.DecodeString(raw)
|
||||
if err != nil {
|
||||
return value
|
||||
}
|
||||
if len(data) < c.aead.NonceSize() {
|
||||
return value
|
||||
}
|
||||
nonce, cipherBytes := data[:c.aead.NonceSize()], data[c.aead.NonceSize():]
|
||||
plain, err := c.aead.Open(nil, nonce, cipherBytes, nil)
|
||||
if err != nil {
|
||||
return value
|
||||
}
|
||||
return string(plain)
|
||||
return c.cipher.Decrypt(value)
|
||||
}
|
||||
|
||||
// IsEncrypted returns true if value carries the encrypted prefix.
|
||||
func (c *CryptoService) IsEncrypted(value string) bool {
|
||||
return strings.HasPrefix(value, encPrefix)
|
||||
return c.cipher.IsEncrypted(value)
|
||||
}
|
||||
|
||||
// MaskAPIKey returns "abcd****wxyz" so the key can be displayed in the
|
||||
|
||||
@@ -66,6 +66,14 @@ var ErrPathOutOfBounds = errors.New("path is outside the allowed roots")
|
||||
// ErrRootMutation protects configured roots such as /media and /downloads.
|
||||
var ErrRootMutation = errors.New("refusing to mutate an allowed root")
|
||||
|
||||
// ResolvePath validates that path sits inside one of the allowed storage roots
|
||||
// and returns its absolute form. Callers use it to accept a server-side path
|
||||
// from the UI (file pickers) without re-implementing the boundary check.
|
||||
func (s *FileManagerService) ResolvePath(path string) (string, error) {
|
||||
abs, _, err := s.requireAllowedPath(path, false)
|
||||
return abs, err
|
||||
}
|
||||
|
||||
// List enumerates a directory under one of the allowed roots, returning up to
|
||||
// maxEntries items sorted by (dir-first, path). Recursive listing is capped by
|
||||
// maxEntries to avoid accidentally walking huge NAS trees from the UI.
|
||||
|
||||
@@ -153,6 +153,37 @@ func TestFileManagerRefusesRootMutation(t *testing.T) {
|
||||
}
|
||||
}
|
||||
|
||||
// TestFileManagerResolvePath 阅读模块的服务器选书/选目录入口靠它做边界校验:
|
||||
// 允许根目录内放行,越界一律拒绝。
|
||||
func TestFileManagerResolvePath(t *testing.T) {
|
||||
root := t.TempDir()
|
||||
svc := newFileManagerTestService(t, root)
|
||||
|
||||
inside := filepath.Join(root, "books", "小说.txt")
|
||||
if err := os.MkdirAll(filepath.Dir(inside), 0o755); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if err := os.WriteFile(inside, []byte("正文"), 0o644); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
got, err := svc.ResolvePath(inside)
|
||||
if err != nil {
|
||||
t.Fatalf("允许根目录内的路径应放行: %v", err)
|
||||
}
|
||||
if got != inside {
|
||||
t.Fatalf("ResolvePath = %q,期望 %q", got, inside)
|
||||
}
|
||||
|
||||
outside := filepath.Join(t.TempDir(), "越界.txt")
|
||||
if err := os.WriteFile(outside, []byte("x"), 0o644); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if _, err := svc.ResolvePath(outside); !errors.Is(err, ErrPathOutOfBounds) {
|
||||
t.Fatalf("越界路径 err = %v, want ErrPathOutOfBounds", err)
|
||||
}
|
||||
}
|
||||
|
||||
func hardlinksUnsupported(t *testing.T, root string) bool {
|
||||
t.Helper()
|
||||
src := filepath.Join(root, "hardlink-probe-src")
|
||||
|
||||
@@ -0,0 +1,44 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"testing"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
)
|
||||
|
||||
// TestSaveAudioConfigPersistsCredits 听书跳过片头/片尾设置要落库,且只能改自己的书。
|
||||
func TestSaveAudioConfigPersistsCredits(t *testing.T) {
|
||||
svc, repos := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
|
||||
book := &model.ReaderBook{UserID: "u1", Name: "宠魅", Type: 1}
|
||||
if err := repos.Reader.CreateBook(ctx, book); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
if err := svc.SaveAudioConfig(ctx, "u1", book.ID, 30, 15); err != nil {
|
||||
t.Fatalf("保存听书设置失败: %v", err)
|
||||
}
|
||||
got, err := svc.GetBook(ctx, book.ID)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got.OpenCredits != 30 || got.CloseCredits != 15 {
|
||||
t.Fatalf("片头/片尾 = %d/%d,期望 30/15", got.OpenCredits, got.CloseCredits)
|
||||
}
|
||||
|
||||
// 0 是合法值(不跳过),必须能写回
|
||||
if err := svc.SaveAudioConfig(ctx, "u1", book.ID, 0, 0); err != nil {
|
||||
t.Fatalf("清零失败: %v", err)
|
||||
}
|
||||
if got, _ = svc.GetBook(ctx, book.ID); got.OpenCredits != 0 || got.CloseCredits != 0 {
|
||||
t.Fatalf("清零后 = %d/%d,期望 0/0", got.OpenCredits, got.CloseCredits)
|
||||
}
|
||||
|
||||
if err := svc.SaveAudioConfig(ctx, "u1", book.ID, -1, 0); err == nil {
|
||||
t.Fatal("负数应被拒绝")
|
||||
}
|
||||
if err := svc.SaveAudioConfig(ctx, "other", book.ID, 10, 10); err == nil {
|
||||
t.Fatal("他人书架应被拒绝")
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,372 @@
|
||||
// 本地/远端有声书的按需转码:源文件是浏览器解不了的格式(典型是 WMA/ASF)时,
|
||||
// 用 ffmpeg 转成 mp3 落到缓存目录,再按 Range 下发。
|
||||
//
|
||||
// 为什么落盘而不是实时流式输出:有声书必须能拖动进度,而 http.ServeContent 的
|
||||
// Range 支持要求可随机读取的文件;实时管道没有长度信息,一旦 seek 就废掉。
|
||||
// 一章通常几 MB,转一次几秒,之后同一章秒开。
|
||||
package reader
|
||||
|
||||
import (
|
||||
"context"
|
||||
"crypto/hmac"
|
||||
"crypto/sha256"
|
||||
"encoding/base64"
|
||||
"encoding/hex"
|
||||
"encoding/json"
|
||||
"errors"
|
||||
"fmt"
|
||||
"net/url"
|
||||
"os"
|
||||
"os/exec"
|
||||
"path/filepath"
|
||||
"sort"
|
||||
"strings"
|
||||
"sync"
|
||||
"time"
|
||||
|
||||
"golang.org/x/sync/singleflight"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/helper"
|
||||
)
|
||||
|
||||
// audioTranscodeFlight 同一章节并发命中时只跑一次 ffmpeg(按缓存路径去重)。
|
||||
var audioTranscodeFlight singleflight.Group
|
||||
|
||||
// 转码互斥表:singleflight 只保证同一 key 不重复执行,prune 与写文件仍需串行。
|
||||
var audioTranscodeMu sync.Mutex
|
||||
|
||||
const (
|
||||
// audioTranscodeDirName 缓存目录名,位于 cache.cache_dir 之下。
|
||||
audioTranscodeDirName = "reader-audio"
|
||||
// maxAudioTranscodeCacheBytes 转码缓存上限,超出按修改时间淘汰最旧的。
|
||||
maxAudioTranscodeCacheBytes = 4 << 30
|
||||
// audioTranscodeKeepRatio 触发淘汰后回落到上限的比例,避免每次写入都淘汰。
|
||||
audioTranscodeKeepRatio = 0.9
|
||||
// maxAudioTranscodeDuration 单章转码超时。
|
||||
maxAudioTranscodeDuration = 15 * time.Minute
|
||||
// audioTranscodeBitrate 语音内容 96k 足够,体积约为原 WMA 的两倍以内。
|
||||
audioTranscodeBitrate = "96k"
|
||||
)
|
||||
|
||||
// needsTranscodeAudioExt 明确需要转码的容器/编码(浏览器都无法直接解码)。
|
||||
// 只列已知有问题的:未列出的格式维持原样直出,避免把本来能播的流也拖去转码。
|
||||
var needsTranscodeAudioExt = map[string]bool{
|
||||
".wma": true, ".asf": true, ".wmv": true, ".ape": true, ".wv": true,
|
||||
".ac3": true, ".dts": true, ".amr": true, ".tta": true, ".dsf": true, ".dff": true,
|
||||
}
|
||||
|
||||
// audioSourceExt 取音频地址的扩展名(小写带点)。
|
||||
//
|
||||
// 既要认本地路径(D:\x\a.wma),也要认带 query 的远端地址
|
||||
// (…/video.wma?acct=…)——后者直接 filepath.Ext 会把 query 一起算进去。
|
||||
func audioSourceExt(source string) string {
|
||||
raw := strings.TrimSpace(source)
|
||||
if raw == "" {
|
||||
return ""
|
||||
}
|
||||
// scheme 长度 >1 才算真 URL:Windows 盘符(D:\…)会被 url.Parse 当成单字符 scheme
|
||||
if u, err := url.Parse(raw); err == nil && len(u.Scheme) > 1 {
|
||||
return strings.ToLower(filepath.Ext(u.Path))
|
||||
}
|
||||
if i := strings.IndexAny(raw, "?#"); i >= 0 {
|
||||
raw = raw[:i]
|
||||
}
|
||||
return strings.ToLower(filepath.Ext(raw))
|
||||
}
|
||||
|
||||
// needsAudioTranscode 判断该音频是否必须转码后才能给浏览器播放。
|
||||
func needsAudioTranscode(source string) bool {
|
||||
return needsTranscodeAudioExt[audioSourceExt(source)]
|
||||
}
|
||||
|
||||
// ffmpegBinary 解析可用的 ffmpeg 可执行文件;找不到返回空串。
|
||||
func (s *ReaderService) ffmpegBinary() string {
|
||||
path := strings.TrimSpace(s.cfg.App.FFmpegPath)
|
||||
if path == "" {
|
||||
path = "ffmpeg"
|
||||
}
|
||||
resolved, err := exec.LookPath(path)
|
||||
if err != nil {
|
||||
return ""
|
||||
}
|
||||
return resolved
|
||||
}
|
||||
|
||||
// ErrAudioTranscodeUnavailable 服务器没有 ffmpeg,无法转码该格式。
|
||||
var ErrAudioTranscodeUnavailable = errors.New("服务器未安装 ffmpeg")
|
||||
|
||||
// transcodeMissingFFmpegError 给前端一条能直接照做的提示。
|
||||
func transcodeMissingFFmpegError(source string) error {
|
||||
ext := strings.TrimPrefix(audioSourceExt(source), ".")
|
||||
if ext == "" {
|
||||
ext = "该"
|
||||
}
|
||||
return fmt.Errorf("这个音频是 %s 格式,浏览器无法直接播放,需要服务器转码;%w,请在设置里配置 app.ffmpeg_path 或安装 ffmpeg",
|
||||
strings.ToUpper(ext), ErrAudioTranscodeUnavailable)
|
||||
}
|
||||
|
||||
// ─── 签名(与本地音频同一套:<audio src> 带不上 JWT) ─────────────────────
|
||||
|
||||
func (s *ReaderService) signAudioTranscode(bookID, source string) string {
|
||||
mac := hmac.New(sha256.New, []byte(s.cfg.Secrets.JWTSecret))
|
||||
mac.Write([]byte(bookID + "|transcode|" + source))
|
||||
return hex.EncodeToString(mac.Sum(nil))[:32]
|
||||
}
|
||||
|
||||
// AudioTranscodeURL 需要转码的音频地址(签名代理)。
|
||||
func (s *ReaderService) AudioTranscodeURL(bookID, source string) string {
|
||||
return "/api/reader/audio/transcode?b=" + url.QueryEscape(bookID) +
|
||||
"&u=" + base64.RawURLEncoding.EncodeToString([]byte(source)) +
|
||||
"&s=" + s.signAudioTranscode(bookID, source)
|
||||
}
|
||||
|
||||
// VerifyAudioTranscodeURL 校验签名并还原原始音频地址(本地路径或远端 URL)。
|
||||
func (s *ReaderService) VerifyAudioTranscodeURL(bookID, encoded, sig string) (string, error) {
|
||||
raw, err := base64.RawURLEncoding.DecodeString(encoded)
|
||||
if err != nil {
|
||||
return "", fmt.Errorf("转码地址解码失败")
|
||||
}
|
||||
expect := s.signAudioTranscode(bookID, string(raw))
|
||||
if !hmac.Equal([]byte(expect), []byte(sig)) {
|
||||
return "", fmt.Errorf("转码签名校验失败")
|
||||
}
|
||||
return string(raw), nil
|
||||
}
|
||||
|
||||
// ─── 缓存与转码 ──────────────────────────────────────────────────────────
|
||||
|
||||
func (s *ReaderService) audioTranscodeDir() (string, error) {
|
||||
base := strings.TrimSpace(s.cfg.Cache.CacheDir)
|
||||
if base == "" {
|
||||
base = filepath.Join(s.cfg.App.DataDir, "cache")
|
||||
}
|
||||
dir := filepath.Join(base, audioTranscodeDirName)
|
||||
if err := os.MkdirAll(dir, 0o750); err != nil {
|
||||
return "", fmt.Errorf("创建转码缓存目录失败: %w", err)
|
||||
}
|
||||
return dir, nil
|
||||
}
|
||||
|
||||
// audioTranscodeCachePath 缓存文件名由(书 + 源地址)哈希决定:同一章重复播放直接命中。
|
||||
func audioTranscodeCachePath(dir, bookID, source string) string {
|
||||
sum := sha256.Sum256([]byte(bookID + "|" + source))
|
||||
return filepath.Join(dir, hex.EncodeToString(sum[:16])+".mp3")
|
||||
}
|
||||
|
||||
// AudioTranscodeCachePath 返回该音轨的转码缓存文件路径(不触发转码),
|
||||
// 供测试与运维排查使用。
|
||||
func (s *ReaderService) AudioTranscodeCachePath(bookID, source string) (string, error) {
|
||||
dir, err := s.audioTranscodeDir()
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
return audioTranscodeCachePath(dir, bookID, source), nil
|
||||
}
|
||||
|
||||
// EnsureTranscodedAudio 确保该音频已有转码结果,返回可 Range 下发的 mp3 路径。
|
||||
func (s *ReaderService) EnsureTranscodedAudio(ctx context.Context, bookID, source string) (string, error) {
|
||||
if strings.TrimSpace(source) == "" {
|
||||
return "", fmt.Errorf("缺少音频地址")
|
||||
}
|
||||
dir, err := s.audioTranscodeDir()
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
dst := audioTranscodeCachePath(dir, bookID, source)
|
||||
if ok := touchCachedAudio(dst); ok {
|
||||
return dst, nil
|
||||
}
|
||||
if s.ffmpegBinary() == "" {
|
||||
return "", transcodeMissingFFmpegError(source)
|
||||
}
|
||||
|
||||
v, err, _ := audioTranscodeFlight.Do(dst, func() (any, error) {
|
||||
// 排队期间别的请求可能已经转好了
|
||||
if ok := touchCachedAudio(dst); ok {
|
||||
return dst, nil
|
||||
}
|
||||
if err := s.transcodeAudioFile(ctx, bookID, source, dst, dir); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
return dst, nil
|
||||
})
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
path, _ := v.(string)
|
||||
if path == "" {
|
||||
return "", fmt.Errorf("转码结果不可用")
|
||||
}
|
||||
return path, nil
|
||||
}
|
||||
|
||||
// touchCachedAudio 命中缓存时刷新访问时间,作为 LRU 依据。
|
||||
func touchCachedAudio(path string) bool {
|
||||
info, err := os.Stat(path)
|
||||
if err != nil || info.IsDir() || info.Size() == 0 {
|
||||
return false
|
||||
}
|
||||
now := time.Now()
|
||||
_ = os.Chtimes(path, now, now)
|
||||
return true
|
||||
}
|
||||
|
||||
// transcodeAudioFile 跑一次 ffmpeg 并原子落盘(先写 .part 再 rename)。
|
||||
func (s *ReaderService) transcodeAudioFile(ctx context.Context, bookID, source, dst, dir string) error {
|
||||
ffmpeg := s.ffmpegBinary()
|
||||
if ffmpeg == "" {
|
||||
return transcodeMissingFFmpegError(source)
|
||||
}
|
||||
remote := isRemoteMediaURL(source)
|
||||
if !remote {
|
||||
if info, err := os.Stat(source); err != nil || info.IsDir() {
|
||||
return fmt.Errorf("音频文件已丢失")
|
||||
}
|
||||
}
|
||||
|
||||
// 转码与淘汰串行,避免边写边删
|
||||
audioTranscodeMu.Lock()
|
||||
pruneAudioTranscodeCache(dir, maxAudioTranscodeCacheBytes)
|
||||
audioTranscodeMu.Unlock()
|
||||
|
||||
tmp := dst + ".part"
|
||||
defer os.Remove(tmp)
|
||||
|
||||
headers := s.transcodeInputHeaders(ctx, bookID, source)
|
||||
// 客户端断开不该杀掉已开始的转码:产物对下次播放仍然有用
|
||||
runCtx, cancel := context.WithTimeout(context.WithoutCancel(ctx), maxAudioTranscodeDuration)
|
||||
defer cancel()
|
||||
|
||||
cmd := exec.CommandContext(runCtx, ffmpeg, buildFFmpegAudioArgs(source, tmp, headers)...)
|
||||
out, err := cmd.CombinedOutput()
|
||||
if err != nil {
|
||||
if runCtx.Err() == context.DeadlineExceeded {
|
||||
return fmt.Errorf("转码超时(超过 %s)", maxAudioTranscodeDuration)
|
||||
}
|
||||
return fmt.Errorf("转码失败: %v: %s", err, truncateForError(string(out), 300))
|
||||
}
|
||||
info, err := os.Stat(tmp)
|
||||
if err != nil || info.Size() == 0 {
|
||||
return fmt.Errorf("转码没有产生有效输出")
|
||||
}
|
||||
if err := os.Rename(tmp, dst); err != nil {
|
||||
return fmt.Errorf("写入转码缓存失败: %w", err)
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
// transcodeInputHeaders 远端音源要带的请求头:浏览器 UA 预设 + 书源请求头 + 登录态。
|
||||
// 本机 /api/strm 这类内网地址不需要,但书源直链类的有声书要靠它才能取到。
|
||||
func (s *ReaderService) transcodeInputHeaders(ctx context.Context, bookID, source string) map[string]string {
|
||||
if !isRemoteMediaURL(source) || s.repo == nil {
|
||||
return nil
|
||||
}
|
||||
headers := map[string]string{}
|
||||
for k, v := range helper.HTTPHeaderPresets() {
|
||||
headers[k] = v
|
||||
}
|
||||
book, err := s.repo.GetBook(ctx, bookID)
|
||||
if err != nil || book == nil || strings.TrimSpace(book.Origin) == "" {
|
||||
return headers
|
||||
}
|
||||
if src, err := s.repo.GetSourceByURL(ctx, book.Origin); err == nil && src != nil && src.Header != "" {
|
||||
var extra map[string]any
|
||||
if json.Unmarshal([]byte(src.Header), &extra) == nil {
|
||||
for k, v := range extra {
|
||||
headers[k] = fmt.Sprintf("%v", v)
|
||||
}
|
||||
}
|
||||
}
|
||||
state := s.newSourceState(ctx, book.Origin)
|
||||
for k, v := range state.LoginHeaderMap() {
|
||||
if !strings.EqualFold(k, "cookie") {
|
||||
headers[k] = v
|
||||
}
|
||||
}
|
||||
if ck := state.CookieForRequest(source); ck != "" {
|
||||
headers["Cookie"] = ck
|
||||
}
|
||||
// 默认 Referer 同上:聚合类书源的 origin 是显示名,拼出来的 Referer 非法,
|
||||
// 会被音源/图床判盗链。只在 origin 是真正的 http(s) 地址时才补。
|
||||
if headers["Referer"] == "" {
|
||||
if referer := sourceReferer(book.Origin); referer != "" {
|
||||
headers["Referer"] = referer
|
||||
}
|
||||
}
|
||||
return headers
|
||||
}
|
||||
|
||||
// buildFFmpegAudioArgs 组装音频转码参数(纯函数,便于测试)。
|
||||
// 只取第一条音频流:WMA 常把专辑封面挂在视频流上,-vn 一并丢掉。
|
||||
func buildFFmpegAudioArgs(source, output string, headers map[string]string) []string {
|
||||
args := []string{"-hide_banner", "-loglevel", "error", "-nostdin", "-y"}
|
||||
if len(headers) > 0 {
|
||||
keys := make([]string, 0, len(headers))
|
||||
for k := range headers {
|
||||
keys = append(keys, k)
|
||||
}
|
||||
sort.Strings(keys)
|
||||
var b strings.Builder
|
||||
for _, k := range keys {
|
||||
b.WriteString(k)
|
||||
b.WriteString(": ")
|
||||
b.WriteString(headers[k])
|
||||
b.WriteString("\r\n")
|
||||
}
|
||||
args = append(args, "-headers", b.String())
|
||||
}
|
||||
return append(args,
|
||||
"-i", source,
|
||||
"-vn", "-map", "0:a:0",
|
||||
"-c:a", "libmp3lame", "-b:a", audioTranscodeBitrate,
|
||||
"-f", "mp3", output,
|
||||
)
|
||||
}
|
||||
|
||||
// pruneAudioTranscodeCache 缓存超过上限时按访问时间淘汰,回落到上限的 90%。
|
||||
func pruneAudioTranscodeCache(dir string, maxBytes int64) {
|
||||
entries, err := os.ReadDir(dir)
|
||||
if err != nil {
|
||||
return
|
||||
}
|
||||
type cacheItem struct {
|
||||
path string
|
||||
size int64
|
||||
mod time.Time
|
||||
}
|
||||
items := make([]cacheItem, 0, len(entries))
|
||||
var total int64
|
||||
for _, e := range entries {
|
||||
if e.IsDir() || !strings.HasSuffix(e.Name(), ".mp3") {
|
||||
continue
|
||||
}
|
||||
info, err := e.Info()
|
||||
if err != nil {
|
||||
continue
|
||||
}
|
||||
items = append(items, cacheItem{filepath.Join(dir, e.Name()), info.Size(), info.ModTime()})
|
||||
total += info.Size()
|
||||
}
|
||||
if total <= maxBytes {
|
||||
return
|
||||
}
|
||||
sort.Slice(items, func(i, j int) bool { return items[i].mod.Before(items[j].mod) })
|
||||
target := int64(float64(maxBytes) * audioTranscodeKeepRatio)
|
||||
for _, it := range items {
|
||||
if total <= target {
|
||||
break
|
||||
}
|
||||
if os.Remove(it.path) == nil {
|
||||
total -= it.size
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// truncateForError 截断 ffmpeg 输出,避免把整段 stderr 塞进响应。
|
||||
func truncateForError(s string, limit int) string {
|
||||
s = strings.TrimSpace(s)
|
||||
if len(s) <= limit {
|
||||
return s
|
||||
}
|
||||
return s[:limit] + "…"
|
||||
}
|
||||
@@ -0,0 +1,257 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"errors"
|
||||
"net/url"
|
||||
"os"
|
||||
"path/filepath"
|
||||
"strings"
|
||||
"testing"
|
||||
"time"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
)
|
||||
|
||||
// TestAudioSourceExt 音频地址取扩展名:本地路径、带 query 的远端地址、
|
||||
// Windows 盘符(会被 url.Parse 当成单字符 scheme)都要认对。
|
||||
func TestAudioSourceExt(t *testing.T) {
|
||||
cases := []struct {
|
||||
in string
|
||||
want string
|
||||
}{
|
||||
{`D:\media\斗破苍穹\001-250\001.wma`, ".wma"},
|
||||
{"/media/books/01.WMA", ".wma"},
|
||||
{"http://127.0.0.1:8080/api/strm/play/cloud115/video.wma?acct=x&path=y", ".wma"},
|
||||
{"https://cdn.example.com/a/b.mp3?token=1#frag", ".mp3"},
|
||||
{"https://cdn.example.com/stream", ""},
|
||||
{"/media/books/01.m4a", ".m4a"},
|
||||
{"", ""},
|
||||
}
|
||||
for _, tc := range cases {
|
||||
if got := audioSourceExt(tc.in); got != tc.want {
|
||||
t.Errorf("audioSourceExt(%q) = %q,期望 %q", tc.in, got, tc.want)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestNeedsAudioTranscode 只对浏览器确定解不了的格式转码,其余维持直出。
|
||||
func TestNeedsAudioTranscode(t *testing.T) {
|
||||
need := []string{
|
||||
`D:\media\001.wma`,
|
||||
"http://127.0.0.1:8080/api/strm/play/cloud115/video.wma?acct=x",
|
||||
"/media/01.asf",
|
||||
"/media/01.ape",
|
||||
}
|
||||
for _, s := range need {
|
||||
if !needsAudioTranscode(s) {
|
||||
t.Errorf("%q 应需要转码", s)
|
||||
}
|
||||
}
|
||||
direct := []string{
|
||||
"/media/01.mp3", "/media/01.m4a", "/media/01.m4b", "/media/01.flac",
|
||||
"/media/01.ogg", "/media/01.opus", "/media/01.wav",
|
||||
"https://cdn.example.com/stream", "",
|
||||
}
|
||||
for _, s := range direct {
|
||||
if needsAudioTranscode(s) {
|
||||
t.Errorf("%q 不该被转码", s)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestAudioTranscodeURLSignature 转码地址的签名可校验、改一处即失效。
|
||||
func TestAudioTranscodeURLSignature(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
source := `D:\media\斗破苍穹\001-250\001.wma`
|
||||
|
||||
raw := svc.AudioTranscodeURL("book-1", source)
|
||||
if !strings.HasPrefix(raw, "/api/reader/audio/transcode?b=book-1&u=") {
|
||||
t.Fatalf("转码地址格式不对: %q", raw)
|
||||
}
|
||||
encoded, sig := queryParam(t, raw, "u"), queryParam(t, raw, "s")
|
||||
got, err := svc.VerifyAudioTranscodeURL("book-1", encoded, sig)
|
||||
if err != nil || got != source {
|
||||
t.Fatalf("签名校验失败: %v / %q", err, got)
|
||||
}
|
||||
if _, err := svc.VerifyAudioTranscodeURL("book-1", encoded, "deadbeef"); err == nil {
|
||||
t.Fatal("错误签名应校验失败")
|
||||
}
|
||||
if _, err := svc.VerifyAudioTranscodeURL("book-2", encoded, sig); err == nil {
|
||||
t.Fatal("换一本书后签名应失效")
|
||||
}
|
||||
}
|
||||
|
||||
// TestAudioTranscodeCachePath 缓存名对(书 + 源地址)稳定且互不冲突。
|
||||
func TestAudioTranscodeCachePath(t *testing.T) {
|
||||
dir := t.TempDir()
|
||||
a := audioTranscodeCachePath(dir, "book-1", "/x/001.wma")
|
||||
b := audioTranscodeCachePath(dir, "book-1", "/x/001.wma")
|
||||
if a != b {
|
||||
t.Fatalf("同一输入应得到同一缓存路径: %q vs %q", a, b)
|
||||
}
|
||||
if !strings.HasSuffix(a, ".mp3") {
|
||||
t.Fatalf("缓存文件应为 mp3: %q", a)
|
||||
}
|
||||
for _, other := range []string{
|
||||
audioTranscodeCachePath(dir, "book-2", "/x/001.wma"),
|
||||
audioTranscodeCachePath(dir, "book-1", "/x/002.wma"),
|
||||
} {
|
||||
if other == a {
|
||||
t.Fatalf("不同书/不同章节不该共用缓存: %q", other)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestEnsureTranscodedAudioWithoutFFmpeg 没装 ffmpeg 时要给出可照做的错误。
|
||||
func TestEnsureTranscodedAudioWithoutFFmpeg(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
svc.cfg.App.FFmpegPath = filepath.Join(t.TempDir(), "definitely-missing-ffmpeg")
|
||||
|
||||
_, err := svc.EnsureTranscodedAudio(t.Context(), "book-1", `D:\media\001.wma`)
|
||||
if !errors.Is(err, ErrAudioTranscodeUnavailable) {
|
||||
t.Fatalf("err = %v,应包含 ErrAudioTranscodeUnavailable", err)
|
||||
}
|
||||
if !strings.Contains(err.Error(), "WMA") || !strings.Contains(err.Error(), "ffmpeg") {
|
||||
t.Fatalf("错误信息应说明格式与解决办法,实际 %q", err.Error())
|
||||
}
|
||||
}
|
||||
|
||||
// TestEnsureTranscodedAudioUsesCache 已有转码结果时直接命中缓存,不依赖 ffmpeg。
|
||||
func TestEnsureTranscodedAudioUsesCache(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
svc.cfg.App.FFmpegPath = filepath.Join(t.TempDir(), "definitely-missing-ffmpeg")
|
||||
|
||||
source := `D:\media\001.wma`
|
||||
dir, err := svc.audioTranscodeDir()
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
want := audioTranscodeCachePath(dir, "book-1", source)
|
||||
if err := os.WriteFile(want, []byte("MP3DATA"), 0o640); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
got, err := svc.EnsureTranscodedAudio(t.Context(), "book-1", source)
|
||||
if err != nil {
|
||||
t.Fatalf("命中缓存时不该报错: %v", err)
|
||||
}
|
||||
if got != want {
|
||||
t.Fatalf("缓存路径 = %q,期望 %q", got, want)
|
||||
}
|
||||
data, err := os.ReadFile(got)
|
||||
if err != nil || string(data) != "MP3DATA" {
|
||||
t.Fatalf("缓存内容被改动了: %q / %v", data, err)
|
||||
}
|
||||
}
|
||||
|
||||
// TestBuildFFmpegAudioArgs ffmpeg 参数:输入输出位置正确,丢掉视频流,
|
||||
// 强制 mp3 编码,请求头按 key 排序保证可复现。
|
||||
func TestBuildFFmpegAudioArgs(t *testing.T) {
|
||||
args := buildFFmpegAudioArgs("/in/01.wma", "/out/01.mp3", nil)
|
||||
if args[0] != "-hide_banner" || args[len(args)-1] != "/out/01.mp3" {
|
||||
t.Fatalf("参数首尾不对: %v", args)
|
||||
}
|
||||
joined := strings.Join(args, " ")
|
||||
for _, want := range []string{"-i /in/01.wma", "-vn", "-map 0:a:0", "-c:a libmp3lame", "-f mp3"} {
|
||||
if !strings.Contains(joined, want) {
|
||||
t.Errorf("参数缺少 %q: %v", want, args)
|
||||
}
|
||||
}
|
||||
|
||||
withHeaders := buildFFmpegAudioArgs("u", "o", map[string]string{
|
||||
"User-Agent": "mebox", "Cookie": "a=1",
|
||||
})
|
||||
idx := indexOf(withHeaders, "-headers")
|
||||
if idx < 0 {
|
||||
t.Fatalf("应带上 -headers: %v", withHeaders)
|
||||
}
|
||||
headers := withHeaders[idx+1]
|
||||
if !strings.HasPrefix(headers, "Cookie: a=1\r\n") || !strings.Contains(headers, "User-Agent: mebox\r\n") {
|
||||
t.Fatalf("请求头内容或顺序不对: %q", headers)
|
||||
}
|
||||
}
|
||||
|
||||
// TestPruneAudioTranscodeCache 超过上限时按访问时间淘汰最旧的,落到 90%。
|
||||
func TestPruneAudioTranscodeCache(t *testing.T) {
|
||||
dir := t.TempDir()
|
||||
base := time.Now().Add(-time.Hour)
|
||||
sizes := []int64{400, 400, 400}
|
||||
names := []string{"old.mp3", "mid.mp3", "new.mp3"}
|
||||
var total int64
|
||||
for i, name := range names {
|
||||
if err := os.WriteFile(filepath.Join(dir, name), make([]byte, sizes[i]), 0o640); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
ts := base.Add(time.Duration(i) * time.Minute)
|
||||
if err := os.Chtimes(filepath.Join(dir, name), ts, ts); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
total += sizes[i]
|
||||
}
|
||||
// 上限设为总量的一半:必须淘汰
|
||||
pruneAudioTranscodeCache(dir, total/2)
|
||||
|
||||
if _, err := os.Stat(filepath.Join(dir, "old.mp3")); !os.IsNotExist(err) {
|
||||
t.Errorf("最旧的缓存应被删除,stat err = %v", err)
|
||||
}
|
||||
if _, err := os.Stat(filepath.Join(dir, "new.mp3")); err != nil {
|
||||
t.Errorf("最新的缓存应保留: %v", err)
|
||||
}
|
||||
}
|
||||
|
||||
// TestLocalAudioTrackWithoutFFmpeg localAudioTrack 在需要转码但没 ffmpeg 时,
|
||||
// 直接给出可读的错误,而不是返回一个注定播不了的地址。
|
||||
func TestLocalAudioTrackWithoutFFmpeg(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
svc.cfg.App.FFmpegPath = filepath.Join(t.TempDir(), "definitely-missing-ffmpeg")
|
||||
|
||||
book := &model.ReaderBook{Base: model.Base{ID: "book-1"}, Type: 1}
|
||||
ch := model.ReaderChapter{Index: 0, Tag: `D:\media\001.wma`}
|
||||
_, transcoding, err := svc.localAudioTrack(book, ch)
|
||||
if err == nil {
|
||||
t.Fatal("缺 ffmpeg 时应报错")
|
||||
}
|
||||
if transcoding {
|
||||
t.Error("报错时不该标记为转码中")
|
||||
}
|
||||
if !strings.Contains(err.Error(), "ffmpeg") {
|
||||
t.Fatalf("错误信息应提到 ffmpeg: %q", err.Error())
|
||||
}
|
||||
}
|
||||
|
||||
// TestLocalAudioTrackPlayableStaysDirect 浏览器能播的格式不转码,仍走本地音频端点。
|
||||
func TestLocalAudioTrackPlayableStaysDirect(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
book := &model.ReaderBook{Base: model.Base{ID: "book-1"}, Type: 1}
|
||||
ch := model.ReaderChapter{Index: 0, Tag: `D:\media\001.mp3`}
|
||||
|
||||
track, transcoding, err := svc.localAudioTrack(book, ch)
|
||||
if err != nil {
|
||||
t.Fatalf("mp3 不该报错: %v", err)
|
||||
}
|
||||
if transcoding {
|
||||
t.Error("mp3 不该标记为转码中")
|
||||
}
|
||||
if !strings.HasPrefix(track, "/api/reader/local/audio?") {
|
||||
t.Fatalf("mp3 应走本地音频端点: %q", track)
|
||||
}
|
||||
}
|
||||
|
||||
// queryParam 取查询参数(签名用例共用)。
|
||||
func queryParam(t *testing.T, raw, key string) string {
|
||||
t.Helper()
|
||||
u, err := url.Parse(raw)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
return u.Query().Get(key)
|
||||
}
|
||||
|
||||
func indexOf(items []string, want string) int {
|
||||
for i, v := range items {
|
||||
if v == want {
|
||||
return i
|
||||
}
|
||||
}
|
||||
return -1
|
||||
}
|
||||
@@ -0,0 +1,149 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"strings"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// 本文件:书籍类型兜底的回归测试。
|
||||
//
|
||||
// 背景:书籍详情页的「加入书架/开始阅读」曾把 origin_type 写死成 0(文本),
|
||||
// 于是听书源的音频书以文本类型落库。正文类型按「规则 JS 声明 > 书架类型 >
|
||||
// 书源类型」取值,书架里的 0 压过书源的 1,播放直链就被当成正文排出来,
|
||||
// 整个阅读页是一屏 URL(用户截图里的现象)。
|
||||
//
|
||||
// 兜底规则:书架类型还停在默认的「文本」时用书源类型;规则 JS 显式声明的
|
||||
// 类型仍然最优先,避免改坏「文本型聚合源提供听书内容」那类源。
|
||||
|
||||
// audioTestSourceJSON 一个最小听书源:正文规则给 JSON 里的播放直链。
|
||||
func audioTestSourceJSON(t *testing.T, base string) string {
|
||||
t.Helper()
|
||||
return `{
|
||||
"bookSourceName": "听书测试源",
|
||||
"bookSourceType": 1,
|
||||
"bookSourceUrl": "` + base + `",
|
||||
"ruleContent": {"content": "$.data.url"}
|
||||
}`
|
||||
}
|
||||
|
||||
// TestAudioBookTypeFallbackBySource 书架类型是默认的文本时按书源类型(音频)渲染,
|
||||
// 并把纠正后的类型写回书架记录。
|
||||
func TestAudioBookTypeFallbackBySource(t *testing.T) {
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
w.Header().Set("Content-Type", "application/json; charset=utf-8")
|
||||
_, _ = w.Write([]byte(`{"data":{"url":"https://cdn.example.com/a.m4a"}}`))
|
||||
}))
|
||||
defer srv.Close()
|
||||
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
sourceID := prepareLoginSource(t, svc, audioTestSourceJSON(t, srv.URL))
|
||||
|
||||
// 复刻详情页写死的 origin_type=0
|
||||
book, err := svc.AddBook(ctx, readerTestUserID, SearchOrigin{
|
||||
SourceID: sourceID,
|
||||
Origin: srv.URL,
|
||||
OriginName: "听书测试源",
|
||||
OriginType: 0,
|
||||
BookURL: srv.URL + "/book/1",
|
||||
}, "测试听书", "某作者", "")
|
||||
if err != nil {
|
||||
t.Fatalf("加入书架失败: %v", err)
|
||||
}
|
||||
if err := svc.SaveChapters(ctx, book.ID, []ChapterInput{
|
||||
{Index: 0, Title: "第一章", URL: srv.URL + "/content"},
|
||||
}); err != nil {
|
||||
t.Fatalf("保存目录失败: %v", err)
|
||||
}
|
||||
|
||||
out, err := svc.GetContentForBook(ctx, readerTestUserID, book.ID, 0)
|
||||
if err != nil {
|
||||
t.Fatalf("取正文失败: %v", err)
|
||||
}
|
||||
if out.Type != "audio" {
|
||||
t.Fatalf("正文类型 = %q(期望 audio:书架类型是默认文本时应当按书源类型兜底),正文 = %.120q",
|
||||
out.Type, out.Content)
|
||||
}
|
||||
if len(out.Tracks) != 1 {
|
||||
t.Fatalf("音轨数 = %d,期望 1", len(out.Tracks))
|
||||
}
|
||||
// 音轨要落到签名代理上(浏览器播放带不上书源的防盗链头)
|
||||
if !strings.HasPrefix(out.Tracks[0], "/api/reader/media?") {
|
||||
t.Fatalf("音轨 = %q,期望签名代理地址", out.Tracks[0])
|
||||
}
|
||||
raw, err := svc.VerifyProxyURL(book.ID, trackQueryValue(t, out.Tracks[0], "u"), trackQueryValue(t, out.Tracks[0], "s"))
|
||||
if err != nil {
|
||||
t.Fatalf("代理地址校验失败: %v", err)
|
||||
}
|
||||
if raw != "https://cdn.example.com/a.m4a" {
|
||||
t.Fatalf("代理还原地址 = %q", raw)
|
||||
}
|
||||
|
||||
// 纠正后的类型要写回书架:否则每读一章都要再兜底一次,注入规则 JS 的
|
||||
// book.type 也一直是错的(书源 JS 会读它分支)。
|
||||
reloaded, err := svc.GetBook(ctx, book.ID)
|
||||
if err != nil {
|
||||
t.Fatalf("重读书籍失败: %v", err)
|
||||
}
|
||||
if reloaded.Type != 1 {
|
||||
t.Fatalf("书架记录类型 = %d,期望 1(音频)", reloaded.Type)
|
||||
}
|
||||
}
|
||||
|
||||
// TestDeclaredTextTypeWinsOverAudioSource 规则 JS 显式声明类型时不套用书源类型兜底:
|
||||
// 文本型聚合源(bookSourceType=0,但听书内容靠 JS 声明)反过来也不会被覆盖成文本。
|
||||
func TestDeclaredTextTypeWinsOverAudioSource(t *testing.T) {
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
w.Header().Set("Content-Type", "application/json; charset=utf-8")
|
||||
_, _ = w.Write([]byte(`{"data":{"url":"https://cdn.example.com/a.m4a"}}`))
|
||||
}))
|
||||
defer srv.Close()
|
||||
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
// 音频书源,但正文规则用 JS 把 book.type 声明成文本(legado 的 BookType 位掩码 8)
|
||||
src := `{
|
||||
"bookSourceName": "声明文本的音频源",
|
||||
"bookSourceType": 1,
|
||||
"bookSourceUrl": "` + srv.URL + `",
|
||||
"ruleContent": {"content": "@js:book.type = 8; result"}
|
||||
}`
|
||||
sourceID := prepareLoginSource(t, svc, src)
|
||||
|
||||
book, err := svc.AddBook(ctx, readerTestUserID, SearchOrigin{
|
||||
SourceID: sourceID, Origin: srv.URL, OriginName: "声明文本的音频源",
|
||||
OriginType: 0, BookURL: srv.URL + "/book/1",
|
||||
}, "声明文本", "某作者", "")
|
||||
if err != nil {
|
||||
t.Fatalf("加入书架失败: %v", err)
|
||||
}
|
||||
if err := svc.SaveChapters(ctx, book.ID, []ChapterInput{
|
||||
{Index: 0, Title: "第一章", URL: srv.URL + "/content"},
|
||||
}); err != nil {
|
||||
t.Fatalf("保存目录失败: %v", err)
|
||||
}
|
||||
|
||||
out, err := svc.GetContentForBook(ctx, readerTestUserID, book.ID, 0)
|
||||
if err != nil {
|
||||
t.Fatalf("取正文失败: %v", err)
|
||||
}
|
||||
if out.Type != "text" {
|
||||
t.Fatalf("正文类型 = %q,期望 text(规则 JS 声明的类型优先于书源类型)", out.Type)
|
||||
}
|
||||
}
|
||||
|
||||
// trackQueryValue 取出代理地址里的查询参数值(测试用)。
|
||||
func trackQueryValue(t *testing.T, raw, key string) string {
|
||||
t.Helper()
|
||||
idx := strings.Index(raw, key+"=")
|
||||
if idx < 0 {
|
||||
t.Fatalf("地址 %q 里没有参数 %s", raw, key)
|
||||
}
|
||||
rest := raw[idx+len(key)+1:]
|
||||
if end := strings.Index(rest, "&"); end >= 0 {
|
||||
rest = rest[:end]
|
||||
}
|
||||
return rest
|
||||
}
|
||||
@@ -0,0 +1,91 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"fmt"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// 本文件:详情 init 链式取值 与 书籍类型翻译 的回归测试。
|
||||
//
|
||||
// 背景(光遇聚合):
|
||||
// - ruleBookInfo.init 是 `<js>…</js>$.data` 组合规则,它的结果会成为后续
|
||||
// name/author/tocUrl 的解析内容(对应 legado
|
||||
// `analyzeRule.setContent(analyzeRule.getElement(infoRule.init))`)。
|
||||
// MeBox 早期把 init 结果当成 JSON 解析 map 再按写死的键名取值,导致
|
||||
// 详情页「书名/作者为空、目录 0 章」。
|
||||
// - 书源在目录规则里用 legado 的 BookType 位掩码给 book.type 赋值
|
||||
// (8=文本 32=音频 64=图片 4=视频),而 MeBox 用 0/1/2/3,必须换算,
|
||||
// 否则听书源会被当成文本,正文渲染成一串裸 URL。
|
||||
|
||||
// TestBookInfoInitResultBecomesContent init 的结果要成为后续字段的解析内容。
|
||||
func TestBookInfoInitResultBecomesContent(t *testing.T) {
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
w.Header().Set("Content-Type", "application/json; charset=utf-8")
|
||||
switch r.URL.Path {
|
||||
case "/detail":
|
||||
_, _ = w.Write([]byte(`{"code":0,"data":{"book_name":"宠魅","author":"某作者","thumb_url":"https://img.example.com/c.jpg","toc_url":"https://example.com/toc"}}`))
|
||||
default:
|
||||
_, _ = w.Write([]byte(`<html><body>book page</body></html>`))
|
||||
}
|
||||
}))
|
||||
defer srv.Close()
|
||||
|
||||
svc, _ := newLoginTestService(t)
|
||||
src := map[string]any{
|
||||
"bookSourceUrl": srv.URL,
|
||||
"bookSourceName": "init 链式测试源",
|
||||
"ruleBookInfo": map[string]any{
|
||||
// 组合规则:先跑 JS 取详情接口,再用 $.data 取出对象
|
||||
"init": fmt.Sprintf(`<js>java.ajax(%q)</js>$.data`, srv.URL+"/detail"),
|
||||
"name": "$.book_name",
|
||||
"author": "$.author",
|
||||
"coverUrl": "$.thumb_url",
|
||||
"tocUrl": "$.toc_url",
|
||||
},
|
||||
}
|
||||
out, _ := json.Marshal(src)
|
||||
sourceID := prepareLoginSource(t, svc, string(out))
|
||||
|
||||
info, err := svc.GetBookInfo(t.Context(), sourceID, "", srv.URL+"/book/1")
|
||||
if err != nil {
|
||||
t.Fatalf("取详情失败: %v", err)
|
||||
}
|
||||
if info.Name != "宠魅" {
|
||||
t.Fatalf("书名 = %q(init 结果未成为解析内容)", info.Name)
|
||||
}
|
||||
if info.Author != "某作者" {
|
||||
t.Fatalf("作者 = %q", info.Author)
|
||||
}
|
||||
if info.CoverURL != "https://img.example.com/c.jpg" {
|
||||
t.Fatalf("封面 = %q", info.CoverURL)
|
||||
}
|
||||
if info.TocURL != "https://example.com/toc" {
|
||||
t.Fatalf("目录地址 = %q", info.TocURL)
|
||||
}
|
||||
}
|
||||
|
||||
// TestNormalizeBookType 把 legado 的 BookType 位掩码换算成 MeBox 的 0/1/2/3。
|
||||
func TestNormalizeBookType(t *testing.T) {
|
||||
cases := map[int]int{
|
||||
0: 0, // 未知/默认 → 文本
|
||||
8: 0, // text
|
||||
32: 1, // audio(听书源默认值)
|
||||
64: 2, // image(漫画源默认值)
|
||||
4: 3, // video(短剧源默认值)
|
||||
1: 0, // 苹果端遗留值不再是音频
|
||||
2: 0,
|
||||
// 组合位:按 音频 > 图片 > 视频 > 文本 的优先级取一个
|
||||
32 | 8: 1,
|
||||
64 | 8: 2,
|
||||
4 | 8: 3,
|
||||
32 | 64: 1,
|
||||
}
|
||||
for in, want := range cases {
|
||||
if got := normalizeBookType(in); got != want {
|
||||
t.Errorf("normalizeBookType(%d) = %d,期望 %d", in, got, want)
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,147 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// 本文件:书架未读章数所需的「网络书籍总章数从目录缓存补齐」逻辑回归测试。
|
||||
// 前端的未读章数 = total_chapter_num -(已读到第几章),所以这里保证 total_chapter_num 可靠。
|
||||
|
||||
func TestListBooksFillsChapterCountFromCache(t *testing.T) {
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
|
||||
book, err := svc.AddBook(ctx, "u1", SearchOrigin{
|
||||
Origin: "https://example.com",
|
||||
OriginName: "测试书源",
|
||||
BookURL: "https://example.com/book/1",
|
||||
}, "测试书", "作者", "")
|
||||
if err != nil {
|
||||
t.Fatalf("加入书架失败: %v", err)
|
||||
}
|
||||
if book.TotalChapterNum != 0 {
|
||||
t.Fatalf("刚加入书架的网文总章数 = %d,期望 0", book.TotalChapterNum)
|
||||
}
|
||||
|
||||
// 目录还没缓存:书架读不到总章数,前端此时不显示未读徽标
|
||||
books, err := svc.ListBooks(ctx, "u1")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if len(books) != 1 || books[0].TotalChapterNum != 0 {
|
||||
t.Fatalf("未缓存目录时书架 = %+v,期望 1 本且总章数 0", books)
|
||||
}
|
||||
|
||||
// 缓存 5 章后,书架应能算出总章数
|
||||
chapters := make([]ChapterInput, 0, 5)
|
||||
for i := 0; i < 5; i++ {
|
||||
chapters = append(chapters, ChapterInput{
|
||||
Index: i,
|
||||
Title: fmt.Sprintf("第 %d 章", i+1),
|
||||
URL: fmt.Sprintf("https://example.com/c/%d", i),
|
||||
})
|
||||
}
|
||||
if err := svc.SaveChapters(ctx, book.ID, chapters); err != nil {
|
||||
t.Fatalf("保存目录失败: %v", err)
|
||||
}
|
||||
|
||||
books, err = svc.ListBooks(ctx, "u1")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if books[0].TotalChapterNum != 5 {
|
||||
t.Fatalf("缓存 5 章后总章数 = %d,期望 5", books[0].TotalChapterNum)
|
||||
}
|
||||
|
||||
// 读到第 2 章后,前端用 dur_chapter_index + 1 算已读;这里确认进度按原样回读
|
||||
if err := svc.SaveProgress(ctx, "u1", book.ID, 1, 0, "第 2 章"); err != nil {
|
||||
t.Fatalf("保存进度失败: %v", err)
|
||||
}
|
||||
books, err = svc.ListBooks(ctx, "u1")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if books[0].DurChapterIndex != 1 || books[0].DurChapterTime == 0 {
|
||||
t.Fatalf("进度 = index %d / time %d,期望 index 1 且有时间戳", books[0].DurChapterIndex, books[0].DurChapterTime)
|
||||
}
|
||||
}
|
||||
|
||||
func TestListBooksChapterCountIsPerUser(t *testing.T) {
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
|
||||
book, err := svc.AddBook(ctx, "u1", SearchOrigin{
|
||||
Origin: "https://example.com",
|
||||
BookURL: "https://example.com/book/1",
|
||||
}, "测试书", "", "")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if err := svc.SaveChapters(ctx, book.ID, []ChapterInput{{Index: 0, Title: "第 1 章", URL: "u"}}); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
// 别人的书架不应该受影响,也不该拿到别人的书
|
||||
other, err := svc.ListBooks(ctx, "u2")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if len(other) != 0 {
|
||||
t.Fatalf("u2 的书架应为空,得到 %+v", other)
|
||||
}
|
||||
}
|
||||
|
||||
func TestWarmUpBookChaptersSkipsWhenCacheExists(t *testing.T) {
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
|
||||
book, err := svc.AddBook(ctx, "u1", SearchOrigin{
|
||||
Origin: "https://example.com",
|
||||
BookURL: "https://example.com/book/1",
|
||||
}, "测试书", "", "")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
chapters := []ChapterInput{
|
||||
{Index: 0, Title: "第 1 章", URL: "https://example.com/c/0"},
|
||||
{Index: 1, Title: "第 2 章", URL: "https://example.com/c/1"},
|
||||
}
|
||||
if err := svc.SaveChapters(ctx, book.ID, chapters); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
// 已有目录缓存时直接返回,不联网、不覆盖缓存
|
||||
svc.WarmUpBookChapters(ctx, "u1", book)
|
||||
|
||||
got, err := svc.ListChapters(ctx, book.ID)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if len(got) != 2 || got[0].Title != "第 1 章" {
|
||||
t.Fatalf("缓存被改动: %+v", got)
|
||||
}
|
||||
}
|
||||
|
||||
func TestWarmUpBookChaptersSkipsLocalBook(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
ctx := t.Context()
|
||||
|
||||
book, err := svc.ImportLocalBook(ctx, "u1", "测试小说.txt", []byte(sampleTXT))
|
||||
if err != nil {
|
||||
t.Fatalf("导入失败: %v", err)
|
||||
}
|
||||
// 本地书籍不应触发联网抓目录(书源为空的地址必然失败),章节保持导入时的结果
|
||||
svc.WarmUpBookChapters(ctx, "u1", book)
|
||||
|
||||
got, err := svc.ListChapters(ctx, book.ID)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if len(got) != 3 {
|
||||
t.Fatalf("本地书籍章节数 = %d,期望 3", len(got))
|
||||
}
|
||||
if book.TotalChapterNum != 3 {
|
||||
t.Fatalf("本地书籍总章数 = %d,期望 3", book.TotalChapterNum)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,239 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"net/url"
|
||||
"sort"
|
||||
"strconv"
|
||||
"strings"
|
||||
"sync"
|
||||
)
|
||||
|
||||
// 本文件:段落评论接口(/para_review)的分页适配。
|
||||
//
|
||||
// 上游聚合站自带的评论页按 page=N 翻页(滑到底请求 page=2、page=3…),但它的列表
|
||||
// 接口早就改成了游标分页:响应里给 next_cursor,请求时要带 cursor=。接口不认识
|
||||
// page,于是每次「加载更多」都原样返回第一页——一条新评论都加不进去,has_more 又
|
||||
// 始终是 true,评论页就一遍遍触发加载更多、反复闪骨架屏。用户看到的就是「段评往
|
||||
// 下滑一闪一闪」,其实评论一条没多。
|
||||
//
|
||||
// 评论页是上游 HTML,改不了,于是在代理层做最小适配:
|
||||
// - 请求带 page=N(N>1)且有上一页给的游标时,换成 cursor= 请求;
|
||||
// - 某页返回的 comment_id 全是已经见过的,就把 has_more 改成 false 收尾,
|
||||
// 免得页面在「加载更多 → 没有新内容」之间空转。
|
||||
//
|
||||
// 只认路径以 /para_review 结尾、带 page 且不带 cursor 的请求;对不上就原样透传,
|
||||
// 不影响别的书源。
|
||||
|
||||
// commentPageState 一个评论列表(同一段落 + 同一排序)的翻页状态。
|
||||
type commentPageState struct {
|
||||
// cursor 上一页响应给的游标:下一页请求用它代替 page。
|
||||
cursor string
|
||||
// seen 已经下发给页面的 comment_id:整页重复时据此收尾。
|
||||
seen map[string]struct{}
|
||||
}
|
||||
|
||||
// commentPager 承载页面里评论接口的翻页适配器(每个承载页面一份)。
|
||||
type commentPager struct {
|
||||
mu sync.Mutex
|
||||
states map[string]*commentPageState
|
||||
}
|
||||
|
||||
func newCommentPager() *commentPager {
|
||||
return &commentPager{states: map[string]*commentPageState{}}
|
||||
}
|
||||
|
||||
// rewriteRequest 按需把 page=N 改写成 cursor=<上一页游标>。
|
||||
//
|
||||
// 返回改写后的地址与状态键(空串表示这个请求不需要适配,调用方也不必观察响应)。
|
||||
func (p *commentPager) rewriteRequest(target string) (string, string) {
|
||||
if p == nil {
|
||||
return target, ""
|
||||
}
|
||||
parsed, err := url.Parse(target)
|
||||
if err != nil || !isCommentListPath(parsed.Path) {
|
||||
return target, ""
|
||||
}
|
||||
query := parsed.Query()
|
||||
pageRaw := strings.TrimSpace(query.Get("page"))
|
||||
if pageRaw == "" || strings.TrimSpace(query.Get("cursor")) != "" {
|
||||
return target, ""
|
||||
}
|
||||
page, err := strconv.Atoi(pageRaw)
|
||||
if err != nil || page < 1 {
|
||||
return target, ""
|
||||
}
|
||||
key := commentPageKey(parsed.Path, query)
|
||||
if page == 1 {
|
||||
// 重新加载 / 换了排序:游标与去重表都从头来。
|
||||
p.mu.Lock()
|
||||
delete(p.states, key)
|
||||
p.mu.Unlock()
|
||||
return target, key
|
||||
}
|
||||
p.mu.Lock()
|
||||
cursor := ""
|
||||
if st := p.states[key]; st != nil {
|
||||
cursor = st.cursor
|
||||
}
|
||||
p.mu.Unlock()
|
||||
if cursor == "" {
|
||||
// 不是从第一页翻过来的(没有游标可用):原样透传,仍然观察响应,
|
||||
// 靠下面的「整页重复」判断让页面停下来。
|
||||
return target, key
|
||||
}
|
||||
query.Set("cursor", cursor)
|
||||
query.Del("page")
|
||||
parsed.RawQuery = query.Encode()
|
||||
return parsed.String(), key
|
||||
}
|
||||
|
||||
// observeResponse 观察评论列表响应:记下游标,并在整页重复时把 has_more 改成 false。
|
||||
func (p *commentPager) observeResponse(key, body string) string {
|
||||
if p == nil || key == "" {
|
||||
return body
|
||||
}
|
||||
trimmed := strings.TrimSpace(body)
|
||||
if !strings.HasPrefix(trimmed, "{") {
|
||||
return body
|
||||
}
|
||||
var env commentListEnvelope
|
||||
if err := json.Unmarshal([]byte(trimmed), &env); err != nil {
|
||||
return body
|
||||
}
|
||||
if env.Data.Comments == nil && env.Data.HasMore == nil {
|
||||
return body
|
||||
}
|
||||
ids := make([]string, 0, len(env.Data.Comments))
|
||||
for _, c := range env.Data.Comments {
|
||||
if id := strings.TrimSpace(c.CommentID); id != "" {
|
||||
ids = append(ids, id)
|
||||
}
|
||||
}
|
||||
|
||||
p.mu.Lock()
|
||||
defer p.mu.Unlock()
|
||||
if p.states == nil {
|
||||
p.states = map[string]*commentPageState{}
|
||||
}
|
||||
st := p.states[key]
|
||||
if st == nil {
|
||||
st = &commentPageState{seen: map[string]struct{}{}}
|
||||
p.states[key] = st
|
||||
}
|
||||
repeated := len(ids) > 0
|
||||
for _, id := range ids {
|
||||
if _, ok := st.seen[id]; !ok {
|
||||
repeated = false
|
||||
break
|
||||
}
|
||||
}
|
||||
if repeated && env.Data.HasMore != nil && *env.Data.HasMore {
|
||||
// 整页都是老评论:接口不认 page,再翻下去也只会拿到同一页。
|
||||
// 改掉 has_more,让页面显示「没有更多了」,别再反复拉取闪骨架屏。
|
||||
if replaced, ok := setFirstJSONBoolFalse(trimmed, "has_more"); ok {
|
||||
return replaced
|
||||
}
|
||||
return body
|
||||
}
|
||||
for _, id := range ids {
|
||||
st.seen[id] = struct{}{}
|
||||
}
|
||||
if env.Data.HasMore != nil && !*env.Data.HasMore {
|
||||
// 已经是最后一页:游标链到头,清掉它,页面万一再要一页也别拿旧游标
|
||||
// 去请求(那种请求会落在下面的「整页重复」判断上收尾)。
|
||||
st.cursor = ""
|
||||
return body
|
||||
}
|
||||
// 游标只在拿到新内容时前进:请求失败/重复时页面会重试同一页,
|
||||
// 游标也跟着重试同一个。
|
||||
if cursor := rawJSONScalar(env.Data.Cursor); cursor != "" {
|
||||
st.cursor = cursor
|
||||
}
|
||||
return body
|
||||
}
|
||||
|
||||
// commentListEnvelope 评论列表响应里我们关心的字段。
|
||||
type commentListEnvelope struct {
|
||||
Data struct {
|
||||
Comments []struct {
|
||||
CommentID string `json:"comment_id"`
|
||||
} `json:"comments"`
|
||||
HasMore *bool `json:"has_more"`
|
||||
Cursor json.RawMessage `json:"next_cursor"`
|
||||
} `json:"data"`
|
||||
}
|
||||
|
||||
// isCommentListPath 判断是否是段落评论列表接口(/para_review)。
|
||||
func isCommentListPath(path string) bool {
|
||||
p := strings.TrimSuffix(strings.ToLower(strings.TrimSpace(path)), "/")
|
||||
return strings.HasSuffix(p, "/para_review")
|
||||
}
|
||||
|
||||
// commentPageKey 生成翻页状态键:接口路径 + 除 page/cursor 之外的查询参数。
|
||||
// 同一段落的不同排序(sort_by)各自一套状态,互不干扰。
|
||||
func commentPageKey(path string, query url.Values) string {
|
||||
rest := make(url.Values, len(query))
|
||||
for k, v := range query {
|
||||
if k == "page" || k == "cursor" {
|
||||
continue
|
||||
}
|
||||
rest[k] = v
|
||||
}
|
||||
keys := make([]string, 0, len(rest))
|
||||
for k := range rest {
|
||||
keys = append(keys, k)
|
||||
}
|
||||
sort.Strings(keys)
|
||||
var b strings.Builder
|
||||
b.WriteString(path)
|
||||
for _, k := range keys {
|
||||
for _, v := range rest[k] {
|
||||
b.WriteString("|" + k + "=" + v)
|
||||
}
|
||||
}
|
||||
return b.String()
|
||||
}
|
||||
|
||||
// rawJSONScalar 取一个 JSON 标量的字面量(游标是毫秒时间戳,可能被上游写成数字)。
|
||||
func rawJSONScalar(raw json.RawMessage) string {
|
||||
s := strings.TrimSpace(string(raw))
|
||||
if s == "" || s == "null" {
|
||||
return ""
|
||||
}
|
||||
if strings.HasPrefix(s, `"`) {
|
||||
var out string
|
||||
if json.Unmarshal(raw, &out) == nil {
|
||||
return strings.TrimSpace(out)
|
||||
}
|
||||
return ""
|
||||
}
|
||||
return s
|
||||
}
|
||||
|
||||
// setFirstJSONBoolFalse 把 JSON 里第一个 `"key":true` 改成 `"key":false`。
|
||||
//
|
||||
// 只做这一处等长替换,不重新序列化整份响应:评论里的 user_id、时间戳动辄十几位,
|
||||
// 走一遍 float64 会变成科学计数法,页面再解析就废了。
|
||||
func setFirstJSONBoolFalse(body, key string) (string, bool) {
|
||||
token := `"` + key + `"`
|
||||
idx := strings.Index(body, token)
|
||||
if idx < 0 {
|
||||
return body, false
|
||||
}
|
||||
i := idx + len(token)
|
||||
for i < len(body) && (body[i] == ' ' || body[i] == '\t' || body[i] == '\n' || body[i] == '\r') {
|
||||
i++
|
||||
}
|
||||
if i >= len(body) || body[i] != ':' {
|
||||
return body, false
|
||||
}
|
||||
i++
|
||||
for i < len(body) && (body[i] == ' ' || body[i] == '\t' || body[i] == '\n' || body[i] == '\r') {
|
||||
i++
|
||||
}
|
||||
if !strings.HasPrefix(body[i:], "true") {
|
||||
return body, false
|
||||
}
|
||||
return body[:i] + "false" + body[i+len("true"):], true
|
||||
}
|
||||
@@ -0,0 +1,187 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"io"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"strings"
|
||||
"sync"
|
||||
"testing"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/service/reader/rule"
|
||||
)
|
||||
|
||||
// 本文件:段落评论翻页适配(browser_comment_page.go)的测试。
|
||||
//
|
||||
// 覆盖两件事:
|
||||
// 1. page=N(N>1)被换成上一页响应给的 cursor;
|
||||
// 2. 整页重复(接口不认 page 时的表现)时把 has_more 改成 false,页面才会收尾。
|
||||
|
||||
const commentParaBase = "https://cmt.example.com/para_review?item_id=X¶=1&source=QQ"
|
||||
|
||||
func TestCommentPagerRewritesPageToCursor(t *testing.T) {
|
||||
p := newCommentPager()
|
||||
|
||||
// 第一页:原样请求,响应里的游标被记下
|
||||
target, key := p.rewriteRequest(commentParaBase + "&page=1")
|
||||
if key == "" {
|
||||
t.Fatal("段落评论接口应被识别")
|
||||
}
|
||||
if target != commentParaBase+"&page=1" {
|
||||
t.Fatalf("第一页不该改写: %q", target)
|
||||
}
|
||||
p.observeResponse(key, `{"code":0,"data":{"comments":[{"comment_id":"c1"}],"has_more":true,"next_cursor":"1500000000000"}}`)
|
||||
|
||||
// 第二页:改成带 cursor 的请求,page 不再出现
|
||||
target, key2 := p.rewriteRequest(commentParaBase + "&page=2")
|
||||
if key2 != key {
|
||||
t.Fatalf("同一段落的状态键应一致: %q vs %q", key2, key)
|
||||
}
|
||||
if !strings.Contains(target, "cursor=1500000000000") || strings.Contains(target, "page=") {
|
||||
t.Fatalf("第二页应换成 cursor 请求: %q", target)
|
||||
}
|
||||
|
||||
// 拿到新评论:游标前进到下一页
|
||||
p.observeResponse(key, `{"code":0,"data":{"comments":[{"comment_id":"c2"}],"has_more":true,"next_cursor":"1400000000000"}}`)
|
||||
target, _ = p.rewriteRequest(commentParaBase + "&page=3")
|
||||
if !strings.Contains(target, "cursor=1400000000000") {
|
||||
t.Fatalf("第三页应使用新的游标: %q", target)
|
||||
}
|
||||
|
||||
// 重新加载(page=1)会把整轮状态清掉:没有游标就原样透传
|
||||
p.rewriteRequest(commentParaBase + "&page=1")
|
||||
target, _ = p.rewriteRequest(commentParaBase + "&page=2")
|
||||
if strings.Contains(target, "cursor=") {
|
||||
t.Fatalf("重置后不该还带着旧游标: %q", target)
|
||||
}
|
||||
}
|
||||
|
||||
func TestCommentPagerStopsOnRepeatedPage(t *testing.T) {
|
||||
p := newCommentPager()
|
||||
_, key := p.rewriteRequest(commentParaBase + "&page=1")
|
||||
first := `{"code":0,"data":{"comments":[{"comment_id":"c1","like_count":562,"user":{"user_id":601948729051219}}],"has_more":true,"next_cursor":"1500000000000"}}`
|
||||
if got := p.observeResponse(key, first); got != first {
|
||||
t.Fatalf("第一页不该被改写: %q", got)
|
||||
}
|
||||
|
||||
// 同一页又被返回一次(接口忽略 page):改 has_more,让页面显示「没有更多了」
|
||||
repeat := `{"code":0,"data":{"comments":[{"comment_id":"c1","like_count":562,"user":{"user_id":601948729051219}}],"has_more":true,"next_cursor":"1500000000000"}}`
|
||||
got := p.observeResponse(key, repeat)
|
||||
if !strings.Contains(got, `"has_more":false`) {
|
||||
t.Fatalf("重复页应把 has_more 改成 false: %q", got)
|
||||
}
|
||||
// 只动 has_more 一处,其它字段(尤其是大整数)保持原样
|
||||
if strings.Replace(got, `"has_more":false`, `"has_more":true`, 1) != first {
|
||||
t.Fatalf("除 has_more 外不应改动响应: %q", got)
|
||||
}
|
||||
// 重复页不再推进游标
|
||||
target, _ := p.rewriteRequest(commentParaBase + "&page=2")
|
||||
if !strings.Contains(target, "cursor=1500000000000") {
|
||||
t.Fatalf("游标不该被重复页顶掉: %q", target)
|
||||
}
|
||||
}
|
||||
|
||||
func TestCommentPagerIgnoresOtherRequests(t *testing.T) {
|
||||
p := newCommentPager()
|
||||
cases := []string{
|
||||
"https://cmt.example.com/book/detail?page=2", // 不是评论接口
|
||||
"https://cmt.example.com/para_review?item_id=X", // 没有 page
|
||||
"https://cmt.example.com/para_review?item_id=X&cursor=1", // 已经在用游标
|
||||
"https://cmt.example.com/para_review?item_id=X&page=0", // page 非法
|
||||
"https://cmt.example.com/para_review?item_id=X&sort_by=1&", // 没有 page
|
||||
}
|
||||
for _, target := range cases {
|
||||
got, key := p.rewriteRequest(target)
|
||||
if got != target || key != "" {
|
||||
t.Fatalf("不该改写的请求被处理了: %q -> %q (key=%q)", target, got, key)
|
||||
}
|
||||
}
|
||||
// 非 JSON 响应原样返回
|
||||
if got := p.observeResponse("k", "<html></html>"); got != "<html></html>" {
|
||||
t.Fatalf("非 JSON 响应被改动: %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
// TestBrowserXHRCommentPagerFollowsCursor 端到端:页面按 page 翻页,
|
||||
// 代理层换成 cursor,接口给出新评论,最后一页重复时收尾。
|
||||
func TestBrowserXHRCommentPagerFollowsCursor(t *testing.T) {
|
||||
svc, _ := newLoginTestService(t)
|
||||
|
||||
var mu sync.Mutex
|
||||
seenCursors := []string{}
|
||||
seenPages := []string{}
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
switch r.URL.Path {
|
||||
case "/page":
|
||||
w.Header().Set("Content-Type", "text/html; charset=utf-8")
|
||||
_, _ = io.WriteString(w, `<html><body>comment page</body></html>`)
|
||||
case "/para_review":
|
||||
q := r.URL.Query()
|
||||
mu.Lock()
|
||||
seenCursors = append(seenCursors, q.Get("cursor"))
|
||||
seenPages = append(seenPages, q.Get("page"))
|
||||
mu.Unlock()
|
||||
w.Header().Set("Content-Type", "application/json; charset=utf-8")
|
||||
// 上游只认 cursor:不认识 page,靠 next_cursor 一页页往前。
|
||||
switch q.Get("cursor") {
|
||||
case "":
|
||||
_, _ = io.WriteString(w, `{"code":0,"data":{"comments":[{"comment_id":"c1"}],"has_more":true,"next_cursor":"1500000000000"}}`)
|
||||
case "1500000000000":
|
||||
_, _ = io.WriteString(w, `{"code":0,"data":{"comments":[{"comment_id":"c2"}],"has_more":true,"next_cursor":"1400000000000"}}`)
|
||||
case "1400000000000":
|
||||
_, _ = io.WriteString(w, `{"code":0,"data":{"comments":[{"comment_id":"c3"}],"has_more":false,"next_cursor":""}}`)
|
||||
default:
|
||||
// 兜底分支:上游把同一页又给了一遍(页面的 page 循环就会走到这里)
|
||||
_, _ = io.WriteString(w, `{"code":0,"data":{"comments":[{"comment_id":"c3"}],"has_more":true,"next_cursor":"1300000000000"}}`)
|
||||
}
|
||||
default:
|
||||
http.NotFound(w, r)
|
||||
}
|
||||
}))
|
||||
defer srv.Close()
|
||||
|
||||
entry, err := svc.registerBrowser(t.Context(), srv.URL, "src-1", readerTestUserID,
|
||||
browserCookieTarget{}, rule.BrowserTask{URL: srv.URL + "/page", Title: "段评"}, browserModeOpen)
|
||||
if err != nil {
|
||||
t.Fatalf("登记承载页面失败: %v", err)
|
||||
}
|
||||
call := func(page string) string {
|
||||
t.Helper()
|
||||
res, err := svc.ProxyBrowserXHR(t.Context(), entry.id, http.MethodGet,
|
||||
srv.URL+"/para_review?item_id=X¶=1&source=QQ&page="+page, nil, "")
|
||||
if err != nil {
|
||||
t.Fatalf("代理请求失败: %v", err)
|
||||
}
|
||||
return res.Body
|
||||
}
|
||||
|
||||
if body := call("1"); !strings.Contains(body, `"has_more":true`) || !strings.Contains(body, "c1") {
|
||||
t.Fatalf("第一页响应异常: %q", body)
|
||||
}
|
||||
if body := call("2"); !strings.Contains(body, "c2") {
|
||||
t.Fatalf("第二页应拿到下一页评论: %q", body)
|
||||
}
|
||||
if body := call("3"); !strings.Contains(body, "c3") || !strings.Contains(body, `"has_more":false`) {
|
||||
t.Fatalf("第三页应是真正的最后一页: %q", body)
|
||||
}
|
||||
// 页面不知道已经到底,还可能再要一页:上游又把最后一页给了一遍,
|
||||
// 适配层要把 has_more 改成 false,页面才会显示「没有更多了」并停下。
|
||||
last := call("4")
|
||||
if !strings.Contains(last, `"has_more":false`) {
|
||||
t.Fatalf("整页重复时应收尾(has_more=false): %q", last)
|
||||
}
|
||||
|
||||
mu.Lock()
|
||||
defer mu.Unlock()
|
||||
if strings.Join(seenPages, ",") != "1,,,4" {
|
||||
t.Fatalf("page 参数只应出现在没有游标可用的请求上: %v", seenPages)
|
||||
}
|
||||
if strings.Join(seenCursors, ",") != ",1500000000000,1400000000000," {
|
||||
t.Fatalf("游标传递不符预期: %v", seenCursors)
|
||||
}
|
||||
var env map[string]any
|
||||
if err := json.Unmarshal([]byte(last), &env); err != nil {
|
||||
t.Fatalf("收尾后的响应不是合法 JSON: %v", err)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,957 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"context"
|
||||
"crypto/hmac"
|
||||
"crypto/rand"
|
||||
"crypto/sha256"
|
||||
"encoding/base64"
|
||||
"encoding/hex"
|
||||
"encoding/json"
|
||||
"errors"
|
||||
"fmt"
|
||||
"io"
|
||||
"net/http"
|
||||
"net/url"
|
||||
"regexp"
|
||||
"strings"
|
||||
"sync"
|
||||
"time"
|
||||
|
||||
"github.com/PuerkitoBio/goquery"
|
||||
"go.uber.org/zap"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/helper"
|
||||
"github.com/truewhile/MeBox/internal/service/reader/rule"
|
||||
)
|
||||
|
||||
// 本文件:书源 JS 的宿主浏览器(java.startBrowser / java.startBrowserAwait)。
|
||||
//
|
||||
// legado 用内置 WebView 承载页面:书源在自己的 JS 里拼一段 HTML(线路设置页
|
||||
// 之类)或给一个网址,让用户点选、登录、过防爬校验,然后(Await 版本)把用户
|
||||
// 操作后的页面源码当成 StrResponse 返回,书源再从中解析出结果写回源变量。
|
||||
//
|
||||
// 服务端没有 WebView,于是把这一段拆成三步:
|
||||
// 1. JS 侧 startBrowserAwait 登记一个待办并阻塞等待;
|
||||
// 2. 前端轮询到待办,在 <iframe> 里承载页面(同源,因此可以直接读回 DOM);
|
||||
// 3. 用户点 √ 后把 iframe 的 DOM 回传,阻塞解除,JS 拿到 body()。
|
||||
//
|
||||
// 页面来源分两种,承载方式不同:
|
||||
// - data:text/html;base64,…:书源自带的内嵌页面,直接解码返回;
|
||||
// - http(s):服务端带书源 Cookie/请求头抓取后返回。这一步很关键——书源的
|
||||
// 登录态存在服务端,浏览器里没有这些 Cookie,若直接让浏览器打开源站,
|
||||
// 「用户后台」这类页面必然是未登录状态。页面里的资源地址会改写到同源
|
||||
// 代理,使后续请求同样带上 Cookie。
|
||||
|
||||
const (
|
||||
// browserAwaitTimeout 单次等待用户操作的上限。
|
||||
browserAwaitTimeout = 10 * time.Minute
|
||||
// browserPageTTL 待办条目与页面链接的有效期。
|
||||
browserPageTTL = 15 * time.Minute
|
||||
// browserMaxPageBytes 服务端抓取页面/资源的大小上限。
|
||||
browserMaxPageBytes = 4 << 20
|
||||
// browserFetchTimeout 抓取待承载页面/资源的超时。
|
||||
browserFetchTimeout = 20 * time.Second
|
||||
// browserModeWait 需要回传 DOM(startBrowserAwait)。
|
||||
browserModeWait = "wait"
|
||||
// browserModeOpen 只展示(startBrowser / showBrowser)。
|
||||
browserModeOpen = "open"
|
||||
)
|
||||
|
||||
// BrowserPage 前端待承载的一个页面(登录对话框轮询用)。
|
||||
type BrowserPage struct {
|
||||
ID string `json:"id"`
|
||||
Title string `json:"title"`
|
||||
// Mode wait = 需回传(点 √ 后回传 DOM);open = 仅展示。
|
||||
Mode string `json:"mode"`
|
||||
// Seq 登记序号(单调递增)。待办表可能同时存多个页面,前端按它取最新的
|
||||
// 那个展示——不能依赖返回顺序,map 遍历是随机的。
|
||||
Seq int64 `json:"seq"`
|
||||
// PageURL 同源承载地址(iframe src),带 HMAC 签名。
|
||||
PageURL string `json:"page_url"`
|
||||
// Refetch 见 rule.BrowserTask.Refetch。
|
||||
Refetch bool `json:"refetch"`
|
||||
// SourceID 发起该请求的书源。
|
||||
SourceID string `json:"source_id"`
|
||||
// TargetURL 原始地址(data: 与超长地址不回传)。
|
||||
TargetURL string `json:"target_url,omitempty"`
|
||||
}
|
||||
|
||||
// BrowserPageSnapshot 承载页面时返回的数据。
|
||||
type BrowserPageSnapshot struct {
|
||||
ID string
|
||||
HTML string
|
||||
SourceID string
|
||||
FinalURL string
|
||||
}
|
||||
|
||||
// pendingBrowser 一个待用户完成的页面。
|
||||
// seq 为登记序号(越大越新,见 ReaderService.browserSeq),供前端挑选最新页面。
|
||||
type pendingBrowser struct {
|
||||
id string
|
||||
seq int64
|
||||
sourceID string
|
||||
sourceURL string
|
||||
userID string
|
||||
request rule.BrowserTask
|
||||
// cookies 是发起这次页面的会话 Cookie 累积区:页面在面板里发出的
|
||||
// 请求(接口代理 / 资源代理)会把 Set-Cookie 写回它,用户点「完成」后
|
||||
// 书源会话就能读到登录凭证。
|
||||
cookies browserCookieTarget
|
||||
// html 已就绪的页面源码(data: 直接解码;http 抓取后资源地址已改写)。
|
||||
html string
|
||||
finalURL string
|
||||
mode string
|
||||
expires time.Time
|
||||
// comments 段落评论接口的翻页适配(上游页面按 page 翻页,接口只认 cursor,
|
||||
// 见 browser_comment_page.go)。
|
||||
comments *commentPager
|
||||
|
||||
done chan struct{}
|
||||
mu sync.Mutex
|
||||
result rule.BrowserResult
|
||||
err error
|
||||
finished bool
|
||||
}
|
||||
|
||||
// browserHost 把服务层的浏览器面板绑定到某个书源会话(实现 rule.BrowserHost)。
|
||||
type browserHost struct {
|
||||
svc *ReaderService
|
||||
sourceURL string
|
||||
sourceID string
|
||||
userID string
|
||||
// state / capture 是这次会话的 Cookie 累积区与 enabledCookieJar 开关。
|
||||
// 面板承载的页面自己发起的请求(ProxyBrowserXHR / 资源代理)会把
|
||||
// 响应 Set-Cookie 写进 state,从而让「页面里登录成功 = 书源会话拿到凭证」。
|
||||
state *sourceState
|
||||
capture bool
|
||||
}
|
||||
|
||||
func (h *browserHost) AwaitBrowser(ctx context.Context, req rule.BrowserTask) (rule.BrowserResult, error) {
|
||||
return h.svc.awaitBrowser(ctx, h.sourceURL, h.sourceID, h.userID, h.cookieTarget(), req)
|
||||
}
|
||||
|
||||
func (h *browserHost) OpenBrowser(ctx context.Context, req rule.BrowserTask) error {
|
||||
return h.svc.openBrowser(ctx, h.sourceURL, h.sourceID, h.userID, h.cookieTarget(), req)
|
||||
}
|
||||
|
||||
// cookieTarget 打包会话状态与开关;capture 为假时返回携带 nil state 的目标。
|
||||
func (h *browserHost) cookieTarget() browserCookieTarget {
|
||||
if !h.capture {
|
||||
return browserCookieTarget{}
|
||||
}
|
||||
return browserCookieTarget{state: h.state}
|
||||
}
|
||||
|
||||
// browserCookieTarget 面板链路要回写 Cookie 的目标会话。
|
||||
type browserCookieTarget struct {
|
||||
state *sourceState
|
||||
}
|
||||
|
||||
// captureFrom 把这次请求(含重定向各跳)下发的 Set-Cookie 写回会话。
|
||||
// ctx 上挂了 sink 时以它为准(逐跳收集,见 cookies.go),最后再补最终响应。
|
||||
// 归属域一律 Cookie 的 Domain 优先、否则该跳地址。
|
||||
func (t browserCookieTarget) captureFrom(ctx context.Context, resp *http.Response) {
|
||||
if t.state == nil || resp == nil {
|
||||
return
|
||||
}
|
||||
sink := cookieSinkFrom(ctx)
|
||||
if sink == nil {
|
||||
sink = &cookieSink{}
|
||||
}
|
||||
if resp.Request != nil {
|
||||
sink.add(resp.Request.URL, resp)
|
||||
}
|
||||
sink.apply(t.state)
|
||||
}
|
||||
|
||||
// ─── 服务层入口 ────────────────────────────────────────────────────────────
|
||||
|
||||
// awaitBrowser 登记待办并阻塞等待用户回传页面内容。
|
||||
func (s *ReaderService) awaitBrowser(ctx context.Context, sourceURL, sourceID, userID string, cookies browserCookieTarget, req rule.BrowserTask) (rule.BrowserResult, error) {
|
||||
entry, err := s.registerBrowser(ctx, sourceURL, sourceID, userID, cookies, req, browserModeWait)
|
||||
if err != nil {
|
||||
return rule.BrowserResult{}, err
|
||||
}
|
||||
defer s.dropBrowser(entry.id)
|
||||
|
||||
timeout := time.NewTimer(browserAwaitTimeout)
|
||||
defer timeout.Stop()
|
||||
select {
|
||||
case <-entry.done:
|
||||
entry.mu.Lock()
|
||||
defer entry.mu.Unlock()
|
||||
if entry.err != nil {
|
||||
return rule.BrowserResult{}, entry.err
|
||||
}
|
||||
return entry.result, nil
|
||||
case <-ctx.Done():
|
||||
return rule.BrowserResult{}, ctx.Err()
|
||||
case <-timeout.C:
|
||||
return rule.BrowserResult{}, fmt.Errorf("等待页面操作超时(%s),请重试", browserAwaitTimeout)
|
||||
}
|
||||
}
|
||||
|
||||
// openBrowser 登记待办但不等待(页面展示给用户即可)。
|
||||
func (s *ReaderService) openBrowser(ctx context.Context, sourceURL, sourceID, userID string, cookies browserCookieTarget, req rule.BrowserTask) error {
|
||||
_, err := s.registerBrowser(ctx, sourceURL, sourceID, userID, cookies, req, browserModeOpen)
|
||||
return err
|
||||
}
|
||||
|
||||
// registerBrowser 准备页面内容并登记待办。
|
||||
func (s *ReaderService) registerBrowser(ctx context.Context, sourceURL, sourceID, userID string, cookies browserCookieTarget, req rule.BrowserTask, mode string) (*pendingBrowser, error) {
|
||||
// 先分配 ID:页面里的资源代理地址需要用它签名。
|
||||
id := newBrowserID()
|
||||
html, finalURL, err := s.prepareBrowserPage(ctx, sourceURL, id, cookies, req)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
if finalURL == "" {
|
||||
finalURL = req.URL
|
||||
}
|
||||
entry := &pendingBrowser{
|
||||
id: id,
|
||||
sourceID: sourceID,
|
||||
sourceURL: sourceURL,
|
||||
userID: userID,
|
||||
request: req,
|
||||
cookies: cookies,
|
||||
html: html,
|
||||
finalURL: finalURL,
|
||||
mode: mode,
|
||||
expires: time.Now().Add(browserPageTTL),
|
||||
done: make(chan struct{}),
|
||||
comments: newCommentPager(),
|
||||
}
|
||||
|
||||
s.browserMu.Lock()
|
||||
if s.browserPending == nil {
|
||||
s.browserPending = map[string]*pendingBrowser{}
|
||||
}
|
||||
s.pruneBrowsersLocked()
|
||||
s.browserSeq++
|
||||
entry.seq = s.browserSeq
|
||||
s.browserPending[id] = entry
|
||||
s.browserMu.Unlock()
|
||||
|
||||
if s.log != nil {
|
||||
s.log.Info("reader:待用户完成页面操作",
|
||||
zap.String("source", sourceID), zap.String("mode", mode), zap.String("title", req.Title))
|
||||
}
|
||||
return entry, nil
|
||||
}
|
||||
|
||||
// browserBridgeScript 注入到承载页面里:父窗口请求时回传当前 DOM,把页面自己的
|
||||
// fetch / XMLHttpRequest 转交给父窗口代为请求,并补上沙箱里不可用的存储对象。
|
||||
//
|
||||
// 为什么不让父窗口直接读 contentDocument:那要求 iframe 同源,而书源页面是
|
||||
// 第三方 HTML,拿到同源权限就能读写 MeBox 自己的页面与 localStorage(JWT)。
|
||||
// 改成 postMessage 后,iframe 可以只开 allow-scripts(不透明源)。
|
||||
//
|
||||
// 为什么要补存储对象:不透明源里访问 localStorage / sessionStorage / document.cookie
|
||||
// 会直接抛 SecurityError,SPA 类的「用户后台」在启动读 token 时就崩了,
|
||||
// 表现是一直停在「正在安全加载」。这里用内存版兜住,既不给出同源权限,
|
||||
// 又让页面能正常跑完启动流程。
|
||||
//
|
||||
// 为什么要代发请求:iframe 是不透明源,页面自己的 XHR 既带不上书源 Cookie,
|
||||
// 也会被 CORS 拦掉——靠接口取数的页面会取不到数据。
|
||||
const browserBridgeScript = `<script data-mebox-browser-bridge="1">(function(){` +
|
||||
`if(window.__meboxBridgeInstalled){return}window.__meboxBridgeInstalled=true;` +
|
||||
// ── 沙箱内存储兜底 ──
|
||||
`function memStore(){var d={};return{` +
|
||||
`getItem:function(k){k=String(k);return Object.prototype.hasOwnProperty.call(d,k)?d[k]:null},` +
|
||||
`setItem:function(k,v){d[String(k)]=String(v)},` +
|
||||
`removeItem:function(k){delete d[String(k)]},` +
|
||||
`clear:function(){d={}},` +
|
||||
`key:function(i){var ks=Object.keys(d);return i>=0&&i<ks.length?ks[i]:null},` +
|
||||
`get length(){return Object.keys(d).length}}}` +
|
||||
`function patchStore(n){try{var s=memStore();Object.defineProperty(window,n,{configurable:true,get:function(){return s}})}catch(e){}}` +
|
||||
`patchStore('localStorage');patchStore('sessionStorage');` +
|
||||
// document.cookie 也要给:这类页面常用它判断是否已登录(真实 Cookie 在服务端,
|
||||
// 浏览器里没有)。把书源在该站点的 Cookie 预置进去,页面才会正常渲染。
|
||||
`var ck={};` +
|
||||
`(function(seed){String(seed||'').split(';').forEach(function(p){var i=p.indexOf('=');if(i>0){ck[p.slice(0,i).trim()]=p.slice(i+1).trim()}})})` + "(`__MEBOX_COOKIE__`)" + `;` +
|
||||
`try{Object.defineProperty(document,'cookie',{configurable:true,` +
|
||||
`get:function(){var o=[];for(var k in ck){o.push(k+'='+ck[k])}return o.join('; ')},` +
|
||||
`set:function(v){try{var p=String(v).split(';')[0].split('=');if(p[0]){ck[p[0].trim()]=p.slice(1).join('=')}}catch(e){}}})}catch(e){}` +
|
||||
// 书源页面自身的基地址:相对地址要按它解析,而不是按承载地址(我们自己的源)
|
||||
`var BASE=` + "`__MEBOX_BASE_URL__`" + `;` +
|
||||
`function abs(u){try{return new URL(u,BASE||location.href).href}catch(e){try{return new URL(u,location.href).href}catch(e2){return u}}}` +
|
||||
// ── DOM 回传 ──
|
||||
`window.addEventListener('message',function(e){` +
|
||||
`if(!e||e.data!=='__mebox_dom__'){return}` +
|
||||
`try{parent.postMessage({__mebox_dom__:true,html:document.documentElement.outerHTML},'*')}catch(err){}` +
|
||||
`},false);` +
|
||||
// ── 请求代理(父窗口持有登录态,由它转交服务端补 Cookie) ──
|
||||
`var seq=0,waiters={};` +
|
||||
`window.addEventListener('message',function(e){` +
|
||||
`var d=e&&e.data;` +
|
||||
`if(!d||d.__mebox_proxy_res__!==true){return}` +
|
||||
`var w=waiters[d.reqId];` +
|
||||
`if(!w){return}delete waiters[d.reqId];w(d);` +
|
||||
`},false);` +
|
||||
`function rpc(method,url,headers,body){` +
|
||||
`return new Promise(function(resolve,reject){` +
|
||||
`var id='r'+(++seq);waiters[id]=resolve;` +
|
||||
`try{parent.postMessage({__mebox_proxy__:true,reqId:id,url:url,method:method,headers:headers||{},body:body||''},'*')}catch(e){delete waiters[id];reject(e);return}` +
|
||||
`setTimeout(function(){if(waiters[id]){delete waiters[id];reject(new Error('mebox proxy timeout'))}},60000);` +
|
||||
`})}` +
|
||||
`function decode(res){return res&&res.base64?atob(res.body||''):((res&&res.body)||'')}` +
|
||||
`var origFetch=window.fetch;` +
|
||||
`if(origFetch){window.fetch=function(input,init){` +
|
||||
`init=init||{};` +
|
||||
`var url=typeof input==='string'?input:(input&&input.url);` +
|
||||
`if(!url){return origFetch.apply(this,arguments)}` +
|
||||
`var method=String(init.method||(input&&input.method)||'GET').toUpperCase();` +
|
||||
`var headers={};` +
|
||||
`try{new Headers(init.headers||(input&&input.headers)||{}).forEach(function(v,k){headers[k]=v})}catch(e){}` +
|
||||
`var body=init.body?String(init.body):'';` +
|
||||
`return rpc(method,abs(url),headers,body).then(function(res){` +
|
||||
`return new Response(decode(res),{status:(res&&res.status)||200,headers:{'Content-Type':(res&&res.contentType)||'text/plain'}})` +
|
||||
`})}}` +
|
||||
`var OrigXHR=window.XMLHttpRequest;` +
|
||||
`function ProxyXHR(){this._h={};this._m='GET';this._u='';this.readyState=0;this.status=0;this.response=null;this.responseText='';this.responseType=''}` +
|
||||
`ProxyXHR.prototype.open=function(m,u){this._m=m;this._u=u;this.readyState=1;this._fire('onreadystatechange')};` +
|
||||
`ProxyXHR.prototype.setRequestHeader=function(k,v){this._h[k]=v};` +
|
||||
`ProxyXHR.prototype._fire=function(n){try{if(this[n])this[n]()}catch(e){}};` +
|
||||
`ProxyXHR.prototype.send=function(body){` +
|
||||
`var self=this;` +
|
||||
`rpc(String(this._m).toUpperCase(),abs(this._u),this._h,body?String(body):'').then(function(res){` +
|
||||
`self.status=(res&&res.status)||200;self.readyState=4;self.responseText=decode(res);` +
|
||||
`if(self.responseType==='json'){try{self.response=JSON.parse(self.responseText)}catch(e){self.response=null}}else{self.response=self.responseText}` +
|
||||
`self._fire('onreadystatechange');self._fire('onload')` +
|
||||
`},function(){self.status=0;self.readyState=4;self._fire('onreadystatechange');self._fire('onerror')})};` +
|
||||
`ProxyXHR.prototype.getResponseHeader=function(){return null};` +
|
||||
`ProxyXHR.prototype.getAllResponseHeaders=function(){return ''};` +
|
||||
`ProxyXHR.prototype.abort=function(){};` +
|
||||
`window.XMLHttpRequest=ProxyXHR;` +
|
||||
`})();</script>`
|
||||
|
||||
// injectBrowserBridge 把回传/代理脚本插进页面 <head>(没有 head 就插在最前面),
|
||||
// 并把书源页面的基地址与 Cookie 填进脚本:前者供相对地址解析,后者供页面判断
|
||||
// 登录态(真实 Cookie 在服务端,页面看不到)。
|
||||
func injectBrowserBridge(html, baseURL, cookie string) string {
|
||||
if strings.Contains(html, `data-mebox-browser-bridge`) {
|
||||
return html
|
||||
}
|
||||
script := strings.ReplaceAll(browserBridgeScript, "__MEBOX_BASE_URL__", jsStringEscape(baseURL))
|
||||
script = strings.ReplaceAll(script, "__MEBOX_COOKIE__", jsStringEscape(cookie))
|
||||
lower := strings.ToLower(html)
|
||||
if i := strings.Index(lower, "<head>"); i >= 0 {
|
||||
pos := i + len("<head>")
|
||||
return html[:pos] + script + html[pos:]
|
||||
}
|
||||
if i := strings.Index(lower, "<html"); i >= 0 {
|
||||
if j := strings.Index(lower[i:], ">"); j >= 0 {
|
||||
pos := i + j + 1
|
||||
return html[:pos] + script + html[pos:]
|
||||
}
|
||||
}
|
||||
return script + html
|
||||
}
|
||||
|
||||
// jsStringEscape 转义要嵌进 JS 模板字符串的地址(反引号 / 反斜杠 / ${)。
|
||||
func jsStringEscape(s string) string {
|
||||
r := strings.NewReplacer("\\", "\\\\", "`", "\\`", "${", "\\${")
|
||||
return r.Replace(s)
|
||||
}
|
||||
|
||||
// prepareBrowserPage 得到 iframe 要承载的页面源码。
|
||||
//
|
||||
// 返回的页面里会注入回传/代理脚本;脚本需要两样东西:
|
||||
// - 基地址:解析页面里的相对地址(书源自带的 data: 页面没有真实地址,
|
||||
// 退化成书源站点地址——书源自己的 BaseUrl() 也指向它);
|
||||
// - Cookie:页面常用 document.cookie 判断登录态,而真实 Cookie 在服务端,
|
||||
// 不预置的话「用户后台」会以为未登录并把浏览器导到 /login。
|
||||
func (s *ReaderService) prepareBrowserPage(ctx context.Context, sourceURL, id string, cookies browserCookieTarget, req rule.BrowserTask) (string, string, error) {
|
||||
// 1) 书源自带 HTML(显式 html 参数或 data: URL)
|
||||
if html := strings.TrimSpace(req.HTML); html != "" {
|
||||
return injectBrowserBridge(html, sourceURL, s.browserCookieHeader(ctx, sourceURL, sourceURL)), req.URL, nil
|
||||
}
|
||||
if html := rule.ParseDataHTML(req.URL); html != "" {
|
||||
return injectBrowserBridge(html, sourceURL, s.browserCookieHeader(ctx, sourceURL, sourceURL)), req.URL, nil
|
||||
}
|
||||
// 2) 外部地址:服务端带书源 Cookie 抓取,并把资源地址改写到代理
|
||||
if !strings.HasPrefix(req.URL, "http://") && !strings.HasPrefix(req.URL, "https://") {
|
||||
return "", "", fmt.Errorf("无法承载该地址: %s", truncateForLog(req.URL, 120))
|
||||
}
|
||||
body, finalURL, contentType, err := s.fetchBrowserPage(ctx, sourceURL, cookies, req.URL)
|
||||
if err != nil {
|
||||
return "", "", fmt.Errorf("打开页面失败: %w", err)
|
||||
}
|
||||
// 空响应没有可展示的内容:以前照样包一层空 <pre> 交给 iframe,面板就是一片白,
|
||||
// 用户看不出是地址有问题还是自己没登录。这里直接报错(如段评地址参数被改写、
|
||||
// 上游按缺参数返回空 body 的情形)。
|
||||
if strings.TrimSpace(body) == "" {
|
||||
return "", "", fmt.Errorf("目标地址没有返回任何内容: %s", truncateForLog(req.URL, 120))
|
||||
}
|
||||
cookie := s.browserCookieHeader(ctx, sourceURL, finalURL)
|
||||
if !strings.Contains(strings.ToLower(contentType), "html") {
|
||||
// 非 HTML(如 JSON 接口):包一层 <pre>,至少让用户看到内容
|
||||
wrapped := "<!DOCTYPE html><html><head><meta charset=\"utf-8\"><title>" + htmlEscape(req.Title) +
|
||||
"</title></head><body><pre style=\"white-space:pre-wrap;word-break:break-all;padding:16px;" +
|
||||
"font:13px/1.6 ui-monospace,monospace\">" + htmlEscape(body) + "</pre></body></html>"
|
||||
return injectBrowserBridge(wrapped, finalURL, cookie), finalURL, nil
|
||||
}
|
||||
return injectBrowserBridge(s.rewriteBrowserHTML(body, finalURL, id), finalURL, cookie), finalURL, nil
|
||||
}
|
||||
|
||||
// browserCookieHeader 取书源在目标站点上的 Cookie 串(页面里预置 document.cookie 用)。
|
||||
func (s *ReaderService) browserCookieHeader(ctx context.Context, sourceURL, target string) string {
|
||||
if s.repo == nil || sourceURL == "" {
|
||||
return ""
|
||||
}
|
||||
return s.newSourceState(ctx, sourceURL).CookieForRequest(target)
|
||||
}
|
||||
|
||||
// fetchBrowserPage 服务端抓取页面(附带书源 Cookie / 登录请求头 / 书源请求头)。
|
||||
func (s *ReaderService) fetchBrowserPage(ctx context.Context, sourceURL string, cookies browserCookieTarget, target string) (string, string, string, error) {
|
||||
contentType, _, data, finalURL, err := s.requestBrowserResource(ctx, sourceURL, cookies, target)
|
||||
if err != nil {
|
||||
return "", "", "", err
|
||||
}
|
||||
return string(data), finalURL, contentType, nil
|
||||
}
|
||||
|
||||
// FetchBrowserAsset 代理拉取页面资源(带书源 Cookie/请求头),供 iframe 内引用。
|
||||
// 返回 (contentType, status, body, error)。按待办 ID 找会话,顺带把响应
|
||||
// Set-Cookie 写回(页面可能靠资源响应续期会话)。
|
||||
func (s *ReaderService) FetchBrowserAsset(ctx context.Context, id, target string) (string, int, []byte, error) {
|
||||
entry := s.lookupBrowser(id)
|
||||
if entry == nil {
|
||||
return "", 0, nil, errors.New("页面已过期,请重新打开")
|
||||
}
|
||||
contentType, status, data, _, err := s.requestBrowserResource(ctx, entry.sourceURL, entry.cookies, target)
|
||||
return contentType, status, data, err
|
||||
}
|
||||
|
||||
// requestBrowserResource 带书源凭据请求一个外部地址。
|
||||
func (s *ReaderService) requestBrowserResource(ctx context.Context, sourceURL string, cookies browserCookieTarget, target string) (string, int, []byte, string, error) {
|
||||
reqCtx, cancel := context.WithTimeout(ctx, browserFetchTimeout)
|
||||
defer cancel()
|
||||
// 挂上 sink,让重定向各跳的 Set-Cookie 也能被收集。
|
||||
if cookies.state != nil {
|
||||
reqCtx = withCookieSink(reqCtx, &cookieSink{})
|
||||
}
|
||||
|
||||
state := s.newSourceState(reqCtx, sourceURL)
|
||||
httpReq, err := http.NewRequestWithContext(reqCtx, http.MethodGet, target, nil)
|
||||
if err != nil {
|
||||
return "", 0, nil, "", err
|
||||
}
|
||||
for k, v := range state.LoginHeaderMap() {
|
||||
if strings.EqualFold(k, "cookie") {
|
||||
continue
|
||||
}
|
||||
httpReq.Header.Set(k, v)
|
||||
}
|
||||
if ck := state.CookieForRequest(target); ck != "" {
|
||||
httpReq.Header.Set("Cookie", ck)
|
||||
}
|
||||
// 书源级请求头(含 Referer / UA),与阅读请求保持一致
|
||||
if s.repo != nil {
|
||||
if src, findErr := s.repo.GetSourceByURL(reqCtx, sourceURL); findErr == nil && src != nil && src.Header != "" {
|
||||
var headers map[string]any
|
||||
if json.Unmarshal([]byte(src.Header), &headers) == nil {
|
||||
for k, v := range headers {
|
||||
if httpReq.Header.Get(k) == "" {
|
||||
httpReq.Header.Set(k, fmt.Sprintf("%v", v))
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
// Accept-Encoding 交给 net/http 管:显式设置会让它放弃自动解压,
|
||||
// 压缩过的页面/资源就会以原始字节回到 iframe(页面直接白屏或乱码)。
|
||||
helper.StripAcceptEncoding(httpReq.Header)
|
||||
resp, err := s.http.Do(httpReq)
|
||||
if err != nil {
|
||||
return "", 0, nil, "", err
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
data, err := io.ReadAll(io.LimitReader(resp.Body, browserMaxPageBytes))
|
||||
if err != nil {
|
||||
return "", 0, nil, "", err
|
||||
}
|
||||
data = helper.DecompressBody(resp, data)
|
||||
finalURL := target
|
||||
if resp.Request != nil && resp.Request.URL != nil {
|
||||
finalURL = resp.Request.URL.String()
|
||||
}
|
||||
// 页面/资源响应下发的 Set-Cookie 也写回会话(含跳转各跳)。
|
||||
cookies.captureFrom(reqCtx, resp)
|
||||
contentType := resp.Header.Get("Content-Type")
|
||||
if contentType == "" {
|
||||
contentType = "application/octet-stream"
|
||||
}
|
||||
return contentType, resp.StatusCode, data, finalURL, nil
|
||||
}
|
||||
|
||||
// BrowserXHRResult 页面内 fetch/XHR 经服务端转发后的响应。
|
||||
type BrowserXHRResult struct {
|
||||
Status int `json:"status"`
|
||||
ContentType string `json:"content_type"`
|
||||
Body string `json:"body"`
|
||||
// Base64 为真时 Body 是 base64(二进制资源)。
|
||||
Base64 bool `json:"base64"`
|
||||
}
|
||||
|
||||
// browserXHRMaxBytes 转发接口响应的大小上限。
|
||||
const browserXHRMaxBytes = 4 << 20
|
||||
|
||||
// ProxyBrowserXHR 以书源身份转发页面内的接口请求。
|
||||
//
|
||||
// 承载页面是不透明源,页面自己的 XHR 既带不上书源 Cookie 也会被 CORS 拦掉,
|
||||
// 所以由父窗口把请求转交进来,这里补上书源凭据再发出去。
|
||||
// 返回的 HTTP 状态/内容类型原样回给页面,让页面自己的逻辑能正常分支。
|
||||
func (s *ReaderService) ProxyBrowserXHR(ctx context.Context, id, method, target string, headers map[string]string, body string) (*BrowserXHRResult, error) {
|
||||
entry := s.lookupBrowser(id)
|
||||
if entry == nil {
|
||||
return nil, errors.New("页面已过期,请重新打开")
|
||||
}
|
||||
if !strings.HasPrefix(target, "http://") && !strings.HasPrefix(target, "https://") {
|
||||
return nil, fmt.Errorf("仅支持 http(s) 地址: %s", truncateForLog(target, 120))
|
||||
}
|
||||
if method == "" {
|
||||
method = http.MethodGet
|
||||
}
|
||||
// 段落评论接口:评论页按 page=N 翻页,而接口只认 cursor(见
|
||||
// browser_comment_page.go)。改写发生在补书源凭据之前,Cookie/Referer
|
||||
// 仍然按目标站点选取。
|
||||
pagerKey := ""
|
||||
if entry.comments != nil {
|
||||
target, pagerKey = entry.comments.rewriteRequest(target)
|
||||
}
|
||||
reqCtx, cancel := context.WithTimeout(ctx, browserFetchTimeout)
|
||||
defer cancel()
|
||||
// 挂上 sink:页面自己发起的请求(如扫码登录的轮询/取票跳转)下发的
|
||||
// Set-Cookie 要能落到书源会话里,否则「面板里登录成功」书源却始终未登录。
|
||||
if entry.cookies.state != nil {
|
||||
reqCtx = withCookieSink(reqCtx, &cookieSink{})
|
||||
}
|
||||
|
||||
httpReq, err := http.NewRequestWithContext(reqCtx, strings.ToUpper(method), target, strings.NewReader(body))
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
// 先放页面自己声明的头,再用书源凭据补缺(Cookie 始终以服务端为准)
|
||||
for k, v := range headers {
|
||||
if strings.EqualFold(k, "cookie") || strings.EqualFold(k, "host") ||
|
||||
strings.EqualFold(k, "content-length") {
|
||||
continue
|
||||
}
|
||||
httpReq.Header.Set(k, v)
|
||||
}
|
||||
state := s.newSourceState(reqCtx, entry.sourceURL)
|
||||
for k, v := range state.LoginHeaderMap() {
|
||||
if strings.EqualFold(k, "cookie") {
|
||||
continue
|
||||
}
|
||||
if httpReq.Header.Get(k) == "" {
|
||||
httpReq.Header.Set(k, v)
|
||||
}
|
||||
}
|
||||
if httpReq.Header.Get("Cookie") == "" {
|
||||
if ck := state.CookieForRequest(target); ck != "" {
|
||||
httpReq.Header.Set("Cookie", ck)
|
||||
}
|
||||
}
|
||||
if s.repo != nil {
|
||||
if src, findErr := s.repo.GetSourceByURL(reqCtx, entry.sourceURL); findErr == nil && src != nil && src.Header != "" {
|
||||
var extra map[string]any
|
||||
if json.Unmarshal([]byte(src.Header), &extra) == nil {
|
||||
for k, v := range extra {
|
||||
if httpReq.Header.Get(k) == "" {
|
||||
httpReq.Header.Set(k, fmt.Sprintf("%v", v))
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
helper.StripAcceptEncoding(httpReq.Header)
|
||||
resp, err := s.http.Do(httpReq)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
data, err := io.ReadAll(io.LimitReader(resp.Body, browserXHRMaxBytes))
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
data = helper.DecompressBody(resp, data)
|
||||
// 页面请求同样把 Set-Cookie 写回书源会话(含跳转各跳)。
|
||||
entry.cookies.captureFrom(reqCtx, resp)
|
||||
contentType := resp.Header.Get("Content-Type")
|
||||
if contentType == "" {
|
||||
contentType = "text/plain"
|
||||
}
|
||||
out := &BrowserXHRResult{Status: resp.StatusCode, ContentType: contentType}
|
||||
if isTextualContent(contentType) {
|
||||
out.Body = string(data)
|
||||
if entry.comments != nil && pagerKey != "" {
|
||||
out.Body = entry.comments.observeResponse(pagerKey, out.Body)
|
||||
}
|
||||
} else {
|
||||
out.Body = base64.StdEncoding.EncodeToString(data)
|
||||
out.Base64 = true
|
||||
}
|
||||
return out, nil
|
||||
}
|
||||
|
||||
// isTextualContent 判断响应是否可以直接当字符串交给页面。
|
||||
func isTextualContent(contentType string) bool {
|
||||
ct := strings.ToLower(contentType)
|
||||
for _, p := range []string{"text/", "json", "javascript", "xml", "html", "x-www-form-urlencoded", "csv"} {
|
||||
if strings.Contains(ct, p) {
|
||||
return true
|
||||
}
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
// ─── 资源改写与代理 ────────────────────────────────────────────────────────
|
||||
|
||||
// cssURLRe 匹配 CSS 里的 url(...) 引用。
|
||||
// Go 的 regexp 不支持反向引用,因此引号用可选的成对字符类近似匹配。
|
||||
var cssURLRe = regexp.MustCompile(`url\(\s*['"]?([^'")]+)['"]?\s*\)`)
|
||||
|
||||
// rewriteBrowserHTML 把页面里的资源/表单/站内链接地址改写到同源代理,
|
||||
// 使页面在 iframe 里的后续请求同样带上书源 Cookie。
|
||||
func (s *ReaderService) rewriteBrowserHTML(body, baseURL, id string) string {
|
||||
base, err := url.Parse(baseURL)
|
||||
if err != nil {
|
||||
return body
|
||||
}
|
||||
doc, err := goquery.NewDocumentFromReader(strings.NewReader(body))
|
||||
if err != nil {
|
||||
return body
|
||||
}
|
||||
rewrite := func(sel, attr string) {
|
||||
doc.Find(sel).Each(func(_ int, node *goquery.Selection) {
|
||||
val, ok := node.Attr(attr)
|
||||
if !ok || strings.TrimSpace(val) == "" {
|
||||
return
|
||||
}
|
||||
if abs := absoluteBrowserURL(base, val); abs != "" {
|
||||
node.SetAttr(attr, s.browserAssetProxyURL(id, abs))
|
||||
}
|
||||
})
|
||||
}
|
||||
rewrite("img[src]", "src")
|
||||
rewrite("script[src]", "src")
|
||||
rewrite("link[href]", "href")
|
||||
rewrite("iframe[src]", "src")
|
||||
rewrite("source[src]", "src")
|
||||
rewrite("video[src]", "src")
|
||||
rewrite("audio[src]", "src")
|
||||
// 表单提交也走代理,避免 POST 丢掉 Cookie
|
||||
rewrite("form[action]", "action")
|
||||
// 站内链接走代理;站外链接保留(用户可能确实想出去)
|
||||
doc.Find("a[href]").Each(func(_ int, node *goquery.Selection) {
|
||||
val, ok := node.Attr("href")
|
||||
if !ok {
|
||||
return
|
||||
}
|
||||
abs := absoluteBrowserURL(base, val)
|
||||
if abs == "" || !sameSite(abs, baseURL) {
|
||||
return
|
||||
}
|
||||
node.SetAttr("href", s.browserAssetProxyURL(id, abs))
|
||||
})
|
||||
// 内联样式里的 url(...)
|
||||
doc.Find("[style]").Each(func(_ int, node *goquery.Selection) {
|
||||
style, ok := node.Attr("style")
|
||||
if !ok || !strings.Contains(style, "url(") {
|
||||
return
|
||||
}
|
||||
node.SetAttr("style", rewriteCSSURLs(style, base, func(abs string) string {
|
||||
return s.browserAssetProxyURL(id, abs)
|
||||
}))
|
||||
})
|
||||
out, err := doc.Html()
|
||||
if err != nil {
|
||||
return body
|
||||
}
|
||||
return "<!DOCTYPE html>" + out
|
||||
}
|
||||
|
||||
// RewriteBrowserCSS 改写 CSS 文本里的 url(...) 引用(资源代理用)。
|
||||
func (s *ReaderService) RewriteBrowserCSS(css, baseURL, id string) string {
|
||||
base, err := url.Parse(baseURL)
|
||||
if err != nil {
|
||||
return css
|
||||
}
|
||||
return rewriteCSSURLs(css, base, func(abs string) string {
|
||||
return s.browserAssetProxyURL(id, abs)
|
||||
})
|
||||
}
|
||||
|
||||
func rewriteCSSURLs(css string, base *url.URL, proxy func(abs string) string) string {
|
||||
return cssURLRe.ReplaceAllStringFunc(css, func(m string) string {
|
||||
sub := cssURLRe.FindStringSubmatch(m)
|
||||
if len(sub) < 2 {
|
||||
return m
|
||||
}
|
||||
abs := absoluteBrowserURL(base, sub[1])
|
||||
if abs == "" {
|
||||
return m
|
||||
}
|
||||
return `url("` + proxy(abs) + `")`
|
||||
})
|
||||
}
|
||||
|
||||
// absoluteBrowserURL 把页面里的相对地址解析成绝对地址;不可代理的协议返回空串。
|
||||
func absoluteBrowserURL(base *url.URL, ref string) string {
|
||||
ref = strings.TrimSpace(ref)
|
||||
if ref == "" || strings.HasPrefix(ref, "#") {
|
||||
return ""
|
||||
}
|
||||
lower := strings.ToLower(ref)
|
||||
for _, p := range []string{"data:", "javascript:", "mailto:", "tel:", "blob:", "about:", "ws:", "wss:"} {
|
||||
if strings.HasPrefix(lower, p) {
|
||||
return ""
|
||||
}
|
||||
}
|
||||
parsed, err := url.Parse(ref)
|
||||
if err != nil {
|
||||
return ""
|
||||
}
|
||||
abs := base.ResolveReference(parsed).String()
|
||||
if !strings.HasPrefix(abs, "http://") && !strings.HasPrefix(abs, "https://") {
|
||||
return ""
|
||||
}
|
||||
return abs
|
||||
}
|
||||
|
||||
// sameSite 判断两个地址是否同站(eTLD+1)。
|
||||
func sameSite(a, b string) bool {
|
||||
return rule.CookieDomain(a) == rule.CookieDomain(b)
|
||||
}
|
||||
|
||||
// ─── 签名与地址 ────────────────────────────────────────────────────────────
|
||||
|
||||
// browserSign 用 JWT 密钥做 HMAC,绑定待办 ID(页面)或 ID+目标地址(资源)。
|
||||
func (s *ReaderService) browserSign(payload string) string {
|
||||
mac := hmac.New(sha256.New, []byte(s.cfg.Secrets.JWTSecret))
|
||||
mac.Write([]byte(payload))
|
||||
return hex.EncodeToString(mac.Sum(nil))[:32]
|
||||
}
|
||||
|
||||
func (s *ReaderService) browserPageSig(id string) string { return s.browserSign("page|" + id) }
|
||||
|
||||
func (s *ReaderService) browserAssetSig(id, target string) string {
|
||||
return s.browserSign("asset|" + id + "|" + target)
|
||||
}
|
||||
|
||||
// browserPageURL 生成 iframe 承载地址。
|
||||
//
|
||||
// 承载页面常常靠 location.search 取值:光遇聚合的段评页就是从 ?item_id¶&source
|
||||
// 读出「哪本书的哪一段、哪个来源」,再据此请求 /para_review。承载地址是我们自己的
|
||||
// 代理路径,页面看到的 query 只有 id/s,原地址的参数全丢了——页面于是按缺参数去
|
||||
// 请求评论接口,气泡数照样显示,点开却一条段评都没有。
|
||||
//
|
||||
// 所以把原地址的 query 原样拼在签名参数之后:页面看到的 search 与原站一致。
|
||||
// 我们自己的 id/s 放最前,Gin 取同名的第一个值,不会被原地址里的同名参数顶掉。
|
||||
func (s *ReaderService) browserPageURL(id, target string) string {
|
||||
u := "/api/reader/browser/page?id=" + url.QueryEscape(id) + "&s=" + s.browserPageSig(id)
|
||||
if q := browserTargetQuery(target); q != "" {
|
||||
u += "&" + q
|
||||
}
|
||||
return u
|
||||
}
|
||||
|
||||
// browserTargetQuery 取原地址的 query(不含 `?`);data: 地址或解析失败时返回空串。
|
||||
func browserTargetQuery(target string) string {
|
||||
target = strings.TrimSpace(target)
|
||||
if target == "" || strings.HasPrefix(strings.ToLower(target), "data:") {
|
||||
return ""
|
||||
}
|
||||
parsed, err := url.Parse(target)
|
||||
if err != nil {
|
||||
return ""
|
||||
}
|
||||
return parsed.RawQuery
|
||||
}
|
||||
|
||||
// browserAssetProxyURL 生成资源代理地址(iframe 页面内引用用)。
|
||||
func (s *ReaderService) browserAssetProxyURL(id, target string) string {
|
||||
return "/api/reader/browser/asset?id=" + url.QueryEscape(id) +
|
||||
"&s=" + s.browserAssetSig(id, target) +
|
||||
"&u=" + base64.RawURLEncoding.EncodeToString([]byte(target))
|
||||
}
|
||||
|
||||
// VerifyBrowserPage 校验页面承载签名,返回待办快照。
|
||||
func (s *ReaderService) VerifyBrowserPage(id, sig string) (*BrowserPageSnapshot, error) {
|
||||
if !hmac.Equal([]byte(s.browserPageSig(id)), []byte(sig)) {
|
||||
return nil, errors.New("页面地址签名校验失败")
|
||||
}
|
||||
entry := s.lookupBrowser(id)
|
||||
if entry == nil {
|
||||
return nil, errors.New("页面已过期,请重新打开")
|
||||
}
|
||||
return &BrowserPageSnapshot{
|
||||
ID: entry.id,
|
||||
HTML: entry.html,
|
||||
SourceID: entry.sourceID,
|
||||
FinalURL: entry.finalURL,
|
||||
}, nil
|
||||
}
|
||||
|
||||
// VerifyBrowserAsset 校验资源代理签名并还原目标地址,返回(目标地址, 书源 URL)。
|
||||
func (s *ReaderService) VerifyBrowserAsset(id, encoded, sig string) (string, string, error) {
|
||||
raw, err := base64.RawURLEncoding.DecodeString(encoded)
|
||||
if err != nil {
|
||||
return "", "", errors.New("资源地址解码失败")
|
||||
}
|
||||
target := string(raw)
|
||||
if !hmac.Equal([]byte(s.browserAssetSig(id, target)), []byte(sig)) {
|
||||
return "", "", errors.New("资源地址签名校验失败")
|
||||
}
|
||||
entry := s.lookupBrowser(id)
|
||||
if entry == nil {
|
||||
return "", "", errors.New("页面已过期,请重新打开")
|
||||
}
|
||||
return target, entry.sourceURL, nil
|
||||
}
|
||||
|
||||
// ─── 待办表 ────────────────────────────────────────────────────────────────
|
||||
|
||||
// PendingBrowserPages 返回该用户在某书源下待用户完成的页面(前端轮询)。
|
||||
// 含 open 模式:startBrowser 只展示不回传,同样需要前端把页面呈现出来。
|
||||
func (s *ReaderService) PendingBrowserPages(userID, sourceID string) []BrowserPage {
|
||||
s.browserMu.Lock()
|
||||
defer s.browserMu.Unlock()
|
||||
s.pruneBrowsersLocked()
|
||||
out := make([]BrowserPage, 0, len(s.browserPending))
|
||||
for _, e := range s.browserPending {
|
||||
if e.sourceID != sourceID {
|
||||
continue
|
||||
}
|
||||
// 已交付/已取消的页面不再返回:否则前端下一次轮询会把它当成新页面弹出来。
|
||||
if e.finished {
|
||||
continue
|
||||
}
|
||||
// 待办是短生命周期对象;未标注用户的老调用路径一律放行。
|
||||
if e.userID != "" && userID != "" && e.userID != userID {
|
||||
continue
|
||||
}
|
||||
out = append(out, s.pageOfLocked(e))
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// ResolveBrowser 用户完成/取消页面后回传结果,解除 JS 侧的阻塞。
|
||||
//
|
||||
// 交付后立刻把条目移出待办表:open 模式(java.startBrowser)没有别的清理
|
||||
// 路径,留着会一直躺到 TTL 到期——光遇聚合的「❇️ 更新书源」就是这么在
|
||||
// 15 分钟内反复串到其它按钮上的。
|
||||
func (s *ReaderService) ResolveBrowser(id, otherUserID, body, finalURL string, cancelled bool) error {
|
||||
entry := s.lookupBrowser(id)
|
||||
if entry == nil {
|
||||
return errors.New("页面已过期或已完成")
|
||||
}
|
||||
// 注意锁序:browserMu 必须在 entry.mu 之外获取(pruneBrowsersLocked 是
|
||||
// browserMu → entry.mu),这里不能持着 entry.mu 去等 browserMu。
|
||||
entry.mu.Lock()
|
||||
if entry.finished {
|
||||
entry.mu.Unlock()
|
||||
return errors.New("该页面已完成")
|
||||
}
|
||||
if cancelled {
|
||||
entry.err = rule.ErrBrowserCancelled
|
||||
} else {
|
||||
resultURL := strings.TrimSpace(finalURL)
|
||||
if resultURL == "" {
|
||||
resultURL = entry.finalURL
|
||||
}
|
||||
entry.result = rule.BrowserResult{URL: resultURL, Body: body}
|
||||
}
|
||||
entry.finished = true
|
||||
if entry.done != nil {
|
||||
close(entry.done)
|
||||
}
|
||||
entry.mu.Unlock()
|
||||
|
||||
s.dropBrowser(id)
|
||||
return nil
|
||||
}
|
||||
|
||||
func (s *ReaderService) lookupBrowser(id string) *pendingBrowser {
|
||||
s.browserMu.Lock()
|
||||
defer s.browserMu.Unlock()
|
||||
s.pruneBrowsersLocked()
|
||||
return s.browserPending[id]
|
||||
}
|
||||
|
||||
func (s *ReaderService) dropBrowser(id string) {
|
||||
s.browserMu.Lock()
|
||||
defer s.browserMu.Unlock()
|
||||
delete(s.browserPending, id)
|
||||
}
|
||||
|
||||
// pruneBrowsersLocked 清掉过期条目并唤醒仍在等待的调用方(避免泄漏协程)。
|
||||
func (s *ReaderService) pruneBrowsersLocked() {
|
||||
now := time.Now()
|
||||
for id, e := range s.browserPending {
|
||||
if now.Before(e.expires) {
|
||||
continue
|
||||
}
|
||||
e.mu.Lock()
|
||||
if !e.finished {
|
||||
e.finished = true
|
||||
e.err = errors.New("页面等待超时")
|
||||
if e.done != nil {
|
||||
close(e.done)
|
||||
}
|
||||
}
|
||||
e.mu.Unlock()
|
||||
delete(s.browserPending, id)
|
||||
}
|
||||
}
|
||||
|
||||
// browserPageOf 生成前端展示用的描述(自行加锁)。
|
||||
// registerBrowser 之后需要把页面直接回给调用方(如打开一条段评)时用它。
|
||||
func (s *ReaderService) browserPageOf(e *pendingBrowser) BrowserPage {
|
||||
s.browserMu.Lock()
|
||||
defer s.browserMu.Unlock()
|
||||
return s.pageOfLocked(e)
|
||||
}
|
||||
|
||||
// pageOfLocked 生成前端展示用的描述(调用方需持有 browserMu)。
|
||||
func (s *ReaderService) pageOfLocked(e *pendingBrowser) BrowserPage {
|
||||
target := e.request.URL
|
||||
if strings.HasPrefix(target, "data:") || len(target) > 512 {
|
||||
target = ""
|
||||
}
|
||||
return BrowserPage{
|
||||
ID: e.id,
|
||||
Title: e.request.Title,
|
||||
Mode: e.mode,
|
||||
Seq: e.seq,
|
||||
PageURL: s.browserPageURL(e.id, e.request.URL),
|
||||
Refetch: e.request.Refetch,
|
||||
SourceID: e.sourceID,
|
||||
TargetURL: target,
|
||||
}
|
||||
}
|
||||
|
||||
// ─── 小工具 ────────────────────────────────────────────────────────────────
|
||||
|
||||
func newBrowserID() string {
|
||||
buf := make([]byte, 16)
|
||||
if _, err := rand.Read(buf); err != nil {
|
||||
return fmt.Sprintf("%d", time.Now().UnixNano())
|
||||
}
|
||||
return hex.EncodeToString(buf)
|
||||
}
|
||||
|
||||
func htmlEscape(s string) string {
|
||||
return strings.NewReplacer("&", "&", "<", "<", ">", ">", `"`, """).Replace(s)
|
||||
}
|
||||
|
||||
func truncateForLog(s string, n int) string {
|
||||
if len(s) <= n {
|
||||
return s
|
||||
}
|
||||
return s[:n] + "…"
|
||||
}
|
||||
@@ -0,0 +1,480 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"net/url"
|
||||
"strings"
|
||||
"testing"
|
||||
"time"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/service/reader/rule"
|
||||
)
|
||||
|
||||
// 本文件:书源宿主浏览器面板的服务层测试(browser_panel.go)。
|
||||
//
|
||||
// 覆盖真实书源依赖的三件事:
|
||||
// 1. startBrowserAwait 登记待办 → 前端轮询到 → 回传 DOM → 阻塞解除;
|
||||
// 2. http(s) 页面由服务端带书源 Cookie 抓取(浏览器里没有这些 Cookie);
|
||||
// 3. 页面里的资源地址被改写到同源代理,且签名校验拦得住伪造。
|
||||
|
||||
// browserPanelSourceJSON 构造一个内联书源:登录面板有「切换线路」按钮,
|
||||
// 其实现与光遇聚合的 getServerSettings 同构——打开内嵌 HTML,回传 DOM 后
|
||||
// 从 DOM 里解析线路写进源变量。
|
||||
func browserPanelSourceJSON(t *testing.T, sourceURL string) string {
|
||||
t.Helper()
|
||||
loginJS := `function switchLine() {
|
||||
let html = '<!DOCTYPE html><html><body><span id="serverValue">线路甲</span></body></html>';
|
||||
let body = java.startBrowserAwait('data:text/html;base64,' + java.base64Encode(html), '线路设置', false).body();
|
||||
let m = body.match(/id="serverValue"\s*>\s*([^<]*?)\s*<\/span>/);
|
||||
source.setVariable(JSON.stringify({线路: m ? m[1] : ''}));
|
||||
return m ? m[1] : '';
|
||||
}`
|
||||
src := map[string]any{
|
||||
"bookSourceUrl": sourceURL,
|
||||
"bookSourceName": "浏览器面板测试源",
|
||||
"loginUrl": loginJS,
|
||||
"loginUi": `[{"name":"切换线路","type":"button","action":"switchLine()"}]`,
|
||||
}
|
||||
out, err := json.Marshal(src)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
return string(out)
|
||||
}
|
||||
|
||||
// querySig 从签名地址里取出 s 参数。
|
||||
func querySig(t *testing.T, raw string) string {
|
||||
t.Helper()
|
||||
u, err := url.Parse(raw)
|
||||
if err != nil {
|
||||
t.Fatalf("承载地址不合法: %q", raw)
|
||||
}
|
||||
return u.Query().Get("s")
|
||||
}
|
||||
|
||||
// waitPending 轮询等待出现一个待办页面。
|
||||
func waitPending(t *testing.T, svc *ReaderService, sourceID string) BrowserPage {
|
||||
t.Helper()
|
||||
deadline := time.Now().Add(5 * time.Second)
|
||||
for time.Now().Before(deadline) {
|
||||
pages := svc.PendingBrowserPages(readerTestUserID, sourceID)
|
||||
if len(pages) > 0 {
|
||||
return pages[0]
|
||||
}
|
||||
time.Sleep(20 * time.Millisecond)
|
||||
}
|
||||
t.Fatal("等待超时:未出现待用户完成的页面")
|
||||
return BrowserPage{}
|
||||
}
|
||||
|
||||
// TestBrowserPanelSwitchLineRoundTrip 端到端:书源阻塞 → 前端拿到页面 →
|
||||
// 回传 DOM → 书源把线路写进源变量。
|
||||
func TestBrowserPanelSwitchLineRoundTrip(t *testing.T) {
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
sourceID := prepareLoginSource(t, svc, browserPanelSourceJSON(t, "https://panel.example.com"))
|
||||
|
||||
type actionResult struct {
|
||||
res *LoginResult
|
||||
err error
|
||||
}
|
||||
done := make(chan actionResult, 1)
|
||||
go func() {
|
||||
res, err := svc.RunLoginAction(ctx, readerTestUserID, sourceID, "switchLine()", nil)
|
||||
done <- actionResult{res: res, err: err}
|
||||
}()
|
||||
|
||||
page := waitPending(t, svc, sourceID)
|
||||
if page.Mode != browserModeWait {
|
||||
t.Fatalf("模式 = %q,期望 wait", page.Mode)
|
||||
}
|
||||
if page.Title != "线路设置" {
|
||||
t.Fatalf("标题 = %q", page.Title)
|
||||
}
|
||||
// data: 地址不应回传给前端展示(又长又没用)
|
||||
if page.TargetURL != "" {
|
||||
t.Fatalf("data: 地址不应回传: %q", page.TargetURL)
|
||||
}
|
||||
|
||||
// 承载页面的签名必须有效,且内容是书源拼的那段 HTML
|
||||
snap, err := svc.VerifyBrowserPage(page.ID, querySig(t, page.PageURL))
|
||||
if err != nil {
|
||||
t.Fatalf("承载页面校验失败: %v", err)
|
||||
}
|
||||
if !strings.Contains(snap.HTML, `id="serverValue"`) {
|
||||
t.Fatalf("承载页面内容异常: %q", snap.HTML)
|
||||
}
|
||||
// 伪造签名必须被拒
|
||||
if _, err := svc.VerifyBrowserPage(page.ID, "deadbeef"); err == nil {
|
||||
t.Fatal("伪造签名不应通过校验")
|
||||
}
|
||||
|
||||
// 模拟用户在页面里改选了「线路乙」后点 √:回传操作后的 DOM
|
||||
if err := svc.ResolveBrowser(page.ID, readerTestUserID,
|
||||
`<html><body><span id="serverValue">线路乙</span></body></html>`, "", false); err != nil {
|
||||
t.Fatalf("回传结果失败: %v", err)
|
||||
}
|
||||
|
||||
select {
|
||||
case r := <-done:
|
||||
if r.err != nil {
|
||||
t.Fatalf("登录动作失败: %v", r.err)
|
||||
}
|
||||
if !r.res.OK {
|
||||
t.Fatalf("动作未成功: %+v", r.res)
|
||||
}
|
||||
case <-time.After(5 * time.Second):
|
||||
t.Fatal("回传后阻塞未解除")
|
||||
}
|
||||
|
||||
// 书源应已把线路写进源变量(真正切换生效)
|
||||
st := svc.newSourceState(ctx, "https://panel.example.com")
|
||||
if !strings.Contains(st.GetVariable(), "线路乙") {
|
||||
t.Fatalf("线路未写入源变量: %q", st.GetVariable())
|
||||
}
|
||||
// 待办应被清理
|
||||
if pages := svc.PendingBrowserPages(readerTestUserID, sourceID); len(pages) != 0 {
|
||||
t.Fatalf("完成后待办未清理: %+v", pages)
|
||||
}
|
||||
}
|
||||
|
||||
// TestBrowserPanelCancelReleasesBlock 取消页面时要解除阻塞(书源走降级分支)。
|
||||
func TestBrowserPanelCancelReleasesBlock(t *testing.T) {
|
||||
svc, _ := newLoginTestService(t)
|
||||
sourceID := prepareLoginSource(t, svc, browserPanelSourceJSON(t, "https://panel.example.com"))
|
||||
|
||||
done := make(chan error, 1)
|
||||
go func() {
|
||||
// action 为空表示执行 login();这里用 switchLine,取消后书源会抛异常
|
||||
_, err := svc.RunLoginAction(t.Context(), readerTestUserID, sourceID, "switchLine()", nil)
|
||||
done <- err
|
||||
}()
|
||||
|
||||
page := waitPending(t, svc, sourceID)
|
||||
if err := svc.ResolveBrowser(page.ID, readerTestUserID, "", "", true); err != nil {
|
||||
t.Fatalf("取消失败: %v", err)
|
||||
}
|
||||
select {
|
||||
case err := <-done:
|
||||
// 书源未 catch,异常应回到调用方;关键是「不能一直卡住」
|
||||
if err == nil {
|
||||
t.Log("取消后动作用空 body 继续执行(书源自行降级)")
|
||||
}
|
||||
case <-time.After(5 * time.Second):
|
||||
t.Fatal("取消后阻塞未解除")
|
||||
}
|
||||
}
|
||||
|
||||
// TestBrowserPageURLKeepsTargetQuery 承载地址必须原样带上目标地址的 query。
|
||||
//
|
||||
// 光遇聚合的段评页从 location.search 读 item_id/para/source,再据此请求评论接口。
|
||||
// 承载地址若只剩我们自己的 id/s,页面就按缺参数去请求:段评气泡上的数字照常有,
|
||||
// 点开却一条评论都没有。
|
||||
func TestBrowserPageURLKeepsTargetQuery(t *testing.T) {
|
||||
svc, _ := newLoginTestService(t)
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
w.Header().Set("Content-Type", "text/html; charset=utf-8")
|
||||
_, _ = w.Write([]byte(`<html><body>comment page</body></html>`))
|
||||
}))
|
||||
defer srv.Close()
|
||||
|
||||
// 目标地址里故意带上与我们同名的 s/id:它们不能顶掉签名参数。
|
||||
target := srv.URL + "/get_para_review?item_id=abc¶=1&source=QQ%E9%98%85%E8%AF%BB&s=decoy&id=decoy"
|
||||
entry, err := svc.registerBrowser(t.Context(), srv.URL, "src-1", readerTestUserID,
|
||||
browserCookieTarget{}, rule.BrowserTask{URL: target, Title: "段评"}, browserModeOpen)
|
||||
if err != nil {
|
||||
t.Fatalf("登记承载页面失败: %v", err)
|
||||
}
|
||||
|
||||
page := svc.browserPageOf(entry)
|
||||
u, err := url.Parse(page.PageURL)
|
||||
if err != nil {
|
||||
t.Fatalf("承载地址不合法: %q", page.PageURL)
|
||||
}
|
||||
q := u.Query()
|
||||
if q.Get("item_id") != "abc" || q.Get("para") != "1" || q.Get("source") != "QQ阅读" {
|
||||
t.Fatalf("承载地址丢了目标参数(段评页会因此取不到评论): %q", page.PageURL)
|
||||
}
|
||||
// 路由按 id/s 取承载的页面快照,同名参数必须仍指向我们自己的待办
|
||||
if q.Get("id") != entry.id || q.Get("s") != svc.browserPageSig(entry.id) {
|
||||
t.Fatalf("同名参数顶掉了签名参数: %q", page.PageURL)
|
||||
}
|
||||
if _, err := svc.VerifyBrowserPage(q.Get("id"), q.Get("s")); err != nil {
|
||||
t.Fatalf("带业务参数后签名校验失败: %v", err)
|
||||
}
|
||||
// 没有 query 的地址(含 data: 页面)不该多出参数
|
||||
for _, plain := range []string{"https://panel.example.com/plain", "data:text/html;base64,AAA"} {
|
||||
if got := svc.browserPageURL("id1", plain); strings.Count(got, "&") != 1 {
|
||||
t.Fatalf("无 query 的地址被拼进了多余参数: %q", got)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestBrowserPanelInjectsSourceCookies http(s) 页面必须由服务端带书源 Cookie
|
||||
// 抓取,否则「用户后台」在浏览器里永远是未登录状态。
|
||||
func TestBrowserPanelInjectsSourceCookies(t *testing.T) {
|
||||
var userCookie string
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
switch r.URL.Path {
|
||||
case "/user":
|
||||
userCookie = r.Header.Get("Cookie")
|
||||
w.Header().Set("Content-Type", "text/html; charset=utf-8")
|
||||
if !strings.Contains(userCookie, "qttoken=") {
|
||||
_, _ = w.Write([]byte(`<html><body>请先登录</body></html>`))
|
||||
return
|
||||
}
|
||||
_, _ = w.Write([]byte(`<html><body><h1>我的账号</h1>` +
|
||||
`<img src="/avatar.png"><a href="/orders">订单</a>` +
|
||||
`<a href="https://other-site.net/x">站外</a></body></html>`))
|
||||
case "/avatar.png":
|
||||
_, _ = w.Write([]byte("PNGDATA"))
|
||||
default:
|
||||
http.NotFound(w, r)
|
||||
}
|
||||
}))
|
||||
defer srv.Close()
|
||||
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
sourceID := prepareLoginSource(t, svc, browserPanelSourceJSON(t, srv.URL))
|
||||
|
||||
// 先造出登录态(等价于书源已登录成功)
|
||||
st := svc.newSourceState(ctx, srv.URL)
|
||||
st.SetCookie(srv.URL, "qttoken=SESSION_abcdef123456")
|
||||
st.flush()
|
||||
|
||||
// 直接驱动服务层:等价于书源调用 startBrowser 打开用户后台
|
||||
if err := svc.openBrowser(ctx, srv.URL, sourceID, readerTestUserID, browserCookieTarget{}, rule.BrowserTask{
|
||||
URL: srv.URL + "/user", Title: "用户后台",
|
||||
}); err != nil {
|
||||
t.Fatalf("打开用户后台失败: %v", err)
|
||||
}
|
||||
if !strings.Contains(userCookie, "qttoken=SESSION_abcdef123456") {
|
||||
t.Fatalf("抓取页面时未带上书源 Cookie: %q", userCookie)
|
||||
}
|
||||
|
||||
page := waitPending(t, svc, sourceID)
|
||||
if page.Mode != browserModeOpen {
|
||||
t.Fatalf("模式 = %q,期望 open", page.Mode)
|
||||
}
|
||||
if page.TargetURL != srv.URL+"/user" {
|
||||
t.Fatalf("原始地址 = %q", page.TargetURL)
|
||||
}
|
||||
snap, err := svc.VerifyBrowserPage(page.ID, querySig(t, page.PageURL))
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
// 已登录内容(证明用的是服务端 Cookie 抓到的页面)
|
||||
if !strings.Contains(snap.HTML, "我的账号") {
|
||||
t.Fatalf("页面不是登录态内容: %q", snap.HTML)
|
||||
}
|
||||
// 资源地址应被改写到同源代理;站外链接保留原样
|
||||
if !strings.Contains(snap.HTML, "/api/reader/browser/asset?") {
|
||||
t.Fatalf("资源地址未改写: %q", snap.HTML)
|
||||
}
|
||||
if !strings.Contains(snap.HTML, "https://other-site.net/x") {
|
||||
t.Fatalf("站外链接不应改写: %q", snap.HTML)
|
||||
}
|
||||
// 注入脚本要带齐三样:DOM 回传、请求代理、以及书源登录态
|
||||
if !strings.Contains(snap.HTML, "data-mebox-browser-bridge") {
|
||||
t.Fatalf("缺少注入脚本: %q", snap.HTML)
|
||||
}
|
||||
if !strings.Contains(snap.HTML, "__mebox_proxy__") {
|
||||
t.Fatal("注入脚本缺少请求代理(页面自己的 fetch/XHR 会因不透明源被 CORS 拦)")
|
||||
}
|
||||
// 页面常用 document.cookie 判断登录态;真实 Cookie 在服务端,
|
||||
// 不预置的话「用户后台」会以为未登录并把浏览器导去 /login
|
||||
if !strings.Contains(snap.HTML, "qttoken=SESSION_abcdef123456") {
|
||||
t.Fatalf("注入脚本未预置书源 Cookie: %q", snap.HTML)
|
||||
}
|
||||
// 相对地址要按书源页面地址解析(沙箱里基地址不是承载地址)
|
||||
if !strings.Contains(snap.HTML, srv.URL) {
|
||||
t.Fatalf("注入脚本未带上页面基地址: %q", snap.HTML)
|
||||
}
|
||||
|
||||
// 资源代理要能校验签名并把 Cookie 带上
|
||||
assetURL := extractAssetURL(t, snap.HTML)
|
||||
u, err := url.Parse(assetURL)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
target, gotSourceURL, err := svc.VerifyBrowserAsset(
|
||||
u.Query().Get("id"), u.Query().Get("u"), u.Query().Get("s"))
|
||||
if err != nil {
|
||||
t.Fatalf("资源签名校验失败: %v", err)
|
||||
}
|
||||
if !strings.HasSuffix(target, "/avatar.png") {
|
||||
t.Fatalf("资源目标 = %q", target)
|
||||
}
|
||||
if gotSourceURL != srv.URL {
|
||||
t.Fatalf("资源关联的书源 = %q", gotSourceURL)
|
||||
}
|
||||
contentType, status, data, err := svc.FetchBrowserAsset(ctx, page.ID, target)
|
||||
if err != nil {
|
||||
t.Fatalf("拉取资源失败: %v", err)
|
||||
}
|
||||
if status != http.StatusOK || string(data) != "PNGDATA" {
|
||||
t.Fatalf("资源内容异常: status=%d body=%q", status, data)
|
||||
}
|
||||
if contentType == "" {
|
||||
t.Fatal("资源缺少 Content-Type")
|
||||
}
|
||||
// 伪造资源签名必须被拒
|
||||
if _, _, err := svc.VerifyBrowserAsset(u.Query().Get("id"), u.Query().Get("u"), "deadbeef"); err == nil {
|
||||
t.Fatal("伪造资源签名不应通过校验")
|
||||
}
|
||||
}
|
||||
|
||||
// TestBrowserPanelXHRProxy 页面内的接口请求要由服务端代发并补上书源 Cookie。
|
||||
// 「用户后台」这类页面靠接口取数,iframe 自己的 XHR 带不上 Cookie 也会被 CORS 拦。
|
||||
func TestBrowserPanelXHRProxy(t *testing.T) {
|
||||
var apiCookie string
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
switch r.URL.Path {
|
||||
case "/user":
|
||||
w.Header().Set("Content-Type", "text/html; charset=utf-8")
|
||||
_, _ = w.Write([]byte(`<html><head></head><body>后台</body></html>`))
|
||||
case "/api/me":
|
||||
apiCookie = r.Header.Get("Cookie")
|
||||
w.Header().Set("Content-Type", "application/json; charset=utf-8")
|
||||
if !strings.Contains(apiCookie, "qttoken=") {
|
||||
w.WriteHeader(http.StatusUnauthorized)
|
||||
_, _ = w.Write([]byte(`{"error":"未登录"}`))
|
||||
return
|
||||
}
|
||||
_, _ = w.Write([]byte(`{"nickname":"tester","vip":true}`))
|
||||
default:
|
||||
http.NotFound(w, r)
|
||||
}
|
||||
}))
|
||||
defer srv.Close()
|
||||
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
sourceID := prepareLoginSource(t, svc, browserPanelSourceJSON(t, srv.URL))
|
||||
st := svc.newSourceState(ctx, srv.URL)
|
||||
st.SetCookie(srv.URL, "qttoken=SESSION_abcdef123456")
|
||||
st.flush()
|
||||
|
||||
if err := svc.openBrowser(ctx, srv.URL, sourceID, readerTestUserID, browserCookieTarget{}, rule.BrowserTask{
|
||||
URL: srv.URL + "/user", Title: "用户后台",
|
||||
}); err != nil {
|
||||
t.Fatalf("打开用户后台失败: %v", err)
|
||||
}
|
||||
page := waitPending(t, svc, sourceID)
|
||||
|
||||
res, err := svc.ProxyBrowserXHR(ctx, page.ID, http.MethodGet, srv.URL+"/api/me",
|
||||
map[string]string{"X-From-Page": "1"}, "")
|
||||
if err != nil {
|
||||
t.Fatalf("转发请求失败: %v", err)
|
||||
}
|
||||
if res.Status != http.StatusOK {
|
||||
t.Fatalf("状态 = %d body=%q", res.Status, res.Body)
|
||||
}
|
||||
if !strings.Contains(apiCookie, "qttoken=SESSION_abcdef123456") {
|
||||
t.Fatalf("转发时未带上书源 Cookie: %q", apiCookie)
|
||||
}
|
||||
if !strings.Contains(res.Body, `"nickname":"tester"`) {
|
||||
t.Fatalf("响应体异常: %q", res.Body)
|
||||
}
|
||||
if !strings.Contains(res.ContentType, "json") {
|
||||
t.Fatalf("内容类型 = %q", res.ContentType)
|
||||
}
|
||||
if res.Base64 {
|
||||
t.Fatal("JSON 响应不应走 base64")
|
||||
}
|
||||
|
||||
// 未知待办:拒绝(链接已过期)
|
||||
if _, err := svc.ProxyBrowserXHR(ctx, "nope", http.MethodGet, srv.URL+"/api/me", nil, ""); err == nil {
|
||||
t.Fatal("未知待办应报错")
|
||||
}
|
||||
// 非 http(s):拒绝,避免被当成任意协议跳板
|
||||
if _, err := svc.ProxyBrowserXHR(ctx, page.ID, http.MethodGet, "file:///etc/passwd", nil, ""); err == nil {
|
||||
t.Fatal("非 http(s) 地址应报错")
|
||||
}
|
||||
}
|
||||
|
||||
// extractAssetURL 从改写后的页面里取出第一个资源代理地址。
|
||||
func extractAssetURL(t *testing.T, html string) string {
|
||||
t.Helper()
|
||||
idx := strings.Index(html, "/api/reader/browser/asset?")
|
||||
if idx < 0 {
|
||||
t.Fatalf("页面里没有资源代理地址: %q", html)
|
||||
}
|
||||
rest := html[idx:]
|
||||
// 到引号/尖括号为止
|
||||
end := strings.IndexAny(rest, `"'< `)
|
||||
if end < 0 {
|
||||
end = len(rest)
|
||||
}
|
||||
return strings.ReplaceAll(rest[:end], "&", "&")
|
||||
}
|
||||
|
||||
// TestBrowserPanelUnknownSourceIsolated 待办按书源隔离,避免串台。
|
||||
func TestBrowserPanelUnknownSourceIsolated(t *testing.T) {
|
||||
svc, _ := newLoginTestService(t)
|
||||
sourceID := prepareLoginSource(t, svc, browserPanelSourceJSON(t, "https://panel.example.com"))
|
||||
|
||||
done := make(chan struct{}, 1)
|
||||
go func() {
|
||||
_, _ = svc.RunLoginAction(t.Context(), readerTestUserID, sourceID, "switchLine()", nil)
|
||||
done <- struct{}{}
|
||||
}()
|
||||
page := waitPending(t, svc, sourceID)
|
||||
|
||||
if pages := svc.PendingBrowserPages(readerTestUserID, "other-source"); len(pages) != 0 {
|
||||
t.Fatalf("其它书源不应看到待办: %+v", pages)
|
||||
}
|
||||
// 收尾,避免 goroutine 挂到超时
|
||||
_ = svc.ResolveBrowser(page.ID, readerTestUserID, "<span id=\"serverValue\">x</span>", "", false)
|
||||
select {
|
||||
case <-done:
|
||||
case <-time.After(5 * time.Second):
|
||||
t.Fatal("未解除阻塞")
|
||||
}
|
||||
}
|
||||
|
||||
// TestBrowserPanelOpenPageDroppedAfterResolve open 模式(java.startBrowser)
|
||||
// 的页面在用户关闭后必须立刻离开待办表。
|
||||
//
|
||||
// 回归:这条日志只删除 wait 模式(awaitBrowser 的 defer),open 模式没有清理
|
||||
// 路径,会一直躺到 15 分钟 TTL 到期。前端轮询到它就又把旧页面弹出来——光遇
|
||||
// 聚合点过「❇️ 更新书源」后,接下来点「书源设置」「段评设置」都会跳到书源更新页。
|
||||
func TestBrowserPanelOpenPageDroppedAfterResolve(t *testing.T) {
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
sourceID := prepareLoginSource(t, svc, browserPanelSourceJSON(t, "https://panel.example.com"))
|
||||
|
||||
// 等价于书源的 renderVersionPage():java.startBrowser(data:..., '光遇书源更新')
|
||||
if err := svc.openBrowser(ctx, "https://panel.example.com", sourceID, readerTestUserID, browserCookieTarget{},
|
||||
rule.BrowserTask{URL: "data:text/html,<html><body>更新</body></html>", Title: "光遇书源更新"}); err != nil {
|
||||
t.Fatalf("登记 open 页面失败: %v", err)
|
||||
}
|
||||
old := waitPending(t, svc, sourceID)
|
||||
if old.Mode != browserModeOpen {
|
||||
t.Fatalf("模式 = %q,期望 open", old.Mode)
|
||||
}
|
||||
|
||||
// 用户关掉页面
|
||||
if err := svc.ResolveBrowser(old.ID, readerTestUserID, "", "", true); err != nil {
|
||||
t.Fatalf("关闭页面失败: %v", err)
|
||||
}
|
||||
if pages := svc.PendingBrowserPages(readerTestUserID, sourceID); len(pages) != 0 {
|
||||
t.Fatalf("关闭后待办未清理: %+v", pages)
|
||||
}
|
||||
|
||||
// 再点别的按钮:待办表里只应剩新页面,且序号更大
|
||||
if err := svc.openBrowser(ctx, "https://panel.example.com", sourceID, readerTestUserID, browserCookieTarget{},
|
||||
rule.BrowserTask{URL: "data:text/html,<html><body>设置</body></html>", Title: "光遇书源设置"}); err != nil {
|
||||
t.Fatalf("登记新页面失败: %v", err)
|
||||
}
|
||||
pages := svc.PendingBrowserPages(readerTestUserID, sourceID)
|
||||
if len(pages) != 1 || pages[0].Title != "光遇书源设置" {
|
||||
t.Fatalf("待办表应只剩新页面: %+v", pages)
|
||||
}
|
||||
if pages[0].Seq <= old.Seq {
|
||||
t.Fatalf("新页面 seq 应更大: %d <= %d", pages[0].Seq, old.Seq)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,107 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"encoding/base64"
|
||||
"fmt"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"strings"
|
||||
"testing"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
)
|
||||
|
||||
// 漫画源(聚合类漫画源就是这么写的)正文规则直接给 <img src="…"> 的 HTML。
|
||||
// 整行当地址会被代理成一堆取不回的「图」,页面上全是破图;必须按标签抽 src。
|
||||
func TestImageRefsInLine(t *testing.T) {
|
||||
cases := []struct {
|
||||
name string
|
||||
line string
|
||||
want []string
|
||||
}{
|
||||
{"纯地址原样返回", "https://cdn.example.com/1.webp", []string{"https://cdn.example.com/1.webp"}},
|
||||
{"双引号 img", `<img src="https://cdn.example.com/2.webp">`, []string{"https://cdn.example.com/2.webp"}},
|
||||
{"单引号 img", `<img src='https://cdn.example.com/3.webp'/>`, []string{"https://cdn.example.com/3.webp"}},
|
||||
{"属性顺序无关", `<img data-x="1" src="https://cdn.example.com/4.webp" class="a">`, []string{"https://cdn.example.com/4.webp"}},
|
||||
{"一行多个标签", `<p><img src="/a/1.webp"><img src='/a/2.webp'></p>`, []string{"/a/1.webp", "/a/2.webp"}},
|
||||
{"img 没有 src 时跳过", `<img alt="空">`, nil},
|
||||
{"非 img 的 HTML 片段跳过", `</div>`, nil},
|
||||
}
|
||||
for _, c := range cases {
|
||||
got := imageRefsInLine(c.line)
|
||||
if strings.Join(got, ",") != strings.Join(c.want, ",") {
|
||||
t.Fatalf("%s: got %q want %q", c.name, got, c.want)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// 端到端:图片型书籍走 GetContentForBook 时,Images 必须是真正的图片地址,
|
||||
// 不能把 <img> 标签本身当成图片地址(那会让代理取回一堆 HTML,前端全破图)。
|
||||
func TestGetContentForBookExtractsImageSrc(t *testing.T) {
|
||||
page := `<html><body><div class="pages">
|
||||
<img src="/img/1.webp"><img src="/img/2.webp">
|
||||
</div></body></html>`
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
w.Header().Set("Content-Type", "text/html; charset=utf-8")
|
||||
_, _ = w.Write([]byte(page))
|
||||
}))
|
||||
defer srv.Close()
|
||||
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
srcJSON := fmt.Sprintf(`{
|
||||
"bookSourceUrl": %q,
|
||||
"bookSourceName": "漫画测试源",
|
||||
"bookSourceType": 64,
|
||||
"ruleContent": { "content": "class.pages@html" }
|
||||
}`, srv.URL)
|
||||
sourceID := importTestSource(t, svc, srcJSON, srv.URL)
|
||||
|
||||
book := &model.ReaderBook{
|
||||
UserID: "u1",
|
||||
Origin: srv.URL,
|
||||
OriginName: "漫画测试源",
|
||||
BookURL: srv.URL + "/book/1",
|
||||
Name: "航海王",
|
||||
Type: 2, // 图片
|
||||
}
|
||||
if err := svc.repo.CreateBook(ctx, book); err != nil {
|
||||
t.Fatalf("创建书籍失败: %v", err)
|
||||
}
|
||||
if err := svc.SaveChapters(ctx, book.ID, []ChapterInput{
|
||||
{Index: 0, Title: "第1话", URL: srv.URL + "/book/1/c1.html"},
|
||||
}); err != nil {
|
||||
t.Fatalf("写入章节失败: %v", err)
|
||||
}
|
||||
_ = sourceID
|
||||
|
||||
out, err := svc.GetContentForBook(ctx, "u1", book.ID, 0)
|
||||
if err != nil {
|
||||
t.Fatalf("取正文失败: %v", err)
|
||||
}
|
||||
if out.Type != "image" {
|
||||
t.Fatalf("类型 = %q,期望 image", out.Type)
|
||||
}
|
||||
if len(out.Images) != 2 {
|
||||
t.Fatalf("图片数 = %d,期望 2(%v)", len(out.Images), out.Images)
|
||||
}
|
||||
for i, img := range out.Images {
|
||||
// 书架维度返回的是签名代理地址(浏览器取图带不上防盗链头),
|
||||
// 把里面的原始地址解出来核对,顺便确认没有被塞进 HTML 标签。
|
||||
if !strings.HasPrefix(img, "/api/reader/media?") {
|
||||
t.Fatalf("第 %d 张不是代理地址:%q", i, img)
|
||||
}
|
||||
encoded := img[strings.Index(img, "u=")+2:]
|
||||
if end := strings.IndexByte(encoded, '&'); end >= 0 {
|
||||
encoded = encoded[:end]
|
||||
}
|
||||
raw, err := base64.RawURLEncoding.DecodeString(encoded)
|
||||
if err != nil {
|
||||
t.Fatalf("第 %d 张地址解不开:%q", i, img)
|
||||
}
|
||||
want := fmt.Sprintf("%s/img/%d.webp", srv.URL, i+1)
|
||||
if string(raw) != want {
|
||||
t.Fatalf("第 %d 张解析出 %q,期望 %q", i, raw, want)
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,494 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"context"
|
||||
"encoding/base64"
|
||||
"fmt"
|
||||
"net"
|
||||
"net/url"
|
||||
"regexp"
|
||||
"sort"
|
||||
"strconv"
|
||||
"strings"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/service/reader/rule"
|
||||
)
|
||||
|
||||
// 本文件:正文里的「段评」锚点解析(对齐 legado 的段评机制)。
|
||||
//
|
||||
// legado 侧段评有两种落在正文里的形态:
|
||||
//
|
||||
// 1. 段落里的 <comment ident="评论地址" count="评论数" />:宿主(苹果/源阅分支)
|
||||
// 直接把它解析成段落旁的评论气泡;
|
||||
// 2. 「改版」宿主不认 <comment>,改由书源在正文规则里把它换成一张内嵌 SVG 图片,
|
||||
// 图片地址后面再跟一段 JSON 配置:
|
||||
// <img src="data:image/svg+xml;base64,…{"click":"showCmt('评论地址','平台','段评')","style":"text"}">
|
||||
// legado 识别的是地址后面的这份配置(ReadBookActivity.oldClickImg),点击时把
|
||||
// 配置里的 click/js 当 JS 执行(showCmt 内部再调 java.showBrowser 打开评论页)。
|
||||
//
|
||||
// MeBox 没有 Android 的 Canvas 排版,正文按行渲染,因此这里把两种形态都归一成
|
||||
// 「某一行的段评气泡」:只摘掉标记本身,正文文字一个字都不动,并把评论地址 / 评论数
|
||||
// 结构化下发给前端;点击气泡时前端复用宿主浏览器(带书源 Cookie)打开评论地址。
|
||||
//
|
||||
// 约定:但凡无法确定评论地址的图片,一律原样保留交给既有链路,绝不猜一个地址出来。
|
||||
|
||||
// ContentComment 正文里的一条段评锚点(随正文一起下发给前端)。
|
||||
type ContentComment struct {
|
||||
// Line 段落行号:Content 按 "\n" 拆分后的 0-based 下标。
|
||||
Line int `json:"line"`
|
||||
// Count 评论数(0 表示未知:前端只画一个不带数字的气泡)。
|
||||
Count int `json:"count"`
|
||||
// URL 评论地址(书源 showCmt 的第一个参数),前端点击时交给宿主浏览器打开。
|
||||
URL string `json:"url"`
|
||||
// Label 评论类型:段评 / 本章说,用作气泡提示。
|
||||
Label string `json:"label,omitempty"`
|
||||
// Block 独占一行的整块评论(章末「本章说」),前端按居中整块渲染。
|
||||
Block bool `json:"block,omitempty"`
|
||||
}
|
||||
|
||||
// commentTagRe 匹配正文里的 <comment …> 标签。属性顺序不定、可能自闭合,
|
||||
// 也可能带一个配对结束标签(`<comment …>…</comment>`),一并吃掉。
|
||||
var commentTagRe = regexp.MustCompile(`(?is)<comment\b[^>]*?/?>(?:</comment>)?`)
|
||||
|
||||
// 下面三条属性正则都允许单/双引号两种写法;捕获组 1 是双引号内容,2 是单引号内容。
|
||||
var (
|
||||
commentIdentRe = regexp.MustCompile(`(?is)\bident\s*=\s*(?:"([^"]*)"|'([^']*)')`)
|
||||
commentCountRe = regexp.MustCompile(`(?is)\bcount\s*=\s*(?:"([^"]*)"|'([^']*)')`)
|
||||
commentOnPressRe = regexp.MustCompile(`(?is)\bonPress\s*=\s*(?:"([^"]*)"|'([^']*)')`)
|
||||
)
|
||||
|
||||
// commentJSURLRe 从 showCmt / showReadingBrowser / startBrowserDp 的调用里取第一个字符串参数。
|
||||
// 苹果分支的 <comment onPress="java.showReadingBrowser('地址',…)" /> 就靠它取地址。
|
||||
// RE2 不支持反向引用,双/单引号两种写法拆成两个捕获组。
|
||||
var commentJSURLRe = regexp.MustCompile(`(?is)(?:showCmt|showReadingBrowser|startBrowserDp)\s*\(\s*(?:"([^"]*)"|'([^']*)')`)
|
||||
|
||||
// imgParamRe 宽容地提取图片地址后面 JSON 配置里的键值对。
|
||||
//
|
||||
// 不用 encoding/json 解析:光遇聚合的「本章说」把配置拼成 {'type':'qtbzs',"click":"…",'style':'FULL'}
|
||||
// (单双引号混用),严格 JSON 解析会失败,而它又是合法 JS 对象字面量。
|
||||
var imgParamRe = regexp.MustCompile(`(?is)["']?(click|js|style|type|width)["']?\s*:\s*(?:"((?:[^"\\]|\\.)*)"|'((?:[^'\\]|\\.)*)')`)
|
||||
|
||||
// svgCountRe 取内嵌段评 SVG 里画在气泡上的数字(书源把评论数直接画进了 SVG 文本)。
|
||||
var svgCountRe = regexp.MustCompile(`(?is)<text[^>]*>\s*([0-9]+)\s*\+?\s*</text>`)
|
||||
|
||||
// wrapMarkupRe 匹配不带内容的块级标签,用来判断一行是不是「只有评论标记」。
|
||||
var wrapMarkupRe = regexp.MustCompile(`(?is)</?(?:p|div|br|hr|h[1-6]|article|dd|dl)\b[^>]*>`)
|
||||
|
||||
// extractContentComments 摘出正文里的段评标记,返回清理后的正文与结构化锚点。
|
||||
//
|
||||
// 只改行内内容、不增删行,因此评论的 Line 与调用方按 "\n" 拆出的下标始终一致。
|
||||
func extractContentComments(content string) (string, []ContentComment) {
|
||||
if content == "" || (!strings.Contains(content, "<comment") && !strings.Contains(content, "<img")) {
|
||||
return content, nil
|
||||
}
|
||||
lines := strings.Split(content, "\n")
|
||||
out := make([]string, len(lines))
|
||||
comments := make([]ContentComment, 0, 4)
|
||||
changed := false
|
||||
for i, line := range lines {
|
||||
text, cs := extractLineComments(line, i)
|
||||
out[i] = text
|
||||
if len(cs) > 0 {
|
||||
changed = true
|
||||
comments = append(comments, cs...)
|
||||
}
|
||||
}
|
||||
if !changed {
|
||||
return content, nil
|
||||
}
|
||||
return strings.Join(out, "\n"), comments
|
||||
}
|
||||
|
||||
// extractLineComments 处理一行:返回去掉段评标记后的文字与该行上的段评。
|
||||
func extractLineComments(line string, lineNo int) (string, []ContentComment) {
|
||||
type marker struct {
|
||||
start, end int
|
||||
comment ContentComment
|
||||
}
|
||||
var markers []marker
|
||||
|
||||
for _, m := range commentTagRe.FindAllStringIndex(line, -1) {
|
||||
if c, ok := parseCommentTag(line[m[0]:m[1]], lineNo); ok {
|
||||
markers = append(markers, marker{m[0], m[1], c})
|
||||
}
|
||||
}
|
||||
for _, tag := range scanImgTags(line) {
|
||||
if c, ok := parseCommentImage(tag.src, lineNo); ok {
|
||||
markers = append(markers, marker{tag.start, tag.end, c})
|
||||
}
|
||||
}
|
||||
if len(markers) == 0 {
|
||||
return line, nil
|
||||
}
|
||||
|
||||
sort.Slice(markers, func(i, j int) bool { return markers[i].start < markers[j].start })
|
||||
comments := make([]ContentComment, 0, len(markers))
|
||||
var b strings.Builder
|
||||
prev := 0
|
||||
for _, mk := range markers {
|
||||
if mk.start < prev { // 与前一个标记重叠(理论上不会出现),跳过避免重复删除
|
||||
continue
|
||||
}
|
||||
b.WriteString(line[prev:mk.start])
|
||||
prev = mk.end
|
||||
c := mk.comment
|
||||
// 摘掉标记后整行只剩标签/空白 → 说明评论独占一行(章末「本章说」)
|
||||
c.Block = isBlankMarkup(line[:mk.start] + line[mk.end:])
|
||||
comments = append(comments, c)
|
||||
}
|
||||
b.WriteString(line[prev:])
|
||||
return b.String(), comments
|
||||
}
|
||||
|
||||
// imgTagMatch 一行里一个 <img> 标签的位置与 src 值。
|
||||
type imgTagMatch struct {
|
||||
start, end int
|
||||
src string
|
||||
}
|
||||
|
||||
// scanImgTags 扫描一行里的所有 <img> 标签,返回标签位置与 src 值。
|
||||
//
|
||||
// 不能用 <img[^>]*src="([^"]*)"> 这类正则:书源把段评 SVG 拼成
|
||||
// <img src="data:…,{"click":"showCmt('…','…')"}"> —— 属性值里的 JSON 又用了双引号,
|
||||
// 正则会在第一个内层引号处截断。这里改为手工扫描:进入花括号后忽略引号,
|
||||
// 只把「花括号之外遇到的引号」当作属性值的结束。
|
||||
func scanImgTags(line string) []imgTagMatch {
|
||||
var out []imgTagMatch
|
||||
lower := strings.ToLower(line)
|
||||
for i := 0; i < len(line); {
|
||||
idx := strings.Index(lower[i:], "<img")
|
||||
if idx < 0 {
|
||||
break
|
||||
}
|
||||
start := i + idx
|
||||
src, end, ok := scanImgTagSrc(line, start)
|
||||
if !ok {
|
||||
i = start + 4
|
||||
continue
|
||||
}
|
||||
out = append(out, imgTagMatch{start: start, end: end, src: src})
|
||||
i = end
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// scanImgTagSrc 从 <img 起始位置扫描出 src 属性值,并返回标签结束下标(含 '>')。
|
||||
func scanImgTagSrc(line string, start int) (string, int, bool) {
|
||||
lower := strings.ToLower(line)
|
||||
i := start + 4 // 跳过 "<img"
|
||||
for {
|
||||
si := strings.Index(lower[i:], "src")
|
||||
if si < 0 {
|
||||
return "", 0, false
|
||||
}
|
||||
si += i
|
||||
q := si + 3
|
||||
for q < len(line) && isHTMLSpace(line[q]) {
|
||||
q++
|
||||
}
|
||||
// 必须是 src= 而不是 srcset= 之类的近似字段
|
||||
if q >= len(line) || line[q] != '=' {
|
||||
i = si + 3
|
||||
continue
|
||||
}
|
||||
q++
|
||||
for q < len(line) && isHTMLSpace(line[q]) {
|
||||
q++
|
||||
}
|
||||
if q >= len(line) || (line[q] != '"' && line[q] != '\'') {
|
||||
i = si + 3
|
||||
continue
|
||||
}
|
||||
quote := line[q]
|
||||
valStart := q + 1
|
||||
depth := 0
|
||||
j := valStart
|
||||
for ; j < len(line); j++ {
|
||||
c := line[j]
|
||||
switch {
|
||||
case c == '{':
|
||||
depth++
|
||||
case c == '}' && depth > 0:
|
||||
depth--
|
||||
case c == quote && depth == 0:
|
||||
// 花括号之外的引号才是属性值的结束
|
||||
gt := strings.IndexByte(line[j:], '>')
|
||||
if gt < 0 {
|
||||
return "", 0, false
|
||||
}
|
||||
return line[valStart:j], j + gt + 1, true
|
||||
}
|
||||
}
|
||||
return "", 0, false
|
||||
}
|
||||
}
|
||||
|
||||
// isHTMLSpace 判断 HTML 属性之间的空白(含换行/制表)。
|
||||
func isHTMLSpace(c byte) bool {
|
||||
return c == ' ' || c == '\t' || c == '\n' || c == '\r'
|
||||
}
|
||||
|
||||
// parseCommentTag 解析宿主侧的 <comment ident count /> 形式。
|
||||
func parseCommentTag(tag string, lineNo int) (ContentComment, bool) {
|
||||
rawURL := attrOf(commentIdentRe, tag)
|
||||
if rawURL == "" {
|
||||
rawURL = attrOf(commentOnPressRe, tag)
|
||||
}
|
||||
rawURL = extractCommentURL(rawURL)
|
||||
if rawURL == "" {
|
||||
return ContentComment{}, false
|
||||
}
|
||||
count, _ := strconv.Atoi(strings.TrimSpace(attrOf(commentCountRe, tag)))
|
||||
return ContentComment{Line: lineNo, Count: count, URL: rawURL, Label: "段评"}, true
|
||||
}
|
||||
|
||||
// parseCommentImage 解析「图片地址 + JSON 配置」形式的段评。
|
||||
//
|
||||
// 只把真正指向评论的气泡当段评:地址要么来自 showCmt/showReadingBrowser,
|
||||
// 要么图片本身就是内嵌的段评 SVG。其它带 click 配置的图片保持原样,避免误伤。
|
||||
func parseCommentImage(src string, lineNo int) (ContentComment, bool) {
|
||||
src = strings.TrimSpace(src)
|
||||
if src == "" {
|
||||
return ContentComment{}, false
|
||||
}
|
||||
base, params := splitImgSrcParams(src)
|
||||
if len(params) == 0 {
|
||||
return ContentComment{}, false
|
||||
}
|
||||
action := params["click"]
|
||||
if action == "" {
|
||||
action = params["js"]
|
||||
}
|
||||
if action == "" {
|
||||
return ContentComment{}, false
|
||||
}
|
||||
isCommentCall := strings.Contains(action, "showCmt") || strings.Contains(action, "showReadingBrowser")
|
||||
if !isCommentCall && !strings.HasPrefix(base, "data:image/svg+xml") {
|
||||
return ContentComment{}, false
|
||||
}
|
||||
rawURL := extractCommentURL(action)
|
||||
if rawURL == "" {
|
||||
return ContentComment{}, false
|
||||
}
|
||||
label := "段评"
|
||||
if strings.Contains(action, "本章说") {
|
||||
label = "本章说"
|
||||
}
|
||||
return ContentComment{
|
||||
Line: lineNo,
|
||||
Count: commentCountFromImage(base),
|
||||
URL: rawURL,
|
||||
Label: label,
|
||||
}, true
|
||||
}
|
||||
|
||||
// splitImgSrcParams 把图片地址拆成(地址, JSON 配置),没有配置时返回 (src, nil)。
|
||||
//
|
||||
// 对应 legado 的 AnalyzeUrl.paramPattern `\s*,\s*(?=\{)`:地址与配置之间用「逗号 + 花括号」分隔。
|
||||
func splitImgSrcParams(src string) (string, map[string]string) {
|
||||
for i := 0; i < len(src); i++ {
|
||||
if src[i] != ',' {
|
||||
continue
|
||||
}
|
||||
j := i + 1
|
||||
for j < len(src) && (src[j] == ' ' || src[j] == '\t' || src[j] == '\n' || src[j] == '\r') {
|
||||
j++
|
||||
}
|
||||
if j >= len(src) || src[j] != '{' {
|
||||
continue
|
||||
}
|
||||
end := strings.LastIndexByte(src, '}')
|
||||
if end <= j {
|
||||
continue
|
||||
}
|
||||
if params := parseImgParams(src[j : end+1]); len(params) > 0 {
|
||||
return strings.TrimSpace(src[:i]), params
|
||||
}
|
||||
}
|
||||
return src, nil
|
||||
}
|
||||
|
||||
// parseImgParams 宽容解析图片配置里的键值对(见 imgParamRe 的说明)。
|
||||
func parseImgParams(raw string) map[string]string {
|
||||
ms := imgParamRe.FindAllStringSubmatch(raw, -1)
|
||||
if len(ms) == 0 {
|
||||
return nil
|
||||
}
|
||||
out := make(map[string]string, len(ms))
|
||||
for _, m := range ms {
|
||||
val := m[2]
|
||||
if val == "" {
|
||||
val = m[3]
|
||||
}
|
||||
val = strings.ReplaceAll(val, `\"`, `"`)
|
||||
val = strings.ReplaceAll(val, `\'`, `'`)
|
||||
out[strings.ToLower(m[1])] = val
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// commentCountFromImage 从内嵌段评 SVG 里取评论数(书源把数字画进了 SVG 文本)。
|
||||
// 取不到时返回 0,前端只画一个不带数字的气泡。
|
||||
func commentCountFromImage(src string) int {
|
||||
decoded, ok := decodeCommentDataURI(src)
|
||||
if !ok {
|
||||
return 0
|
||||
}
|
||||
m := svgCountRe.FindStringSubmatch(decoded)
|
||||
if m == nil {
|
||||
return 0
|
||||
}
|
||||
n, err := strconv.Atoi(m[1])
|
||||
if err != nil || n < 0 {
|
||||
return 0
|
||||
}
|
||||
return n
|
||||
}
|
||||
|
||||
// decodeCommentDataURI 解出 data: URI 的文本内容(支持 base64 与明文两种)。
|
||||
func decodeCommentDataURI(s string) (string, bool) {
|
||||
if !strings.HasPrefix(s, "data:") {
|
||||
return "", false
|
||||
}
|
||||
comma := strings.IndexByte(s, ',')
|
||||
if comma < 0 {
|
||||
return "", false
|
||||
}
|
||||
meta, payload := s[:comma], s[comma+1:]
|
||||
if strings.HasSuffix(strings.ToLower(meta), ";base64") {
|
||||
b, err := base64.StdEncoding.DecodeString(payload)
|
||||
if err != nil {
|
||||
return "", false
|
||||
}
|
||||
return string(b), true
|
||||
}
|
||||
return payload, true
|
||||
}
|
||||
|
||||
// extractCommentURL 从 showCmt/showReadingBrowser 调用或裸地址里取评论地址。
|
||||
func extractCommentURL(js string) string {
|
||||
js = strings.TrimSpace(js)
|
||||
if js == "" {
|
||||
return ""
|
||||
}
|
||||
if m := commentJSURLRe.FindStringSubmatch(js); m != nil {
|
||||
return trimCommentURL(firstSubmatch(m))
|
||||
}
|
||||
if strings.HasPrefix(js, "http://") || strings.HasPrefix(js, "https://") {
|
||||
return trimCommentURL(js)
|
||||
}
|
||||
return ""
|
||||
}
|
||||
|
||||
// attrOf 取属性值,双引号优先、单引号兜底。
|
||||
func attrOf(re *regexp.Regexp, s string) string {
|
||||
m := re.FindStringSubmatch(s)
|
||||
if m == nil {
|
||||
return ""
|
||||
}
|
||||
return firstSubmatch(m)
|
||||
}
|
||||
|
||||
// firstSubmatch 返回「双引号 / 单引号」两个捕获组里非空的那个。
|
||||
func firstSubmatch(m []string) string {
|
||||
if len(m) > 1 && m[1] != "" {
|
||||
return m[1]
|
||||
}
|
||||
if len(m) > 2 {
|
||||
return m[2]
|
||||
}
|
||||
return ""
|
||||
}
|
||||
|
||||
// trimCommentURL 去空白并还原 HTML 实体。
|
||||
// 苹果/源阅分支会把地址里的 & 转成 &(paraForiOS),直接拿去请求会 404。
|
||||
func trimCommentURL(raw string) string {
|
||||
r := strings.NewReplacer(
|
||||
""", `"`, """, `"`,
|
||||
"'", "'", "'", "'",
|
||||
"<", "<", ">", ">",
|
||||
"&", "&",
|
||||
)
|
||||
return strings.TrimSpace(r.Replace(raw))
|
||||
}
|
||||
|
||||
// isBlankMarkup 判断一段文字去掉空块级标签/空白后是否为空。
|
||||
func isBlankMarkup(s string) bool {
|
||||
s = wrapMarkupRe.ReplaceAllString(s, "")
|
||||
s = strings.ReplaceAll(s, " ", " ")
|
||||
s = strings.ReplaceAll(s, "\u3000", " ")
|
||||
return strings.TrimSpace(s) == ""
|
||||
}
|
||||
|
||||
// ─── 打开段评(宿主浏览器承载) ─────────────────────────────────────────────
|
||||
|
||||
// OpenContentComment 在宿主浏览器里打开一条段评。
|
||||
//
|
||||
// 书源点击段评气泡时执行的是它自己的 showCmt(...):先 java.ajax 取评论页,
|
||||
// 再 java.showBrowser 展示。MeBox 不跑书源 JS,而是把 showCmt 的第一个参数
|
||||
// (评论地址)解析出来(见本文件),复用承载登录页的那套宿主浏览器打开:
|
||||
// 服务端带书源 Cookie 抓取评论页,前端 iframe 展示。这样评论接口需要的登录态
|
||||
// 仍然生效——前端直接 window.open 会以未登录身份访问。
|
||||
//
|
||||
// 返回的 BrowserPage 直接给前端渲染,不必走轮询(段评只需展示、无需回传 DOM)。
|
||||
func (s *ReaderService) OpenContentComment(ctx context.Context, userID, bookID, rawURL, title string) (*BrowserPage, error) {
|
||||
if s.repo == nil {
|
||||
return nil, fmt.Errorf("阅读服务未就绪")
|
||||
}
|
||||
target, err := validateCommentURL(rawURL)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
book, err := s.repo.GetBook(ctx, bookID)
|
||||
if err != nil || book == nil {
|
||||
return nil, fmt.Errorf("书籍不存在或已移出书架")
|
||||
}
|
||||
sourceID := ""
|
||||
if src, findErr := s.repo.GetSourceByURL(ctx, book.Origin); findErr == nil && src != nil {
|
||||
sourceID = src.ID
|
||||
}
|
||||
if strings.TrimSpace(title) == "" {
|
||||
title = "段评"
|
||||
}
|
||||
entry, err := s.registerBrowser(ctx, book.Origin, sourceID, userID,
|
||||
// 段评是「只展示」的页面,没有登录会话可回写 Cookie。
|
||||
browserCookieTarget{},
|
||||
rule.BrowserTask{URL: target, Title: title}, browserModeOpen)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
page := s.browserPageOf(entry)
|
||||
return &page, nil
|
||||
}
|
||||
|
||||
// validateCommentURL 校验评论地址:只允许 http(s),并挡住内网地址。
|
||||
//
|
||||
// 段评地址来自书源正文,会带着书源登录态由服务端去抓;不做限制就等于给了一个
|
||||
// 「以书源身份请求任意地址」的口子(SSRF)。
|
||||
func validateCommentURL(raw string) (string, error) {
|
||||
u := strings.TrimSpace(raw)
|
||||
if u == "" {
|
||||
return "", fmt.Errorf("评论地址为空")
|
||||
}
|
||||
parsed, err := url.Parse(u)
|
||||
if err != nil || parsed.Host == "" || (parsed.Scheme != "http" && parsed.Scheme != "https") {
|
||||
return "", fmt.Errorf("评论地址无效:%s", truncateForLog(u, 80))
|
||||
}
|
||||
if isPrivateHost(parsed.Hostname()) {
|
||||
return "", fmt.Errorf("评论地址指向内网,已拒绝")
|
||||
}
|
||||
return u, nil
|
||||
}
|
||||
|
||||
// isPrivateHost 判断主机名是否是回环/内网/链路本地地址。
|
||||
func isPrivateHost(host string) bool {
|
||||
h := strings.ToLower(strings.Trim(host, "[]"))
|
||||
if h == "" || h == "localhost" || strings.HasSuffix(h, ".localhost") {
|
||||
return true
|
||||
}
|
||||
ip := net.ParseIP(h)
|
||||
if ip == nil {
|
||||
return false
|
||||
}
|
||||
return ip.IsLoopback() || ip.IsPrivate() || ip.IsUnspecified() || ip.IsLinkLocalUnicast()
|
||||
}
|
||||
@@ -0,0 +1,342 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"encoding/base64"
|
||||
"fmt"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"strings"
|
||||
"testing"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
)
|
||||
|
||||
// 段评解析单测:覆盖 legado 的两种正文形态(<comment /> 与「图片地址 + JSON 配置」),
|
||||
// 以及「绝不能吞掉或破坏正文文字」这条硬约束。
|
||||
|
||||
// svgCommentSrc 按光遇聚合 createSvg 的写法拼一张带评论数气泡图:
|
||||
// data:image/svg+xml;base64,<svg>,{"click":"showCmt('地址','平台','段评')","style":"text"}
|
||||
//
|
||||
// 注意第二个逗号后面直接跟 JSON,且 JSON 自身用双引号——它嵌在双引号属性里,
|
||||
// 正是会打断普通正则的「非法 HTML」,解析必须扛得住。
|
||||
func svgCommentSrc(t *testing.T, count int, click string, style string) string {
|
||||
t.Helper()
|
||||
svg := fmt.Sprintf(
|
||||
`<svg xmlns='http://www.w3.org/2000/svg'><text x='1' y='2'>%d</text></svg>`, count)
|
||||
b64 := base64.StdEncoding.EncodeToString([]byte(svg))
|
||||
return fmt.Sprintf(`data:image/svg+xml;base64,%s,{"click":%q,"style":%q}`, b64, click, style)
|
||||
}
|
||||
|
||||
func TestExtractContentCommentsLegacyTag(t *testing.T) {
|
||||
content := `<p>第一段<comment ident="https://cdn.example.com/c/1" count="12" /></p>
|
||||
<p>第二段</p>`
|
||||
text, comments := extractContentComments(content)
|
||||
if text != "<p>第一段</p>\n<p>第二段</p>" {
|
||||
t.Fatalf("正文被破坏:%q", text)
|
||||
}
|
||||
if len(comments) != 1 {
|
||||
t.Fatalf("评论数 = %d,期望 1", len(comments))
|
||||
}
|
||||
c := comments[0]
|
||||
if c.Line != 0 || c.Count != 12 || c.URL != "https://cdn.example.com/c/1" || c.Label != "段评" {
|
||||
t.Fatalf("解析结果不符:%+v", c)
|
||||
}
|
||||
if c.Block {
|
||||
t.Fatalf("段落内的段评不应标记为整块:%+v", c)
|
||||
}
|
||||
}
|
||||
|
||||
func TestExtractContentCommentsIOSOnPress(t *testing.T) {
|
||||
content := `<p>正文<comment count="7" onPress="java.showReadingBrowser('https://cdn.example.com/c/2?a=1&b=2','番茄段评')"></comment></p>`
|
||||
text, comments := extractContentComments(content)
|
||||
if text != "<p>正文</p>" {
|
||||
t.Fatalf("正文被破坏:%q", text)
|
||||
}
|
||||
if len(comments) != 1 {
|
||||
t.Fatalf("评论数 = %d,期望 1", len(comments))
|
||||
}
|
||||
// & 必须还原成 &,否则请求评论接口会 404
|
||||
if comments[0].URL != "https://cdn.example.com/c/2?a=1&b=2" || comments[0].Count != 7 {
|
||||
t.Fatalf("解析结果不符:%+v", comments[0])
|
||||
}
|
||||
}
|
||||
|
||||
func TestExtractContentCommentsInlineImage(t *testing.T) {
|
||||
src := svgCommentSrc(t, 3, "showCmt('https://v1.example.com/get_review?book_id=1&item_id=2','番茄','段评')", "text")
|
||||
content := `<p>正文内容<img src="` + src + `"></p>
|
||||
<p>下一段</p>`
|
||||
text, comments := extractContentComments(content)
|
||||
if text != "<p>正文内容</p>\n<p>下一段</p>" {
|
||||
t.Fatalf("正文被破坏:%q", text)
|
||||
}
|
||||
if len(comments) != 1 {
|
||||
t.Fatalf("评论数 = %d,期望 1", len(comments))
|
||||
}
|
||||
c := comments[0]
|
||||
if c.Line != 0 || c.Count != 3 || c.Label != "段评" {
|
||||
t.Fatalf("解析结果不符:%+v", c)
|
||||
}
|
||||
// 地址里的 & 原样保留(JSON 里的 & 不是实体)
|
||||
if c.URL != "https://v1.example.com/get_review?book_id=1&item_id=2" {
|
||||
t.Fatalf("评论地址 = %q", c.URL)
|
||||
}
|
||||
}
|
||||
|
||||
func TestExtractContentCommentsChapterEndBlock(t *testing.T) {
|
||||
// 章末「本章说」:单独一行,配置用单引号键、双引号值(合法 JS 对象字面量,非法 JSON)
|
||||
line := `<img src="https://v1.example.com/chapter_review_svg?book_id=1&img=1,` +
|
||||
`{'type':'qtbzs',` +
|
||||
`"click":"showCmt('https://v1.example.com/get_review?book_id=1','番茄','本章说' )",` +
|
||||
`'style':'FULL'}">`
|
||||
content := "<p>正文最后一段</p>\n" + line
|
||||
text, comments := extractContentComments(content)
|
||||
if text != "<p>正文最后一段</p>\n" {
|
||||
t.Fatalf("正文被破坏:%q", text)
|
||||
}
|
||||
if len(comments) != 1 {
|
||||
t.Fatalf("评论数 = %d,期望 1", len(comments))
|
||||
}
|
||||
c := comments[0]
|
||||
if c.Line != 1 || c.Label != "本章说" || !c.Block {
|
||||
t.Fatalf("解析结果不符:%+v", c)
|
||||
}
|
||||
if c.URL != "https://v1.example.com/get_review?book_id=1" {
|
||||
t.Fatalf("评论地址 = %q", c.URL)
|
||||
}
|
||||
}
|
||||
|
||||
func TestExtractContentCommentsCountOver99(t *testing.T) {
|
||||
// 书源把 >99 的评论数显示成 "99+",解析要取到 99
|
||||
src := svgCommentSrc(t, 99, "showCmt('https://x.example.com/c','番茄','段评')", "text")
|
||||
// 手工把文本改成 99+
|
||||
src = strings.Replace(src, ">99<", ">99+<", 1)
|
||||
_, comments := extractContentComments(`<p>x<img src="` + src + `"></p>`)
|
||||
if len(comments) != 1 || comments[0].Count != 99 {
|
||||
t.Fatalf("期望解析出 99,得到 %+v", comments)
|
||||
}
|
||||
}
|
||||
|
||||
// 没有任何段评标记时,正文必须逐字节不变(不能因为「看起来像 HTML」就动手)。
|
||||
func TestExtractContentCommentsPreservesPlainContent(t *testing.T) {
|
||||
content := `<p>普通段落</p>
|
||||
<p>带插图的段落<img src="https://cdn.example.com/a.jpg"></p>
|
||||
<p>带 onclick 图但非段评<img src="https://cdn.example.com/b.jpg,{'click':'openPhoto(1)'}"></p>`
|
||||
text, comments := extractContentComments(content)
|
||||
if text != content {
|
||||
t.Fatalf("正文被改动:\n got %q\nwant %q", text, content)
|
||||
}
|
||||
if len(comments) != 0 {
|
||||
t.Fatalf("不应解析出段评:%+v", comments)
|
||||
}
|
||||
}
|
||||
|
||||
func TestExtractContentCommentsEmpty(t *testing.T) {
|
||||
if text, comments := extractContentComments(""); text != "" || comments != nil {
|
||||
t.Fatalf("空正文应原样返回,得到 %q / %+v", text, comments)
|
||||
}
|
||||
if text, comments := extractContentComments("<p>只有普通正文</p>"); text != "<p>只有普通正文</p>" || comments != nil {
|
||||
t.Fatalf("无标记正文应原样返回,得到 %q / %+v", text, comments)
|
||||
}
|
||||
}
|
||||
|
||||
func TestSplitImgSrcParams(t *testing.T) {
|
||||
base, params := splitImgSrcParams(`data:image/svg+xml;base64,QUJD,{"click":"showCmt('u','x')","style":"text"}`)
|
||||
if base != "data:image/svg+xml;base64,QUJD" {
|
||||
t.Fatalf("base = %q", base)
|
||||
}
|
||||
if params["click"] != "showCmt('u','x')" || params["style"] != "text" {
|
||||
t.Fatalf("params = %+v", params)
|
||||
}
|
||||
if base, params := splitImgSrcParams("https://x.example.com/a.jpg"); params != nil || base != "https://x.example.com/a.jpg" {
|
||||
t.Fatalf("无配置时不应拆出参数:%q / %+v", base, params)
|
||||
}
|
||||
}
|
||||
|
||||
func TestValidateCommentURL(t *testing.T) {
|
||||
bad := []string{
|
||||
"", "not-a-url", "file:///etc/passwd", "javascript:alert(1)",
|
||||
"http://localhost/c", "http://127.0.0.1/c", "http://192.168.1.1/c", "http://[::1]/c",
|
||||
}
|
||||
for _, u := range bad {
|
||||
if _, err := validateCommentURL(u); err == nil {
|
||||
t.Fatalf("%q 应被拒绝", u)
|
||||
}
|
||||
}
|
||||
good := "https://v1.qingtian618.com/get_review?book_id=1&item_id=2"
|
||||
if got, err := validateCommentURL(good); err != nil || got != good {
|
||||
t.Fatalf("合法地址被拒:%q / %v", got, err)
|
||||
}
|
||||
}
|
||||
|
||||
// 端到端:书源正文里带段评标记时,GetContentForBook 应清掉标记、
|
||||
// 下发结构化 Comments,且正文文字完好。
|
||||
func TestGetContentForBookExtractsComments(t *testing.T) {
|
||||
src := svgCommentSrc(t, 5, "showCmt('https://v1.example.com/get_review?book_id=1','番茄','段评')", "text")
|
||||
// 服务端按 hex 还原规则返回值里的正文;这里让正文规则直接返回一段放进 data 信封的 JSON。
|
||||
pageJSON := fmt.Sprintf(`{"content":%q}`, "<p>第一段<img src=\""+src+"\"></p>\n<p>第二段</p>")
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
w.Header().Set("Content-Type", "application/json; charset=utf-8")
|
||||
_, _ = w.Write([]byte(pageJSON))
|
||||
}))
|
||||
defer srv.Close()
|
||||
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
srcJSON := fmt.Sprintf(`{
|
||||
"bookSourceUrl": %q,
|
||||
"bookSourceName": "段评测试源",
|
||||
"bookSourceType": 0,
|
||||
"ruleContent": { "content": "$.content" }
|
||||
}`, srv.URL)
|
||||
importTestSource(t, svc, srcJSON, srv.URL)
|
||||
|
||||
book := &model.ReaderBook{
|
||||
UserID: "u1",
|
||||
Origin: srv.URL,
|
||||
OriginName: "段评测试源",
|
||||
BookURL: srv.URL + "/book/1",
|
||||
Name: "测试书",
|
||||
Type: 0,
|
||||
}
|
||||
if err := svc.repo.CreateBook(ctx, book); err != nil {
|
||||
t.Fatalf("创建书籍失败: %v", err)
|
||||
}
|
||||
if err := svc.SaveChapters(ctx, book.ID, []ChapterInput{
|
||||
{Index: 0, Title: "第一章", URL: srv.URL + "/book/1/c1.html"},
|
||||
}); err != nil {
|
||||
t.Fatalf("写入章节失败: %v", err)
|
||||
}
|
||||
|
||||
out, err := svc.GetContentForBook(ctx, "u1", book.ID, 0)
|
||||
if err != nil {
|
||||
t.Fatalf("取正文失败: %v", err)
|
||||
}
|
||||
if out.Type != "text" {
|
||||
t.Fatalf("类型 = %q,期望 text", out.Type)
|
||||
}
|
||||
if strings.Contains(out.Content, "<img") || strings.Contains(out.Content, "showCmt") {
|
||||
t.Fatalf("段评标记没有清干净:%q", out.Content)
|
||||
}
|
||||
// 块级标签折成换行后,段落文字一个字都不少(多余空行已压掉)。
|
||||
if out.Content != "第一段\n第二段" {
|
||||
t.Fatalf("正文被破坏:%q", out.Content)
|
||||
}
|
||||
if strings.Contains(out.Content, "<p>") || strings.Contains(out.Content, "</p>") {
|
||||
t.Fatalf("块级标签没有折行:%q", out.Content)
|
||||
}
|
||||
if len(out.Comments) != 1 {
|
||||
t.Fatalf("段评数 = %d,期望 1", len(out.Comments))
|
||||
}
|
||||
c := out.Comments[0]
|
||||
if c.Line != 0 || c.Count != 5 || c.URL != "https://v1.example.com/get_review?book_id=1" {
|
||||
t.Fatalf("段评内容不符:%+v", c)
|
||||
}
|
||||
}
|
||||
|
||||
// setupTextChapterBook 搭一个「正文规则直接返回指定字符串」的文本源 + 一本书,
|
||||
// 返回服务实例与书籍 ID,供正文链路回归测试复用。
|
||||
func setupTextChapterBook(t *testing.T, content string) (*ReaderService, string) {
|
||||
t.Helper()
|
||||
pageJSON := fmt.Sprintf(`{"content":%q}`, content)
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
w.Header().Set("Content-Type", "application/json; charset=utf-8")
|
||||
_, _ = w.Write([]byte(pageJSON))
|
||||
}))
|
||||
t.Cleanup(srv.Close)
|
||||
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
srcJSON := fmt.Sprintf(`{
|
||||
"bookSourceUrl": %q,
|
||||
"bookSourceName": "正文链路测试源",
|
||||
"bookSourceType": 0,
|
||||
"ruleContent": { "content": "$.content" }
|
||||
}`, srv.URL)
|
||||
importTestSource(t, svc, srcJSON, srv.URL)
|
||||
|
||||
book := &model.ReaderBook{
|
||||
UserID: "u1",
|
||||
Origin: srv.URL,
|
||||
OriginName: "正文链路测试源",
|
||||
BookURL: srv.URL + "/book/1",
|
||||
Name: "测试书",
|
||||
Type: 0,
|
||||
}
|
||||
if err := svc.repo.CreateBook(ctx, book); err != nil {
|
||||
t.Fatalf("创建书籍失败: %v", err)
|
||||
}
|
||||
if err := svc.SaveChapters(ctx, book.ID, []ChapterInput{
|
||||
{Index: 0, Title: "第一章", URL: srv.URL + "/book/1/c1.html"},
|
||||
}); err != nil {
|
||||
t.Fatalf("写入章节失败: %v", err)
|
||||
}
|
||||
return svc, book.ID
|
||||
}
|
||||
|
||||
// TestGetContentForBookKeepsTextWithStandaloneImage 回归问题 A:带 <p> 正文的章节
|
||||
// 混入一行独立普通插图时,仍然必须是文本类型、正文一个字都不能丢
|
||||
// (旧实现把 <p>正文</p> 当孤立标签跳过,整章被误判成图片章清空)。
|
||||
func TestGetContentForBookKeepsTextWithStandaloneImage(t *testing.T) {
|
||||
content := "<p>第一段正文</p>\n" +
|
||||
"<p>第二段正文</p>\n" +
|
||||
`<img src="https://cdn.example.com/pic.jpg">`
|
||||
svc, bookID := setupTextChapterBook(t, content)
|
||||
|
||||
out, err := svc.GetContentForBook(t.Context(), "u1", bookID, 0)
|
||||
if err != nil {
|
||||
t.Fatalf("取正文失败: %v", err)
|
||||
}
|
||||
if out.Type != "text" {
|
||||
t.Fatalf("类型 = %q,期望 text(正文不能被当成图片章)", out.Type)
|
||||
}
|
||||
if !strings.Contains(out.Content, "第一段正文") || !strings.Contains(out.Content, "第二段正文") {
|
||||
t.Fatalf("正文文字丢失:%q", out.Content)
|
||||
}
|
||||
if !strings.HasPrefix(out.Content, "第一段正文\n第二段正文\n"+imgMarkerPrefix) {
|
||||
t.Fatalf("正文结构不符:%q", out.Content)
|
||||
}
|
||||
if len(out.Images) != 0 {
|
||||
t.Fatalf("文本类型的 Images 应为空,得到 %v", out.Images)
|
||||
}
|
||||
}
|
||||
|
||||
// TestGetContentForBookStripsBlockTagsKeepsText 回归问题 B:段评开启时正文是
|
||||
// <p>正文<img …,{click}> 形态,下发/显示的文本里不能残留块级标签,且文字一字不少。
|
||||
func TestGetContentForBookStripsBlockTagsKeepsText(t *testing.T) {
|
||||
src := svgCommentSrc(t, 3, "showCmt('https://v1.example.com/get_review?book_id=1','番茄','段评')", "text")
|
||||
content := "<p>第一段正文</p>\n" +
|
||||
"<p>第二段正文<img src=\"" + src + "\"></p>\n" +
|
||||
"<p>第三段<br>折行后的文字</p>\n" +
|
||||
`<img src="https://cdn.example.com/pic.jpg">`
|
||||
svc, bookID := setupTextChapterBook(t, content)
|
||||
|
||||
out, err := svc.GetContentForBook(t.Context(), "u1", bookID, 0)
|
||||
if err != nil {
|
||||
t.Fatalf("取正文失败: %v", err)
|
||||
}
|
||||
if out.Type != "text" {
|
||||
t.Fatalf("类型 = %q,期望 text", out.Type)
|
||||
}
|
||||
t.Logf("最终下发/显示的正文 = %q", out.Content)
|
||||
t.Logf("段评锚点 = %+v", out.Comments)
|
||||
for _, tag := range []string{"<p>", "</p>", "<br>", "<img", "showCmt"} {
|
||||
if strings.Contains(out.Content, tag) {
|
||||
t.Fatalf("显示文本里残留了 %q:%q", tag, out.Content)
|
||||
}
|
||||
}
|
||||
// 正文文字一字不少(含 <br> 折出来的那一段)。
|
||||
for _, want := range []string{"第一段正文", "第二段正文", "折行后的文字"} {
|
||||
if !strings.Contains(out.Content, want) {
|
||||
t.Fatalf("正文文字丢失 %q:%q", want, out.Content)
|
||||
}
|
||||
}
|
||||
if !strings.Contains(out.Content, imgMarkerPrefix) {
|
||||
t.Fatalf("普通插图应保留成 %s 标记:%q", imgMarkerPrefix, out.Content)
|
||||
}
|
||||
if len(out.Comments) != 1 {
|
||||
t.Fatalf("段评数 = %d,期望 1", len(out.Comments))
|
||||
}
|
||||
if out.Comments[0].URL != "https://v1.example.com/get_review?book_id=1" {
|
||||
t.Fatalf("段评地址不符:%+v", out.Comments[0])
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,55 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"strings"
|
||||
"testing"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
)
|
||||
|
||||
// TestContentRuleJSReceivesJoinedString 覆盖拷贝漫画(优++)的正文规则写法:
|
||||
//
|
||||
// $..url
|
||||
// <js>result.split("\n").map(x=>'<img src="'+x+'">').join("\n")</js>
|
||||
//
|
||||
// JSON 段先命中「多张图片的 url 列表」,JS 段拿到的必须是按 \n 拼好的字符串
|
||||
// (对应 legado BookContent 用 analyzeRule.getString 求值正文规则),
|
||||
// 喂成数组就会 TypeError: Object has no member 'split',整章图片全取不到。
|
||||
func TestContentRuleJSReceivesJoinedString(t *testing.T) {
|
||||
svc, _ := newLoginTestService(t)
|
||||
|
||||
const srcURL = "https://www.mangacopy.com/"
|
||||
const contentRule = "$..url\n<js>result.split(\"\\n\").map(x=>'<img src=\"'+x+'\">').join(\"\\n\")</js>"
|
||||
header := ""
|
||||
src := &model.ReaderBookSource{Name: "拷贝漫画(优++)", SourceURL: srcURL}
|
||||
bs := &BookSource{
|
||||
BookSourceURL: srcURL,
|
||||
RuleContent: &ContentRule{Content: strPtr(contentRule)},
|
||||
Header: &header,
|
||||
}
|
||||
|
||||
body := `{"results":{"chapter":{"contents":[` +
|
||||
`{"url":"https://img.example.com/1.webp"},` +
|
||||
`{"url":"https://img.example.com/2.webp"}]}}}`
|
||||
|
||||
sess := svc.newSession(t.Context(), src, bs)
|
||||
defer sess.close()
|
||||
ar := sess.newAnalyzer("", 0, body, "https://api.mangacopy.com/api/v3/chapter/x")
|
||||
|
||||
got, err := ar.GetString(contentRule, nil, false)
|
||||
if err != nil {
|
||||
t.Fatalf("正文规则求值失败: %v", err)
|
||||
}
|
||||
want := "<img src=\"https://img.example.com/1.webp\">\n<img src=\"https://img.example.com/2.webp\">"
|
||||
if got != want {
|
||||
t.Fatalf("正文 = %q\n期望 %q", got, want)
|
||||
}
|
||||
// 图片提取(漫画分支):每行一个 <img>,应抽出 2 个真实地址
|
||||
refs := 0
|
||||
for _, line := range strings.Split(got, "\n") {
|
||||
refs += len(imageRefsInLine(line))
|
||||
}
|
||||
if refs != 2 {
|
||||
t.Fatalf("从正文抽出 %d 张图,期望 2", refs)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,181 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"strings"
|
||||
"sync"
|
||||
"testing"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/service/reader/rule"
|
||||
)
|
||||
|
||||
// 本文件:响应 Set-Cookie 回写会话的回归测试(cookies.go)。
|
||||
//
|
||||
// 背景:扫码登录类书源把凭证放在**跳转链中间那一跳**的 Set-Cookie 里
|
||||
// (B站 的 crossDomain 票据地址就是如此),并且签到站点会声明业务域的
|
||||
// Domain(Domain=.bilibili.com)。之前的实现只看最终响应、按请求地址归档,
|
||||
// 于是凭证永远进不了会话,表现为「面板显示已授权,书源却始终未登录」。
|
||||
|
||||
// cookieCaptureServer 构造跳转链:/redeem 下发声明了 Domain 的凭证并 302,
|
||||
// /done 只回 200。这样「中间那一跳的 Set-Cookie」与「最终响应的 Set-Cookie」
|
||||
// 能被分开断言。
|
||||
func cookieCaptureServer(t *testing.T) *httptest.Server {
|
||||
t.Helper()
|
||||
mux := http.NewServeMux()
|
||||
mux.HandleFunc("/redeem", func(w http.ResponseWriter, r *http.Request) {
|
||||
http.SetCookie(w, &http.Cookie{
|
||||
Name: "SESSDATA", Value: "cred-abc", Domain: ".example.com", Path: "/",
|
||||
})
|
||||
http.Redirect(w, r, "/done", http.StatusFound)
|
||||
})
|
||||
mux.HandleFunc("/done", func(w http.ResponseWriter, r *http.Request) {
|
||||
http.SetCookie(w, &http.Cookie{Name: "biz", Value: "1", Path: "/"})
|
||||
_, _ = w.Write([]byte(`{"ok":true}`))
|
||||
})
|
||||
mux.HandleFunc("/plain", func(w http.ResponseWriter, r *http.Request) {
|
||||
_, _ = w.Write([]byte(`<html><body>面板</body></html>`))
|
||||
})
|
||||
srv := httptest.NewServer(mux)
|
||||
t.Cleanup(srv.Close)
|
||||
return srv
|
||||
}
|
||||
|
||||
// TestExecuteCapturesRedirectHopCookies 主链路:跳转中间那一跳的 Set-Cookie
|
||||
// 必须被收下,且按 Cookie 自己的 Domain 归档(而不是按请求地址)。
|
||||
func TestExecuteCapturesRedirectHopCookies(t *testing.T) {
|
||||
srv := cookieCaptureServer(t)
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
state := svc.newSourceState(ctx, srv.URL)
|
||||
|
||||
req := &rule.Request{Method: "GET", URL: srv.URL + "/redeem", Headers: map[string]string{}}
|
||||
if _, _, _, err := svc.executeWithState(ctx, req, state, true); err != nil {
|
||||
t.Fatalf("请求失败: %v", err)
|
||||
}
|
||||
|
||||
// 凭证声明的是 .example.com,就应该落在 example.com 这个域桶里。
|
||||
if got := state.GetCookie("https://example.com/"); !strings.Contains(got, "SESSDATA=cred-abc") {
|
||||
t.Fatalf("跳转中间那一跳的凭证没有按 Domain 归档: %q", got)
|
||||
}
|
||||
// 最终响应的 Cookie 按该跳地址归档。
|
||||
if got := state.GetCookie(srv.URL); !strings.Contains(got, "biz=1") {
|
||||
t.Fatalf("最终响应的 Cookie 没有被收下: %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
// TestExecuteSkipsCookieCaptureWhenJarDisabled enabledCookieJar=false 时
|
||||
// 主链路不应自动累积 Cookie。
|
||||
func TestExecuteSkipsCookieCaptureWhenJarDisabled(t *testing.T) {
|
||||
srv := cookieCaptureServer(t)
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
state := svc.newSourceState(ctx, srv.URL)
|
||||
|
||||
req := &rule.Request{Method: "GET", URL: srv.URL + "/redeem", Headers: map[string]string{}}
|
||||
if _, _, _, err := svc.executeWithState(ctx, req, state, false); err != nil {
|
||||
t.Fatalf("请求失败: %v", err)
|
||||
}
|
||||
if got := state.GetCookie("https://example.com/"); got != "" {
|
||||
t.Fatalf("关掉 CookieJar 后不应自动累积: %q", got)
|
||||
}
|
||||
if got := state.GetCookie(srv.URL); got != "" {
|
||||
t.Fatalf("关掉 CookieJar 后不应自动累积: %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
// TestProxyBrowserXHRCapturesCookies 面板链路:页面自己发起的请求(经
|
||||
// ProxyBrowserXHR 转发)拿到的 Set-Cookie 也要写回书源会话并能落库。
|
||||
// 这正是「页面里登录成功、书源却始终未登录」的那个缺口。
|
||||
func TestProxyBrowserXHRCapturesCookies(t *testing.T) {
|
||||
srv := cookieCaptureServer(t)
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
const sourceURL = "https://api.example.com"
|
||||
const sourceID = "panel-source-1"
|
||||
state := svc.newSourceState(ctx, sourceURL)
|
||||
|
||||
// 承载页面本身不下发凭证(/plain),确保断言只针对代理请求。
|
||||
if err := svc.openBrowser(ctx, sourceURL, sourceID, readerTestUserID,
|
||||
browserCookieTarget{state: state},
|
||||
rule.BrowserTask{URL: srv.URL + "/plain", Title: "面板"}); err != nil {
|
||||
t.Fatalf("登记页面失败: %v", err)
|
||||
}
|
||||
pages := svc.PendingBrowserPages(readerTestUserID, sourceID)
|
||||
if len(pages) != 1 {
|
||||
t.Fatalf("应登记 1 个待办页面,实际 %d", len(pages))
|
||||
}
|
||||
|
||||
// 页面请求票据地址:302 时下发凭证,最终响应只带回 JSON。
|
||||
if _, err := svc.ProxyBrowserXHR(ctx, pages[0].ID, http.MethodGet, srv.URL+"/redeem", nil, ""); err != nil {
|
||||
t.Fatalf("代理请求失败: %v", err)
|
||||
}
|
||||
if got := state.GetCookie("https://example.com/"); !strings.Contains(got, "SESSDATA=cred-abc") {
|
||||
t.Fatalf("面板请求的凭证没有写回会话: %q", got)
|
||||
}
|
||||
|
||||
// 登录态必须能持久化:重新读一条会话也应看到这份 Cookie。
|
||||
state.flush()
|
||||
reloaded := svc.newSourceState(ctx, sourceURL)
|
||||
if got := reloaded.GetCookie("https://example.com/"); !strings.Contains(got, "SESSDATA=cred-abc") {
|
||||
t.Fatalf("面板拿到的凭证没有落库: %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
// TestProxyBrowserXHRWithoutSessionSkipsCapture 没有会话可回写时(如段评面板)
|
||||
// 代理请求照常工作,只是不做 Cookie 累积。
|
||||
func TestProxyBrowserXHRWithoutSessionSkipsCapture(t *testing.T) {
|
||||
srv := cookieCaptureServer(t)
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
const sourceID = "panel-source-2"
|
||||
|
||||
if err := svc.openBrowser(ctx, srv.URL, sourceID, readerTestUserID,
|
||||
browserCookieTarget{},
|
||||
rule.BrowserTask{URL: srv.URL + "/plain", Title: "面板"}); err != nil {
|
||||
t.Fatalf("登记页面失败: %v", err)
|
||||
}
|
||||
pages := svc.PendingBrowserPages(readerTestUserID, sourceID)
|
||||
if len(pages) != 1 {
|
||||
t.Fatalf("应登记 1 个待办页面,实际 %d", len(pages))
|
||||
}
|
||||
res, err := svc.ProxyBrowserXHR(ctx, pages[0].ID, http.MethodGet, srv.URL+"/done", nil, "")
|
||||
if err != nil {
|
||||
t.Fatalf("代理请求应照常成功: %v", err)
|
||||
}
|
||||
if res.Status != http.StatusOK {
|
||||
t.Fatalf("代理状态 = %d", res.Status)
|
||||
}
|
||||
}
|
||||
|
||||
// TestBrowserHostCookieTargetRespectsJarFlag 面板链路同样受 enabledCookieJar 控制。
|
||||
func TestBrowserHostCookieTargetRespectsJarFlag(t *testing.T) {
|
||||
st := &sourceState{cookies: map[string]string{}}
|
||||
if got := (&browserHost{state: st, capture: false}).cookieTarget(); got.state != nil {
|
||||
t.Fatal("enabledCookieJar=false 时不应把会话交给面板回写 Cookie")
|
||||
}
|
||||
if got := (&browserHost{state: st, capture: true}).cookieTarget(); got.state != st {
|
||||
t.Fatal("开启时应把会话交给面板回写 Cookie")
|
||||
}
|
||||
}
|
||||
|
||||
// TestSourceStateCookieConcurrency 面板里的资源/接口代理是并发回写 Cookie 的,
|
||||
// 必须与书源 JS 的读取安全共存(-race 下能抓到缺锁)。
|
||||
func TestSourceStateCookieConcurrency(t *testing.T) {
|
||||
st := &sourceState{cookies: map[string]string{}}
|
||||
var wg sync.WaitGroup
|
||||
for i := 0; i < 16; i++ {
|
||||
wg.Add(1)
|
||||
go func(n int) {
|
||||
defer wg.Done()
|
||||
url := "https://example.com/"
|
||||
st.SetCookie(url, "k=1")
|
||||
_ = st.GetCookie(url)
|
||||
_ = st.snapshotCookies()
|
||||
}(i)
|
||||
}
|
||||
wg.Wait()
|
||||
if got := st.GetCookie("https://example.com/"); !strings.Contains(got, "k=1") {
|
||||
t.Fatalf("并发写入后应能读到 Cookie: %q", got)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,148 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"context"
|
||||
"errors"
|
||||
"net/http"
|
||||
"net/url"
|
||||
"strings"
|
||||
"sync"
|
||||
)
|
||||
|
||||
// 本文件:把响应下发的 Set-Cookie 收进书源会话。
|
||||
//
|
||||
// 之前只有 executeWithState 收 Set-Cookie,而且只看**最终**响应、按请求地址归档。
|
||||
// 对「靠跳转链下发凭证」的登录源(典型是扫码登录:票据地址一跳 302 带
|
||||
// Set-Cookie,再跳到业务域名)这三个做法都会漏:
|
||||
//
|
||||
// 1. 凭证在中间某一跳的 Set-Cookie 里,只看最终响应拿不到;
|
||||
// 2. 签发站点(如 passport.biligame.com)声明的 Domain 可能是业务域
|
||||
// (Domain=.bilibili.com),按请求地址归档会存进错误的域桶,
|
||||
// 之后 api.bilibili.com 的请求取不到这份 Cookie;
|
||||
// 3. 浏览器面板里的页面自己发起的请求走 ProxyBrowserXHR,完全不落库,
|
||||
// 页面在面板里「登录成功」但书源会话始终是空的。
|
||||
//
|
||||
// 因此统一成:一个 cookieSink 收集一次请求(含各跳)的 Set-Cookie,
|
||||
// 再按「Cookie 自身的 Domain 优先、否则该跳请求地址」写入会话。
|
||||
|
||||
// capturedCookie 一条待写入会话的 Cookie 及其归属地址。
|
||||
type capturedCookie struct {
|
||||
// url 归属地址:Cookie 声明了 Domain 时用 Domain(去掉前导点),
|
||||
// 否则用下发它的那一跳的请求地址。
|
||||
url string
|
||||
cookie string
|
||||
}
|
||||
|
||||
// cookieSink 收集一次请求(含重定向各跳)下发的 Set-Cookie。
|
||||
// 并发安全:面板里的多个资源/接口请求会同时走这里。
|
||||
type cookieSink struct {
|
||||
mu sync.Mutex
|
||||
items []capturedCookie
|
||||
}
|
||||
|
||||
// add 收集一份响应里的 Set-Cookie。reqURL 是产生该响应(或跳转前那一跳)的地址。
|
||||
func (s *cookieSink) add(reqURL *url.URL, resp *http.Response) {
|
||||
if s == nil || resp == nil {
|
||||
return
|
||||
}
|
||||
items := make([]capturedCookie, 0, len(resp.Cookies()))
|
||||
for _, ck := range resp.Cookies() {
|
||||
if ck == nil || ck.Name == "" {
|
||||
continue
|
||||
}
|
||||
target := ""
|
||||
if d := strings.TrimSpace(ck.Domain); d != "" {
|
||||
target = strings.TrimPrefix(d, ".")
|
||||
} else if reqURL != nil {
|
||||
target = reqURL.String()
|
||||
}
|
||||
if target == "" {
|
||||
continue
|
||||
}
|
||||
items = append(items, capturedCookie{url: target, cookie: ck.Name + "=" + ck.Value})
|
||||
}
|
||||
if len(items) == 0 {
|
||||
return
|
||||
}
|
||||
s.mu.Lock()
|
||||
s.items = append(s.items, items...)
|
||||
s.mu.Unlock()
|
||||
}
|
||||
|
||||
// drain 取出并清空已收集的 Cookie。
|
||||
func (s *cookieSink) drain() []capturedCookie {
|
||||
if s == nil {
|
||||
return nil
|
||||
}
|
||||
s.mu.Lock()
|
||||
defer s.mu.Unlock()
|
||||
out := s.items
|
||||
s.items = nil
|
||||
return out
|
||||
}
|
||||
|
||||
// apply 把收集到的 Cookie 写入会话并清空(会话为 nil 时直接丢弃)。
|
||||
func (s *cookieSink) apply(state *sourceState) {
|
||||
if s == nil {
|
||||
return
|
||||
}
|
||||
items := s.drain()
|
||||
if state == nil || len(items) == 0 {
|
||||
return
|
||||
}
|
||||
for _, it := range items {
|
||||
state.SetCookie(it.url, it.cookie)
|
||||
}
|
||||
}
|
||||
|
||||
// ─── 请求级挂载:sink 放进 context,由 CheckRedirect 逐跳收集 ──────────────
|
||||
|
||||
type cookieSinkKey struct{}
|
||||
|
||||
// withCookieSink 在请求 context 上挂一个 sink;sink 为 nil 时原样返回。
|
||||
func withCookieSink(ctx context.Context, sink *cookieSink) context.Context {
|
||||
if sink == nil {
|
||||
return ctx
|
||||
}
|
||||
return context.WithValue(ctx, cookieSinkKey{}, sink)
|
||||
}
|
||||
|
||||
// cookieSinkFrom 取出请求 context 上的 sink(没有则 nil)。
|
||||
func cookieSinkFrom(ctx context.Context) *cookieSink {
|
||||
if ctx == nil {
|
||||
return nil
|
||||
}
|
||||
sink, _ := ctx.Value(cookieSinkKey{}).(*cookieSink)
|
||||
return sink
|
||||
}
|
||||
|
||||
// captureRedirectCookies 作为 http.Client.CheckRedirect:
|
||||
// 每发生一次跳转,就把「上一跳响应」里的 Set-Cookie 收进 context 上的 sink。
|
||||
//
|
||||
// 没有挂 sink 时是纯透传;跳转次数上限与 net/http 的默认实现保持一致
|
||||
// (10 次),所以装上它不改变既有链路的行为。
|
||||
func captureRedirectCookies(req *http.Request, via []*http.Request) error {
|
||||
if req != nil {
|
||||
if sink := cookieSinkFrom(req.Context()); sink != nil && req.Response != nil {
|
||||
// via 的最后一跳才是产生 req.Response 的那个请求。
|
||||
prev := req
|
||||
if n := len(via); n > 0 {
|
||||
prev = via[n-1]
|
||||
}
|
||||
sink.add(prev.URL, req.Response)
|
||||
}
|
||||
}
|
||||
if len(via) >= 10 {
|
||||
return errors.New("stopped after 10 redirects")
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
// enableCookieCapture 给 reader 的 HTTP 客户端装上「逐跳收集 Set-Cookie」。
|
||||
// NewReaderService 调用一次即可:没有挂 sink 的请求完全不受影响。
|
||||
func enableCookieCapture(c *http.Client) {
|
||||
if c == nil {
|
||||
return
|
||||
}
|
||||
c.CheckRedirect = captureRedirectCookies
|
||||
}
|
||||
@@ -0,0 +1,38 @@
|
||||
package reader
|
||||
|
||||
import "testing"
|
||||
|
||||
// 聚合类书源(如「光遇聚合」)在没有封面时会把搜索参数信封当封面返回,
|
||||
// normalizeCoverURL 必须拦住这类非图片地址,避免前端 <img> 显示破图。
|
||||
func TestNormalizeCoverURL(t *testing.T) {
|
||||
cases := []struct {
|
||||
name string
|
||||
in string
|
||||
want string
|
||||
}{
|
||||
{"empty", "", ""},
|
||||
{"blank", " ", ""},
|
||||
{"https", "https://img.example.com/a.jpg", "https://img.example.com/a.jpg"},
|
||||
{"http", "http://img.example.com/a.jpg", "http://img.example.com/a.jpg"},
|
||||
{"uppercase scheme", "HTTPS://img.example.com/a.jpg", "HTTPS://img.example.com/a.jpg"},
|
||||
{"trimmed", " https://img.example.com/a.jpg ", "https://img.example.com/a.jpg"},
|
||||
{"inline image data uri", "data:image/png;base64,iVBORw0KGgo=", "data:image/png;base64,iVBORw0KGgo="},
|
||||
{"local asset url", "/api/reader/local/asset?b=1&p=2&s=3", "/api/reader/local/asset?b=1&p=2&s=3"},
|
||||
{
|
||||
"search params envelope",
|
||||
"data:;base64,eyJrZXkiOiLlhajnkIPpq5jmraYiLCJ0YWIiOiLlsI/or7QiLCJzb3VyY2VzS2V5Ijoi5YWo6YOoIiwicGFnZSI6MSwiZGlzYWJsZWRfc291cmNlcyI6IjAifQ==",
|
||||
"",
|
||||
},
|
||||
{"text data uri", "data:text/plain;base64,aGk=", ""},
|
||||
{"source name", "光遇聚合", ""},
|
||||
{"javascript url", "javascript:alert(1)", ""},
|
||||
}
|
||||
|
||||
for _, tc := range cases {
|
||||
t.Run(tc.name, func(t *testing.T) {
|
||||
if got := normalizeCoverURL(tc.in); got != tc.want {
|
||||
t.Fatalf("normalizeCoverURL(%q) = %q, want %q", tc.in, got, tc.want)
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,7 @@
|
||||
package reader
|
||||
|
||||
import "encoding/base64"
|
||||
|
||||
func base64EncodeStr(s string) string {
|
||||
return base64.StdEncoding.EncodeToString([]byte(s))
|
||||
}
|
||||
@@ -0,0 +1,111 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"context"
|
||||
"fmt"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
)
|
||||
|
||||
// 书架分组(对应 legado BookGroup;存储见 model.ReaderBookGroups)。
|
||||
//
|
||||
// 语义与影视模块的媒体库标签一致:整份替换、一本书只归一个组、组内顺序即展示顺序。
|
||||
// 额外的两条收敛放在服务层,因为它们需要知道书籍是否存在:
|
||||
// - 只保留该用户书架上的书(书被移出书架后分组里不留死 ID);
|
||||
// - 一本书只归一个组(越靠前的分组优先)。
|
||||
// 空分组会保留:用户可能就是先建好分组再往里放书。
|
||||
|
||||
// GetBookGroups 读用户的书架分组。没有分组时返回空切片(非 nil),方便前端直接遍历。
|
||||
func (s *ReaderService) GetBookGroups(ctx context.Context, userID string) ([]model.BookGroupSet, error) {
|
||||
if userID == "" {
|
||||
return []model.BookGroupSet{}, nil
|
||||
}
|
||||
row, err := s.repo.GetBookGroups(ctx, userID)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
groups := model.NormalizeBookGroups(row.DecodeBookGroups())
|
||||
if len(groups) == 0 {
|
||||
return []model.BookGroupSet{}, nil
|
||||
}
|
||||
owned, err := s.ownedBookIDs(ctx, userID)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
return filterBookGroups(groups, owned), nil
|
||||
}
|
||||
|
||||
// SetBookGroups 覆盖保存用户的书架分组,落库前做过收敛与归属过滤。
|
||||
func (s *ReaderService) SetBookGroups(ctx context.Context, userID string, groups []model.BookGroupSet) ([]model.BookGroupSet, error) {
|
||||
if userID == "" {
|
||||
return nil, fmt.Errorf("缺少用户信息")
|
||||
}
|
||||
owned, err := s.ownedBookIDs(ctx, userID)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
normalized := dedupeBookGroups(filterBookGroups(model.NormalizeBookGroups(groups), owned))
|
||||
raw, err := model.EncodeBookGroups(normalized)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
if err := s.repo.SaveBookGroups(ctx, &model.ReaderBookGroups{UserID: userID, Groups: raw}); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
if len(normalized) == 0 {
|
||||
return []model.BookGroupSet{}, nil
|
||||
}
|
||||
return normalized, nil
|
||||
}
|
||||
|
||||
// ownedBookIDs 返回该用户书架上的书籍 ID 集合。
|
||||
func (s *ReaderService) ownedBookIDs(ctx context.Context, userID string) (map[string]struct{}, error) {
|
||||
books, err := s.repo.ListBooks(ctx, userID)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
out := make(map[string]struct{}, len(books))
|
||||
for i := range books {
|
||||
out[books[i].ID] = struct{}{}
|
||||
}
|
||||
return out, nil
|
||||
}
|
||||
|
||||
// filterBookGroups 只保留 owned 里的书籍 ID;空分组原样保留。
|
||||
func filterBookGroups(groups []model.BookGroupSet, owned map[string]struct{}) []model.BookGroupSet {
|
||||
if len(groups) == 0 {
|
||||
return nil
|
||||
}
|
||||
out := make([]model.BookGroupSet, 0, len(groups))
|
||||
for _, group := range groups {
|
||||
ids := make([]string, 0, len(group.BookIDs))
|
||||
for _, id := range group.BookIDs {
|
||||
if _, ok := owned[id]; ok {
|
||||
ids = append(ids, id)
|
||||
}
|
||||
}
|
||||
out = append(out, model.BookGroupSet{Name: group.Name, BookIDs: ids})
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// dedupeBookGroups 收敛成「一本书只归一个组」:越靠前的分组优先,后面的重复项被移除。
|
||||
func dedupeBookGroups(groups []model.BookGroupSet) []model.BookGroupSet {
|
||||
if len(groups) == 0 {
|
||||
return nil
|
||||
}
|
||||
claimed := make(map[string]struct{})
|
||||
out := make([]model.BookGroupSet, 0, len(groups))
|
||||
for _, group := range groups {
|
||||
ids := make([]string, 0, len(group.BookIDs))
|
||||
for _, id := range group.BookIDs {
|
||||
if _, ok := claimed[id]; ok {
|
||||
continue
|
||||
}
|
||||
claimed[id] = struct{}{}
|
||||
ids = append(ids, id)
|
||||
}
|
||||
out = append(out, model.BookGroupSet{Name: group.Name, BookIDs: ids})
|
||||
}
|
||||
return out
|
||||
}
|
||||
@@ -0,0 +1,119 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"compress/gzip"
|
||||
"encoding/json"
|
||||
"fmt"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"strings"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// 本文件:规则层必须拿到「解压后」的响应体。
|
||||
//
|
||||
// 回归背景:HTTPHeaderPresets 曾显式写 `Accept-Encoding: gzip, deflate, br`。
|
||||
// net/http 只在调用方「没有」设置该头时才会自动补 gzip 并透明解压,显式设置会
|
||||
// 让它原样交出压缩字节。于是凡是走了压缩的上游(静态 config.json、CDN 页面等)
|
||||
// 都会把 gzip 字节喂给书源的 JSON.parse,报
|
||||
// SyntaxError: invalid character '\x1f' looking for beginning of value
|
||||
// 表现出来就是「获取最新配置失败:SyntaxError: Unexpected end of JSON input」,
|
||||
// 并且书源会误判为线路故障,逐条切换全部线路后仍然失败。
|
||||
|
||||
// gzipJSONServer 返回一个「客户端支持压缩时才压缩」的站点,模拟真实 CDN。
|
||||
func gzipJSONServer(t *testing.T, payload string) (*httptest.Server, *string) {
|
||||
t.Helper()
|
||||
var seenAcceptEncoding string
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
seenAcceptEncoding = r.Header.Get("Accept-Encoding")
|
||||
w.Header().Set("Content-Type", "application/json; charset=utf-8")
|
||||
if strings.Contains(seenAcceptEncoding, "gzip") {
|
||||
w.Header().Set("Content-Encoding", "gzip")
|
||||
gz := gzip.NewWriter(w)
|
||||
_, _ = gz.Write([]byte(payload))
|
||||
_ = gz.Close()
|
||||
return
|
||||
}
|
||||
_, _ = w.Write([]byte(payload))
|
||||
}))
|
||||
return srv, &seenAcceptEncoding
|
||||
}
|
||||
|
||||
// readConfigSourceJSON 构造一个用 java.ajax 取配置并 JSON.parse 的书源。
|
||||
func readConfigSourceJSON(t *testing.T, serverURL, target string, sourceHeader string) string {
|
||||
t.Helper()
|
||||
loginJS := fmt.Sprintf(`function readConfig() {
|
||||
let data = java.ajax(%q);
|
||||
let js = JSON.parse(String(data));
|
||||
java.longToast('配置版本=' + js.version);
|
||||
}`, target)
|
||||
src := map[string]any{
|
||||
"bookSourceUrl": serverURL,
|
||||
"bookSourceName": "压缩测试源",
|
||||
"loginUrl": loginJS,
|
||||
"loginUi": `[{"name":"读配置","type":"button","action":"readConfig()"}]`,
|
||||
}
|
||||
if sourceHeader != "" {
|
||||
src["header"] = sourceHeader
|
||||
}
|
||||
out, err := json.Marshal(src)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
return string(out)
|
||||
}
|
||||
|
||||
// TestReaderJSSeesDecompressedBody 上游压缩的响应必须先解压再进规则层。
|
||||
func TestReaderJSSeesDecompressedBody(t *testing.T) {
|
||||
const payload = `{"version":"20260926"}`
|
||||
srv, seenAE := gzipJSONServer(t, payload)
|
||||
defer srv.Close()
|
||||
|
||||
svc, _ := newLoginTestService(t)
|
||||
sourceID := prepareLoginSource(t, svc, readConfigSourceJSON(t,
|
||||
srv.URL, srv.URL+"/static/source_config/config.json", ""))
|
||||
|
||||
res, err := svc.RunLoginAction(t.Context(), readerTestUserID, sourceID, "readConfig()", nil)
|
||||
if err != nil {
|
||||
t.Fatalf("动作执行失败: %v", err)
|
||||
}
|
||||
if !res.OK {
|
||||
t.Fatalf("动作未成功: %+v", res)
|
||||
}
|
||||
joined := strings.Join(res.Toasts, "\n")
|
||||
if !strings.Contains(joined, "配置版本=20260926") {
|
||||
t.Fatalf("规则层拿到的不是解压后的 JSON(压缩字节泄漏到 JSON.parse): %v", res.Toasts)
|
||||
}
|
||||
// 必须仍然协商压缩,否则等于用「不压缩」回避问题,真实站点该压还是压
|
||||
if !strings.Contains(*seenAE, "gzip") {
|
||||
t.Fatalf("请求应携带 gzip(由 net/http 自动添加),实际 %q", *seenAE)
|
||||
}
|
||||
}
|
||||
|
||||
// TestReaderStripsAcceptEncodingFromSourceHeader 书源 header 字段里的
|
||||
// Accept-Encoding 也必须被清掉:显式设置会让 net/http 放弃解压,
|
||||
// 而且一旦服务端选了 brotli,我们用标准库根本解不出来。
|
||||
func TestReaderStripsAcceptEncodingFromSourceHeader(t *testing.T) {
|
||||
const payload = `{"version":"20260926"}`
|
||||
srv, seenAE := gzipJSONServer(t, payload)
|
||||
defer srv.Close()
|
||||
|
||||
svc, _ := newLoginTestService(t)
|
||||
sourceID := prepareLoginSource(t, svc, readConfigSourceJSON(t,
|
||||
srv.URL, srv.URL+"/static/source_config/config.json",
|
||||
`{"Accept-Encoding":"br, gzip"}`))
|
||||
|
||||
res, err := svc.RunLoginAction(t.Context(), readerTestUserID, sourceID, "readConfig()", nil)
|
||||
if err != nil {
|
||||
t.Fatalf("动作执行失败: %v", err)
|
||||
}
|
||||
if !res.OK {
|
||||
t.Fatalf("动作未成功: %+v", res)
|
||||
}
|
||||
if !strings.Contains(strings.Join(res.Toasts, "\n"), "配置版本=20260926") {
|
||||
t.Fatalf("规则层未拿到解压后的 JSON: %v", res.Toasts)
|
||||
}
|
||||
if strings.Contains(*seenAE, "br") {
|
||||
t.Fatalf("书源 header 里的 Accept-Encoding 未被清理(服务端可能回 brotli): %q", *seenAE)
|
||||
}
|
||||
}
|
||||
File diff suppressed because it is too large
Load Diff
@@ -0,0 +1,957 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"archive/zip"
|
||||
"bytes"
|
||||
"net/url"
|
||||
"os"
|
||||
"path/filepath"
|
||||
"strings"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// 本文件:本地书籍(TXT / EPUB)导入与阅读的回归测试。
|
||||
|
||||
// newLocalBookService 复用登录测试的建库逻辑,另给一个临时 DataDir,
|
||||
// 避免本地书籍文件写到仓库目录里。
|
||||
func newLocalBookService(t *testing.T) *ReaderService {
|
||||
t.Helper()
|
||||
svc, _ := newLoginTestService(t)
|
||||
svc.cfg.App.DataDir = t.TempDir()
|
||||
return svc
|
||||
}
|
||||
|
||||
const sampleTXT = `书名:测试小说
|
||||
作者:某某
|
||||
|
||||
第1章 开端
|
||||
第一章的正文内容。
|
||||
这里还有一行。
|
||||
|
||||
第2章 发展
|
||||
第二章的正文内容。
|
||||
|
||||
第3章 结局
|
||||
第三章的正文内容。
|
||||
`
|
||||
|
||||
func TestSplitTXTChapters(t *testing.T) {
|
||||
chapters := splitTXTChapters(sampleTXT)
|
||||
if len(chapters) != 3 {
|
||||
t.Fatalf("章节数 = %d,期望 3:%+v", len(chapters), chapters)
|
||||
}
|
||||
wantTitles := []string{"第1章 开端", "第2章 发展", "第3章 结局"}
|
||||
for i, c := range chapters {
|
||||
if c.Title != wantTitles[i] {
|
||||
t.Errorf("第 %d 章标题 = %q,期望 %q", i+1, c.Title, wantTitles[i])
|
||||
}
|
||||
}
|
||||
// 第一章要从文件头开始,把书名/作者并进去
|
||||
if chapters[0].Start != 0 {
|
||||
t.Errorf("第一章起点 = %d,期望 0", chapters[0].Start)
|
||||
}
|
||||
if chapters[len(chapters)-1].End != len(sampleTXT) {
|
||||
t.Errorf("末章终点 = %d,期望 %d", chapters[len(chapters)-1].End, len(sampleTXT))
|
||||
}
|
||||
// 区间必须首尾相接且不重叠
|
||||
for i := 1; i < len(chapters); i++ {
|
||||
if chapters[i].Start != chapters[i-1].End {
|
||||
t.Errorf("第 %d/%d 章区间不连续:%d vs %d", i, i+1, chapters[i-1].End, chapters[i].Start)
|
||||
}
|
||||
}
|
||||
// 每章正文要能切出来且包含该章内容
|
||||
body := sampleTXT[chapters[1].Start:chapters[1].End]
|
||||
if !strings.Contains(body, "第二章的正文内容") {
|
||||
t.Errorf("第二章区间内容不对: %q", body)
|
||||
}
|
||||
if strings.Contains(body, "第三章") {
|
||||
t.Errorf("第二章区间串到了第三章: %q", body)
|
||||
}
|
||||
}
|
||||
|
||||
func TestSplitTXTChaptersFallback(t *testing.T) {
|
||||
chapters := splitTXTChapters("没有任何章节标记的一段文字。\n第二行。")
|
||||
if len(chapters) != 1 || chapters[0].Title != "全文" {
|
||||
t.Fatalf("无章标记应整本当一章: %+v", chapters)
|
||||
}
|
||||
if chapters[0].Start != 0 || chapters[0].End != len("没有任何章节标记的一段文字。\n第二行。") {
|
||||
t.Fatalf("整本区间不对: %+v", chapters[0])
|
||||
}
|
||||
}
|
||||
|
||||
func TestDecodeTextFileGBK(t *testing.T) {
|
||||
// "第1章 开端" 的 GBK 编码
|
||||
gbk := []byte{0xB5, 0xDA, '1', 0xD5, 0xC2, ' ', 0xBF, 0xAA, 0xB6, 0xCB}
|
||||
got, charset := decodeTextFile(gbk)
|
||||
if charset != "gbk" {
|
||||
t.Fatalf("字符集 = %q,期望 gbk", charset)
|
||||
}
|
||||
if got != "第1章 开端" {
|
||||
t.Fatalf("解码结果 = %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
func TestImportLocalBookAndRead(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
ctx := t.Context()
|
||||
|
||||
book, err := svc.ImportLocalBook(ctx, "u1", "测试小说.txt", []byte(sampleTXT))
|
||||
if err != nil {
|
||||
t.Fatalf("导入失败: %v", err)
|
||||
}
|
||||
if !book.IsLocal || book.LocalPath == "" {
|
||||
t.Fatalf("导入后应标记为本地书籍: %+v", book)
|
||||
}
|
||||
if book.Name != "测试小说" {
|
||||
t.Errorf("书名 = %q", book.Name)
|
||||
}
|
||||
if book.TotalChapterNum != 3 {
|
||||
t.Errorf("总章数 = %d,期望 3", book.TotalChapterNum)
|
||||
}
|
||||
|
||||
chapters, err := svc.ListChapters(ctx, book.ID)
|
||||
if err != nil || len(chapters) != 3 {
|
||||
t.Fatalf("目录 = %d 章, err=%v", len(chapters), err)
|
||||
}
|
||||
|
||||
got, err := svc.LocalChapterContent(ctx, "u1", book.ID, 1)
|
||||
if err != nil {
|
||||
t.Fatalf("读第二章失败: %v", err)
|
||||
}
|
||||
if got.Type != "text" {
|
||||
t.Errorf("type = %q", got.Type)
|
||||
}
|
||||
if !strings.Contains(got.Content, "第二章的正文内容") {
|
||||
t.Errorf("第二章正文 = %q", got.Content)
|
||||
}
|
||||
if strings.Contains(got.Content, "第三章") || strings.Contains(got.Content, "第一章的正文") {
|
||||
t.Errorf("第二章正文串章了: %q", got.Content)
|
||||
}
|
||||
// 书名等前言应落在第一章
|
||||
first, err := svc.LocalChapterContent(ctx, "u1", book.ID, 0)
|
||||
if err != nil {
|
||||
t.Fatalf("读第一章失败: %v", err)
|
||||
}
|
||||
if !strings.Contains(first.Content, "第一章的正文内容") {
|
||||
t.Errorf("第一章正文 = %q", first.Content)
|
||||
}
|
||||
|
||||
// 越界与越权
|
||||
if _, err := svc.LocalChapterContent(ctx, "u1", book.ID, 9); err == nil {
|
||||
t.Error("越界章节应报错")
|
||||
}
|
||||
if _, err := svc.LocalChapterContent(ctx, "other", book.ID, 0); err == nil {
|
||||
t.Error("他人书架应拒绝")
|
||||
}
|
||||
|
||||
// 书架列表要带 is_local
|
||||
books, err := svc.ListBooks(ctx, "u1")
|
||||
if err != nil || len(books) != 1 || !books[0].IsLocal {
|
||||
t.Fatalf("书架列表未标记本地书籍: %+v, err=%v", books, err)
|
||||
}
|
||||
|
||||
// 重复导入同名文件:覆盖更新且保留进度
|
||||
if err := svc.SaveProgress(ctx, "u1", book.ID, 1, 0, "第2章 发展"); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
again, err := svc.ImportLocalBook(ctx, "u1", "测试小说.txt", []byte(sampleTXT))
|
||||
if err != nil {
|
||||
t.Fatalf("重复导入失败: %v", err)
|
||||
}
|
||||
if again.ID != book.ID {
|
||||
t.Errorf("同名导入应覆盖同一本书,得到新 ID %s", again.ID)
|
||||
}
|
||||
if again.DurChapterIndex != 1 {
|
||||
t.Errorf("章节数不变时应保留进度,得到 %d", again.DurChapterIndex)
|
||||
}
|
||||
|
||||
// 移出书架要删掉落盘文件
|
||||
path, err := svc.localFilePath(again)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if err := svc.RemoveBook(ctx, "u1", again.ID); err != nil {
|
||||
t.Fatalf("移出书架失败: %v", err)
|
||||
}
|
||||
if _, err := os.ReadFile(path); err == nil {
|
||||
t.Error("移出书架后本地文件应被删除")
|
||||
}
|
||||
}
|
||||
|
||||
// epubFixture 造一个最小可用的 EPUB(NCX 目录 + 两个 XHTML 章节)。
|
||||
func epubFixture(t *testing.T) []byte {
|
||||
t.Helper()
|
||||
var buf bytes.Buffer
|
||||
zw := zip.NewWriter(&buf)
|
||||
add := func(name, body string) {
|
||||
w, err := zw.Create(name)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if _, err := w.Write([]byte(body)); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
}
|
||||
add("mimetype", "application/epub+zip")
|
||||
add("META-INF/container.xml", `<?xml version="1.0"?>
|
||||
<container version="1.0" xmlns="urn:oasis:names:tc:opendocument:xmlns:container">
|
||||
<rootfiles><rootfile full-path="OEBPS/content.opf" media-type="application/oebps-package+xml"/></rootfiles>
|
||||
</container>`)
|
||||
add("OEBPS/content.opf", `<?xml version="1.0"?>
|
||||
<package xmlns="http://www.idpf.org/2007/opf" version="2.0">
|
||||
<manifest>
|
||||
<item id="ncx" href="toc.ncx" media-type="application/x-dtbncx+xml"/>
|
||||
<item id="c1" href="ch1.xhtml" media-type="application/xhtml+xml"/>
|
||||
<item id="c2" href="ch2.xhtml" media-type="application/xhtml+xml"/>
|
||||
</manifest>
|
||||
<spine toc="ncx">
|
||||
<itemref idref="c1"/><itemref idref="c2"/>
|
||||
</spine>
|
||||
</package>`)
|
||||
add("OEBPS/toc.ncx", `<?xml version="1.0"?>
|
||||
<ncx xmlns="http://www.daisy.org/z3986/2005/ncx/" version="2005-1">
|
||||
<navMap>
|
||||
<navPoint id="n1"><navLabel><text>第一章 起风</text></navLabel><content src="ch1.xhtml"/></navPoint>
|
||||
<navPoint id="n2"><navLabel><text>第二章 落雨</text></navLabel><content src="ch2.xhtml"/></navPoint>
|
||||
</navMap>
|
||||
</ncx>`)
|
||||
add("OEBPS/ch1.xhtml", `<html xmlns="http://www.w3.org/1999/xhtml"><head><title>t</title></head>
|
||||
<body><h1>第一章 起风</h1><p>第一段 & 实体。</p><p>第二段。</p><script>var x=1;</script></body></html>`)
|
||||
add("OEBPS/ch2.xhtml", `<html xmlns="http://www.w3.org/1999/xhtml"><body><p>第二章正文。</p><br/><p>又一段。</p></body></html>`)
|
||||
if err := zw.Close(); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
return buf.Bytes()
|
||||
}
|
||||
|
||||
func TestImportEPUBAndRead(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
ctx := t.Context()
|
||||
|
||||
book, err := svc.ImportLocalBook(ctx, "u1", "测试电子书.epub", epubFixture(t))
|
||||
if err != nil {
|
||||
t.Fatalf("导入 EPUB 失败: %v", err)
|
||||
}
|
||||
if book.TotalChapterNum != 2 {
|
||||
t.Fatalf("章数 = %d,期望 2", book.TotalChapterNum)
|
||||
}
|
||||
chapters, err := svc.ListChapters(ctx, book.ID)
|
||||
if err != nil || len(chapters) != 2 {
|
||||
t.Fatalf("目录 = %+v, err=%v", chapters, err)
|
||||
}
|
||||
if chapters[0].Title != "第一章 起风" || chapters[1].Title != "第二章 落雨" {
|
||||
t.Fatalf("章节标题应取自 NCX: %q / %q", chapters[0].Title, chapters[1].Title)
|
||||
}
|
||||
// 标题不要留在正文里
|
||||
first, err := svc.LocalChapterContent(ctx, "u1", book.ID, 0)
|
||||
if err != nil {
|
||||
t.Fatalf("读第一章失败: %v", err)
|
||||
}
|
||||
if !strings.Contains(first.Content, "第一段 & 实体。") {
|
||||
t.Errorf("实体未解码: %q", first.Content)
|
||||
}
|
||||
if !strings.Contains(first.Content, "第一段 & 实体。\n第二段。") {
|
||||
t.Errorf("段落换行不对: %q", first.Content)
|
||||
}
|
||||
if strings.Contains(first.Content, "var x=1") {
|
||||
t.Errorf("script 应被剔除: %q", first.Content)
|
||||
}
|
||||
second, err := svc.LocalChapterContent(ctx, "u1", book.ID, 1)
|
||||
if err != nil || !strings.Contains(second.Content, "第二章正文。") {
|
||||
t.Fatalf("读第二章失败: %v / %q", err, second.Content)
|
||||
}
|
||||
}
|
||||
|
||||
// TestImportEPUBWithImage 端到端:导入带图片的 EPUB,章节正文里图片应变成签名
|
||||
// 地址,且该地址能取回原始图片字节。
|
||||
func TestImportEPUBWithImage(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
ctx := t.Context()
|
||||
|
||||
var buf bytes.Buffer
|
||||
zw := zip.NewWriter(&buf)
|
||||
add := func(name, body string) {
|
||||
w, err := zw.Create(name)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if _, err := w.Write([]byte(body)); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
}
|
||||
imgBytes := []byte("\xff\xd8\xff\xe0fakejpeg")
|
||||
add("mimetype", "application/epub+zip")
|
||||
add("META-INF/container.xml", `<?xml version="1.0"?><container version="1.0"
|
||||
xmlns="urn:oasis:names:tc:opendocument:xmlns:container"><rootfiles>
|
||||
<rootfile full-path="OEBPS/content.opf" media-type="application/oebps-package+xml"/></rootfiles></container>`)
|
||||
add("OEBPS/content.opf", `<?xml version="1.0"?><package xmlns="http://www.idpf.org/2007/opf" version="3.0">
|
||||
<manifest>
|
||||
<item id="nav" href="nav.xhtml" media-type="application/xhtml+xml" properties="nav"/>
|
||||
<item id="c1" href="ch1.xhtml" media-type="application/xhtml+xml"/>
|
||||
<item id="i1" href="images/pic.jpg" media-type="image/jpeg"/>
|
||||
</manifest>
|
||||
<spine><itemref idref="c1"/></spine></package>`)
|
||||
add("OEBPS/nav.xhtml", `<html xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops">
|
||||
<body><nav epub:type="toc"><ol><li><a href="ch1.xhtml">插图章</a></li></ol></nav></body></html>`)
|
||||
add("OEBPS/ch1.xhtml", `<html xmlns="http://www.w3.org/1999/xhtml"><body><h1>插图章</h1>
|
||||
<p>图片前。</p><img src="images/pic.jpg" alt=""/><p>图片后。</p></body></html>`)
|
||||
w, err := zw.Create("OEBPS/images/pic.jpg")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if _, err := w.Write(imgBytes); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if err := zw.Close(); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
book, err := svc.ImportLocalBook(ctx, "u1", "带图.epub", buf.Bytes())
|
||||
if err != nil {
|
||||
t.Fatalf("导入失败: %v", err)
|
||||
}
|
||||
chapters, err := svc.ListChapters(ctx, book.ID)
|
||||
if err != nil || len(chapters) != 1 || chapters[0].Title != "插图章" {
|
||||
t.Fatalf("目录 = %+v, err=%v", chapters, err)
|
||||
}
|
||||
content, err := svc.LocalChapterContent(ctx, "u1", book.ID, 0)
|
||||
if err != nil {
|
||||
t.Fatalf("读正文失败: %v", err)
|
||||
}
|
||||
if !strings.Contains(content.Content, "图片前。") || !strings.Contains(content.Content, "图片后。") {
|
||||
t.Fatalf("正文缺内容: %q", content.Content)
|
||||
}
|
||||
var asset string
|
||||
for _, line := range strings.Split(content.Content, "\n") {
|
||||
if strings.HasPrefix(line, imgMarkerPrefix) {
|
||||
asset = strings.TrimPrefix(line, imgMarkerPrefix)
|
||||
}
|
||||
}
|
||||
if !strings.HasPrefix(asset, "/api/reader/local/asset?") {
|
||||
t.Fatalf("图片未换成签名地址: %q(正文 %q)", asset, content.Content)
|
||||
}
|
||||
u, err := url.Parse(asset)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
entry, err := svc.VerifyLocalAssetURL(u.Query().Get("b"), u.Query().Get("p"), u.Query().Get("s"))
|
||||
if err != nil {
|
||||
t.Fatalf("资源签名校验失败: %v", err)
|
||||
}
|
||||
data, ct, err := svc.ReadLocalAsset(ctx, book.ID, entry)
|
||||
if err != nil {
|
||||
t.Fatalf("读图片失败: %v", err)
|
||||
}
|
||||
if !bytes.Equal(data, imgBytes) {
|
||||
t.Fatalf("图片字节不一致: %q", data)
|
||||
}
|
||||
if ct != "image/jpeg" {
|
||||
t.Fatalf("Content-Type = %q", ct)
|
||||
}
|
||||
}
|
||||
|
||||
func TestHTMLToText(t *testing.T) {
|
||||
got := htmlToText("<div>甲</div>\n<p>乙<br/>丙</p><style>p{}</style>")
|
||||
if got != "甲\n乙\n丙" {
|
||||
t.Fatalf("htmlToText = %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
// TestParseEPUBTitlesHandlesNestedNavPoint 网上不少 EPUB(Epubor 导出)navPoint
|
||||
// 漏了闭合标签,标题会整棵挂进上一个 navPoint。按固定层级解会丢一大半标题,
|
||||
// 这里要求深层节点也能收全。
|
||||
func TestParseEPUBTitlesHandlesNestedNavPoint(t *testing.T) {
|
||||
raw := []byte(`<?xml version="1.0" encoding="UTF-8"?>
|
||||
<ncx xmlns="http://www.daisy.org/z3986/2005/ncx/" version="2005-1">
|
||||
<docTitle><text>UnKnown</text></docTitle>
|
||||
<navMap>
|
||||
<navPoint id="id1"><navLabel><text>目录</text></navLabel><content src="text00000.html"/></navPoint>
|
||||
<navPoint id="id2"><navLabel><text>小狗钱钱1</text></navLabel><content src="text00002.html"/>
|
||||
<navPoint id="id3"><navLabel><text>童话与理财</text></navLabel><content src="text00005.html"/></navPoint>
|
||||
<navPoint id="id4"><navLabel><text>前言</text></navLabel><content src="text00006.html"/></navPoint>
|
||||
</navPoint>
|
||||
</navMap>
|
||||
</ncx>`)
|
||||
titles := parseEPUBTitles(raw)
|
||||
want := map[string]string{
|
||||
"text00000.html": "目录",
|
||||
"text00002.html": "小狗钱钱1",
|
||||
"text00005.html": "童话与理财",
|
||||
"text00006.html": "前言",
|
||||
}
|
||||
for href, title := range want {
|
||||
if titles[href] != title {
|
||||
t.Errorf("%s 标题 = %q,期望 %q(全部:%v)", href, titles[href], title, titles)
|
||||
}
|
||||
}
|
||||
if titles["UnKnown"] != "" {
|
||||
t.Errorf("docTitle 不该被当成章节标题:%v", titles)
|
||||
}
|
||||
}
|
||||
|
||||
// TestParseEPUBTitlesNav 兼容 EPUB3 的 nav 目录(含带锚点与嵌套 ol 的情况)。
|
||||
func TestParseEPUBTitlesNav(t *testing.T) {
|
||||
raw := []byte(`<?xml version="1.0" encoding="UTF-8"?>
|
||||
<html xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops">
|
||||
<body>
|
||||
<nav epub:type="toc"><h1>目录</h1><ol>
|
||||
<li><a href="ch1.xhtml">第一章 起风</a></li>
|
||||
<li><a href="ch2.xhtml#s1">第二章 落雨</a>
|
||||
<ol><li><a href="ch3.xhtml">第三章 天晴</a></li></ol>
|
||||
</li>
|
||||
</ol></nav>
|
||||
</body></html>`)
|
||||
titles := parseEPUBTitles(raw)
|
||||
for href, title := range map[string]string{
|
||||
"ch1.xhtml": "第一章 起风",
|
||||
"ch2.xhtml": "第二章 落雨",
|
||||
"ch3.xhtml": "第三章 天晴",
|
||||
} {
|
||||
if titles[href] != title {
|
||||
t.Errorf("%s 标题 = %q,期望 %q(全部:%v)", href, titles[href], title, titles)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestEpubHTMLToTextWithImages 图片要变成 [img]条目 标记行,且相对路径按正文所在
|
||||
// 目录解析;标记行不能带缩进(前端据此渲染图片)。
|
||||
func TestEpubHTMLToTextWithImages(t *testing.T) {
|
||||
src := `<html><body><h1>第一章</h1>
|
||||
<p>正文一。</p>
|
||||
<img src="Image00024.jpg" style="width:100%;height:100%;" />
|
||||
<p>正文二<img src='sub/pic.png'/>尾巴。</p>
|
||||
<p><img src="http://cdn.example.com/a.jpg"/></p>
|
||||
</body></html>`
|
||||
got := epubHTMLToText(src, "OEBPS")
|
||||
want := "第一章\n正文一。\n" + imgMarkerPrefix + filepath.Join("OEBPS", "Image00024.jpg") +
|
||||
"\n正文二\n" + imgMarkerPrefix + filepath.Join("OEBPS", "sub", "pic.png") + "\n尾巴。\n" +
|
||||
imgMarkerPrefix + "http://cdn.example.com/a.jpg"
|
||||
if got != want {
|
||||
t.Fatalf("epubHTMLToText =\n%q\n期望\n%q", got, want)
|
||||
}
|
||||
}
|
||||
|
||||
// TestRewriteLocalImages 标记里的条目要换成签名地址,且重复处理不会二次改写。
|
||||
func TestRewriteLocalImages(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
text := imgMarkerPrefix + filepath.Join("OEBPS", "a.jpg") + "\n正文\n" + imgMarkerPrefix + "http://x/b.png"
|
||||
out := svc.rewriteLocalImages("book-1", text)
|
||||
lines := strings.Split(out, "\n")
|
||||
if !strings.HasPrefix(lines[0], imgMarkerPrefix+"/api/reader/local/asset?b=book-1&p=") {
|
||||
t.Fatalf("相对路径未改写:%q", lines[0])
|
||||
}
|
||||
if !strings.HasSuffix(lines[2], "http://x/b.png") {
|
||||
t.Fatalf("外链图片不该改写:%q", lines[2])
|
||||
}
|
||||
// 再跑一次应保持不变
|
||||
if again := svc.rewriteLocalImages("book-1", out); again != out {
|
||||
t.Fatalf("重复改写改变了结果:%q", again)
|
||||
}
|
||||
// 签名可校验,改一个字符就不认
|
||||
entry := filepath.Join("OEBPS", "a.jpg")
|
||||
u := svc.LocalAssetURL("book-1", entry)
|
||||
parsed, err := url.Parse(u)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
got, err := svc.VerifyLocalAssetURL("book-1", parsed.Query().Get("p"), parsed.Query().Get("s"))
|
||||
if err != nil || got != entry {
|
||||
t.Fatalf("签名校验失败: %v / %q", err, got)
|
||||
}
|
||||
if _, err := svc.VerifyLocalAssetURL("book-1", parsed.Query().Get("p"), "deadbeef"); err == nil {
|
||||
t.Fatal("错误签名应校验失败")
|
||||
}
|
||||
if _, err := svc.VerifyLocalAssetURL("book-2", parsed.Query().Get("p"), parsed.Query().Get("s")); err == nil {
|
||||
t.Fatal("换一本书后签名应失效")
|
||||
}
|
||||
}
|
||||
|
||||
// epubWithCoverFixture 造一个带封面的最小 EPUB。opfCover 决定封面怎么写:
|
||||
// EPUB3 用 properties="cover-image",EPUB2 用 <meta name="cover">;都为 false
|
||||
// 时封面只能靠文件名兜底。
|
||||
func epubWithCoverFixture(t *testing.T, epub3, epub2 bool) ([]byte, []byte) {
|
||||
t.Helper()
|
||||
var buf bytes.Buffer
|
||||
zw := zip.NewWriter(&buf)
|
||||
add := func(name, body string) {
|
||||
w, err := zw.Create(name)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if _, err := w.Write([]byte(body)); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
}
|
||||
coverBytes := []byte("\xff\xd8\xff\xe0fakecoverjpeg")
|
||||
add("mimetype", "application/epub+zip")
|
||||
add("META-INF/container.xml", `<?xml version="1.0"?>
|
||||
<container version="1.0" xmlns="urn:oasis:names:tc:opendocument:xmlns:container">
|
||||
<rootfiles><rootfile full-path="OEBPS/content.opf" media-type="application/oebps-package+xml"/></rootfiles>
|
||||
</container>`)
|
||||
meta := ""
|
||||
if epub2 {
|
||||
meta = `<metadata><meta name="cover" content="cover-img"/></metadata>`
|
||||
}
|
||||
props := ""
|
||||
if epub3 {
|
||||
props = ` properties="cover-image"`
|
||||
}
|
||||
add("OEBPS/content.opf", `<?xml version="1.0"?>
|
||||
<package xmlns="http://www.idpf.org/2007/opf" version="3.0">`+meta+`
|
||||
<manifest>
|
||||
<item id="c1" href="ch1.xhtml" media-type="application/xhtml+xml"/>
|
||||
<item id="cover-img" href="images/cover.jpg" media-type="image/jpeg"`+props+`/>
|
||||
</manifest>
|
||||
<spine><itemref idref="c1"/></spine>
|
||||
</package>`)
|
||||
add("OEBPS/ch1.xhtml", `<html xmlns="http://www.w3.org/1999/xhtml"><body><h1>正文</h1><p>内容。</p></body></html>`)
|
||||
w, err := zw.Create("OEBPS/images/cover.jpg")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if _, err := w.Write(coverBytes); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if err := zw.Close(); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
return buf.Bytes(), coverBytes
|
||||
}
|
||||
|
||||
// TestImportEPUBCover 封面要能从 OPF 声明(EPUB3 / EPUB2)与文件名兜底三条路
|
||||
// 解析出来,并转成可访问的签名资源地址。
|
||||
func TestImportEPUBCover(t *testing.T) {
|
||||
ctx := t.Context()
|
||||
cases := []struct {
|
||||
name string
|
||||
epub3, epub2 bool
|
||||
}{
|
||||
{"EPUB3 cover-image", true, false},
|
||||
{"EPUB2 meta cover", false, true},
|
||||
{"文件名兜底", false, false},
|
||||
}
|
||||
for _, tc := range cases {
|
||||
t.Run(tc.name, func(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
raw, coverBytes := epubWithCoverFixture(t, tc.epub3, tc.epub2)
|
||||
book, err := svc.ImportLocalBook(ctx, "u1", "有封面.epub", raw)
|
||||
if err != nil {
|
||||
t.Fatalf("导入失败: %v", err)
|
||||
}
|
||||
if !strings.HasPrefix(book.CoverURL, "/api/reader/local/asset?") {
|
||||
t.Fatalf("封面未生成签名地址: %q", book.CoverURL)
|
||||
}
|
||||
u, err := url.Parse(book.CoverURL)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
entry, err := svc.VerifyLocalAssetURL(u.Query().Get("b"), u.Query().Get("p"), u.Query().Get("s"))
|
||||
if err != nil {
|
||||
t.Fatalf("封面签名校验失败: %v", err)
|
||||
}
|
||||
data, ct, err := svc.ReadLocalAsset(ctx, book.ID, entry)
|
||||
if err != nil {
|
||||
t.Fatalf("读封面失败: %v", err)
|
||||
}
|
||||
if !bytes.Equal(data, coverBytes) {
|
||||
t.Fatalf("封面字节不一致: %q", data)
|
||||
}
|
||||
if ct != "image/jpeg" {
|
||||
t.Fatalf("封面 MIME = %q,期望 image/jpeg", ct)
|
||||
}
|
||||
|
||||
// 落库后书架列表/详情也要带上封面,重开服务不丢
|
||||
books, err := svc.ListBooks(ctx, "u1")
|
||||
if err != nil || len(books) != 1 || books[0].CoverURL != book.CoverURL {
|
||||
t.Fatalf("书架封面 = %+v, err=%v", books, err)
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
// TestBackfillLocalCoverOnListBooks 早期导入的本地 EPUB 没存封面,
|
||||
// 加载书架时应该自动补上并落库,不必让用户重新导入。
|
||||
func TestBackfillLocalCoverOnListBooks(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
ctx := t.Context()
|
||||
|
||||
raw, coverBytes := epubWithCoverFixture(t, true, false)
|
||||
book, err := svc.ImportLocalBook(ctx, "u1", "老书.epub", raw)
|
||||
if err != nil {
|
||||
t.Fatalf("导入失败: %v", err)
|
||||
}
|
||||
// 模拟修复前导入的旧数据:库里没有封面
|
||||
book.CoverURL = ""
|
||||
if err := svc.repo.UpdateBook(ctx, book); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
books, err := svc.ListBooks(ctx, "u1")
|
||||
if err != nil || len(books) != 1 {
|
||||
t.Fatalf("书架 = %+v, err=%v", books, err)
|
||||
}
|
||||
restored := books[0].CoverURL
|
||||
if !strings.HasPrefix(restored, "/api/reader/local/asset?") {
|
||||
t.Fatalf("书架加载未回填封面: %q", restored)
|
||||
}
|
||||
|
||||
// 已经写回数据库,下次不用再解析
|
||||
stored, err := svc.repo.GetBook(ctx, book.ID)
|
||||
if err != nil || stored.CoverURL != restored {
|
||||
t.Fatalf("封面未落库: %+v, err=%v", stored, err)
|
||||
}
|
||||
|
||||
// 回填出来的地址必须真能取到封面图
|
||||
u, err := url.Parse(restored)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
entry, err := svc.VerifyLocalAssetURL(u.Query().Get("b"), u.Query().Get("p"), u.Query().Get("s"))
|
||||
if err != nil {
|
||||
t.Fatalf("回填封面签名校验失败: %v", err)
|
||||
}
|
||||
data, _, err := svc.ReadLocalAsset(ctx, book.ID, entry)
|
||||
if err != nil || !bytes.Equal(data, coverBytes) {
|
||||
t.Fatalf("回填封面取回失败: %v / %q", err, data)
|
||||
}
|
||||
}
|
||||
|
||||
// TestBackfillLocalCoverSkipsTXTAndCoveredBooks TXT 没有内嵌图片,已有封面的书
|
||||
// 也不该被重复处理。
|
||||
func TestBackfillLocalCoverSkipsTXTAndCoveredBooks(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
ctx := t.Context()
|
||||
|
||||
txt, err := svc.ImportLocalBook(ctx, "u1", "小说.txt", []byte(sampleTXT))
|
||||
if err != nil {
|
||||
t.Fatalf("导入 TXT 失败: %v", err)
|
||||
}
|
||||
if svc.BackfillLocalCover(ctx, txt) {
|
||||
t.Error("TXT 不该被回填封面")
|
||||
}
|
||||
|
||||
raw, _ := epubWithCoverFixture(t, true, false)
|
||||
epub, err := svc.ImportLocalBook(ctx, "u1", "已带封面.epub", raw)
|
||||
if err != nil {
|
||||
t.Fatalf("导入 EPUB 失败: %v", err)
|
||||
}
|
||||
if svc.BackfillLocalCover(ctx, epub) {
|
||||
t.Error("已有封面的书不该被回填")
|
||||
}
|
||||
}
|
||||
|
||||
// TestImportLocalBookFromPath 服务器选书:原地引用不复制,章节可读,
|
||||
// 移出书架不删除源文件。
|
||||
func TestImportLocalBookFromPath(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
ctx := t.Context()
|
||||
|
||||
srcDir := t.TempDir()
|
||||
srcPath := filepath.Join(srcDir, "服务器上的书.txt")
|
||||
if err := os.WriteFile(srcPath, []byte(sampleTXT), 0o640); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
book, err := svc.ImportLocalBookFromPath(ctx, "u1", srcPath)
|
||||
if err != nil {
|
||||
t.Fatalf("从服务器路径导入失败: %v", err)
|
||||
}
|
||||
if !book.LocalExternal {
|
||||
t.Fatal("应标记为原地引用")
|
||||
}
|
||||
if book.LocalPath != srcPath {
|
||||
t.Fatalf("LocalPath = %q,期望源文件路径 %q", book.LocalPath, srcPath)
|
||||
}
|
||||
if book.TotalChapterNum != 3 {
|
||||
t.Fatalf("章节数 = %d,期望 3", book.TotalChapterNum)
|
||||
}
|
||||
if book.Name != "服务器上的书" {
|
||||
t.Fatalf("书名 = %q", book.Name)
|
||||
}
|
||||
|
||||
content, err := svc.LocalChapterContent(ctx, "u1", book.ID, 0)
|
||||
if err != nil {
|
||||
t.Fatalf("读正文失败: %v", err)
|
||||
}
|
||||
if !strings.Contains(content.Content, "第一章的正文内容。") {
|
||||
t.Fatalf("正文不对: %q", content.Content)
|
||||
}
|
||||
|
||||
// 不应把文件复制进 data/reader/local
|
||||
localDir, err := svc.localBooksDir()
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
entries, err := os.ReadDir(localDir)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if len(entries) != 0 {
|
||||
t.Fatalf("原地引用不该产生托管副本,目录里有 %d 个文件", len(entries))
|
||||
}
|
||||
|
||||
// 移出书架只解除引用,源文件必须还在
|
||||
if err := svc.RemoveBook(ctx, "u1", book.ID); err != nil {
|
||||
t.Fatalf("移出书架失败: %v", err)
|
||||
}
|
||||
if _, err := os.Stat(srcPath); err != nil {
|
||||
t.Fatalf("原地引用的源文件被删除了: %v", err)
|
||||
}
|
||||
}
|
||||
|
||||
// TestImportLocalBookFromPathRejectsUnsupported 只接受 TXT / EPUB,目录与超限文件要报错。
|
||||
func TestImportLocalBookFromPathRejectsUnsupported(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
ctx := t.Context()
|
||||
|
||||
dir := t.TempDir()
|
||||
pdf := filepath.Join(dir, "book.pdf")
|
||||
if err := os.WriteFile(pdf, []byte("%PDF-1.4"), 0o640); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if _, err := svc.ImportLocalBookFromPath(ctx, "u1", pdf); err == nil {
|
||||
t.Fatal("PDF 不该被接受")
|
||||
}
|
||||
if _, err := svc.ImportLocalBookFromPath(ctx, "u1", dir); err == nil {
|
||||
t.Fatal("目录不该被接受")
|
||||
}
|
||||
if _, err := svc.ImportLocalBookFromPath(ctx, "u1", filepath.Join(dir, "nope.txt")); err == nil {
|
||||
t.Fatal("不存在的文件应报错")
|
||||
}
|
||||
}
|
||||
|
||||
// TestImportLocalAudioDir 目录导入有声书:音频与 .strm 成为章节,按相对路径排序,
|
||||
// 其它文件忽略;.strm 指向远端走媒体代理,指向本地文件走本地音频端点。
|
||||
func TestImportLocalAudioDir(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
ctx := t.Context()
|
||||
|
||||
base := t.TempDir()
|
||||
dir := filepath.Join(base, "有声书")
|
||||
shared := filepath.Join(base, "共享")
|
||||
for _, d := range []string{filepath.Join(dir, "00-cd1"), shared} {
|
||||
if err := os.MkdirAll(d, 0o750); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
}
|
||||
write := func(path, body string) {
|
||||
if err := os.WriteFile(path, []byte(body), 0o640); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
}
|
||||
// 相对路径排序:00-cd1/ 子目录在前,其后是同级文件
|
||||
write(filepath.Join(dir, "00-cd1", "01-正文.mp3"), "mp3-bytes-2")
|
||||
write(filepath.Join(dir, "01-开场.mp3"), "mp3-bytes")
|
||||
write(filepath.Join(dir, "02-远端.strm"), "https://cdn.example.com/a.mp3\n")
|
||||
localTarget := filepath.Join(shared, "尾巴.flac")
|
||||
write(localTarget, "flac-bytes")
|
||||
write(filepath.Join(dir, "03-本地.strm"), localTarget+"\n")
|
||||
// 非音频文件应被忽略
|
||||
write(filepath.Join(dir, "cover.jpg"), "jpg")
|
||||
write(filepath.Join(dir, "readme.txt"), "txt")
|
||||
|
||||
book, err := svc.ImportLocalAudioDir(ctx, "u1", dir)
|
||||
if err != nil {
|
||||
t.Fatalf("导入有声书目录失败: %v", err)
|
||||
}
|
||||
if book.Type != 1 {
|
||||
t.Fatalf("Type = %d,期望 1(音频)", book.Type)
|
||||
}
|
||||
if !book.LocalExternal || book.LocalPath != dir {
|
||||
t.Fatalf("应原地引用目录,得到 LocalPath=%q external=%v", book.LocalPath, book.LocalExternal)
|
||||
}
|
||||
if book.Name != "有声书" {
|
||||
t.Fatalf("书名应取目录名,得到 %q", book.Name)
|
||||
}
|
||||
if book.TotalChapterNum != 4 {
|
||||
t.Fatalf("章节数 = %d,期望 4(忽略 cover.jpg / readme.txt)", book.TotalChapterNum)
|
||||
}
|
||||
|
||||
chapters, err := svc.ListChapters(ctx, book.ID)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
wantTitles := []string{"01-正文", "01-开场", "02-远端", "03-本地"}
|
||||
for i, want := range wantTitles {
|
||||
if chapters[i].Title != want {
|
||||
t.Fatalf("第 %d 章标题 = %q,期望 %q(完整目录 %+v)", i, chapters[i].Title, want, chapters)
|
||||
}
|
||||
}
|
||||
|
||||
// 本地音频文件 → 本地音频流端点
|
||||
first, err := svc.LocalChapterContent(ctx, "u1", book.ID, 1)
|
||||
if err != nil {
|
||||
t.Fatalf("读第 2 章失败: %v", err)
|
||||
}
|
||||
if first.Type != "audio" || len(first.Tracks) != 1 {
|
||||
t.Fatalf("第 2 章应为单轨音频: %+v", first)
|
||||
}
|
||||
if !strings.HasPrefix(first.Tracks[0], "/api/reader/local/audio?") {
|
||||
t.Fatalf("本地音频应走本地流端点: %q", first.Tracks[0])
|
||||
}
|
||||
|
||||
// .strm 指向远端 → 媒体代理
|
||||
remote, err := svc.LocalChapterContent(ctx, "u1", book.ID, 2)
|
||||
if err != nil {
|
||||
t.Fatalf("读第 3 章失败: %v", err)
|
||||
}
|
||||
if !strings.HasPrefix(remote.Tracks[0], "/api/reader/media?") {
|
||||
t.Fatalf("远端 .strm 应走媒体代理: %q", remote.Tracks[0])
|
||||
}
|
||||
|
||||
// .strm 指向本地文件 → 本地音频流端点
|
||||
local, err := svc.LocalChapterContent(ctx, "u1", book.ID, 3)
|
||||
if err != nil {
|
||||
t.Fatalf("读第 4 章失败: %v", err)
|
||||
}
|
||||
if !strings.HasPrefix(local.Tracks[0], "/api/reader/local/audio?") {
|
||||
t.Fatalf("本地 .strm 应走本地流端点: %q", local.Tracks[0])
|
||||
}
|
||||
u, err := url.Parse(local.Tracks[0])
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
resolved, err := svc.VerifyLocalAudioURL(u.Query().Get("b"), u.Query().Get("p"), u.Query().Get("s"))
|
||||
if err != nil {
|
||||
t.Fatalf("音频签名校验失败: %v", err)
|
||||
}
|
||||
if resolved != localTarget {
|
||||
t.Fatalf("解析出的音频路径 = %q,期望 %q", resolved, localTarget)
|
||||
}
|
||||
}
|
||||
|
||||
// TestImportLocalAudioDirNoAudio 目录里没有音频时应给出明确错误。
|
||||
func TestImportLocalAudioDirNoAudio(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
dir := t.TempDir()
|
||||
if err := os.WriteFile(filepath.Join(dir, "a.txt"), []byte("x"), 0o640); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if _, err := svc.ImportLocalAudioDir(t.Context(), "u1", dir); err == nil {
|
||||
t.Fatal("没有音频文件时应报错")
|
||||
}
|
||||
}
|
||||
|
||||
// TestLocalAudioURLSignature 音频流地址的签名可校验、改一处即失效。
|
||||
func TestLocalAudioURLSignature(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
path := filepath.Join("media", "有声书", "01.mp3")
|
||||
|
||||
raw := svc.LocalAudioURL("book-1", path)
|
||||
u, err := url.Parse(raw)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if !strings.HasPrefix(raw, "/api/reader/local/audio?b=book-1&p=") {
|
||||
t.Fatalf("音频地址格式不对: %q", raw)
|
||||
}
|
||||
got, err := svc.VerifyLocalAudioURL("book-1", u.Query().Get("p"), u.Query().Get("s"))
|
||||
if err != nil || got != path {
|
||||
t.Fatalf("签名校验失败: %v / %q", err, got)
|
||||
}
|
||||
if _, err := svc.VerifyLocalAudioURL("book-1", u.Query().Get("p"), "deadbeef"); err == nil {
|
||||
t.Fatal("错误签名应校验失败")
|
||||
}
|
||||
if _, err := svc.VerifyLocalAudioURL("book-2", u.Query().Get("p"), u.Query().Get("s")); err == nil {
|
||||
t.Fatal("换一本书后签名应失效")
|
||||
}
|
||||
}
|
||||
|
||||
// TestIsRemoteMediaURL 区分远端地址与本地路径。
|
||||
func TestIsRemoteMediaURL(t *testing.T) {
|
||||
remote := []string{
|
||||
"https://cdn.example.com/a.mp3",
|
||||
"http://x/y.flac",
|
||||
"webdav://host/a.mp3",
|
||||
"/api/strm/play/local/abc",
|
||||
}
|
||||
for _, raw := range remote {
|
||||
if !isRemoteMediaURL(raw) {
|
||||
t.Errorf("%q 应判为远端", raw)
|
||||
}
|
||||
}
|
||||
local := []string{
|
||||
"本地音频.flac",
|
||||
`D:\media\有声书\01.mp3`,
|
||||
"sub/02.mp3",
|
||||
}
|
||||
for _, raw := range local {
|
||||
if isRemoteMediaURL(raw) {
|
||||
t.Errorf("%q 应判为本地路径", raw)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestImportEPUBWithoutCover 没有封面图时不应硬凑,CoverURL 保持为空。
|
||||
func TestImportEPUBWithoutCover(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
book, err := svc.ImportLocalBook(t.Context(), "u1", "无封面.epub", epubFixture(t))
|
||||
if err != nil {
|
||||
t.Fatalf("导入失败: %v", err)
|
||||
}
|
||||
if book.CoverURL != "" {
|
||||
t.Fatalf("无封面书籍不该有 cover_url: %q", book.CoverURL)
|
||||
}
|
||||
}
|
||||
|
||||
// TestImportEPUBCoverFallbackOnBrokenDeclaration 声明的封面 id 指不到条目时,
|
||||
// 应退回“文件名像封面”的图片,而不是直接没有封面。
|
||||
func TestImportEPUBCoverFallbackOnBrokenDeclaration(t *testing.T) {
|
||||
svc := newLocalBookService(t)
|
||||
coverBytes := []byte("\xff\xd8\xff\xe0fallbackjpeg")
|
||||
var buf bytes.Buffer
|
||||
zw := zip.NewWriter(&buf)
|
||||
add := func(name, body string) {
|
||||
w, err := zw.Create(name)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if _, err := w.Write([]byte(body)); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
}
|
||||
add("mimetype", "application/epub+zip")
|
||||
add("META-INF/container.xml", `<?xml version="1.0"?>
|
||||
<container version="1.0" xmlns="urn:oasis:names:tc:opendocument:xmlns:container">
|
||||
<rootfiles><rootfile full-path="OEBPS/content.opf" media-type="application/oebps-package+xml"/></rootfiles>
|
||||
</container>`)
|
||||
// meta 指向一个 manifest 里不存在的 id
|
||||
add("OEBPS/content.opf", `<?xml version="1.0"?>
|
||||
<package xmlns="http://www.idpf.org/2007/opf" version="2.0">
|
||||
<metadata><meta name="cover" content="not-exist"/></metadata>
|
||||
<manifest>
|
||||
<item id="c1" href="ch1.xhtml" media-type="application/xhtml+xml"/>
|
||||
<item id="pic" href="images/封面.png" media-type="image/png"/>
|
||||
</manifest>
|
||||
<spine><itemref idref="c1"/></spine>
|
||||
</package>`)
|
||||
add("OEBPS/ch1.xhtml", `<html xmlns="http://www.w3.org/1999/xhtml"><body><p>内容。</p></body></html>`)
|
||||
// 声明失效,只能靠文件名(含“封面”)兜底
|
||||
add("OEBPS/images/封面.png", string(coverBytes))
|
||||
if err := zw.Close(); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
book, err := svc.ImportLocalBook(t.Context(), "u1", "兜底.epub", buf.Bytes())
|
||||
if err != nil {
|
||||
t.Fatalf("导入失败: %v", err)
|
||||
}
|
||||
if !strings.HasPrefix(book.CoverURL, "/api/reader/local/asset?") {
|
||||
t.Fatalf("应兜底找到封面: %q", book.CoverURL)
|
||||
}
|
||||
u, err := url.Parse(book.CoverURL)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
entry, err := svc.VerifyLocalAssetURL(u.Query().Get("b"), u.Query().Get("p"), u.Query().Get("s"))
|
||||
if err != nil {
|
||||
t.Fatalf("封面签名校验失败: %v", err)
|
||||
}
|
||||
if filepath.Base(entry) != "封面.png" {
|
||||
t.Fatalf("兜底应命中封面.png,实际 %q", entry)
|
||||
}
|
||||
data, _, err := svc.ReadLocalAsset(t.Context(), book.ID, entry)
|
||||
if err != nil || !bytes.Equal(data, coverBytes) {
|
||||
t.Fatalf("读兜底封面失败: %v / %q", err, data)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,300 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"context"
|
||||
"encoding/json"
|
||||
"fmt"
|
||||
"strings"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/service/reader/rule"
|
||||
)
|
||||
|
||||
// 本文件:书源登录(对应 legado SourceLoginDialog / SourceLoginViewModel)。
|
||||
//
|
||||
// legado 的登录模型:
|
||||
// - loginUrl 是一段 JS,既是登录逻辑,也是 loginUi 各按钮 action 的函数库;
|
||||
// - loginUi 是一份表单描述(RowUi 数组:text / password / button / toggle / select);
|
||||
// - 点按钮时执行 "loginUrl + '\n' + action",作用域里 result 是"表单当前值",
|
||||
// 返回值被丢弃,只有抛异常才算失败;
|
||||
// - 点"登录"确认按钮时执行 loginUrl 里的 login() 函数(不存在则报错)。
|
||||
//
|
||||
// 服务端无法弹窗,因此把 java.toast / java.startBrowser 的调用收集起来回传前端。
|
||||
|
||||
// LoginField 登录表单的一个控件(对应 legado RowUi)。
|
||||
type LoginField struct {
|
||||
Name string `json:"name"`
|
||||
Type string `json:"type"` // text / password / button / toggle / select
|
||||
Action string `json:"action,omitempty"`
|
||||
Chars []string `json:"chars,omitempty"`
|
||||
Default string `json:"default,omitempty"`
|
||||
// ViewName 按钮/标签文案;legado 中它本身可以是 JS 表达式。
|
||||
ViewName string `json:"viewName,omitempty"`
|
||||
Style map[string]any `json:"style,omitempty"`
|
||||
}
|
||||
|
||||
// SourceLoginInfo 书源登录界面描述与当前状态。
|
||||
type SourceLoginInfo struct {
|
||||
SourceID string `json:"source_id"`
|
||||
SourceName string `json:"source_name"`
|
||||
// HasLoginJS loginUrl 提供了登录逻辑。
|
||||
HasLoginJS bool `json:"has_login_js"`
|
||||
// LoginJS 登录逻辑 JS 原文(前端只作展示/调试,不执行)。
|
||||
LoginJS string `json:"login_js,omitempty"`
|
||||
// Fields 登录表单控件。
|
||||
Fields []LoginField `json:"fields"`
|
||||
// Values 已保存的登录信息(表单回填)。
|
||||
Values map[string]string `json:"values"`
|
||||
// Cookies 当前已保存的 Cookie(domain → cookie 串)。
|
||||
Cookies map[string]string `json:"cookies"`
|
||||
// Variable 源变量 JSON 原文(供变量编辑器)。
|
||||
Variable string `json:"variable"`
|
||||
// VariableComment 源变量说明(书源作者写的填写指引)。
|
||||
VariableComment string `json:"variable_comment,omitempty"`
|
||||
// LoggedIn 是否已具备登录态(存了登录信息或 Cookie)。
|
||||
LoggedIn bool `json:"logged_in"`
|
||||
// LoginFieldsHint loginUi 解析或执行的失败原因(不阻断展示)。
|
||||
Error string `json:"error,omitempty"`
|
||||
}
|
||||
|
||||
// LoginResult 登录动作执行结果。
|
||||
type LoginResult struct {
|
||||
OK bool `json:"ok"`
|
||||
Error string `json:"error,omitempty"`
|
||||
Toasts []string `json:"toasts,omitempty"`
|
||||
Browsers []rule.BrowserRequest `json:"browsers,omitempty"`
|
||||
// UIRefresh 书源通过 java.reLoginView / refreshExplore / upLoginData
|
||||
// 要求重新渲染登录表单(前端据此重建 loginUi)。
|
||||
UIRefresh bool `json:"ui_refresh,omitempty"`
|
||||
// Values 执行后的登录信息(可能与执行前不同,如 checkStatus 回填邮箱)。
|
||||
Values map[string]string `json:"values"`
|
||||
Cookies map[string]string `json:"cookies"`
|
||||
LoggedIn bool `json:"logged_in"`
|
||||
}
|
||||
|
||||
// GetSourceLogin 返回书源登录界面描述与当前登录状态。
|
||||
func (s *ReaderService) GetSourceLogin(ctx context.Context, userID, sourceID string) (*SourceLoginInfo, error) {
|
||||
src, bs, err := s.loadSource(ctx, sourceID)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
sess := s.newSession(ctx, src, bs)
|
||||
sess.userID = userID
|
||||
defer sess.close()
|
||||
state := sess.state
|
||||
|
||||
info := &SourceLoginInfo{
|
||||
SourceID: src.ID,
|
||||
SourceName: firstNonEmpty(src.Name, bs.BookSourceName),
|
||||
HasLoginJS: strings.TrimSpace(SPtr(bs.LoginURL)) != "",
|
||||
LoginJS: bs.LoginJS(),
|
||||
Fields: []LoginField{},
|
||||
Values: map[string]string{},
|
||||
Cookies: state.snapshotCookies(),
|
||||
Variable: state.GetVariable(),
|
||||
VariableComment: SPtr(bs.VariableComment),
|
||||
}
|
||||
if m, err := parseLoginInfoValues(state.GetLoginInfo()); err == nil {
|
||||
info.Values = m
|
||||
}
|
||||
// 登录态以「是否拿到凭证(Cookie)」为准:
|
||||
// 只填过表单并不等于已登录,否则失败的登录也会显示为已登录。
|
||||
info.LoggedIn = len(info.Cookies) > 0
|
||||
|
||||
fields, err := s.resolveLoginFields(sess, bs, info.Values)
|
||||
if err != nil {
|
||||
info.Error = err.Error()
|
||||
// loginUi 解析失败时若只有 loginUrl,仍可用「打开登录页」方式登录。
|
||||
if SPtr(bs.LoginUI) == "" {
|
||||
info.Error = ""
|
||||
}
|
||||
} else {
|
||||
info.Fields = fields
|
||||
}
|
||||
return info, nil
|
||||
}
|
||||
|
||||
// resolveLoginFields 解析 loginUi:直接是 JSON 数组时直接解析;
|
||||
// 是 @js:/<js> 时先执行得到 JSON(对应 legado evalUiJs)。
|
||||
// values 为已保存的登录信息,作为 result 注入(legado 的 loginUi JS 会读它)。
|
||||
func (s *ReaderService) resolveLoginFields(sess *sourceSession, bs *BookSource, values map[string]string) ([]LoginField, error) {
|
||||
raw := strings.TrimSpace(SPtr(bs.LoginUI))
|
||||
if raw == "" {
|
||||
return nil, fmt.Errorf("书源未配置登录界面(loginUi)")
|
||||
}
|
||||
if isJSWrapped(raw) {
|
||||
runner := sess.runner("", 0)
|
||||
v, err := runner.EvalAction(bs.LoginJS()+"\n"+stripJSWrapper(raw), loginBindings(sess, values))
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("loginUi JS 执行失败: %w", err)
|
||||
}
|
||||
raw = strings.TrimSpace(anyToStr(v))
|
||||
}
|
||||
var fields []LoginField
|
||||
if err := json.Unmarshal([]byte(raw), &fields); err != nil {
|
||||
return nil, fmt.Errorf("loginUi 不是合法的 JSON 表单: %w", err)
|
||||
}
|
||||
return fields, nil
|
||||
}
|
||||
|
||||
// RunLoginAction 执行一个登录动作。
|
||||
//
|
||||
// action 为 loginUi 里某个控件的 action(如 "login(true)" / "checkStatus()");
|
||||
// fields 为前端提交的表单值,会与已保存的登录信息合并后作为 result 传入。
|
||||
// action 为空时执行 loginUrl 里的 login()(即 legado 的「确认登录」)。
|
||||
func (s *ReaderService) RunLoginAction(ctx context.Context, userID, sourceID, action string, fields map[string]string) (*LoginResult, error) {
|
||||
src, bs, err := s.loadSource(ctx, sourceID)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
sess := s.newSession(ctx, src, bs)
|
||||
sess.userID = userID
|
||||
// 登录动作里才注入宿主浏览器:书源的「切换线路」「用户后台」等按钮
|
||||
// 依赖 java.startBrowserAwait 打开页面并等待用户操作。
|
||||
sess.browserEnabled = true
|
||||
defer sess.close()
|
||||
state := sess.state
|
||||
|
||||
loginJS := bs.LoginJS()
|
||||
if loginJS == "" {
|
||||
return nil, fmt.Errorf("书源未配置登录逻辑(loginUrl)")
|
||||
}
|
||||
|
||||
// 合并登录信息:已保存值 + 本次提交值(对应 legado getLoginData)。
|
||||
values := map[string]string{}
|
||||
if m, err := parseLoginInfoValues(state.GetLoginInfo()); err == nil {
|
||||
values = m
|
||||
}
|
||||
for k, v := range fields {
|
||||
values[k] = v
|
||||
}
|
||||
// 若提供了新的表单值,先持久化(legado 在调用登录函数前先存 loginInfo)。
|
||||
if len(fields) > 0 {
|
||||
if b, err := json.Marshal(values); err == nil {
|
||||
state.SetLoginInfo(string(b))
|
||||
}
|
||||
}
|
||||
|
||||
body := loginJS + "\n"
|
||||
if strings.TrimSpace(action) == "" {
|
||||
// 对应 legado:login() 必须由书源实现
|
||||
body += "if (typeof login=='function'){ login.apply(this); } else { throw('Function login not implements!!!'); }"
|
||||
} else {
|
||||
body += action
|
||||
}
|
||||
|
||||
runner := sess.runner("", 0)
|
||||
_, runErr := runner.EvalAction(body, loginBindings(sess, values))
|
||||
|
||||
// 无论成功失败都要落库:Cookie/变量可能已被部分改写(如已拿到 token 但后续步骤报错)。
|
||||
state.flush()
|
||||
|
||||
res := &LoginResult{
|
||||
OK: runErr == nil,
|
||||
Toasts: state.toasts,
|
||||
Browsers: state.browsers,
|
||||
UIRefresh: state.UIRefreshRequested(),
|
||||
Values: values,
|
||||
Cookies: state.snapshotCookies(),
|
||||
}
|
||||
if runErr != nil {
|
||||
res.Error = runErr.Error()
|
||||
}
|
||||
// 登录态以是否拿到凭证(Cookie)为准,避免"只存了表单"被显示成已登录。
|
||||
res.LoggedIn = len(res.Cookies) > 0
|
||||
return res, nil
|
||||
}
|
||||
|
||||
// SetSourceVariable 覆盖保存书源变量(前端变量编辑器)。
|
||||
// 保存后书源 JS 的 getVariable() 即可读到。
|
||||
func (s *ReaderService) SetSourceVariable(ctx context.Context, sourceID, variable string) error {
|
||||
if strings.TrimSpace(variable) != "" && !json.Valid([]byte(variable)) {
|
||||
return fmt.Errorf("变量必须是合法 JSON")
|
||||
}
|
||||
src, bs, err := s.loadSource(ctx, sourceID)
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
sess := s.newSession(ctx, src, bs)
|
||||
defer sess.close()
|
||||
sess.state.SetVariable(variable)
|
||||
sess.state.flush()
|
||||
return nil
|
||||
}
|
||||
|
||||
// SetSourceLoginInfo 直接覆盖保存登录信息(前端表单保存,不触发登录动作)。
|
||||
func (s *ReaderService) SetSourceLoginInfo(ctx context.Context, sourceID string, fields map[string]string) error {
|
||||
src, bs, err := s.loadSource(ctx, sourceID)
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
sess := s.newSession(ctx, src, bs)
|
||||
defer sess.close()
|
||||
b, err := json.Marshal(fields)
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
sess.state.SetLoginInfo(string(b))
|
||||
sess.state.flush()
|
||||
return nil
|
||||
}
|
||||
|
||||
// ClearSourceLogin 清除登录态:登录信息与全部 Cookie(对应 legado logout)。
|
||||
func (s *ReaderService) ClearSourceLogin(ctx context.Context, sourceID string) error {
|
||||
src, bs, err := s.loadSource(ctx, sourceID)
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
sess := s.newSession(ctx, src, bs)
|
||||
defer sess.close()
|
||||
sess.state.SetLoginInfo("")
|
||||
sess.state.SetLoginHeader("")
|
||||
sess.state.clearCookies()
|
||||
sess.state.flush()
|
||||
return nil
|
||||
}
|
||||
|
||||
// loginBindings 构造登录动作的 JS 作用域绑定。
|
||||
// 对应 legado:result 为表单数据;book/chapter/isLongClick 一并提供。
|
||||
func loginBindings(sess *sourceSession, values map[string]string) map[string]any {
|
||||
if values == nil {
|
||||
values = map[string]string{}
|
||||
}
|
||||
return map[string]any{
|
||||
"result": values,
|
||||
"book": nil,
|
||||
"chapter": nil,
|
||||
"isLongClick": false,
|
||||
"sourceUrl": sess.srcURL(),
|
||||
}
|
||||
}
|
||||
|
||||
func parseLoginInfoValues(raw string) (map[string]string, error) {
|
||||
out := map[string]string{}
|
||||
if strings.TrimSpace(raw) == "" {
|
||||
return out, nil
|
||||
}
|
||||
if err := json.Unmarshal([]byte(raw), &out); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
return out, nil
|
||||
}
|
||||
|
||||
// isJSWrapped 判断规则串是否带 @js: / <js> 包裹。
|
||||
func isJSWrapped(s string) bool {
|
||||
s = strings.TrimSpace(s)
|
||||
return strings.HasPrefix(s, "@js:") || strings.HasPrefix(s, "<js>")
|
||||
}
|
||||
|
||||
func anyToStr(v any) string {
|
||||
if v == nil {
|
||||
return ""
|
||||
}
|
||||
if s, ok := v.(string); ok {
|
||||
return s
|
||||
}
|
||||
b, err := json.Marshal(v)
|
||||
if err != nil {
|
||||
return fmt.Sprintf("%v", v)
|
||||
}
|
||||
return string(b)
|
||||
}
|
||||
@@ -0,0 +1,407 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"context"
|
||||
"encoding/json"
|
||||
"fmt"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"strings"
|
||||
"testing"
|
||||
|
||||
"github.com/glebarez/sqlite"
|
||||
"go.uber.org/zap"
|
||||
"gorm.io/gorm"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/config"
|
||||
"github.com/truewhile/MeBox/internal/database"
|
||||
"github.com/truewhile/MeBox/internal/repository"
|
||||
)
|
||||
|
||||
// 本文件:登录类书源的服务层链路测试。
|
||||
// 覆盖「登录 → Cookie 落库 → 后续请求自动携带 Cookie → 登出清理」。
|
||||
|
||||
// readerTestUserID 登录接口按用户隔离浏览器待办,测试里统一用一个固定用户。
|
||||
const readerTestUserID = "test-user"
|
||||
|
||||
// loginTestServer 模拟一个需要登录的书源站点:
|
||||
// - POST /login_api 校验账号密码并下发会话 Cookie
|
||||
// - GET /search 读取 Cookie,无 Cookie 返回 401(模拟登录后才能搜索)
|
||||
func loginTestServer(t *testing.T) *httptest.Server {
|
||||
t.Helper()
|
||||
return httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
switch {
|
||||
case r.URL.Path == "/login_api":
|
||||
var body map[string]string
|
||||
_ = json.NewDecoder(r.Body).Decode(&body)
|
||||
if body["register_email"] != "user@example.com" || body["password"] != "pw123456" {
|
||||
_, _ = w.Write([]byte(`{"code":1,"msg":"账号或密码错误"}`))
|
||||
return
|
||||
}
|
||||
http.SetCookie(w, &http.Cookie{Name: "qttoken", Value: "SESSION_abcdef123456", Path: "/"})
|
||||
_, _ = w.Write([]byte(`{"code":0,"key":"SESSION_abcdef123456"}`))
|
||||
case r.URL.Path == "/search":
|
||||
if !strings.Contains(r.Header.Get("Cookie"), "qttoken=SESSION_abcdef123456") {
|
||||
w.WriteHeader(http.StatusUnauthorized)
|
||||
_, _ = w.Write([]byte(`<html><body>未登录</body></html>`))
|
||||
return
|
||||
}
|
||||
_, _ = w.Write([]byte(`<html><body>
|
||||
<div class="item"><h3><a href="/book/9">会员专享书</a></h3><span class="author">作者</span></div>
|
||||
</body></html>`))
|
||||
default:
|
||||
http.NotFound(w, r)
|
||||
}
|
||||
}))
|
||||
}
|
||||
|
||||
func loginTestSourceJSON(t *testing.T, server string) string {
|
||||
t.Helper()
|
||||
// loginUrl 是登录逻辑:读 result 里的表单值 → 调登录接口 → 写入 Cookie。
|
||||
// BaseUrl() 在真实书源里由 jsLib 提供,这里一并定义。
|
||||
loginJS := `function BaseUrl() { return baseUrl; }
|
||||
function login(flag) {
|
||||
var payload = JSON.stringify({register_email: result['邮箱'], password: result['密码']});
|
||||
var res = java.ajax(BaseUrl() + '/login_api,{"method":"POST","headers":{"Content-Type":"application/json"},"body":' + JSON.stringify(payload) + '}');
|
||||
var data = JSON.parse(res);
|
||||
if (data.code == 0) {
|
||||
setAllCookies('qttoken=' + data.key);
|
||||
java.toast('登录成功');
|
||||
return true;
|
||||
}
|
||||
java.toast(data.msg || '登录失败');
|
||||
return false;
|
||||
}
|
||||
function setAllCookies(ck) { cookie.setCookie(BaseUrl(), ck); }`
|
||||
|
||||
loginUI := []map[string]any{
|
||||
{"name": "邮箱", "type": "text"},
|
||||
{"name": "密码", "type": "password"},
|
||||
{"name": "登录", "type": "button", "action": "login(true)"},
|
||||
}
|
||||
uiJSON, err := json.Marshal(loginUI)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
src := map[string]any{
|
||||
"bookSourceUrl": server,
|
||||
"bookSourceName": "登录源",
|
||||
"bookSourceType": 0,
|
||||
"enabledCookieJar": true,
|
||||
"loginUrl": loginJS,
|
||||
"loginUi": string(uiJSON),
|
||||
"searchUrl": server + "/search",
|
||||
"ruleSearch": map[string]any{
|
||||
"bookList": "class.item",
|
||||
"name": "tag.h3@tag.a@text",
|
||||
"bookUrl": "tag.h3@tag.a@href",
|
||||
"author": "class.author@text",
|
||||
},
|
||||
}
|
||||
out, err := json.Marshal(src)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
return string(out)
|
||||
}
|
||||
|
||||
func newLoginTestService(t *testing.T) (*ReaderService, *repository.Container) {
|
||||
t.Helper()
|
||||
// 唯一库名避免同包测试互相污染;cache=shared 让连接池共享同一份内存库
|
||||
// (多源搜索会并发写会话状态,而 :memory: 下每个连接各自一个库,
|
||||
// 并发写入对后续读取不可见,会造成测试假失败)。
|
||||
dsn := fmt.Sprintf("file:%s?mode=memory&cache=shared", strings.ReplaceAll(t.Name(), "/", "_"))
|
||||
db, err := gorm.Open(sqlite.Open(dsn), &gorm.Config{})
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if err := database.AutoMigrate(db); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
repos := repository.New(db)
|
||||
cfg := &config.Config{}
|
||||
cfg.Secrets.JWTSecret = "test-secret"
|
||||
svc := NewReaderService(cfg, zap.NewNop(), repos)
|
||||
return svc, repos
|
||||
}
|
||||
|
||||
// prepareLoginSource 导入测试书源并返回其 ID。
|
||||
func prepareLoginSource(t *testing.T, svc *ReaderService, sourceJSON string) string {
|
||||
t.Helper()
|
||||
if _, err := svc.ImportSources(t.Context(), sourceJSON); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
srcs, err := svc.ListSources(t.Context())
|
||||
if err != nil || len(srcs) == 0 {
|
||||
t.Fatalf("导入后应能读到书源: %v", err)
|
||||
}
|
||||
return srcs[0].ID
|
||||
}
|
||||
|
||||
// TestSourceLoginEndToEnd 登录成功后:
|
||||
// - Cookie 落库(可再次读出)
|
||||
// - 搜索请求自动携带 Cookie 并通过鉴权
|
||||
// - 登出后 Cookie 清除,搜索重新变成未登录
|
||||
func TestSourceLoginEndToEnd(t *testing.T) {
|
||||
srv := loginTestServer(t)
|
||||
defer srv.Close()
|
||||
svc, repos := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
sourceID := prepareLoginSource(t, svc, loginTestSourceJSON(t, srv.URL))
|
||||
|
||||
// ── 登录前:未鉴权,站点返回未登录页 → 搜不到书 ──
|
||||
if books, _, err := svc.Search(ctx, "会员", nil, 1); err != nil {
|
||||
t.Fatal(err)
|
||||
} else if len(books) != 0 {
|
||||
t.Fatalf("未登录时不应搜到结果: %+v", books)
|
||||
}
|
||||
|
||||
// ── 登录 ──
|
||||
res, err := svc.RunLoginAction(ctx, readerTestUserID, sourceID, "", map[string]string{
|
||||
"邮箱": "user@example.com", "密码": "pw123456",
|
||||
})
|
||||
if err != nil {
|
||||
t.Fatalf("登录失败: %v", err)
|
||||
}
|
||||
if !res.OK {
|
||||
t.Fatalf("登录未成功: %+v", res)
|
||||
}
|
||||
if !res.LoggedIn {
|
||||
t.Fatalf("登录后应处于已登录态: %+v", res)
|
||||
}
|
||||
if len(res.Cookies) == 0 {
|
||||
t.Fatalf("登录后应有 Cookie 落库: %+v", res)
|
||||
}
|
||||
|
||||
// ── 登录态应落库(换一个 service 实例仍可读到)──
|
||||
svc2 := NewReaderService(svc.cfg, zap.NewNop(), repos)
|
||||
info, err := svc2.GetSourceLogin(ctx, readerTestUserID, sourceID)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if !info.LoggedIn || len(info.Cookies) == 0 {
|
||||
t.Fatalf("新实例未能读到已持久化的登录态: %+v", info)
|
||||
}
|
||||
if info.Values["邮箱"] != "user@example.com" {
|
||||
t.Fatalf("登录表单值未持久化: %+v", info.Values)
|
||||
}
|
||||
|
||||
// ── 登录后搜索:应携带 Cookie 并成功 ──
|
||||
books, skipped, err := svc.Search(ctx, "会员", nil, 1)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if len(skipped) > 0 {
|
||||
t.Fatalf("已登录后搜索不应失败: %+v", skipped)
|
||||
}
|
||||
if len(books) != 1 || books[0].Name != "会员专享书" {
|
||||
t.Fatalf("搜索结果异常: %+v", books)
|
||||
}
|
||||
|
||||
// ── 登出:Cookie 清除,搜索重新未登录 ──
|
||||
if err := svc.ClearSourceLogin(ctx, sourceID); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
info, err = svc.GetSourceLogin(ctx, readerTestUserID, sourceID)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if info.LoggedIn || len(info.Cookies) != 0 {
|
||||
t.Fatalf("登出后不应残留登录态: %+v", info)
|
||||
}
|
||||
if books, _, err := svc.Search(ctx, "会员", nil, 1); err != nil {
|
||||
t.Fatal(err)
|
||||
} else if len(books) != 0 {
|
||||
t.Fatalf("登出后不应还能搜到结果: %+v", books)
|
||||
}
|
||||
}
|
||||
|
||||
// TestSourceLoginWrongPassword 密码错误时登录动作应报失败并给出书源提示。
|
||||
func TestSourceLoginWrongPassword(t *testing.T) {
|
||||
srv := loginTestServer(t)
|
||||
defer srv.Close()
|
||||
svc, _ := newLoginTestService(t)
|
||||
sourceID := prepareLoginSource(t, svc, loginTestSourceJSON(t, srv.URL))
|
||||
|
||||
res, err := svc.RunLoginAction(t.Context(), readerTestUserID, sourceID, "", map[string]string{
|
||||
"邮箱": "user@example.com", "密码": "wrong",
|
||||
})
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if res.LoggedIn {
|
||||
t.Fatalf("密码错误不应处于已登录态: %+v", res)
|
||||
}
|
||||
if len(res.Toasts) == 0 {
|
||||
t.Fatalf("应把书源的失败提示回传:%+v", res)
|
||||
}
|
||||
if !strings.Contains(strings.Join(res.Toasts, " "), "密码错误") {
|
||||
t.Fatalf("提示语未透传: %v", res.Toasts)
|
||||
}
|
||||
}
|
||||
|
||||
// TestSourceLoginInfo_ExposesUIFields 登录界面描述应包含 loginUi 的控件定义。
|
||||
func TestSourceLoginInfo_ExposesUIFields(t *testing.T) {
|
||||
srv := loginTestServer(t)
|
||||
defer srv.Close()
|
||||
svc, _ := newLoginTestService(t)
|
||||
sourceID := prepareLoginSource(t, svc, loginTestSourceJSON(t, srv.URL))
|
||||
|
||||
info, err := svc.GetSourceLogin(t.Context(), readerTestUserID, sourceID)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if !info.HasLoginJS {
|
||||
t.Fatal("应识别到 loginUrl")
|
||||
}
|
||||
var names []string
|
||||
for _, f := range info.Fields {
|
||||
names = append(names, f.Name)
|
||||
}
|
||||
for _, want := range []string{"邮箱", "密码", "登录"} {
|
||||
if !containsStr(names, want) {
|
||||
t.Fatalf("loginUi 字段缺失 %q: %v", want, names)
|
||||
}
|
||||
}
|
||||
// 密码字段类型应保留,前端据此用 password 输入框
|
||||
for _, f := range info.Fields {
|
||||
if f.Name == "密码" && f.Type != "password" {
|
||||
t.Fatalf("密码字段类型 = %q", f.Type)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestSourceStateEncryptedAtRest 登录信息与 Cookie 应加密落库。
|
||||
func TestSourceStateEncryptedAtRest(t *testing.T) {
|
||||
srv := loginTestServer(t)
|
||||
defer srv.Close()
|
||||
svc, repos := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
sourceID := prepareLoginSource(t, svc, loginTestSourceJSON(t, srv.URL))
|
||||
|
||||
if _, err := svc.RunLoginAction(ctx, readerTestUserID, sourceID, "", map[string]string{
|
||||
"邮箱": "user@example.com", "密码": "pw123456",
|
||||
}); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
src, err := repos.Reader.GetSource(ctx, sourceID)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
st, err := repos.Reader.GetSourceState(ctx, src.SourceURL)
|
||||
if err != nil || st == nil {
|
||||
t.Fatalf("未找到会话状态: %v", err)
|
||||
}
|
||||
if !strings.HasPrefix(st.LoginInfo, "enc:v1:") {
|
||||
t.Fatalf("登录信息应加密存储,实际: %q", st.LoginInfo)
|
||||
}
|
||||
if !strings.HasPrefix(st.Cookies, "enc:v1:") {
|
||||
t.Fatalf("Cookie 应加密存储,实际: %q", st.Cookies)
|
||||
}
|
||||
if strings.Contains(st.LoginInfo, "pw123456") {
|
||||
t.Fatal("明文密码出现在库中")
|
||||
}
|
||||
}
|
||||
|
||||
// TestSourceVariableRoundTrip 源变量可通过接口读写,并被书源 JS getVariable 读到。
|
||||
func TestSourceVariableRoundTrip(t *testing.T) {
|
||||
srv := loginTestServer(t)
|
||||
defer srv.Close()
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
sourceID := prepareLoginSource(t, svc, loginTestSourceJSON(t, srv.URL))
|
||||
|
||||
if err := svc.SetSourceVariable(ctx, sourceID, `{"线路":"https://v2.example.com"}`); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
info, err := svc.GetSourceLogin(ctx, readerTestUserID, sourceID)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
var m map[string]string
|
||||
if err := json.Unmarshal([]byte(info.Variable), &m); err != nil {
|
||||
t.Fatalf("变量回读失败: %q", info.Variable)
|
||||
}
|
||||
if m["线路"] != "https://v2.example.com" {
|
||||
t.Fatalf("变量值不符: %+v", m)
|
||||
}
|
||||
|
||||
// 非法 JSON 应被拒绝
|
||||
if err := svc.SetSourceVariable(ctx, sourceID, "not-json"); err == nil {
|
||||
t.Fatal("非法 JSON 变量应被拒绝")
|
||||
}
|
||||
}
|
||||
|
||||
func containsStr(list []string, want string) bool {
|
||||
for _, s := range list {
|
||||
if s == want {
|
||||
return true
|
||||
}
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
// TestEnabledCookieJarGating enabledCookieJar=false 时不再自动保存响应 Set-Cookie,
|
||||
// 但书源 JS 主动 cookie.setCookie 写入的仍应保留(对应 legado 语义)。
|
||||
func TestEnabledCookieJarGating(t *testing.T) {
|
||||
// 站点在响应里下发 Set-Cookie
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
http.SetCookie(w, &http.Cookie{Name: "auto", Value: "from-response", Path: "/"})
|
||||
_, _ = w.Write([]byte(`<html><body><div class="item"><h3><a href="/book/1">书</a></h3></div></body></html>`))
|
||||
}))
|
||||
defer srv.Close()
|
||||
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
|
||||
// 造一个 searchUrl 指向该站点、enabledCookieJar=false 的书源
|
||||
srcJSON := `{
|
||||
"bookSourceUrl": "` + srv.URL + `",
|
||||
"bookSourceName": "无 CookieJar 源",
|
||||
"bookSourceType": 0,
|
||||
"enabledCookieJar": false,
|
||||
"searchUrl": "` + srv.URL + `/search",
|
||||
"ruleSearch": {"bookList":"class.item","name":"tag.h3@tag.a@text","bookUrl":"tag.h3@tag.a@href"}
|
||||
}`
|
||||
sourceID := prepareLoginSource(t, svc, srcJSON)
|
||||
|
||||
if _, _, err := svc.Search(ctx, "任意", nil, 1); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
// 自动捕获被关闭:不应出现 auto=from-response
|
||||
src, _ := reposReaderURL(ctx, svc, sourceID)
|
||||
st := svc.newSourceState(ctx, src)
|
||||
if got := st.GetCookie(srv.URL); strings.Contains(got, "auto=from-response") {
|
||||
t.Fatalf("enabledCookieJar=false 时不应自动保存 Set-Cookie: %q", got)
|
||||
}
|
||||
|
||||
// 对照:开启时应当捕获
|
||||
srcJSONOn := strings.Replace(srcJSON, `"enabledCookieJar": false`, `"enabledCookieJar": true`, 1)
|
||||
srv2 := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
http.SetCookie(w, &http.Cookie{Name: "auto", Value: "from-response", Path: "/"})
|
||||
_, _ = w.Write([]byte(`<html><body><div class="item"><h3><a href="/book/1">书</a></h3></div></body></html>`))
|
||||
}))
|
||||
defer srv2.Close()
|
||||
srcJSONOn = strings.ReplaceAll(srcJSONOn, srv.URL, srv2.URL)
|
||||
// prepareLoginSource 返回列表首个书源,这里按 URL 精确定位刚导入的对照源
|
||||
if _, err := svc.ImportSources(ctx, srcJSONOn); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if _, _, err := svc.Search(ctx, "任意", nil, 1); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got := svc.newSourceState(ctx, srv2.URL).GetCookie(srv2.URL); !strings.Contains(got, "auto=from-response") {
|
||||
t.Fatalf("enabledCookieJar 默认开启时应捕获 Set-Cookie: %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
// reposReaderURL 取书源的 SourceURL(会话状态按它索引)。
|
||||
func reposReaderURL(ctx context.Context, svc *ReaderService, sourceID string) (string, error) {
|
||||
src, err := svc.repo.GetSource(ctx, sourceID)
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
return src.SourceURL, nil
|
||||
}
|
||||
@@ -0,0 +1,190 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"context"
|
||||
"fmt"
|
||||
"math"
|
||||
"strings"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
)
|
||||
|
||||
// 阅读器偏好(每个用户一条,见 model.ReaderProfile)。
|
||||
//
|
||||
// 这些值原本只存在浏览器 localStorage,属于设备级;现在按用户落库,让主题、
|
||||
// 排版、听书、书架展示设置能跨设备一致。服务端只做「存储 + 收敛」,
|
||||
// 具体语义(哪些主题、怎么排版)仍由前端决定,所以这里只校验取值范围与枚举,
|
||||
// 不解释它们。
|
||||
|
||||
// 取值范围与前端 readerSettings store 里的 clamp 完全一致,两边改动需同步。
|
||||
const (
|
||||
readerFontSizeMin, readerFontSizeMax = 14, 32
|
||||
readerLineHeightMin, readerLineHeightMax = 1.4, 2.6
|
||||
readerParaSpacingMin, readerParaSpacingMax = 0, 32
|
||||
readerAudioSpeedMin, readerAudioSpeedMax = 0.5, 3
|
||||
readerAudioTimerMin, readerAudioTimerMax = 0, 180
|
||||
readerGridColumnsMin, readerGridColumnsMax = 0, 6
|
||||
|
||||
// readerThemeIDDefault 主题标识的自愈值:themeId 不枚举校验(前端可以自由
|
||||
// 新增主题),只限制长度,空值时回落到内置的「预设1」。
|
||||
readerThemeIDDefault = "preset1"
|
||||
readerThemeIDMaxLen = 32
|
||||
)
|
||||
|
||||
// 前端定义的枚举取值;未知值一律回落成默认项,避免脏值把阅读器置于不可用状态。
|
||||
var (
|
||||
readerPageModes = []string{"page", "scroll"}
|
||||
readerShelfLayout = []string{"grid", "list", "compact"}
|
||||
readerShelfSorts = []string{"recent", "update", "mixed", "name", "author", "manual"}
|
||||
)
|
||||
|
||||
// ReaderSettings 阅读器偏好的读写载荷(与前端 ReaderSettingsProfile 字段一一对应)。
|
||||
type ReaderSettings struct {
|
||||
ThemeID string `json:"theme_id"`
|
||||
Night bool `json:"night"`
|
||||
PageMode string `json:"page_mode"`
|
||||
FontSize int `json:"font_size"`
|
||||
LineHeight float64 `json:"line_height"`
|
||||
ParagraphSpacing int `json:"paragraph_spacing"`
|
||||
|
||||
AudioSpeed float64 `json:"audio_speed"`
|
||||
AudioTimerMinutes int `json:"audio_timer_minutes"`
|
||||
|
||||
ShelfLayout string `json:"shelf_layout"`
|
||||
ShelfGridColumns int `json:"shelf_grid_columns"`
|
||||
ShelfSort string `json:"shelf_sort"`
|
||||
ShelfShowUnread bool `json:"shelf_show_unread"`
|
||||
ShelfShowUpdateTime bool `json:"shelf_show_update_time"`
|
||||
}
|
||||
|
||||
// GetReaderSettings 读用户阅读器偏好。
|
||||
// 该用户还没保存过时返回 (nil, nil):由前端用它本地的值播种,避免这里再维护一份默认值。
|
||||
func (s *ReaderService) GetReaderSettings(ctx context.Context, userID string) (*ReaderSettings, error) {
|
||||
if userID == "" {
|
||||
return nil, nil
|
||||
}
|
||||
row, err := s.repo.GetReaderProfile(ctx, userID)
|
||||
if err != nil || row == nil {
|
||||
return nil, err
|
||||
}
|
||||
out := settingsFromProfile(*row)
|
||||
return &out, nil
|
||||
}
|
||||
|
||||
// SaveReaderSettings 覆盖保存用户阅读器偏好,落库前先做范围收敛。
|
||||
func (s *ReaderService) SaveReaderSettings(ctx context.Context, userID string, in ReaderSettings) (*ReaderSettings, error) {
|
||||
if userID == "" {
|
||||
return nil, fmt.Errorf("缺少用户信息")
|
||||
}
|
||||
sanitized := sanitizeReaderSettings(in)
|
||||
row := &model.ReaderProfile{UserID: userID}
|
||||
applySettingsToProfile(row, sanitized)
|
||||
if err := s.repo.SaveReaderProfile(ctx, row); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
out := settingsFromProfile(*row)
|
||||
return &out, nil
|
||||
}
|
||||
|
||||
// sanitizeReaderSettings 把入参收敛到合法范围:数值夹到区间、枚举未知即回落默认、
|
||||
// 主题标识去空白并限长。
|
||||
func sanitizeReaderSettings(in ReaderSettings) ReaderSettings {
|
||||
return ReaderSettings{
|
||||
ThemeID: sanitizeThemeID(in.ThemeID),
|
||||
Night: in.Night,
|
||||
PageMode: oneOf(in.PageMode, readerPageModes, readerPageModes[0]),
|
||||
FontSize: clampInt(in.FontSize, readerFontSizeMin, readerFontSizeMax),
|
||||
LineHeight: roundTo1(clampFloat(in.LineHeight, readerLineHeightMin, readerLineHeightMax)),
|
||||
ParagraphSpacing: clampInt(in.ParagraphSpacing, readerParaSpacingMin, readerParaSpacingMax),
|
||||
|
||||
AudioSpeed: roundTo1(clampFloat(in.AudioSpeed, readerAudioSpeedMin, readerAudioSpeedMax)),
|
||||
AudioTimerMinutes: clampInt(in.AudioTimerMinutes, readerAudioTimerMin, readerAudioTimerMax),
|
||||
|
||||
ShelfLayout: oneOf(in.ShelfLayout, readerShelfLayout, readerShelfLayout[0]),
|
||||
ShelfGridColumns: clampInt(in.ShelfGridColumns, readerGridColumnsMin, readerGridColumnsMax),
|
||||
ShelfSort: oneOf(in.ShelfSort, readerShelfSorts, readerShelfSorts[0]),
|
||||
ShelfShowUnread: in.ShelfShowUnread,
|
||||
ShelfShowUpdateTime: in.ShelfShowUpdateTime,
|
||||
}
|
||||
}
|
||||
|
||||
func sanitizeThemeID(v string) string {
|
||||
v = strings.TrimSpace(v)
|
||||
if v == "" || len(v) > readerThemeIDMaxLen {
|
||||
return readerThemeIDDefault
|
||||
}
|
||||
return v
|
||||
}
|
||||
|
||||
func settingsFromProfile(p model.ReaderProfile) ReaderSettings {
|
||||
return ReaderSettings{
|
||||
ThemeID: p.ThemeID,
|
||||
Night: p.Night,
|
||||
PageMode: p.PageMode,
|
||||
FontSize: p.FontSize,
|
||||
LineHeight: p.LineHeight,
|
||||
ParagraphSpacing: p.ParagraphSpacing,
|
||||
|
||||
AudioSpeed: p.AudioSpeed,
|
||||
AudioTimerMinutes: p.AudioTimerMinutes,
|
||||
|
||||
ShelfLayout: p.ShelfLayout,
|
||||
ShelfGridColumns: p.ShelfGridColumns,
|
||||
ShelfSort: p.ShelfSort,
|
||||
ShelfShowUnread: p.ShelfShowUnread,
|
||||
ShelfShowUpdateTime: p.ShelfShowUpdateTime,
|
||||
}
|
||||
}
|
||||
|
||||
func applySettingsToProfile(p *model.ReaderProfile, in ReaderSettings) {
|
||||
p.ThemeID = in.ThemeID
|
||||
p.Night = in.Night
|
||||
p.PageMode = in.PageMode
|
||||
p.FontSize = in.FontSize
|
||||
p.LineHeight = in.LineHeight
|
||||
p.ParagraphSpacing = in.ParagraphSpacing
|
||||
|
||||
p.AudioSpeed = in.AudioSpeed
|
||||
p.AudioTimerMinutes = in.AudioTimerMinutes
|
||||
|
||||
p.ShelfLayout = in.ShelfLayout
|
||||
p.ShelfGridColumns = in.ShelfGridColumns
|
||||
p.ShelfSort = in.ShelfSort
|
||||
p.ShelfShowUnread = in.ShelfShowUnread
|
||||
p.ShelfShowUpdateTime = in.ShelfShowUpdateTime
|
||||
}
|
||||
|
||||
func clampInt(v, lo, hi int) int {
|
||||
if v < lo {
|
||||
return lo
|
||||
}
|
||||
if v > hi {
|
||||
return hi
|
||||
}
|
||||
return v
|
||||
}
|
||||
|
||||
func clampFloat(v, lo, hi float64) float64 {
|
||||
// JSON 不会给出 NaN,这里是防御性的(例如将来接了别的调用方)
|
||||
if math.IsNaN(v) || v < lo {
|
||||
return lo
|
||||
}
|
||||
if v > hi {
|
||||
return hi
|
||||
}
|
||||
return v
|
||||
}
|
||||
|
||||
func roundTo1(v float64) float64 {
|
||||
return math.Round(v*10) / 10
|
||||
}
|
||||
|
||||
// oneOf 返回 v(当它属于 allowed 时),否则返回 fallback。
|
||||
func oneOf(v string, allowed []string, fallback string) string {
|
||||
for _, a := range allowed {
|
||||
if v == a {
|
||||
return v
|
||||
}
|
||||
}
|
||||
return fallback
|
||||
}
|
||||
@@ -0,0 +1,134 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"io"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"net/url"
|
||||
"sync/atomic"
|
||||
"testing"
|
||||
"time"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/config"
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
)
|
||||
|
||||
// TestProxyURLSignRoundTrip 验证媒体代理签名:往返还原 + 防篡改。
|
||||
func TestProxyURLSignRoundTrip(t *testing.T) {
|
||||
s := &ReaderService{cfg: &config.Config{Secrets: config.SecretsConfig{JWTSecret: "test-secret"}}}
|
||||
const raw = "https://cdn.example.com/audio/ep1.mp3?token=abc"
|
||||
signed := s.ProxyURL("book-1", raw)
|
||||
|
||||
parsed, err := url.Parse(signed)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
q := parsed.Query()
|
||||
if q.Get("b") != "book-1" {
|
||||
t.Fatalf("bookID = %q", q.Get("b"))
|
||||
}
|
||||
restored, err := s.VerifyProxyURL(q.Get("b"), q.Get("u"), q.Get("s"))
|
||||
if err != nil {
|
||||
t.Fatalf("verify failed: %v", err)
|
||||
}
|
||||
if restored != raw {
|
||||
t.Fatalf("restored = %q", restored)
|
||||
}
|
||||
// 篡改签名必须被拒绝
|
||||
if _, err := s.VerifyProxyURL(q.Get("b"), q.Get("u"), "00000000000000000000000000000000"); err == nil {
|
||||
t.Fatal("expected signature rejection")
|
||||
}
|
||||
// 换书签必须被拒绝
|
||||
if _, err := s.VerifyProxyURL("book-2", q.Get("u"), q.Get("s")); err == nil {
|
||||
t.Fatal("expected cross-book signature rejection")
|
||||
}
|
||||
// 已是代理地址的不再二次包裹
|
||||
if again := s.ProxyURL("book-1", signed); again != signed {
|
||||
t.Fatalf("double wrap: %q", again)
|
||||
}
|
||||
}
|
||||
|
||||
// TestSourceReferer 验证默认 Referer 只在 origin 是真正的 http(s) 书源地址时
|
||||
// 才生成。聚合类书源的 origin 是显示名(如「光遇聚合」),拿它当 Referer 会被
|
||||
// 图床判盗链,返回一张「请到本网站阅读」的占位图。
|
||||
func TestSourceReferer(t *testing.T) {
|
||||
cases := []struct {
|
||||
name string
|
||||
origin string
|
||||
want string
|
||||
}{
|
||||
{"普通书源地址", "https://www.example.com", "https://www.example.com/"},
|
||||
{"带尾斜杠", "https://www.example.com/", "https://www.example.com/"},
|
||||
{"带子路径", "https://www.example.com/site", "https://www.example.com/site/"},
|
||||
{"聚合书源显示名", "光遇聚合", ""},
|
||||
{"空值", "", ""},
|
||||
{"只有空白", " ", ""},
|
||||
{"非 http 协议", "ftp://example.com", ""},
|
||||
{"没有主机", "https://", ""},
|
||||
}
|
||||
for _, c := range cases {
|
||||
if got := sourceReferer(c.origin); got != c.want {
|
||||
t.Errorf("%s: sourceReferer(%q) = %q,期望 %q", c.name, c.origin, got, c.want)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestFetchMediaRetriesTransientFailure 验证图床偶发 403(并发限流)会被重试掉,
|
||||
// 否则漫画一屏并发取图时总有几页留成破图。
|
||||
func TestFetchMediaRetriesTransientFailure(t *testing.T) {
|
||||
var hits int32
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
if atomic.AddInt32(&hits, 1) < 3 {
|
||||
w.WriteHeader(http.StatusForbidden)
|
||||
return
|
||||
}
|
||||
w.Header().Set("Content-Type", "image/webp")
|
||||
_, _ = w.Write([]byte("real-image-bytes"))
|
||||
}))
|
||||
defer srv.Close()
|
||||
|
||||
s := &ReaderService{http: &http.Client{Timeout: 10 * time.Second}}
|
||||
book := &model.ReaderBook{Origin: srv.URL}
|
||||
|
||||
resp, err := s.FetchMedia(t.Context(), book, srv.URL+"/1.webp", "")
|
||||
if err != nil {
|
||||
t.Fatalf("FetchMedia 失败: %v", err)
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
if resp.StatusCode != http.StatusOK {
|
||||
t.Fatalf("状态码 = %d,期望 200(前两次 403 应被重试掉)", resp.StatusCode)
|
||||
}
|
||||
data, _ := io.ReadAll(resp.Body)
|
||||
if string(data) != "real-image-bytes" {
|
||||
t.Fatalf("body = %q", data)
|
||||
}
|
||||
if got := atomic.LoadInt32(&hits); got != 3 {
|
||||
t.Fatalf("上游请求次数 = %d,期望 3", got)
|
||||
}
|
||||
}
|
||||
|
||||
// TestFetchMediaDoesNotRetryNotFound 验证确定性的 4xx(如图片不存在)不做无谓重试,
|
||||
// 避免首屏白白多等两轮退避。
|
||||
func TestFetchMediaDoesNotRetryNotFound(t *testing.T) {
|
||||
var hits int32
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
atomic.AddInt32(&hits, 1)
|
||||
w.WriteHeader(http.StatusNotFound)
|
||||
}))
|
||||
defer srv.Close()
|
||||
|
||||
s := &ReaderService{http: &http.Client{Timeout: 10 * time.Second}}
|
||||
book := &model.ReaderBook{Origin: srv.URL}
|
||||
|
||||
resp, err := s.FetchMedia(t.Context(), book, srv.URL+"/missing.webp", "")
|
||||
if err != nil {
|
||||
t.Fatalf("FetchMedia 失败: %v", err)
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
if resp.StatusCode != http.StatusNotFound {
|
||||
t.Fatalf("状态码 = %d,期望 404", resp.StatusCode)
|
||||
}
|
||||
if got := atomic.LoadInt32(&hits); got != 1 {
|
||||
t.Fatalf("上游请求次数 = %d,期望 1(404 不该重试)", got)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,111 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"context"
|
||||
"strconv"
|
||||
"strings"
|
||||
"sync"
|
||||
"time"
|
||||
)
|
||||
|
||||
// sourceRateLimiter 单源限速。
|
||||
//
|
||||
// 书源用 concurrentRate 声明「请求该源的最小间隔」,风控严格的站点靠它避免被封
|
||||
// (不实现的话,搜索/目录/正文连着打过去很容易被拦,表现为「有时能用有时不能用」)。
|
||||
// 对应 legado 的 ConcurrentRateLimiter,支持两种写法:
|
||||
//
|
||||
// "1000" 每次请求至少间隔 1000ms
|
||||
// "3/1000" 每 1000ms 最多 3 次(折算成 333ms 间隔)
|
||||
//
|
||||
// 实现为「按源排槽位」:同一书源的请求被排成固定间隔,天然串行。
|
||||
type sourceRateLimiter struct {
|
||||
mu sync.Mutex
|
||||
slots map[string]time.Time
|
||||
}
|
||||
|
||||
// maxRateGapMs 允许的最大间隔,避免书源写出离谱的值把请求卡死。
|
||||
const maxRateGapMs = 30000
|
||||
|
||||
func newSourceRateLimiter() *sourceRateLimiter {
|
||||
return &sourceRateLimiter{slots: map[string]time.Time{}}
|
||||
}
|
||||
|
||||
// wait 按源限速等待;key 为空或 rate 无法解析时不等待。
|
||||
func (l *sourceRateLimiter) wait(ctx context.Context, key, rate string) {
|
||||
gap := parseConcurrentRate(rate)
|
||||
if gap <= 0 || key == "" {
|
||||
return
|
||||
}
|
||||
now := time.Now()
|
||||
l.mu.Lock()
|
||||
slot := now
|
||||
if next, ok := l.slots[key]; ok && next.After(now) {
|
||||
slot = next
|
||||
}
|
||||
l.slots[key] = slot.Add(gap)
|
||||
l.mu.Unlock()
|
||||
sleepWithContext(ctx, slot.Sub(now))
|
||||
}
|
||||
|
||||
// parseConcurrentRate 解析 concurrentRate,返回两次请求的最小间隔;无法解析返回 0。
|
||||
func parseConcurrentRate(rate string) time.Duration {
|
||||
rate = strings.TrimSpace(rate)
|
||||
if rate == "" {
|
||||
return 0
|
||||
}
|
||||
gapMs := 0
|
||||
if i := strings.IndexByte(rate, '/'); i > 0 {
|
||||
count, errCount := strconv.Atoi(strings.TrimSpace(rate[:i]))
|
||||
interval, errInterval := strconv.Atoi(strings.TrimSpace(rate[i+1:]))
|
||||
if errCount != nil || errInterval != nil || count <= 0 || interval <= 0 {
|
||||
return 0
|
||||
}
|
||||
gapMs = interval / count
|
||||
} else {
|
||||
n, err := strconv.Atoi(rate)
|
||||
if err != nil || n <= 0 {
|
||||
return 0
|
||||
}
|
||||
gapMs = n
|
||||
}
|
||||
if gapMs <= 0 {
|
||||
return 0
|
||||
}
|
||||
if gapMs > maxRateGapMs {
|
||||
gapMs = maxRateGapMs
|
||||
}
|
||||
return time.Duration(gapMs) * time.Millisecond
|
||||
}
|
||||
|
||||
const (
|
||||
// maxRequestAttempts 单次请求最大尝试次数(含首次),兜住书源里写很大的 retry。
|
||||
maxRequestAttempts = 5
|
||||
// requestRetryDelay 请求重试退避基数,实际等待为 attempt 倍。
|
||||
requestRetryDelay = 300 * time.Millisecond
|
||||
)
|
||||
|
||||
// retryableStatus 该状态码是否值得重试:限流(403/429)与上游抖动(5xx)可重试,
|
||||
// 404/400 这类确定性错误不重试。规则请求与媒体代理共用。
|
||||
func retryableStatus(code int) bool {
|
||||
switch code {
|
||||
case 403, 408, 425, 429, 500, 502, 503, 504:
|
||||
return true
|
||||
default:
|
||||
return false
|
||||
}
|
||||
}
|
||||
|
||||
// sleepWithContext 可被 ctx 取消的等待;返回 false 表示上下文已结束。
|
||||
func sleepWithContext(ctx context.Context, d time.Duration) bool {
|
||||
if d <= 0 {
|
||||
return true
|
||||
}
|
||||
timer := time.NewTimer(d)
|
||||
defer timer.Stop()
|
||||
select {
|
||||
case <-ctx.Done():
|
||||
return false
|
||||
case <-timer.C:
|
||||
return true
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,73 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"testing"
|
||||
"time"
|
||||
)
|
||||
|
||||
// TestParseConcurrentRate 书源 concurrentRate 的两种写法。
|
||||
func TestParseConcurrentRate(t *testing.T) {
|
||||
cases := []struct {
|
||||
in string
|
||||
want time.Duration
|
||||
}{
|
||||
{"", 0},
|
||||
{"0", 0},
|
||||
{"abc", 0},
|
||||
{"1000", 1000 * time.Millisecond},
|
||||
{" 500 ", 500 * time.Millisecond},
|
||||
{"3/1000", 333 * time.Millisecond},
|
||||
{"1/2000", 2000 * time.Millisecond},
|
||||
{"3/0", 0},
|
||||
{"0/1000", 0},
|
||||
{"999999", time.Duration(maxRateGapMs) * time.Millisecond}, // 上限夹紧
|
||||
}
|
||||
for _, c := range cases {
|
||||
if got := parseConcurrentRate(c.in); got != c.want {
|
||||
t.Errorf("parseConcurrentRate(%q) = %v,期望 %v", c.in, got, c.want)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestRateLimiterSpacesRequests 同一个源连续两次请求之间至少间隔一个周期。
|
||||
func TestRateLimiterSpacesRequests(t *testing.T) {
|
||||
l := newSourceRateLimiter()
|
||||
start := time.Now()
|
||||
l.wait(t.Context(), "src", "120") // 120ms
|
||||
l.wait(t.Context(), "src", "120")
|
||||
if elapsed := time.Since(start); elapsed < 110*time.Millisecond {
|
||||
t.Fatalf("两次请求只隔了 %v,没有按 concurrentRate 限速", elapsed)
|
||||
}
|
||||
// 不同源互不影响
|
||||
start = time.Now()
|
||||
l.wait(t.Context(), "other", "500")
|
||||
if elapsed := time.Since(start); elapsed > 50*time.Millisecond {
|
||||
t.Fatalf("换源后不该等待,实际 %v", elapsed)
|
||||
}
|
||||
}
|
||||
|
||||
// TestRateLimiterIgnoresEmptyRate 未声明 concurrentRate 时不引入任何等待。
|
||||
func TestRateLimiterIgnoresEmptyRate(t *testing.T) {
|
||||
l := newSourceRateLimiter()
|
||||
start := time.Now()
|
||||
for i := 0; i < 5; i++ {
|
||||
l.wait(t.Context(), "src", "")
|
||||
}
|
||||
if elapsed := time.Since(start); elapsed > 50*time.Millisecond {
|
||||
t.Fatalf("未声明限速却等待了 %v", elapsed)
|
||||
}
|
||||
}
|
||||
|
||||
// TestRetryableStatus 只有可恢复的失败才重试。
|
||||
func TestRetryableStatus(t *testing.T) {
|
||||
for _, code := range []int{403, 429, 500, 502, 503, 504} {
|
||||
if !retryableStatus(code) {
|
||||
t.Errorf("状态码 %d 应当可重试", code)
|
||||
}
|
||||
}
|
||||
for _, code := range []int{200, 301, 400, 401, 404, 410} {
|
||||
if retryableStatus(code) {
|
||||
t.Errorf("状态码 %d 不该重试", code)
|
||||
}
|
||||
}
|
||||
}
|
||||
File diff suppressed because it is too large
Load Diff
@@ -0,0 +1,299 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"net/url"
|
||||
"strings"
|
||||
"testing"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
)
|
||||
|
||||
// TestRewriteContentImageMarkers 文本型漫画源正文里的 <img> 要转成前端认识的 [img] 标记,
|
||||
// 且地址要绝对化、过签名代理(有些图需要书源防盗链头才取得到)。
|
||||
func TestRewriteContentImageMarkers(t *testing.T) {
|
||||
proxy := func(u string) string { return "/proxy?u=" + url.QueryEscape(u) }
|
||||
content := "第一章开始\n" +
|
||||
"<img src=\"https://img.example.com/1.webp\">\n" +
|
||||
"<img src='/img/2.webp'/>\n" +
|
||||
"中间夹了图片 <img src=\"https://img.example.com/3.webp\"> 的这一行不动\n" +
|
||||
"结束"
|
||||
got := rewriteContentImageMarkers(content, "https://site.example.com/ch/1", proxy)
|
||||
lines := strings.Split(got, "\n")
|
||||
if len(lines) != 5 {
|
||||
t.Fatalf("行数 = %d,期望 5\n%s", len(lines), got)
|
||||
}
|
||||
if lines[0] != "第一章开始" || lines[4] != "结束" {
|
||||
t.Fatalf("普通文本行被改动: %q / %q", lines[0], lines[4])
|
||||
}
|
||||
if !strings.HasPrefix(lines[1], imgMarkerPrefix) {
|
||||
t.Fatalf("第 2 行没有变成 [img] 标记: %q", lines[1])
|
||||
}
|
||||
if !strings.Contains(lines[1], url.QueryEscape("https://img.example.com/1.webp")) {
|
||||
t.Fatalf("第 2 行地址没走代理: %q", lines[1])
|
||||
}
|
||||
// 相对地址按章节地址绝对化
|
||||
if !strings.Contains(lines[2], url.QueryEscape("https://site.example.com/img/2.webp")) {
|
||||
t.Fatalf("相对图片地址没有绝对化: %q", lines[2])
|
||||
}
|
||||
// 段落中间的内嵌图片保持原样,避免破坏文本源语义
|
||||
if !strings.Contains(lines[3], "<img src=") {
|
||||
t.Fatalf("段落中间的 <img> 不该被改写: %q", lines[3])
|
||||
}
|
||||
}
|
||||
|
||||
// TestRewriteContentImageMarkersWithoutProxy 不需要代理时(如预览链路)直接落到绝对地址。
|
||||
func TestRewriteContentImageMarkersWithoutProxy(t *testing.T) {
|
||||
got := rewriteContentImageMarkers(`<img src="https://img.example.com/1.webp">`, "https://site.example.com/ch/1", nil)
|
||||
if got != imgMarkerPrefix+"https://img.example.com/1.webp" {
|
||||
t.Fatalf("got %q", got)
|
||||
}
|
||||
// 没有 <img> 的内容原样返回(含 null 字节之外的普通文本)
|
||||
const plain = "纯文本\n第二行"
|
||||
if got := rewriteContentImageMarkers(plain, "", nil); got != plain {
|
||||
t.Fatalf("无图内容被改动: %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
// TestImageMarkersOnly 整章都是 [img] 标记时判定为图片章(前端才能走漫画阅读器、
|
||||
// 双页铺开);只要掺了能读的正文就必须保持文本,不能把文字吃掉。
|
||||
func TestImageMarkersOnly(t *testing.T) {
|
||||
cases := []struct {
|
||||
name string
|
||||
content string
|
||||
wantOK bool
|
||||
wantURLs []string
|
||||
}{
|
||||
{
|
||||
name: "整章都是标记",
|
||||
content: imgMarkerPrefix + "/a.jpg\n" + imgMarkerPrefix + "/b.jpg\n",
|
||||
wantOK: true,
|
||||
wantURLs: []string{"/a.jpg", "/b.jpg"},
|
||||
},
|
||||
{
|
||||
name: "夹着空行 / 孤立 html 标签 / 首尾空白",
|
||||
content: "\n" + imgMarkerPrefix + "/a.jpg\n \n<div>\n</div>\n< br >\n" +
|
||||
" " + imgMarkerPrefix + " /b.jpg \n",
|
||||
wantOK: true,
|
||||
wantURLs: []string{"/a.jpg", "/b.jpg"},
|
||||
},
|
||||
{
|
||||
name: "掺了正文就保持文本(图 + 长段落)",
|
||||
content: imgMarkerPrefix + "/a.jpg\n" + "第一句话。\n" + imgMarkerPrefix + "/b.jpg\n",
|
||||
wantOK: false,
|
||||
},
|
||||
{
|
||||
name: "纯文本",
|
||||
content: "第一章\n正文正文\n",
|
||||
wantOK: false,
|
||||
},
|
||||
{
|
||||
name: "空正文",
|
||||
content: "\n\n \n",
|
||||
wantOK: false,
|
||||
},
|
||||
{
|
||||
name: "标记地址为空",
|
||||
content: imgMarkerPrefix + "/a.jpg\n" + imgMarkerPrefix + " \n",
|
||||
wantOK: false,
|
||||
},
|
||||
{
|
||||
// 「<3」这类不是标签,是有内容的一行,不能当成空白忽略
|
||||
name: "非标签的尖括号行",
|
||||
content: imgMarkerPrefix + "/a.jpg\n3 < 5\n",
|
||||
wantOK: false,
|
||||
},
|
||||
{
|
||||
// 回归:<p>正文</p> 是「标签里裹着正文」,绝不能当成孤立标签跳过,
|
||||
// 否则整章正文会被误判成图片章清空。
|
||||
name: "标签里裹着正文的行",
|
||||
content: imgMarkerPrefix + "/a.jpg\n<p>第一章正文</p>\n<p>第二章正文</p>\n",
|
||||
wantOK: false,
|
||||
},
|
||||
}
|
||||
for _, c := range cases {
|
||||
got, ok := imageMarkersOnly(c.content)
|
||||
if ok != c.wantOK {
|
||||
t.Errorf("%s: ok = %v,期望 %v(urls=%v)", c.name, ok, c.wantOK, got)
|
||||
continue
|
||||
}
|
||||
if !ok {
|
||||
continue
|
||||
}
|
||||
if len(got) != len(c.wantURLs) {
|
||||
t.Errorf("%s: urls = %v,期望 %v", c.name, got, c.wantURLs)
|
||||
continue
|
||||
}
|
||||
for i := range got {
|
||||
if got[i] != c.wantURLs[i] {
|
||||
t.Errorf("%s: urls[%d] = %q,期望 %q", c.name, i, got[i], c.wantURLs[i])
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestIsHTMLTagOnly 「没有可读文字的标签行」才算空行:标签里裹着正文的行必须放行,
|
||||
// 否则 imageMarkersOnly 会把整章正文当图片丢掉。
|
||||
func TestIsHTMLTagOnly(t *testing.T) {
|
||||
blank := []string{"<div>", "</div>", "< br >", "<br/>", "<p></p>", "<hr />", "</p >", "<a> </a>", "<p>"}
|
||||
for _, s := range blank {
|
||||
if !isHTMLTagOnly(s) {
|
||||
t.Errorf("%q 应视为没有可读文字的标签行", s)
|
||||
}
|
||||
}
|
||||
text := []string{
|
||||
"<p>正文</p>", "<p>正文</p></p>", "<div>文字</div>", "<p>第一段</p><p>第二段</p>",
|
||||
"正文 <b>粗</b>", "3 < 5", "<3>", "a", "", " ",
|
||||
}
|
||||
for _, s := range text {
|
||||
if isHTMLTagOnly(s) {
|
||||
t.Errorf("%q 含可读文字/不是标签,不能当空行", s)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestNormalizeContentBlocks 块级标签折成换行、行内标签与文字原样保留。
|
||||
func TestNormalizeContentBlocks(t *testing.T) {
|
||||
cases := []struct{ in, want string }{
|
||||
{"<p>第一段</p>\n<p>第二段</p>", "第一段\n第二段"},
|
||||
{"<p>第一段<br>折行文字</p>", "第一段\n折行文字"},
|
||||
{"<P CLASS='x'>大写带属性</P>", "大写带属性"},
|
||||
{"<div>块</div><br/><br />", "块"},
|
||||
// 行内插图原样保留(不能吃掉普通 <img>)
|
||||
{`<p>正文<img src="https://cdn.example.com/a.jpg"></p>`, `正文<img src="https://cdn.example.com/a.jpg">`},
|
||||
// 纯文本原样返回
|
||||
{"第一段。\n第二段。", "第一段。\n第二段。"},
|
||||
{"", ""},
|
||||
}
|
||||
for _, c := range cases {
|
||||
if got := normalizeContentBlocks(c.in); got != c.want {
|
||||
t.Errorf("normalizeContentBlocks(%q) = %q,期望 %q", c.in, got, c.want)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestChapterContentEmpty 判断「什么都没取到」的正文。
|
||||
//
|
||||
// 聚合类书源把所有线路试完会返回空串(光遇聚合的 request() 就是这么写的),
|
||||
// 这种空结果不能再当成功下发(前端会渲染成白页并缓存下来)。
|
||||
func TestChapterContentEmpty(t *testing.T) {
|
||||
cases := []struct {
|
||||
name string
|
||||
in ChapterContent
|
||||
want bool
|
||||
}{
|
||||
{"空正文", ChapterContent{Type: "text"}, true},
|
||||
{"只有空白字符", ChapterContent{Type: "text", Content: " \n\t "}, true},
|
||||
{"正常正文", ChapterContent{Type: "text", Content: "第一章 世界大变"}, false},
|
||||
// 整章正文只有一个本章说气泡时正文文字为空,但不能算抓取失败。
|
||||
{"只有段评气泡", ChapterContent{Type: "text", Comments: []ContentComment{{Line: 0, Count: 3}}}, false},
|
||||
{"音频无音轨", ChapterContent{Type: "audio"}, true},
|
||||
{"音频有音轨", ChapterContent{Type: "audio", Tracks: []string{"https://cdn.example.com/a.m4a"}}, false},
|
||||
{"漫画无图", ChapterContent{Type: "image"}, true},
|
||||
{"漫画有图", ChapterContent{Type: "image", Images: []string{"https://cdn.example.com/1.jpg"}}, false},
|
||||
}
|
||||
for _, c := range cases {
|
||||
if got := chapterContentEmpty(&c.in); got != c.want {
|
||||
t.Errorf("%s: chapterContentEmpty = %v,期望 %v", c.name, got, c.want)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestSearchCheckKeyWord 校验关键字的取值规则(对应 legado getCheckKeyword):
|
||||
// 含 http/::/++/-- 的值是地址或扩展标记,不当作关键字。
|
||||
func TestSearchCheckKeyWord(t *testing.T) {
|
||||
mk := func(kw string) *BookSource {
|
||||
return &BookSource{RuleSearch: &SearchRule{CheckKeyWord: &kw}}
|
||||
}
|
||||
cases := []struct {
|
||||
in string
|
||||
want string
|
||||
}{
|
||||
{"登录武林系统", "登录武林系统"},
|
||||
{" 空格清理 ", "空格清理"},
|
||||
{"", ""},
|
||||
{"http://example.com", ""},
|
||||
{"a::b", ""},
|
||||
{"a++b", ""},
|
||||
{"a--b", ""},
|
||||
}
|
||||
for _, c := range cases {
|
||||
if got := mk(c.in).SearchCheckKeyWord(); got != c.want {
|
||||
t.Errorf("SearchCheckKeyWord(%q) = %q,期望 %q", c.in, got, c.want)
|
||||
}
|
||||
}
|
||||
if (&BookSource{}).SearchCheckKeyWord() != "" {
|
||||
t.Fatal("没有 ruleSearch 时应返回空")
|
||||
}
|
||||
}
|
||||
|
||||
// TestSearchReportsCheckKeyWordMiss 搜索 0 条且响应里没有书源声明的校验关键字时,
|
||||
// 必须给出「疑似风控」的原因,而不是静默返回 0 条 —— 用户才知道是源的问题。
|
||||
func TestSearchReportsCheckKeyWordMiss(t *testing.T) {
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
w.Header().Set("Content-Type", "application/json")
|
||||
// 风控/失效时站点返回的就是这种没有结果的响应
|
||||
_, _ = fmt.Fprint(w, `{"code":200,"results":{"list":[]}}`)
|
||||
}))
|
||||
defer srv.Close()
|
||||
|
||||
kw := "登录武林系统"
|
||||
header := ""
|
||||
src := &model.ReaderBookSource{Name: "校验关键字测试源", SourceURL: srv.URL}
|
||||
bs := &BookSource{
|
||||
BookSourceURL: srv.URL,
|
||||
Header: &header,
|
||||
SearchURL: strPtr(srv.URL + "/search?q={{key}}"),
|
||||
RuleSearch: &SearchRule{
|
||||
BookList: strPtr("$.results.list[*]"),
|
||||
Name: strPtr("$.name"),
|
||||
BookURL: strPtr("$.url"),
|
||||
CheckKeyWord: &kw,
|
||||
},
|
||||
}
|
||||
|
||||
svc, _ := newLoginTestService(t)
|
||||
_, err := svc.searchInSource(t.Context(), src, bs, "火影", 1)
|
||||
if err == nil {
|
||||
t.Fatal("0 结果且校验关键字缺失时应返回错误")
|
||||
}
|
||||
if !strings.Contains(err.Error(), kw) {
|
||||
t.Fatalf("错误信息里应带上校验关键字,实际: %v", err)
|
||||
}
|
||||
}
|
||||
|
||||
// TestSearchKeepsEmptyResultWhenKeywordPresent 响应里有关键字(站点正常,只是没有匹配的书)
|
||||
// 时不该报错,只是结果为空。
|
||||
func TestSearchKeepsEmptyResultWhenKeywordPresent(t *testing.T) {
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
w.Header().Set("Content-Type", "application/json")
|
||||
_, _ = fmt.Fprint(w, `{"code":200,"hint":"登录武林系统","results":{"list":[]}}`)
|
||||
}))
|
||||
defer srv.Close()
|
||||
|
||||
kw := "登录武林系统"
|
||||
header := ""
|
||||
src := &model.ReaderBookSource{Name: "校验关键字测试源", SourceURL: srv.URL}
|
||||
bs := &BookSource{
|
||||
BookSourceURL: srv.URL,
|
||||
Header: &header,
|
||||
SearchURL: strPtr(srv.URL + "/search?q={{key}}"),
|
||||
RuleSearch: &SearchRule{
|
||||
BookList: strPtr("$.results.list[*]"),
|
||||
Name: strPtr("$.name"),
|
||||
BookURL: strPtr("$.url"),
|
||||
CheckKeyWord: &kw,
|
||||
},
|
||||
}
|
||||
|
||||
svc, _ := newLoginTestService(t)
|
||||
books, err := svc.searchInSource(t.Context(), src, bs, "火影", 1)
|
||||
if err != nil {
|
||||
t.Fatalf("站点正常时不该报错: %v", err)
|
||||
}
|
||||
if len(books) != 0 {
|
||||
t.Fatalf("结果数 = %d,期望 0", len(books))
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,281 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"io"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"strings"
|
||||
"testing"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/service/reader/rule"
|
||||
)
|
||||
|
||||
const e2eBookHTML = `<!DOCTYPE html>
|
||||
<html><body>
|
||||
<div class="box" id="main">
|
||||
<div class="item"><h3><a href="/book/1">斗破苍穹</a></h3><span class="author">天蚕土豆</span></div>
|
||||
<div class="item"><h3><a href="/book/2">凡人修仙传</a></h3><span class="author">忘语</span></div>
|
||||
</div>
|
||||
</body></html>`
|
||||
|
||||
const e2eBookInfoHTML = `<html><body>
|
||||
<div class="info"><h1>斗破苍穹</h1><span class="author">天蚕土豆</span>
|
||||
<p class="intro">三十年河东三十年河西</p>
|
||||
<a class="toc" href="/book/1/toc.html">查看目录</a></div>
|
||||
</body></html>`
|
||||
|
||||
const e2eTocHTML = `<html><body>
|
||||
<ul class="chapters">
|
||||
<li class="vol">第一卷</li>
|
||||
<li><a href="/book/1/c1.html">第一章 陨落的天才</a></li>
|
||||
<li><a href="/book/1/c2.html">第二章 斗气大陆</a></li>
|
||||
</ul>
|
||||
</body></html>`
|
||||
|
||||
const e2eContentHTML = `<html><body><div id="content"> 魂殿来犯,<br>萧炎浴火重生。 </div>
|
||||
<a class="next" href="/book/1/c1_2.html">下一页</a></body></html>`
|
||||
|
||||
const e2eContentPage2HTML = `<html><body><div id="content">少女微凉的手掌传来。</div></body></html>`
|
||||
|
||||
// e2eServer 模拟一个完整的书源站点:搜索/详情/目录/正文(含 nextContentUrl 翻页)。
|
||||
func e2eServer() *httptest.Server {
|
||||
return httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
w.Header().Set("Content-Type", "text/html; charset=utf-8")
|
||||
switch {
|
||||
case strings.HasPrefix(r.URL.Path, "/search/"):
|
||||
_, _ = w.Write([]byte(e2eBookHTML))
|
||||
case strings.HasPrefix(r.URL.Path, "/book/1/toc"):
|
||||
_, _ = w.Write([]byte(e2eTocHTML))
|
||||
case strings.HasPrefix(r.URL.Path, "/book/1/c1_2"):
|
||||
_, _ = w.Write([]byte(e2eContentPage2HTML))
|
||||
case strings.HasPrefix(r.URL.Path, "/book/1/c"):
|
||||
_, _ = w.Write([]byte(e2eContentHTML))
|
||||
case strings.HasPrefix(r.URL.Path, "/book/"):
|
||||
_, _ = w.Write([]byte(e2eBookInfoHTML))
|
||||
default:
|
||||
http.NotFound(w, r)
|
||||
}
|
||||
}))
|
||||
}
|
||||
|
||||
// e2eSource 对齐 legado 书源 JSON 结构,覆盖 jsoup 全部四段规则。
|
||||
func e2eSourceJSON(server string) string {
|
||||
return fmt.Sprintf(`{
|
||||
"bookSourceUrl": %q,
|
||||
"bookSourceName": "测试源",
|
||||
"bookSourceType": 0,
|
||||
"searchUrl": "%s/search/{{key}}/1.html",
|
||||
"ruleSearch": {
|
||||
"bookList": "class.item",
|
||||
"name": "tag.h3@tag.a@text",
|
||||
"bookUrl": "tag.h3@tag.a@href",
|
||||
"author": "class.author@text"
|
||||
},
|
||||
"ruleBookInfo": {
|
||||
"name": "class.info@tag.h1@text",
|
||||
"author": "class.info@class.author@text",
|
||||
"intro": "class.info@class.intro@text",
|
||||
"tocUrl": "class.info@class.toc@href"
|
||||
},
|
||||
"ruleToc": {
|
||||
"chapterList": "class.chapters@tag.li",
|
||||
"chapterName": "tag.a@text",
|
||||
"chapterUrl": "tag.a@href",
|
||||
"isVolume": "tag.a@text"
|
||||
},
|
||||
"ruleContent": {
|
||||
"content": "id.content@textNodes",
|
||||
"nextContentUrl": "class.next@href"
|
||||
}
|
||||
}`, server, server)
|
||||
}
|
||||
|
||||
// newE2EEngine 执行与 ReaderService 相同的链路(不含 DB):
|
||||
// ParseAnalyzeUrl → execute → AnalyzeRule。
|
||||
type e2eEngine struct {
|
||||
server string
|
||||
client *http.Client
|
||||
}
|
||||
|
||||
func (e *e2eEngine) fetch(t *testing.T, urlRule, key string, page int) (*rule.AnalyzeRule, string) {
|
||||
t.Helper()
|
||||
req, err := rule.ParseAnalyzeUrl(urlRule, key, page, e.server)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if req.Unsupported != nil {
|
||||
t.Fatalf("unsupported: %v", req.Unsupported)
|
||||
}
|
||||
httpReq, _ := http.NewRequest("GET", req.URL, nil)
|
||||
for k, v := range req.Headers {
|
||||
httpReq.Header.Set(k, v)
|
||||
}
|
||||
resp, err := e.client.Do(httpReq)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
body, _ := io.ReadAll(resp.Body)
|
||||
ar := rule.NewAnalyzeRule()
|
||||
ar.SetContent(string(body), resp.Request.URL.String())
|
||||
return ar, resp.Request.URL.String()
|
||||
}
|
||||
|
||||
func TestEndToEndSourceChain(t *testing.T) {
|
||||
srv := e2eServer()
|
||||
defer srv.Close()
|
||||
|
||||
bs, err := ParseBookSource(e2eSourceJSON(srv.URL))
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
engine := &e2eEngine{server: srv.URL, client: srv.Client()}
|
||||
|
||||
// ── 搜索 ──
|
||||
ar, _ := engine.fetch(t, SPtr(bs.SearchURL), "斗罗", 1)
|
||||
sr := bs.RuleSearch
|
||||
els, err := ar.GetElements(SPtr(sr.BookList))
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if len(els) != 2 {
|
||||
t.Fatalf("search elements = %d", len(els))
|
||||
}
|
||||
name, _ := ar.GetString(SPtr(sr.Name), els[0], false)
|
||||
if name != "斗破苍穹" {
|
||||
t.Fatalf("search name = %q", name)
|
||||
}
|
||||
bookURL, _ := ar.GetString(SPtr(sr.BookURL), els[0], true)
|
||||
if !strings.Contains(bookURL, "/book/1") {
|
||||
t.Fatalf("bookURL = %q", bookURL)
|
||||
}
|
||||
|
||||
// ── 详情 ──
|
||||
ar, _ = engine.fetch(t, bookURL, "", 0)
|
||||
bir := bs.RuleBookInfo
|
||||
if got, _ := ar.GetString(SPtr(bir.Name), nil, false); got != "斗破苍穹" {
|
||||
t.Fatalf("info name = %q", got)
|
||||
}
|
||||
if got, _ := ar.GetString(SPtr(bir.Intro), nil, false); got != "三十年河东三十年河西" {
|
||||
t.Fatalf("info intro = %q", got)
|
||||
}
|
||||
tocURL, _ := ar.GetString(SPtr(bir.TocURL), nil, true)
|
||||
if !strings.Contains(tocURL, "/book/1/toc") {
|
||||
t.Fatalf("tocURL = %q", tocURL)
|
||||
}
|
||||
|
||||
// ── 目录 ──
|
||||
ar, _ = engine.fetch(t, tocURL, "", 0)
|
||||
tr := bs.RuleToc
|
||||
chEls, err := ar.GetElements(SPtr(tr.ChapterList))
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if len(chEls) != 3 {
|
||||
t.Fatalf("chapters = %d", len(chEls))
|
||||
}
|
||||
var chapters []TocChapter
|
||||
for i, el := range chEls {
|
||||
title, _ := ar.GetString(SPtr(tr.ChapterName), el, false)
|
||||
if title == "" {
|
||||
continue // 卷名行没有 <a>,取不到标题(服务层同样跳过)
|
||||
}
|
||||
url, _ := ar.GetString(SPtr(tr.ChapterURL), el, true)
|
||||
chapters = append(chapters, TocChapter{Index: i, Title: title, URL: url})
|
||||
}
|
||||
if len(chapters) != 2 {
|
||||
t.Fatalf("chapters = %d", len(chapters))
|
||||
}
|
||||
if chapters[0].Title != "第一章 陨落的天才" {
|
||||
t.Fatalf("chapter0 = %+v", chapters[0])
|
||||
}
|
||||
|
||||
// ── 正文(含 nextContentUrl 翻页合并) ──
|
||||
var parts []string
|
||||
url := chapters[0].URL
|
||||
for i := 0; i < 5; i++ {
|
||||
ar, finalURL := engine.fetch(t, url, "", 0)
|
||||
list, err := ar.GetStringList(SPtr(bs.RuleContent.Content), nil, false)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
parts = append(parts, strings.Join(list, "\n"))
|
||||
next, _ := ar.GetString(SPtr(bs.RuleContent.NextContentURL), nil, true)
|
||||
if next == "" || next == url || next == finalURL {
|
||||
break
|
||||
}
|
||||
url = next
|
||||
}
|
||||
content := strings.Join(parts, "\n")
|
||||
if !strings.Contains(content, "萧炎浴火重生") || !strings.Contains(content, "少女微凉的手掌") {
|
||||
t.Fatalf("content = %q", content)
|
||||
}
|
||||
}
|
||||
|
||||
// TestGetContentForBookEmptyContentFails 书源一条内容都没返回时不能当成功下发。
|
||||
//
|
||||
// 回归:聚合类书源把七条线路全试完会返回空串(光遇聚合的 request() 就是这么写的),
|
||||
// 以前这种空正文会当成功下发,前端渲染成一张白页并缓存下来,读者只能干等。
|
||||
func TestGetContentForBookEmptyContentFails(t *testing.T) {
|
||||
svc, bookID := setupTextChapterBook(t, "")
|
||||
|
||||
_, err := svc.GetContentForBook(t.Context(), "u1", bookID, 0)
|
||||
if err == nil {
|
||||
t.Fatal("空正文应当报错,实际当成功下发了")
|
||||
}
|
||||
if !strings.Contains(err.Error(), "正文为空") {
|
||||
t.Fatalf("错误信息 = %q,期望提示正文为空", err.Error())
|
||||
}
|
||||
}
|
||||
|
||||
// ─── 纯函数测试:导入识别 / 搜索合并 ────────────────────────────────────────
|
||||
|
||||
func TestParseSourcePayload(t *testing.T) {
|
||||
// 数组
|
||||
arr := `[{"bookSourceUrl":"http://a.com","bookSourceName":"A"},{"bookSourceUrl":"http://b.com","bookSourceName":"B"}]`
|
||||
if got := ParseSourcePayload(arr); len(got) != 2 {
|
||||
t.Fatalf("array payload = %d", len(got))
|
||||
}
|
||||
// 单对象
|
||||
single := `{"bookSourceUrl":"http://a.com","bookSourceName":"A"}`
|
||||
if got := ParseSourcePayload(single); len(got) != 1 {
|
||||
t.Fatalf("single payload = %d", len(got))
|
||||
}
|
||||
// Base64
|
||||
b64 := base64StdEncode(single)
|
||||
if got := ParseSourcePayload(b64); len(got) != 1 {
|
||||
t.Fatalf("base64 payload = %d", len(got))
|
||||
}
|
||||
// UTF-8 BOM(Windows 导出文件常见)
|
||||
if got := ParseSourcePayload("\uFEFF" + arr); len(got) != 2 {
|
||||
t.Fatalf("bom array payload = %d", len(got))
|
||||
}
|
||||
}
|
||||
|
||||
func TestMergeSearchResults(t *testing.T) {
|
||||
hit := func(name, author, origin string) searchHit {
|
||||
return searchHit{book: SearchBook{
|
||||
Name: name, Author: author,
|
||||
Origins: []SearchOrigin{{OriginName: origin, BookURL: "u"}},
|
||||
}}
|
||||
}
|
||||
hits := []searchHit{
|
||||
hit("遮天", "辰东", "源C"),
|
||||
hit("斗破苍穹", "天蚕土豆", "源B"),
|
||||
hit("斗破苍穹", "天蚕土豆", "源A"),
|
||||
hit("斗罗大陆", "唐家三少", "源D"),
|
||||
}
|
||||
merged := mergeSearchResults(hits, "斗")
|
||||
if len(merged) != 3 {
|
||||
t.Fatalf("merged = %d", len(merged))
|
||||
}
|
||||
// 精确命中「斗破苍穹」应排第一(tier 0),且合并两源
|
||||
if merged[0].Name != "斗破苍穹" || len(merged[0].Origins) != 2 {
|
||||
t.Fatalf("first = %+v", merged[0])
|
||||
}
|
||||
}
|
||||
|
||||
func base64StdEncode(s string) string {
|
||||
return base64EncodeStr(s)
|
||||
}
|
||||
@@ -0,0 +1,134 @@
|
||||
package reader
|
||||
|
||||
import (
|
||||
"context"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"strings"
|
||||
"testing"
|
||||
|
||||
"github.com/truewhile/MeBox/internal/model"
|
||||
)
|
||||
|
||||
// 本文件:「更新目录」的结果分档回归测试。
|
||||
//
|
||||
// 背景:书架提示曾经把「抓到了目录但没有新章节」也算成失败。完结书每次检查都没有新章节,
|
||||
// 于是书架天天报「N 本失败」,看起来像书源坏了。现在必须区分三档:
|
||||
// 抓到新章节 / 抓到但没变(已是最新)/ 抓取或写入失败。
|
||||
|
||||
// refreshTocForTest 跑一次「更新目录」并返回汇总。
|
||||
func refreshTocForTest(t *testing.T, svc *ReaderService, ctx context.Context) *TocRefreshResult {
|
||||
t.Helper()
|
||||
res, err := svc.RefreshBooksToc(ctx, "u1")
|
||||
if err != nil {
|
||||
t.Fatalf("更新目录失败: %v", err)
|
||||
}
|
||||
return res
|
||||
}
|
||||
|
||||
// newRefreshTocServer 建一个目录页内容可切换的测试源站点。
|
||||
func newRefreshTocServer(tocHandler *http.HandlerFunc) *httptest.Server {
|
||||
return httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
w.Header().Set("Content-Type", "text/html; charset=utf-8")
|
||||
switch {
|
||||
case strings.HasPrefix(r.URL.Path, "/book/1/toc"):
|
||||
(*tocHandler)(w, r)
|
||||
case strings.HasPrefix(r.URL.Path, "/book/"):
|
||||
_, _ = w.Write([]byte(e2eBookInfoHTML))
|
||||
default:
|
||||
http.NotFound(w, r)
|
||||
}
|
||||
}))
|
||||
}
|
||||
|
||||
func TestRefreshBooksTocSeparatesUnchangedFromFailed(t *testing.T) {
|
||||
tocHandler := http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
|
||||
_, _ = w.Write([]byte(e2eTocHTML))
|
||||
})
|
||||
srv := newRefreshTocServer(&tocHandler)
|
||||
defer srv.Close()
|
||||
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
importTestSource(t, svc, e2eSourceJSON(srv.URL), srv.URL)
|
||||
|
||||
book := &model.ReaderBook{
|
||||
UserID: "u1",
|
||||
Origin: srv.URL,
|
||||
OriginName: "测试源",
|
||||
BookURL: srv.URL + "/book/1",
|
||||
TocURL: srv.URL + "/book/1/toc.html",
|
||||
Name: "斗破苍穹",
|
||||
}
|
||||
if err := svc.repo.CreateBook(ctx, book); err != nil {
|
||||
t.Fatalf("创建书籍失败: %v", err)
|
||||
}
|
||||
|
||||
// 1) 首次刷新:抓到 2 章并落库。书架原先没有目录缓存,没有可比基准,算「已是最新」。
|
||||
res := refreshTocForTest(t, svc, ctx)
|
||||
if res.Total != 1 || res.Updated != 0 || res.Unchanged != 1 || res.Failed != 0 {
|
||||
t.Fatalf("首次刷新 = %+v,期望 total 1 / updated 0 / unchanged 1 / failed 0", res)
|
||||
}
|
||||
chapters, err := svc.ListChapters(ctx, book.ID)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if len(chapters) != 2 {
|
||||
t.Fatalf("目录缓存 = %d 章,期望 2", len(chapters))
|
||||
}
|
||||
|
||||
// 2) 站点多出一章:这才算「有新章节」。
|
||||
tocHandler = func(w http.ResponseWriter, _ *http.Request) {
|
||||
_, _ = w.Write([]byte(strings.Replace(
|
||||
e2eTocHTML, "</ul>", `<li><a href="/book/1/c3.html">第三章 药老</a></li></ul>`, 1,
|
||||
)))
|
||||
}
|
||||
res = refreshTocForTest(t, svc, ctx)
|
||||
if res.Updated != 1 || res.Unchanged != 0 || res.Failed != 0 {
|
||||
t.Fatalf("出现新章节时 = %+v,期望 updated 1 / unchanged 0 / failed 0", res)
|
||||
}
|
||||
|
||||
// 3) 站点返回风控响应体(HTTP 200 但不是目录):规则解析成空列表且不报错,
|
||||
// 这种静默失败必须记失败,不能报成「已是最新」。
|
||||
tocHandler = func(w http.ResponseWriter, _ *http.Request) {
|
||||
_, _ = w.Write([]byte(`{"code":210,"message":"請到官網更新最新APP,請下載安裝正版之後等待1小時。"}`))
|
||||
}
|
||||
res = refreshTocForTest(t, svc, ctx)
|
||||
if res.Failed != 1 || res.Unchanged != 0 || res.Updated != 0 {
|
||||
t.Fatalf("风控响应体 = %+v,期望 updated 0 / unchanged 0 / failed 1", res)
|
||||
}
|
||||
// 抓失败不能动已有缓存,否则阅读器会连本来能看的章节都读不到。
|
||||
chapters, err = svc.ListChapters(ctx, book.ID)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if len(chapters) != 3 {
|
||||
t.Fatalf("抓失败后目录缓存 = %d 章,期望保持 3 章", len(chapters))
|
||||
}
|
||||
|
||||
// 4) 站点 404:抓取失败。
|
||||
tocHandler = func(w http.ResponseWriter, r *http.Request) { http.NotFound(w, r) }
|
||||
res = refreshTocForTest(t, svc, ctx)
|
||||
if res.Failed != 1 || res.Unchanged != 0 {
|
||||
t.Fatalf("站点 404 = %+v,期望 unchanged 0 / failed 1", res)
|
||||
}
|
||||
}
|
||||
|
||||
// 本地书籍与没有书源信息的书籍不参与刷新,汇总里也不该出现。
|
||||
func TestRefreshBooksTocSkipsLocalAndSourcelessBooks(t *testing.T) {
|
||||
svc, _ := newLoginTestService(t)
|
||||
ctx := t.Context()
|
||||
|
||||
local := &model.ReaderBook{UserID: "u1", Name: "本地书", LocalPath: "/tmp/a.txt", BookURL: "x", Origin: "y"}
|
||||
sourceless := &model.ReaderBook{UserID: "u1", Name: "无书源", BookURL: "", Origin: ""}
|
||||
for _, b := range []*model.ReaderBook{local, sourceless} {
|
||||
if err := svc.repo.CreateBook(ctx, b); err != nil {
|
||||
t.Fatalf("创建书籍失败: %v", err)
|
||||
}
|
||||
}
|
||||
|
||||
res := refreshTocForTest(t, svc, ctx)
|
||||
if res.Total != 0 || res.Updated != 0 || res.Unchanged != 0 || res.Failed != 0 {
|
||||
t.Fatalf("跳过本地/无源书籍时 = %+v,期望全 0", res)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,661 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"net/url"
|
||||
"regexp"
|
||||
"strings"
|
||||
|
||||
"github.com/PaesslerAG/jsonpath"
|
||||
"golang.org/x/net/html"
|
||||
)
|
||||
|
||||
// 本文件对应 AnalyzeRule.kt 主类。
|
||||
|
||||
// AnalyzeRule 解析规则获取结果(对应 AnalyzeRule)。
|
||||
type AnalyzeRule struct {
|
||||
content any
|
||||
baseUrl string
|
||||
redirectURL *url.URL
|
||||
isJSON bool
|
||||
isRegex bool
|
||||
|
||||
// jsRunner 由 P2 阶段的 goja 引擎注入;nil 时 JS 规则报 ErrJsUnsupported。
|
||||
jsRunner func(js string, result any) (any, error)
|
||||
|
||||
// 变量层级(对应 chapter → book → ruleData → source)
|
||||
chapterVars map[string]string
|
||||
bookVars map[string]string
|
||||
vars map[string]string
|
||||
chapterTitle string
|
||||
chapterIndex int
|
||||
bookName string
|
||||
// bookMeta 书籍元数据(对应 legado 规则 JS 里的 Book 实体字段)。
|
||||
bookMeta map[string]any
|
||||
// bookCustom 书籍自定义变量(对应 legado Book.variableMap),
|
||||
// 由规则 JS 的 book.getVariable / book.putVariable 读写。
|
||||
bookCustom map[string]string
|
||||
// bookTypeOverride 书源在规则 JS 里给 book.type 赋的值
|
||||
// (听书/漫画/短剧源靠它声明书籍类型),由服务层读回。
|
||||
bookTypeOverride *int
|
||||
sourceGetter func(key string) string
|
||||
sourcePutter func(key, value string)
|
||||
|
||||
ruleCache map[string][]*SourceRule
|
||||
}
|
||||
|
||||
// NewAnalyzeRule 创建解析器。
|
||||
func NewAnalyzeRule() *AnalyzeRule {
|
||||
return &AnalyzeRule{ruleCache: map[string][]*SourceRule{}}
|
||||
}
|
||||
|
||||
// SetJSRunner 注入 JS 执行器(P2)。
|
||||
func (a *AnalyzeRule) SetJSRunner(runner func(js string, result any) (any, error)) {
|
||||
a.jsRunner = runner
|
||||
}
|
||||
|
||||
// SetContent 对应 setContent(content, baseUrl)。
|
||||
func (a *AnalyzeRule) SetContent(content any, baseUrl string) *AnalyzeRule {
|
||||
a.content = content
|
||||
switch content.(type) {
|
||||
case *html.Node:
|
||||
a.isJSON = false
|
||||
default:
|
||||
a.isJSON = LooksLikeJSON(anyToString(content))
|
||||
}
|
||||
if baseUrl != "" {
|
||||
a.baseUrl = baseUrl
|
||||
}
|
||||
return a
|
||||
}
|
||||
|
||||
// SetBaseUrl 对应 setBaseUrl。
|
||||
func (a *AnalyzeRule) SetBaseUrl(baseUrl string) *AnalyzeRule {
|
||||
if baseUrl != "" {
|
||||
a.baseUrl = baseUrl
|
||||
}
|
||||
return a
|
||||
}
|
||||
|
||||
// SetRedirectUrl 对应 setRedirectUrl。
|
||||
func (a *AnalyzeRule) SetRedirectUrl(u string) *AnalyzeRule {
|
||||
if strings.HasPrefix(u, "data:") {
|
||||
return a
|
||||
}
|
||||
if parsed, err := url.Parse(u); err == nil {
|
||||
a.redirectURL = parsed
|
||||
}
|
||||
return a
|
||||
}
|
||||
|
||||
// SetChapterContext 设置章节上下文(title 与章节级变量存储)。
|
||||
func (a *AnalyzeRule) SetChapterContext(title string, vars map[string]string) {
|
||||
a.chapterTitle = title
|
||||
if vars != nil {
|
||||
a.chapterVars = vars
|
||||
}
|
||||
}
|
||||
|
||||
// SetChapterIndex 设置当前章节下标(规则 JS 的 chapter.index)。
|
||||
func (a *AnalyzeRule) SetChapterIndex(i int) { a.chapterIndex = i }
|
||||
|
||||
// SetBookMeta 注入书籍元数据(legado 的 Book 实体字段),
|
||||
// 供规则 JS 里的 `book` 对象读取(name/author/coverUrl/durChapterIndex…)。
|
||||
func (a *AnalyzeRule) SetBookMeta(meta map[string]any) {
|
||||
if len(meta) == 0 {
|
||||
return
|
||||
}
|
||||
if a.bookMeta == nil {
|
||||
a.bookMeta = make(map[string]any, len(meta))
|
||||
}
|
||||
for k, v := range meta {
|
||||
a.bookMeta[k] = v
|
||||
}
|
||||
}
|
||||
|
||||
// SetBookCustomVars 注入书籍自定义变量(对应 legado Book.variableMap),
|
||||
// 规则 JS 通过 book.getVariable / book.putVariable 读写。
|
||||
func (a *AnalyzeRule) SetBookCustomVars(vars map[string]string) {
|
||||
if vars != nil {
|
||||
a.bookCustom = vars
|
||||
}
|
||||
}
|
||||
|
||||
// SetBookType 记录书源声明的书籍类型(legado Book.type)。
|
||||
func (a *AnalyzeRule) SetBookType(t int) {
|
||||
v := t
|
||||
a.bookTypeOverride = &v
|
||||
}
|
||||
|
||||
// BookTypeOverride 返回书源在规则 JS 里声明的书籍类型;
|
||||
// ok 为 false 表示书源没有声明(应沿用书架记录里的类型)。
|
||||
func (a *AnalyzeRule) BookTypeOverride() (int, bool) {
|
||||
if a.bookTypeOverride == nil {
|
||||
return 0, false
|
||||
}
|
||||
return *a.bookTypeOverride, true
|
||||
}
|
||||
|
||||
// bookTypeValue 供 book.type 读取:优先书源本次声明的值,否则用元数据里的。
|
||||
func (a *AnalyzeRule) bookTypeValue() any {
|
||||
if a.bookTypeOverride != nil {
|
||||
return *a.bookTypeOverride
|
||||
}
|
||||
if v, ok := a.bookMeta["type"]; ok {
|
||||
return v
|
||||
}
|
||||
return 0
|
||||
}
|
||||
|
||||
// SetBookContext 设置书籍上下文(name 与书籍级变量存储)。
|
||||
func (a *AnalyzeRule) SetBookContext(name string, vars map[string]string) {
|
||||
a.bookName = name
|
||||
if vars != nil {
|
||||
a.bookVars = vars
|
||||
}
|
||||
}
|
||||
|
||||
// SetSourceVariables 注入书源级变量读写(source.variableMap)。
|
||||
func (a *AnalyzeRule) SetSourceVariables(getter func(key string) string, putter func(key, value string)) {
|
||||
a.sourceGetter = getter
|
||||
a.sourcePutter = putter
|
||||
}
|
||||
|
||||
// jsonpathGet 包装 PaesslerAG/jsonpath.Get。
|
||||
func jsonpathGet(path string, root any) (v any, err error) {
|
||||
defer func() {
|
||||
if r := recover(); r != nil {
|
||||
v, err = nil, ErrJsUnsupported
|
||||
}
|
||||
}()
|
||||
return jsonpath.Get(path, root)
|
||||
}
|
||||
|
||||
// ─── 变量存取(对应 put/get) ───────────────────────────────────────────────
|
||||
|
||||
// Put 对应 put(key, value):chapter → book → 局部 → source。
|
||||
func (a *AnalyzeRule) Put(key, value string) string {
|
||||
switch {
|
||||
case a.chapterVars != nil:
|
||||
a.chapterVars[key] = value
|
||||
case a.bookVars != nil:
|
||||
a.bookVars[key] = value
|
||||
default:
|
||||
if a.vars == nil {
|
||||
a.vars = map[string]string{}
|
||||
}
|
||||
a.vars[key] = value
|
||||
}
|
||||
return value
|
||||
}
|
||||
|
||||
// Get 对应 get(key):特殊键 bookName/title 优先取上下文。
|
||||
func (a *AnalyzeRule) Get(key string) string {
|
||||
switch key {
|
||||
case "bookName":
|
||||
if a.bookName != "" {
|
||||
return a.bookName
|
||||
}
|
||||
case "title":
|
||||
if a.chapterTitle != "" {
|
||||
return a.chapterTitle
|
||||
}
|
||||
}
|
||||
for _, store := range []map[string]string{a.chapterVars, a.bookVars, a.vars} {
|
||||
if store != nil {
|
||||
if v, ok := store[key]; ok && v != "" {
|
||||
return v
|
||||
}
|
||||
}
|
||||
}
|
||||
if a.sourceGetter != nil {
|
||||
if v := a.sourceGetter(key); v != "" {
|
||||
return v
|
||||
}
|
||||
}
|
||||
return ""
|
||||
}
|
||||
|
||||
// ─── JS ────────────────────────────────────────────────────────────────────
|
||||
|
||||
func (a *AnalyzeRule) evalJS(js string, result any) (any, error) {
|
||||
if a.jsRunner == nil {
|
||||
return nil, ErrJsUnsupported
|
||||
}
|
||||
return a.jsRunner(js, result)
|
||||
}
|
||||
|
||||
// ─── 规则拆分缓存 ──────────────────────────────────────────────────────────
|
||||
|
||||
func (a *AnalyzeRule) splitSourceRuleCached(ruleStr string) []*SourceRule {
|
||||
if ruleStr == "" {
|
||||
return nil
|
||||
}
|
||||
if cached, ok := a.ruleCache[ruleStr]; ok {
|
||||
return cached
|
||||
}
|
||||
rules := SplitSourceRule(ruleStr, false, a.isJSON, &a.isRegex)
|
||||
a.ruleCache[ruleStr] = rules
|
||||
return rules
|
||||
}
|
||||
|
||||
func (a *AnalyzeRule) putRule(putMap map[string]string) error {
|
||||
for k, v := range putMap {
|
||||
s, err := a.GetString(v, nil, false)
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
a.Put(k, s)
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
// makeDeps 构造 MakeUpRule 展开内嵌 {{...}} 需要的执行环境。
|
||||
//
|
||||
// content 是当前正在解析的元素。legado 的内嵌规则({{$.x}} / {{@x}})在**当前元素**
|
||||
// 上求值:书源普遍用 `{{$.status}},{{$.score}}` 这种模板拼 kind、用
|
||||
// `{{$.source}} {{$.last_chapter_title}}` 拼最新章节。若拿整份响应去求值,
|
||||
// 这些模板会全部取空——表现为 kind=",,,"、"最新"为空(聚合源尤其明显)。
|
||||
func (a *AnalyzeRule) makeDeps(content any) *RuleDeps {
|
||||
return &RuleDeps{
|
||||
JS: a.evalJS,
|
||||
Rule: func(rule string) (string, error) { return a.GetString(rule, content, false) },
|
||||
Get: a.Get,
|
||||
}
|
||||
}
|
||||
|
||||
// ─── 主流程 ────────────────────────────────────────────────────────────────
|
||||
|
||||
// GetStringList 对应 getStringList(rule, mContent, isUrl)。
|
||||
func (a *AnalyzeRule) GetStringList(ruleStr string, mContent any, isUrl bool) ([]string, error) {
|
||||
if ruleStr == "" {
|
||||
return nil, nil
|
||||
}
|
||||
ruleList := a.splitSourceRuleCached(ruleStr)
|
||||
return a.getStringListRules(ruleList, mContent, isUrl)
|
||||
}
|
||||
|
||||
func (a *AnalyzeRule) getStringListRules(ruleList []*SourceRule, mContent any, isUrl bool) ([]string, error) {
|
||||
var result any
|
||||
content := mContent
|
||||
if content == nil {
|
||||
content = a.content
|
||||
}
|
||||
if content != nil && len(ruleList) > 0 {
|
||||
result = content
|
||||
for _, sourceRule := range ruleList {
|
||||
if err := a.putRule(sourceRule.putMap); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
resolved, err := sourceRule.MakeUpRule(result, a.makeDeps(result))
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
if result == nil {
|
||||
continue
|
||||
}
|
||||
rule := resolved.Rule
|
||||
if rule != "" || resolved.ReplaceRegex == "" {
|
||||
switch sourceRule.Mode {
|
||||
case ModeWebJs:
|
||||
return nil, ErrWebJSUnsupported
|
||||
case ModeJs:
|
||||
result, err = a.evalJS(rule, result)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
case ModeJson:
|
||||
result = newJSONAnalyzer(result).getStringList(rule)
|
||||
case ModeXPath:
|
||||
result = newXPathAnalyzer(result).getStringList(rule)
|
||||
case ModeDefault:
|
||||
result = newJsoupAnalyzer(result).getStringList(rule)
|
||||
default:
|
||||
result = rule
|
||||
}
|
||||
}
|
||||
if resolved.ReplaceRegex != "" {
|
||||
if lst, ok := result.([]string); ok {
|
||||
out := make([]string, 0, len(lst))
|
||||
for _, item := range lst {
|
||||
out = append(out, applyReplaceRegex(item, resolved))
|
||||
}
|
||||
result = out
|
||||
} else {
|
||||
result = applyReplaceRegex(resultString(result), resolved)
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
if result == nil {
|
||||
return nil, nil
|
||||
}
|
||||
if s, ok := result.(string); ok {
|
||||
result = strings.Split(s, "\n")
|
||||
}
|
||||
if isUrl {
|
||||
// 这里必须接受任何切片形态:书源的下一页地址/章节地址规则常常由 JS 生成,
|
||||
// goja 导出的是 []any。早期只认 []string,JS 给的地址数组会被静默丢掉,
|
||||
// 表现为「nextTocUrl 明明有规则却永远不翻页」。
|
||||
var urlList []string
|
||||
for _, u := range resultStrings(result) {
|
||||
abs := a.absolutize(u)
|
||||
if abs != "" && !containsStr(urlList, abs) {
|
||||
urlList = append(urlList, abs)
|
||||
}
|
||||
}
|
||||
return urlList, nil
|
||||
}
|
||||
switch t := result.(type) {
|
||||
case []string:
|
||||
return t, nil
|
||||
case []any:
|
||||
out := make([]string, len(t))
|
||||
for i, v := range t {
|
||||
out[i] = anyToString(v)
|
||||
}
|
||||
return out, nil
|
||||
case []*html.Node:
|
||||
out := make([]string, len(t))
|
||||
for i, v := range t {
|
||||
out[i] = outerHTML(v)
|
||||
}
|
||||
return out, nil
|
||||
default:
|
||||
return []string{anyToString(result)}, nil
|
||||
}
|
||||
}
|
||||
|
||||
// resultStrings 把任意分析结果归一成字符串切片([]string / []any / []*html.Node / 单值)。
|
||||
func resultStrings(result any) []string {
|
||||
switch t := result.(type) {
|
||||
case nil:
|
||||
return nil
|
||||
case []string:
|
||||
return t
|
||||
case []any:
|
||||
out := make([]string, 0, len(t))
|
||||
for _, v := range t {
|
||||
out = append(out, anyToString(v))
|
||||
}
|
||||
return out
|
||||
case []*html.Node:
|
||||
out := make([]string, 0, len(t))
|
||||
for _, v := range t {
|
||||
out = append(out, outerHTML(v))
|
||||
}
|
||||
return out
|
||||
case string:
|
||||
return strings.Split(t, "\n")
|
||||
default:
|
||||
return []string{anyToString(result)}
|
||||
}
|
||||
}
|
||||
|
||||
// GetString 对应 getString(rule, mContent, isUrl)。
|
||||
func (a *AnalyzeRule) GetString(ruleStr string, mContent any, isUrl bool) (string, error) {
|
||||
if ruleStr == "" {
|
||||
return "", nil
|
||||
}
|
||||
ruleList := a.splitSourceRuleCached(ruleStr)
|
||||
return a.getStringRules(ruleList, mContent, isUrl)
|
||||
}
|
||||
|
||||
func (a *AnalyzeRule) getStringRules(ruleList []*SourceRule, mContent any, isUrl bool) (string, error) {
|
||||
var result any
|
||||
content := mContent
|
||||
if content == nil {
|
||||
content = a.content
|
||||
}
|
||||
if content != nil && len(ruleList) > 0 {
|
||||
result = content
|
||||
for _, sourceRule := range ruleList {
|
||||
if err := a.putRule(sourceRule.putMap); err != nil {
|
||||
return "", err
|
||||
}
|
||||
resolved, err := sourceRule.MakeUpRule(result, a.makeDeps(result))
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
if result == nil {
|
||||
continue
|
||||
}
|
||||
rule := resolved.Rule
|
||||
if rule != "" || resolved.ReplaceRegex == "" {
|
||||
switch sourceRule.Mode {
|
||||
case ModeWebJs:
|
||||
return "", ErrWebJSUnsupported
|
||||
case ModeJs:
|
||||
result, err = a.evalJS(rule, result)
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
case ModeJson:
|
||||
result = newJSONAnalyzer(result).getString(rule)
|
||||
case ModeXPath:
|
||||
result = newXPathAnalyzer(result).getString(rule)
|
||||
case ModeDefault:
|
||||
if isUrl {
|
||||
result = newJsoupAnalyzer(result).getString0(rule)
|
||||
} else {
|
||||
result = newJsoupAnalyzer(result).getString(rule)
|
||||
}
|
||||
default:
|
||||
result = rule
|
||||
}
|
||||
}
|
||||
if result != nil && resolved.ReplaceRegex != "" {
|
||||
result = applyReplaceRegex(resultString(result), resolved)
|
||||
}
|
||||
}
|
||||
}
|
||||
if result == nil {
|
||||
result = ""
|
||||
}
|
||||
str := unescapeRuleEntities(resultString(result))
|
||||
if isUrl {
|
||||
if strings.TrimSpace(str) == "" {
|
||||
// 对应 legado:取值为空时回退 baseUrl,让相对地址还能解析。
|
||||
// 但 baseUrl 未必是地址:聚合类书源(如「光遇聚合」)的搜索请求地址
|
||||
// 本身就是 data:;base64,... 参数信封,直接回退会把信封当成封面/书址
|
||||
// 返回,前端 <img> 只能显示破图。故仅在 baseUrl 是 http(s) 地址时回退。
|
||||
if base := strings.TrimSpace(a.baseUrl); isAbsURL(base) {
|
||||
return base, nil
|
||||
}
|
||||
return "", nil
|
||||
}
|
||||
return a.absolutize(str), nil
|
||||
}
|
||||
return str, nil
|
||||
}
|
||||
|
||||
// GetElement 对应 getElement(ruleStr)。
|
||||
func (a *AnalyzeRule) GetElement(ruleStr string) (any, error) {
|
||||
if ruleStr == "" {
|
||||
return nil, nil
|
||||
}
|
||||
var result any
|
||||
content := a.content
|
||||
ruleList := SplitSourceRule(ruleStr, true, a.isJSON, &a.isRegex)
|
||||
if content != nil && len(ruleList) > 0 {
|
||||
result = content
|
||||
for _, sourceRule := range ruleList {
|
||||
if err := a.putRule(sourceRule.putMap); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
resolved, err := sourceRule.MakeUpRule(result, a.makeDeps(result))
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
if result == nil {
|
||||
continue
|
||||
}
|
||||
rule := resolved.Rule
|
||||
switch sourceRule.Mode {
|
||||
case ModeRegex:
|
||||
result = regexGetElement(resultString(result), splitNotBlankAndTrim(rule, "&&"), 0)
|
||||
case ModeWebJs:
|
||||
return nil, ErrWebJSUnsupported
|
||||
case ModeJs:
|
||||
result, err = a.evalJS(rule, result)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
case ModeJson:
|
||||
result = newJSONAnalyzer(result).getObject(rule)
|
||||
case ModeXPath:
|
||||
result = newXPathAnalyzer(result).getElements(rule)
|
||||
default:
|
||||
result = newJsoupAnalyzer(result).getElements(rule)
|
||||
}
|
||||
if resolved.ReplaceRegex != "" {
|
||||
result = applyReplaceRegex(resultString(result), resolved)
|
||||
}
|
||||
}
|
||||
}
|
||||
return result, nil
|
||||
}
|
||||
|
||||
// GetElements 对应 getElements(ruleStr):列表获取。
|
||||
// 注意:与 Kotlin 一致,这里不做 makeUpRule(无 ## 正则段处理)。
|
||||
func (a *AnalyzeRule) GetElements(ruleStr string) ([]any, error) {
|
||||
var result any
|
||||
content := a.content
|
||||
ruleList := SplitSourceRule(ruleStr, true, a.isJSON, &a.isRegex)
|
||||
if content != nil && len(ruleList) > 0 {
|
||||
result = content
|
||||
for _, sourceRule := range ruleList {
|
||||
if err := a.putRule(sourceRule.putMap); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
if result == nil {
|
||||
continue
|
||||
}
|
||||
rule := sourceRule.Rule
|
||||
var err error
|
||||
switch sourceRule.Mode {
|
||||
case ModeRegex:
|
||||
result = regexGetElements(resultString(result), splitNotBlankAndTrim(rule, "&&"), 0)
|
||||
case ModeWebJs:
|
||||
return nil, ErrWebJSUnsupported
|
||||
case ModeJs:
|
||||
result, err = a.evalJS(rule, result)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
case ModeJson:
|
||||
result = newJSONAnalyzer(result).getList(rule)
|
||||
case ModeXPath:
|
||||
result = newXPathAnalyzer(result).getElements(rule)
|
||||
default:
|
||||
result = newJsoupAnalyzer(result).getElements(rule)
|
||||
}
|
||||
}
|
||||
}
|
||||
if result != nil {
|
||||
switch t := result.(type) {
|
||||
case []any:
|
||||
return t, nil
|
||||
case []string:
|
||||
out := make([]any, len(t))
|
||||
for i, v := range t {
|
||||
out[i] = v
|
||||
}
|
||||
return out, nil
|
||||
case []*html.Node:
|
||||
out := make([]any, len(t))
|
||||
for i, v := range t {
|
||||
out[i] = v
|
||||
}
|
||||
return out, nil
|
||||
}
|
||||
}
|
||||
return []any{}, nil
|
||||
}
|
||||
|
||||
// ─── 工具 ──────────────────────────────────────────────────────────────────
|
||||
|
||||
func (a *AnalyzeRule) absolutize(u string) string {
|
||||
if strings.TrimSpace(u) == "" {
|
||||
return ""
|
||||
}
|
||||
if a.redirectURL != nil {
|
||||
return GetAbsoluteURLParsed(a.redirectURL, u)
|
||||
}
|
||||
return GetAbsoluteURL(a.baseUrl, u)
|
||||
}
|
||||
|
||||
// applyReplaceRegex 对应 replaceRegex(result, resolved)。
|
||||
func applyReplaceRegex(result string, r ResolvedSourceRule) string {
|
||||
if r.ReplaceRegex == "" {
|
||||
return result
|
||||
}
|
||||
if r.ReplaceFirst {
|
||||
return regexReplaceFirstOnFirstMatch(r.ReplaceRegex, result, r.Replacement)
|
||||
}
|
||||
return regexReplaceAll(r.ReplaceRegex, result, r.Replacement)
|
||||
}
|
||||
|
||||
// entityRefRe 匹配一个 HTML 字符引用:命名实体(&、 )、十进制(')、
|
||||
// 十六进制(')。分号可选——旧式简写( 、&)在 HTML 文本里同样成立。
|
||||
var entityRefRe = regexp.MustCompile(`&(?:[a-zA-Z][a-zA-Z0-9]{1,31}|#[0-9]{1,7}|#[xX][0-9a-fA-F]{1,6});?`)
|
||||
|
||||
// unescapeRuleEntities 还原规则结果里的 HTML 实体(`&` → `&`、`'` → `'`)。
|
||||
//
|
||||
// 唯一的例外是「不带分号的旧式简写后面紧跟 = 或字母数字」这一种:HTML5 里这类写法
|
||||
// 属于历史匹配,而规则结果并不是纯 HTML 文本——正文、地址、查询串都在同一份字符串里。
|
||||
// 按浏览器语义无差别还原会把查询串吃掉:光遇聚合的段评地址
|
||||
// `…?item_id=X¶=1&source=QQ阅读` 会变成 `…?item_id=X¶=1&source=QQ阅读`,
|
||||
// 上游按缺参数处理返回空页面,段评面板就只剩一片空白(见 reader/comment.go)。
|
||||
func unescapeRuleEntities(s string) string {
|
||||
if !strings.Contains(s, "&") {
|
||||
return s
|
||||
}
|
||||
var b strings.Builder
|
||||
last := 0
|
||||
for _, loc := range entityRefRe.FindAllStringIndex(s, -1) {
|
||||
match := s[loc[0]:loc[1]]
|
||||
if !strings.HasSuffix(match, ";") && loc[1] < len(s) && isEntityStopByte(s[loc[1]]) {
|
||||
continue
|
||||
}
|
||||
decoded := html.UnescapeString(match)
|
||||
if decoded == match {
|
||||
continue // 不是已知实体,原样保留
|
||||
}
|
||||
b.WriteString(s[last:loc[0]])
|
||||
b.WriteString(decoded)
|
||||
last = loc[1]
|
||||
}
|
||||
if last == 0 {
|
||||
return s
|
||||
}
|
||||
b.WriteString(s[last:])
|
||||
return b.String()
|
||||
}
|
||||
|
||||
// isEntityStopByte 判断旧式简写实体后面这个字符是否让它失去实体语义
|
||||
// (HTML5 规定不带分号的引用后跟 = 或字母数字时按普通文本处理)。
|
||||
func isEntityStopByte(c byte) bool {
|
||||
return c == '=' || c >= '0' && c <= '9' || c >= 'a' && c <= 'z' || c >= 'A' && c <= 'Z'
|
||||
}
|
||||
|
||||
// resultString 对应 Kotlin 的 result.toString()。
|
||||
func resultString(result any) string {
|
||||
switch t := result.(type) {
|
||||
case nil:
|
||||
return ""
|
||||
case string:
|
||||
return t
|
||||
case *html.Node:
|
||||
return outerHTML(t)
|
||||
case []string:
|
||||
return strings.Join(t, "\n")
|
||||
case []*html.Node:
|
||||
var sb strings.Builder
|
||||
for _, n := range t {
|
||||
sb.WriteString(outerHTML(n))
|
||||
}
|
||||
return sb.String()
|
||||
default:
|
||||
return anyToString(result)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,80 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"strings"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// 规则结果里的地址带查询串。html.UnescapeString 会按 HTML5 旧式简写把 ¶ 吃成
|
||||
// 「¶」,参数名变成 ¶ 之后上游按缺参数返回空页面——光遇聚合的段评面板一片空白
|
||||
// 就是这个原因(见 reader/comment.go、reader/browser_panel.go)。
|
||||
func TestUnescapeRuleEntitiesKeepsQueryString(t *testing.T) {
|
||||
cases := []struct {
|
||||
name string
|
||||
in string
|
||||
want string
|
||||
}{
|
||||
{
|
||||
name: "段评地址里的 ¶ 不能被当成实体",
|
||||
in: "https://v1.qingtian618.com/get_para_review?item_id=X¶=1&source=QQ%E9%98%85%E8%AF%BB",
|
||||
want: "https://v1.qingtian618.com/get_para_review?item_id=X¶=1&source=QQ%E9%98%85%E8%AF%BB",
|
||||
},
|
||||
{
|
||||
name: "¬ 开头的参数名同样不能被吃掉",
|
||||
in: "https://example.com/api?book=1¬ify=1",
|
||||
want: "https://example.com/api?book=1¬ify=1",
|
||||
},
|
||||
{
|
||||
name: "带分号的实体照旧还原",
|
||||
in: "url?a=1&b=2&c=3",
|
||||
want: "url?a=1&b=2&c=3",
|
||||
},
|
||||
{
|
||||
name: "数字实体照旧还原",
|
||||
in: "'引号' 'x' ©",
|
||||
want: "'引号' 'x' ©",
|
||||
},
|
||||
{
|
||||
name: "文字里的旧式简写(后面不是 = 或字母数字)照旧还原",
|
||||
in: "第一段 第二段© 2026×",
|
||||
want: "第一段\u00a0第二段© 2026×",
|
||||
},
|
||||
{
|
||||
name: "不是实体的 & 原样保留",
|
||||
in: "a & b &foo; c",
|
||||
want: "a & b &foo; c",
|
||||
},
|
||||
{
|
||||
name: "不含 & 的字符串直接放行",
|
||||
in: "魔法大陆。",
|
||||
want: "魔法大陆。",
|
||||
},
|
||||
}
|
||||
for _, c := range cases {
|
||||
t.Run(c.name, func(t *testing.T) {
|
||||
if got := unescapeRuleEntities(c.in); got != c.want {
|
||||
t.Fatalf("unescapeRuleEntities(%q) = %q,期望 %q", c.in, got, c.want)
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
// 端到端:走 GetString 的规则结果同样不能把段评地址里的 ¶ 吃掉
|
||||
// (光遇聚合的正文规则是 <js>…</js>$.content 形态,最终取 $.content)。
|
||||
func TestGetStringCommentURLKeepsParaParam(t *testing.T) {
|
||||
const url = "https://v1.qingtian618.com/get_para_review?item_id=YmlkPTYxNDc4MiZjaWQ9MQ¶=11&source=QQ%E9%98%85%E8%AF%BB"
|
||||
payload := `{"content":"<p>正文<comment ident=\"` + url + `\" count=\"39\" /></p>"}`
|
||||
|
||||
a := NewAnalyzeRule()
|
||||
a.SetContent(payload, "https://v1.qingtian618.com")
|
||||
got, err := a.GetString("$.content", nil, false)
|
||||
if err != nil {
|
||||
t.Fatalf("GetString: %v", err)
|
||||
}
|
||||
if !strings.Contains(got, "¶=11") {
|
||||
t.Fatalf("段评地址里的 ¶=11 被改写了: %q", got)
|
||||
}
|
||||
if strings.Contains(got, "\u00b6") {
|
||||
t.Fatalf("段评地址里出现了 ¶(HTML 旧式实体还原的产物): %q", got)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,60 @@
|
||||
package rule
|
||||
|
||||
import "testing"
|
||||
|
||||
// 聚合类书源(如「光遇聚合」)的搜索请求地址本身就是 data:;base64,... 参数信封。
|
||||
// 规则取值为空时 legado 语义会回退 baseUrl;若此时的 baseUrl 不是 http(s) 地址,
|
||||
// 回退结果会把封面/书址污染成非图片文本(前端 <img> 直接破图)。
|
||||
func TestGetStringURLFallbackSkipsNonHTTPBaseURL(t *testing.T) {
|
||||
const payload = `{"data":[{"book_name":"全球高武","author":"老鹰吃小鸡"}]}`
|
||||
const envelope = "data:;base64,eyJrZXkiOiLlhajnkIPpq5jmraYiLCJ0YWIiOiLlsI/or7QiLCJzb3VyY2VzS2V5Ijoi5YWo6YOoIiwicGFnZSI6MSwiZGlzYWJsZWRfc291cmNlcyI6IjAifQ=="
|
||||
|
||||
elementOf := func(t *testing.T, content, baseURL string) (*AnalyzeRule, any) {
|
||||
t.Helper()
|
||||
a := NewAnalyzeRule()
|
||||
a.SetContent(content, baseURL)
|
||||
els, err := a.GetElements("$.data")
|
||||
if err != nil {
|
||||
t.Fatalf("GetElements: %v", err)
|
||||
}
|
||||
if len(els) != 1 {
|
||||
t.Fatalf("GetElements: want 1 element, got %d", len(els))
|
||||
}
|
||||
return a, els[0]
|
||||
}
|
||||
|
||||
t.Run("空取值不会把 data: 信封当地址返回", func(t *testing.T) {
|
||||
a, el := elementOf(t, payload, envelope)
|
||||
got, err := a.GetString("$.thumb_url", el, true)
|
||||
if err != nil {
|
||||
t.Fatalf("GetString: %v", err)
|
||||
}
|
||||
if got != "" {
|
||||
t.Fatalf("封面应为空,实际 = %q", got)
|
||||
}
|
||||
})
|
||||
|
||||
t.Run("baseUrl 是 http(s) 时仍按 legado 语义回退", func(t *testing.T) {
|
||||
const base = "https://www.example.com/search?q=x"
|
||||
a, el := elementOf(t, payload, base)
|
||||
got, err := a.GetString("$.thumb_url", el, true)
|
||||
if err != nil {
|
||||
t.Fatalf("GetString: %v", err)
|
||||
}
|
||||
if got != base {
|
||||
t.Fatalf("want base fallback %q, got %q", base, got)
|
||||
}
|
||||
})
|
||||
|
||||
t.Run("规则有取值时原样返回", func(t *testing.T) {
|
||||
const withCover = `{"data":[{"book_name":"书","thumb_url":"https://img.example.com/c.jpg"}]}`
|
||||
a, el := elementOf(t, withCover, envelope)
|
||||
got, err := a.GetString("$.thumb_url", el, true)
|
||||
if err != nil {
|
||||
t.Fatalf("GetString: %v", err)
|
||||
}
|
||||
if got != "https://img.example.com/c.jpg" {
|
||||
t.Fatalf("封面 = %q", got)
|
||||
}
|
||||
})
|
||||
}
|
||||
@@ -0,0 +1,40 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"strings"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// TestGetStringListURLFromJSArray URL 类规则(nextTocUrl / chapterUrl / coverUrl…)
|
||||
// 由 JS 生成时,goja 导出的是 []any。早期实现只认 []string,JS 给的地址数组会被
|
||||
// 静默丢掉,表现为「书源写了 nextTocUrl 却永远不翻页」。
|
||||
func TestGetStringListURLFromJSArray(t *testing.T) {
|
||||
ar := NewAnalyzeRule()
|
||||
ar.SetContent(`{"total":399,"limit":100}`, "https://api.example.com/toc?limit=100&offset=0")
|
||||
ar.SetJSRunner(func(js string, result any) (any, error) {
|
||||
return NewJSRunner(JSConfig{BaseURL: "https://api.example.com/toc?limit=100&offset=0"}).Run(nil, js, result, "https://api.example.com/toc?limit=100&offset=0")
|
||||
})
|
||||
|
||||
// JS 返回数组(真实书源就是这种写法:一次给出全部后续页)
|
||||
rule := `<js>['https://api.example.com/toc?offset=100','https://api.example.com/toc?offset=200']</js>`
|
||||
urls, err := ar.GetStringList(rule, nil, true)
|
||||
if err != nil {
|
||||
t.Fatalf("err=%v", err)
|
||||
}
|
||||
if len(urls) != 2 {
|
||||
t.Fatalf("拿到 %d 个地址,期望 2(JS 数组被丢掉了): %v", len(urls), urls)
|
||||
}
|
||||
if urls[0] != "https://api.example.com/toc?offset=100" || urls[1] != "https://api.example.com/toc?offset=200" {
|
||||
t.Fatalf("地址不对: %v", urls)
|
||||
}
|
||||
|
||||
// 相对地址要按当前页绝对化
|
||||
ruleRel := `<js>['/toc?offset=300']</js>`
|
||||
urls, err = ar.GetStringList(ruleRel, nil, true)
|
||||
if err != nil {
|
||||
t.Fatalf("err=%v", err)
|
||||
}
|
||||
if len(urls) != 1 || !strings.HasPrefix(urls[0], "https://api.example.com/toc?offset=300") {
|
||||
t.Fatalf("相对地址没有绝对化: %v", urls)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,595 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"crypto/hmac"
|
||||
"crypto/md5"
|
||||
"crypto/sha1"
|
||||
"crypto/sha256"
|
||||
"crypto/sha512"
|
||||
"encoding/base64"
|
||||
"encoding/hex"
|
||||
"fmt"
|
||||
"hash"
|
||||
"strings"
|
||||
|
||||
"github.com/dop251/goja"
|
||||
"github.com/google/uuid"
|
||||
)
|
||||
|
||||
// 本文件对应 legado JsExtensions / JsEncodeUtils / RegexJsExtensions 中
|
||||
// 注入为 `java` 对象的函数。函数名与 Kotlin 版一一对应(存量书源硬编码),
|
||||
// 实现全部为 Go 原生;浏览器/文件/压缩包类能力明确抛出不支持错误。
|
||||
|
||||
func newJavaObject(vm *goja.Runtime, r *JSRunner, a *AnalyzeRule) *goja.Object {
|
||||
o := vm.NewObject()
|
||||
set := func(k string, v any) {
|
||||
if err := o.Set(k, v); err != nil {
|
||||
panic(vm.ToValue(err.Error()))
|
||||
}
|
||||
}
|
||||
|
||||
bridgeErr := func(name string, err error) goja.Value {
|
||||
panic(vm.ToValue(fmt.Sprintf("java.%s: %v", name, err)))
|
||||
}
|
||||
|
||||
// ── 网络(对应 JsExtensions.ajax/ajaxAll/connect/get/post/head) ──
|
||||
fetchStr := func(name string, urlVal goja.Value, extraHeaders map[string]any) string {
|
||||
if r.cfg.Fetch == nil {
|
||||
bridgeErr(name, ErrJsUnsupported)
|
||||
}
|
||||
urlStr := urlVal.String()
|
||||
if arr, ok := urlVal.Export().([]any); ok && len(arr) > 0 {
|
||||
urlStr = fmt.Sprintf("%v", arr[0]) // 对应 legado:List 取第一个
|
||||
}
|
||||
req, err := ParseAnalyzeUrlWithJS(urlStr, "", 0, r.cfg.BaseURL, r)
|
||||
if err != nil {
|
||||
bridgeErr(name, err)
|
||||
}
|
||||
if req.Unsupported != nil {
|
||||
bridgeErr(name, req.Unsupported)
|
||||
}
|
||||
for k, v := range extraHeaders {
|
||||
if _, exists := req.Headers[k]; !exists {
|
||||
req.Headers[k] = fmt.Sprintf("%v", v)
|
||||
}
|
||||
}
|
||||
body, _, _, err := r.fetch(req)
|
||||
if err != nil {
|
||||
bridgeErr(name, err)
|
||||
}
|
||||
return body
|
||||
}
|
||||
|
||||
fetchResp := func(name string, args []goja.Value, method, body string) *goja.Object {
|
||||
if r.cfg.Fetch == nil {
|
||||
bridgeErr(name, ErrJsUnsupported)
|
||||
}
|
||||
urlStr := ""
|
||||
if len(args) > 0 {
|
||||
urlStr = args[0].String()
|
||||
}
|
||||
req, err := ParseAnalyzeUrlWithJS(urlStr, "", 0, r.cfg.BaseURL, r)
|
||||
if err != nil {
|
||||
bridgeErr(name, err)
|
||||
}
|
||||
if method == "POST" {
|
||||
req.Method = "POST"
|
||||
req.Body = body
|
||||
trimmed := strings.TrimSpace(body)
|
||||
if strings.HasPrefix(trimmed, "{") || strings.HasPrefix(trimmed, "[") {
|
||||
req.IsJSON = true
|
||||
} else if body != "" {
|
||||
req.IsForm = true
|
||||
req.Body = encodeParams(body, req.Charset, false)
|
||||
}
|
||||
}
|
||||
if len(args) > 1 && !goja.IsUndefined(args[1]) && !goja.IsNull(args[1]) {
|
||||
if extra, ok := args[1].Export().(map[string]any); ok {
|
||||
for k, v := range extra {
|
||||
req.Headers[k] = fmt.Sprintf("%v", v)
|
||||
}
|
||||
}
|
||||
}
|
||||
if req.Unsupported != nil {
|
||||
bridgeErr(name, req.Unsupported)
|
||||
}
|
||||
respBody, finalURL, code, err := r.fetch(req)
|
||||
if err != nil {
|
||||
bridgeErr(name, err)
|
||||
}
|
||||
return newResponseObject(vm, respBody, code, finalURL, nil)
|
||||
}
|
||||
|
||||
set("ajax", func(call goja.FunctionCall) goja.Value {
|
||||
extra := map[string]any(nil)
|
||||
if len(call.Arguments) > 1 && !goja.IsUndefined(call.Arguments[1]) && !goja.IsNull(call.Arguments[1]) {
|
||||
if m, ok := call.Arguments[1].Export().(map[string]any); ok {
|
||||
extra = m
|
||||
}
|
||||
}
|
||||
return vm.ToValue(fetchStr("ajax", call.Arguments[0], extra))
|
||||
})
|
||||
set("ajaxAll", func(call goja.FunctionCall) goja.Value {
|
||||
var urls []any
|
||||
if len(call.Arguments) > 0 {
|
||||
if arr, ok := call.Arguments[0].Export().([]any); ok {
|
||||
urls = arr
|
||||
}
|
||||
}
|
||||
out := make([]any, 0, len(urls))
|
||||
for _, u := range urls {
|
||||
out = append(out, fetchStr("ajaxAll", vm.ToValue(u), nil))
|
||||
}
|
||||
return vm.ToValue(out)
|
||||
})
|
||||
set("connect", func(call goja.FunctionCall) goja.Value {
|
||||
return fetchResp("connect", call.Arguments, "GET", "")
|
||||
})
|
||||
// get 双语义(对应 Kotlin 重载):get(url, headers) 走网络,get(key) 读变量
|
||||
set("get", func(call goja.FunctionCall) goja.Value {
|
||||
if len(call.Arguments) >= 2 && !goja.IsUndefined(call.Arguments[1]) && !goja.IsNull(call.Arguments[1]) {
|
||||
return fetchResp("get", call.Arguments, "GET", "")
|
||||
}
|
||||
key := stringArg(call, 0)
|
||||
if a != nil {
|
||||
return vm.ToValue(a.Get(key))
|
||||
}
|
||||
return vm.ToValue(r.vars[key])
|
||||
})
|
||||
set("post", func(call goja.FunctionCall) goja.Value {
|
||||
body := stringArgOr(call, 1, "")
|
||||
return fetchResp("post", call.Arguments, "POST", body)
|
||||
})
|
||||
set("head", func(call goja.FunctionCall) goja.Value {
|
||||
return fetchResp("head", call.Arguments, "HEAD", "")
|
||||
})
|
||||
|
||||
// ── 规则回调(桥回当前 AnalyzeRule,对应 AnalyzeRule.getString 等公有方法) ──
|
||||
ruleStr := func(name string, args []goja.Value) (string, error) {
|
||||
if a == nil {
|
||||
return "", fmt.Errorf("无规则上下文")
|
||||
}
|
||||
mContent := any(nil)
|
||||
if len(args) > 1 && !goja.IsUndefined(args[1]) && !goja.IsNull(args[1]) {
|
||||
mContent = args[1].Export()
|
||||
}
|
||||
isURL := len(args) > 2 && args[2].ToBoolean()
|
||||
return a.GetString(args[0].String(), mContent, isURL)
|
||||
}
|
||||
set("getString", func(call goja.FunctionCall) goja.Value {
|
||||
s, err := ruleStr("getString", call.Arguments)
|
||||
if err != nil {
|
||||
bridgeErr("getString", err)
|
||||
}
|
||||
return vm.ToValue(s)
|
||||
})
|
||||
set("getStringList", func(call goja.FunctionCall) goja.Value {
|
||||
if a == nil {
|
||||
return goja.Null()
|
||||
}
|
||||
list, err := a.GetStringList(stringArgOr(call, 0, ""), nil, false)
|
||||
if err != nil {
|
||||
bridgeErr("getStringList", err)
|
||||
}
|
||||
return vm.ToValue(list)
|
||||
})
|
||||
set("getElements", func(call goja.FunctionCall) goja.Value {
|
||||
if a == nil {
|
||||
return goja.Null()
|
||||
}
|
||||
els, err := a.GetElements(stringArgOr(call, 0, ""))
|
||||
if err != nil {
|
||||
bridgeErr("getElements", err)
|
||||
}
|
||||
out := make([]any, 0, len(els))
|
||||
for _, el := range els {
|
||||
out = append(out, resultString(el))
|
||||
}
|
||||
return vm.ToValue(out)
|
||||
})
|
||||
set("put", func(call goja.FunctionCall) goja.Value {
|
||||
key := stringArg(call, 0)
|
||||
val := stringArgOr(call, 1, "")
|
||||
if a != nil {
|
||||
return vm.ToValue(a.Put(key, val))
|
||||
}
|
||||
r.vars[key] = val
|
||||
return vm.ToValue(val)
|
||||
})
|
||||
|
||||
// ── 编码(对应 JsEncodeUtils / JsExtensions) ──
|
||||
set("md5Encode", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(md5Hex(stringArg(call, 0), false))
|
||||
})
|
||||
set("md5Encode16", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(md5Hex(stringArg(call, 0), true))
|
||||
})
|
||||
set("base64Decode", func(call goja.FunctionCall) goja.Value {
|
||||
s, err := base64DecodeString(stringArg(call, 0))
|
||||
if err != nil {
|
||||
bridgeErr("base64Decode", err)
|
||||
}
|
||||
return vm.ToValue(s)
|
||||
})
|
||||
set("base64DecodeToByteArray", func(call goja.FunctionCall) goja.Value {
|
||||
b, err := base64DecodeBytes(stringArg(call, 0))
|
||||
if err != nil {
|
||||
bridgeErr("base64DecodeToByteArray", err)
|
||||
}
|
||||
return vm.ToValue(vm.NewArrayBuffer(b))
|
||||
})
|
||||
set("base64Encode", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(base64.StdEncoding.EncodeToString([]byte(stringArg(call, 0))))
|
||||
})
|
||||
set("hexDecodeToString", func(call goja.FunctionCall) goja.Value {
|
||||
b, err := hex.DecodeString(strings.TrimSpace(stringArg(call, 0)))
|
||||
if err != nil {
|
||||
bridgeErr("hexDecodeToString", err)
|
||||
}
|
||||
return vm.ToValue(string(b))
|
||||
})
|
||||
set("hexDecodeToByteArray", func(call goja.FunctionCall) goja.Value {
|
||||
b, err := hex.DecodeString(strings.TrimSpace(stringArg(call, 0)))
|
||||
if err != nil {
|
||||
bridgeErr("hexDecodeToByteArray", err)
|
||||
}
|
||||
return vm.ToValue(vm.NewArrayBuffer(b))
|
||||
})
|
||||
set("hexEncodeToString", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(hex.EncodeToString([]byte(stringArg(call, 0))))
|
||||
})
|
||||
set("strToBytes", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(vm.NewArrayBuffer([]byte(stringArg(call, 0))))
|
||||
})
|
||||
set("bytesToStr", func(call goja.FunctionCall) goja.Value {
|
||||
if len(call.Arguments) == 0 || goja.IsUndefined(call.Arguments[0]) || goja.IsNull(call.Arguments[0]) {
|
||||
return vm.ToValue("")
|
||||
}
|
||||
if buf, ok := call.Arguments[0].Export().([]byte); ok {
|
||||
return vm.ToValue(string(buf))
|
||||
}
|
||||
return vm.ToValue(call.Arguments[0].String())
|
||||
})
|
||||
set("encodeURI", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(javaURLEncode(stringArg(call, 0), stringArgOr(call, 1, "utf-8")))
|
||||
})
|
||||
set("timeFormat", func(call goja.FunctionCall) goja.Value {
|
||||
ts := int64(call.Arguments[0].ToFloat())
|
||||
format := stringArgOr(call, 1, "yyyy-MM-dd HH:mm")
|
||||
return vm.ToValue(javaTimeFormat(ts, format, 0))
|
||||
})
|
||||
set("timeFormatUTC", func(call goja.FunctionCall) goja.Value {
|
||||
ts := int64(call.Arguments[0].ToFloat())
|
||||
format := stringArgOr(call, 1, "yyyy-MM-dd HH:mm")
|
||||
sh := int64(0)
|
||||
if len(call.Arguments) > 2 {
|
||||
sh = int64(call.Arguments[2].ToFloat())
|
||||
}
|
||||
return vm.ToValue(javaTimeFormat(ts, format, sh))
|
||||
})
|
||||
|
||||
// ── 加解密(对应 JsEncodeUtils) ──
|
||||
set("createSymmetricCrypto", func(call goja.FunctionCall) goja.Value {
|
||||
transformation := stringArgOr(call, 0, "AES/CBC/PKCS5Padding")
|
||||
key := stringArgOr(call, 1, "")
|
||||
iv := stringArgOr(call, 2, "")
|
||||
cipher, err := newSymmetricCipher(transformation, key, iv)
|
||||
if err != nil {
|
||||
bridgeErr("createSymmetricCrypto", err)
|
||||
}
|
||||
return newCipherObject(vm, cipher)
|
||||
})
|
||||
aesVariant := func(mode string) func(goja.FunctionCall) goja.Value {
|
||||
return func(call goja.FunctionCall) goja.Value {
|
||||
data := stringArg(call, 0)
|
||||
key := stringArgOr(call, 1, "")
|
||||
defaultTrans := mode
|
||||
trans := stringArgOr(call, 2, defaultTrans)
|
||||
iv := stringArgOr(call, 3, "")
|
||||
cipher, err := newSymmetricCipher(trans, key, iv)
|
||||
if err != nil {
|
||||
bridgeErr(mode, err)
|
||||
}
|
||||
out, err := cipher.decryptAuto(data)
|
||||
if err != nil {
|
||||
bridgeErr(mode, err)
|
||||
}
|
||||
return vm.ToValue(string(out))
|
||||
}
|
||||
}
|
||||
set("aesDecodeToString", aesVariant("AES/CBC/PKCS5Padding"))
|
||||
set("aesBase64DecodeToString", aesVariant("AES/CBC/PKCS5Padding"))
|
||||
set("aesBase64DecodeToByteArray", aesVariant("AES/CBC/PKCS5Padding"))
|
||||
set("desDecodeToString", aesVariant("DES/CBC/PKCS5Padding"))
|
||||
set("desBase64DecodeToString", aesVariant("DES/CBC/PKCS5Padding"))
|
||||
encryptStr := func(data, key, trans, iv string) string {
|
||||
cipher, err := newSymmetricCipher(trans, key, iv)
|
||||
if err != nil {
|
||||
bridgeErr("encrypt", err)
|
||||
}
|
||||
out, err := cipher.encrypt([]byte(data))
|
||||
if err != nil {
|
||||
bridgeErr("encrypt", err)
|
||||
}
|
||||
return string(out)
|
||||
}
|
||||
set("aesEncodeToString", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(encryptStr(stringArg(call, 0), stringArgOr(call, 1, ""), stringArgOr(call, 2, "AES/CBC/PKCS5Padding"), stringArgOr(call, 3, "")))
|
||||
})
|
||||
set("desEncodeToString", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(encryptStr(stringArg(call, 0), stringArgOr(call, 1, ""), stringArgOr(call, 2, "DES/CBC/PKCS5Padding"), stringArgOr(call, 3, "")))
|
||||
})
|
||||
encryptB64 := func(data, key, trans, iv string) string {
|
||||
cipher, err := newSymmetricCipher(trans, key, iv)
|
||||
if err != nil {
|
||||
bridgeErr("encrypt", err)
|
||||
}
|
||||
out, err := cipher.encrypt([]byte(data))
|
||||
if err != nil {
|
||||
bridgeErr("encrypt", err)
|
||||
}
|
||||
return base64.StdEncoding.EncodeToString(out)
|
||||
}
|
||||
set("aesEncodeToBase64String", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(encryptB64(stringArg(call, 0), stringArgOr(call, 1, ""), stringArgOr(call, 2, "AES/CBC/PKCS5Padding"), stringArgOr(call, 3, "")))
|
||||
})
|
||||
set("desEncodeToBase64String", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(encryptB64(stringArg(call, 0), stringArgOr(call, 1, ""), stringArgOr(call, 2, "DES/CBC/PKCS5Padding"), stringArgOr(call, 3, "")))
|
||||
})
|
||||
set("tripleDESDecodeStr", func(call goja.FunctionCall) goja.Value {
|
||||
data := stringArg(call, 0)
|
||||
key := stringArgOr(call, 1, "")
|
||||
trans := "DESede/CBC/PKCS5Padding"
|
||||
var ivv string
|
||||
switch {
|
||||
case len(call.Arguments) > 4: // (data, key, mode, padding, iv)
|
||||
trans = "DESede/" + stringArgOr(call, 2, "CBC") + "/" + stringArgOr(call, 3, "PKCS5Padding")
|
||||
ivv = stringArgOr(call, 4, "")
|
||||
case len(call.Arguments) > 2: // (data, key, transformation, iv?)
|
||||
trans = stringArgOr(call, 2, trans)
|
||||
ivv = stringArgOr(call, 3, "")
|
||||
}
|
||||
cipher, err := newSymmetricCipher(trans, key, ivv)
|
||||
if err != nil {
|
||||
bridgeErr("tripleDESDecodeStr", err)
|
||||
}
|
||||
out, err := cipher.decryptAuto(data)
|
||||
if err != nil {
|
||||
bridgeErr("tripleDESDecodeStr", err)
|
||||
}
|
||||
return vm.ToValue(string(out))
|
||||
})
|
||||
set("digestHex", func(call goja.FunctionCall) goja.Value {
|
||||
h, err := digestHash(stringArgOr(call, 1, "MD5"))
|
||||
if err != nil {
|
||||
bridgeErr("digestHex", err)
|
||||
}
|
||||
h.Write([]byte(stringArg(call, 0)))
|
||||
return vm.ToValue(hex.EncodeToString(h.Sum(nil)))
|
||||
})
|
||||
set("digestBase64Str", func(call goja.FunctionCall) goja.Value {
|
||||
h, err := digestHash(stringArgOr(call, 1, "MD5"))
|
||||
if err != nil {
|
||||
bridgeErr("digestBase64Str", err)
|
||||
}
|
||||
h.Write([]byte(stringArg(call, 0)))
|
||||
return vm.ToValue(base64.StdEncoding.EncodeToString(h.Sum(nil)))
|
||||
})
|
||||
set("HMacHex", func(call goja.FunctionCall) goja.Value {
|
||||
h, err := hmacHash(stringArgOr(call, 1, "HmacSHA256"), []byte(stringArgOr(call, 2, "")))
|
||||
if err != nil {
|
||||
bridgeErr("HMacHex", err)
|
||||
}
|
||||
h.Write([]byte(stringArg(call, 0)))
|
||||
return vm.ToValue(hex.EncodeToString(h.Sum(nil)))
|
||||
})
|
||||
set("HMacBase64", func(call goja.FunctionCall) goja.Value {
|
||||
h, err := hmacHash(stringArgOr(call, 1, "HmacSHA256"), []byte(stringArgOr(call, 2, "")))
|
||||
if err != nil {
|
||||
bridgeErr("HMacBase64", err)
|
||||
}
|
||||
h.Write([]byte(stringArg(call, 0)))
|
||||
return vm.ToValue(base64.StdEncoding.EncodeToString(h.Sum(nil)))
|
||||
})
|
||||
|
||||
// ── 其他 ──
|
||||
set("log", func(call goja.FunctionCall) goja.Value {
|
||||
if r.cfg.Log != nil {
|
||||
r.cfg.Log(call.Arguments[0].String())
|
||||
}
|
||||
return goja.Null()
|
||||
})
|
||||
set("getWebViewUA", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(defaultWebViewUA)
|
||||
})
|
||||
// 简繁转换(对应 legado ChineseUtils → quick-transfer)。这里按单字映射实现,
|
||||
// 见 zh_convert.go 的说明。
|
||||
set("t2s", func(call goja.FunctionCall) goja.Value { return vm.ToValue(zhToSimplified(stringArg(call, 0))) })
|
||||
set("s2t", func(call goja.FunctionCall) goja.Value { return vm.ToValue(zhToTraditional(stringArg(call, 0))) })
|
||||
set("htmlFormat", func(call goja.FunctionCall) goja.Value { return vm.ToValue(stringArg(call, 0)) })
|
||||
// randomUUID:部分书源拿它生成设备标识/请求 ID(缺失会让整条规则抛异常)。
|
||||
set("randomUUID", func(call goja.FunctionCall) goja.Value { return vm.ToValue(uuid.NewString()) })
|
||||
// toNumChapter:章节标题里的数字规整成阿拉伯数字(见 zh_number.go)。
|
||||
set("toNumChapter", func(call goja.FunctionCall) goja.Value { return vm.ToValue(numChapter(stringArg(call, 0))) })
|
||||
// toURL(url[, baseUrl]):对应 legado JsURL(host/origin/pathname/searchParams)。
|
||||
set("toURL", func(call goja.FunctionCall) goja.Value {
|
||||
base := ""
|
||||
if len(call.Arguments) > 1 && !goja.IsUndefined(call.Arguments[1]) && !goja.IsNull(call.Arguments[1]) {
|
||||
base = call.Arguments[1].String()
|
||||
}
|
||||
obj, err := newJSURLObject(vm, stringArg(call, 0), base)
|
||||
if err != nil {
|
||||
bridgeErr("toURL", err)
|
||||
}
|
||||
return obj
|
||||
})
|
||||
// logType:调试用,打印值的类型(对应 legado 的日志实现)。
|
||||
set("logType", func(call goja.FunctionCall) goja.Value {
|
||||
if r.cfg.Log != nil {
|
||||
if len(call.Arguments) == 0 || goja.IsNull(call.Arguments[0]) || goja.IsUndefined(call.Arguments[0]) {
|
||||
r.cfg.Log("null")
|
||||
} else {
|
||||
r.cfg.Log(fmt.Sprintf("%T", call.Arguments[0].Export()))
|
||||
}
|
||||
}
|
||||
return goja.Null()
|
||||
})
|
||||
|
||||
// ── 书源会话状态(legado 中 `java` 与 `source` 是同一对象) ──
|
||||
if r.state != nil {
|
||||
bindSourceState(vm, set, r.state, r.cfg.SourceProps)
|
||||
}
|
||||
|
||||
// ── 宿主交互:提示、浏览器页面、登录界面刷新 ──
|
||||
//
|
||||
// legado 用内置 WebView 承载页面:startBrowser / startBrowserAwait 打开一个
|
||||
// 页面让用户完成防爬校验、登录或参数选择,Await 版本还会把用户操作后的
|
||||
// 页面源码作为 StrResponse 返回。真实的聚合类书源(如光遇聚合)把「线路
|
||||
// 切换」「用户后台」「书源设置」全部建在这两个函数上,因此这里必须真正
|
||||
// 把页面交给前端,而不是记录个地址就算完。
|
||||
if r.state != nil {
|
||||
toast := func(call goja.FunctionCall) goja.Value {
|
||||
if len(call.Arguments) > 0 {
|
||||
r.state.Toast(call.Arguments[0].String())
|
||||
}
|
||||
return goja.Null()
|
||||
}
|
||||
set("toast", toast)
|
||||
set("longToast", toast)
|
||||
|
||||
// startBrowser(url, title[, html]):展示页面,不等待用户完成。
|
||||
set("startBrowser", func(call goja.FunctionCall) goja.Value {
|
||||
if err := r.openBrowser(parseBrowserArgs(call)); err != nil {
|
||||
bridgeErr("startBrowser", err)
|
||||
}
|
||||
return goja.Null()
|
||||
})
|
||||
// showBrowser(url, html, preloadJs, config):legado 中同样是「打开即返回」
|
||||
// 的对话框,只是参数含义不同,这里统一按展示处理。
|
||||
set("showBrowser", func(call goja.FunctionCall) goja.Value {
|
||||
req := parseBrowserArgs(call)
|
||||
if req.Title == "" {
|
||||
req.Title = stringArgOr(call, 2, "")
|
||||
}
|
||||
if err := r.openBrowser(req); err != nil {
|
||||
bridgeErr("showBrowser", err)
|
||||
}
|
||||
return goja.Null()
|
||||
})
|
||||
// startBrowserAwait(url, title[, refetchAfterSuccess][, html]):
|
||||
// 展示页面并阻塞等待用户完成后回传页面内容(对应 legado StrResponse)。
|
||||
set("startBrowserAwait", func(call goja.FunctionCall) goja.Value {
|
||||
req := parseBrowserArgs(call)
|
||||
if len(call.Arguments) > 2 && !goja.IsUndefined(call.Arguments[2]) && !goja.IsNull(call.Arguments[2]) {
|
||||
if _, isStr := call.Arguments[2].Export().(string); !isStr {
|
||||
req.Refetch = call.Arguments[2].ToBoolean()
|
||||
}
|
||||
}
|
||||
res, err := r.awaitBrowser(req)
|
||||
if err != nil {
|
||||
bridgeErr("startBrowserAwait", err)
|
||||
}
|
||||
finalURL := res.URL
|
||||
if finalURL == "" {
|
||||
finalURL = req.URL
|
||||
}
|
||||
return newResponseObject(vm, res.Body, 200, finalURL, nil)
|
||||
})
|
||||
|
||||
// reLoginView / refreshExplore:请求宿主重新渲染登录表单。
|
||||
// MeBox 每次登录动作后都会重新拉取 loginUi 并重建表单,因此这是真实
|
||||
// 生效的信号,而不是静默空实现。
|
||||
for _, name := range []string{"reLoginView", "refreshExplore"} {
|
||||
set(name, func(call goja.FunctionCall) goja.Value {
|
||||
r.requestUIRefresh()
|
||||
return goja.Null()
|
||||
})
|
||||
}
|
||||
// upLoginData(data):书源把服务端返回的值回填进登录表单。
|
||||
set("upLoginData", func(call goja.FunctionCall) goja.Value {
|
||||
r.applyLoginData(call.Arguments)
|
||||
return goja.Null()
|
||||
})
|
||||
}
|
||||
// 设备标识:部分源用 deviceID/androidId 做"是否支持该环境"探测,
|
||||
// 成功返回会让源走安卓分支,这里统一以异常告知不支持并回退到通用分支。
|
||||
for _, name := range []string{"deviceID", "androidId"} {
|
||||
set(name, func(call goja.FunctionCall) goja.Value {
|
||||
panic(vm.ToValue("java." + name + ": 服务端无设备标识"))
|
||||
})
|
||||
}
|
||||
// 说明:下面这些名字在服务端刻意不定义,调用即抛异常,以对齐 legado 的
|
||||
// 真实可用面,避免「静默空实现」让书源误判运行环境:
|
||||
// qread / showReadingBrowser / startBrowserDp —— legado 中并不存在
|
||||
// (例如 checkEnv 用 java.qread() 探测「轻阅」,旧实现返回成功会让书源
|
||||
// 误认为运行在轻阅里,从而跳过它自己的降级分支)
|
||||
// open / searchBook —— legado 中仅做原生界面跳转(打开搜索页 / 登录页),
|
||||
// 服务端没有对应页面,谎报成功会让书源走错分支
|
||||
|
||||
// 需要真正无头浏览器/本地文件系统的能力:明确抛出不支持,并把原因说清楚
|
||||
// (书源调试面板会把这句原样显示出来,用户能立刻判断该不该继续折腾这个源)。
|
||||
unsupported := func(name, reason string) func(goja.FunctionCall) goja.Value {
|
||||
return func(call goja.FunctionCall) goja.Value {
|
||||
panic(vm.ToValue(fmt.Sprintf("java.%s 服务端不支持:%s", name, reason)))
|
||||
}
|
||||
}
|
||||
for _, name := range []string{
|
||||
"webView", "webViewGetSource", "webViewGetOverrideUrl",
|
||||
"openVideoPlayer", "getVerificationCode",
|
||||
} {
|
||||
set(name, unsupported(name, "需要无头浏览器(WebView)"))
|
||||
}
|
||||
for _, name := range []string{
|
||||
"importScript", "cacheFile", "downloadFile", "readFile", "readTxtFile", "deleteFile",
|
||||
"unzipFile", "un7zFile", "unrarFile", "unArchiveFile", "getTxtInFolder",
|
||||
"getZipStringContent", "getZipByteArrayContent",
|
||||
"getRarStringContent", "get7zStringContent",
|
||||
} {
|
||||
set(name, unsupported(name, "需要本地文件/压缩包能力"))
|
||||
}
|
||||
|
||||
return o
|
||||
}
|
||||
|
||||
const defaultWebViewUA = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.0.0 Safari/537.36"
|
||||
|
||||
func md5Hex(s string, short bool) string {
|
||||
sum := md5.Sum([]byte(s))
|
||||
h := hex.EncodeToString(sum[:])
|
||||
if short {
|
||||
return h[8:24]
|
||||
}
|
||||
return h
|
||||
}
|
||||
|
||||
func digestHash(algorithm string) (hash.Hash, error) {
|
||||
switch strings.ToUpper(strings.NewReplacer("-", "", "_", "").Replace(algorithm)) {
|
||||
case "MD5":
|
||||
return md5.New(), nil
|
||||
case "SHA1":
|
||||
return sha1.New(), nil
|
||||
case "SHA256":
|
||||
return sha256.New(), nil
|
||||
case "SHA384":
|
||||
return sha512.New384(), nil
|
||||
case "SHA512":
|
||||
return sha512.New(), nil
|
||||
default:
|
||||
return nil, fmt.Errorf("不支持的摘要算法: %s", algorithm)
|
||||
}
|
||||
}
|
||||
|
||||
func hmacHash(algorithm string, key []byte) (hash.Hash, error) {
|
||||
switch strings.ToLower(strings.NewReplacer("-", "", "_", "").Replace(algorithm)) {
|
||||
case "hmacmd5":
|
||||
return hmac.New(md5.New, key), nil
|
||||
case "hmacsha1":
|
||||
return hmac.New(sha1.New, key), nil
|
||||
case "hmacsha256":
|
||||
return hmac.New(sha256.New, key), nil
|
||||
case "hmacsha512":
|
||||
return hmac.New(sha512.New, key), nil
|
||||
default:
|
||||
return nil, fmt.Errorf("不支持的 HMAC 算法: %s", algorithm)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,231 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"context"
|
||||
"encoding/base64"
|
||||
"fmt"
|
||||
"net/url"
|
||||
"strings"
|
||||
|
||||
"github.com/dop251/goja"
|
||||
)
|
||||
|
||||
// 本文件:书源 JS 的宿主浏览器能力契约。
|
||||
//
|
||||
// legado 用内置 WebView 实现 java.startBrowser / java.startBrowserAwait:
|
||||
// 打开一个页面让用户完成防爬校验、登录或参数选择,并(Await 版本)把用户
|
||||
// 操作后的页面源码作为 StrResponse 返回给书源解析。真实书源把整套「线路
|
||||
// 切换」「用户后台」「书源设置」都建在这两个函数上,因此服务端必须提供
|
||||
// 等价能力,否则这些按钮只会走到书源自己的 catch 分支。
|
||||
//
|
||||
// MeBox 没有 WebView,改由服务层把页面交给前端网页承载:
|
||||
// - AwaitBrowser:阻塞等待前端回传 DOM(对应 startBrowserAwait)
|
||||
// - OpenBrowser: 只展示不等待(对应 startBrowser / showBrowser)
|
||||
|
||||
// BrowserTask 一个需要用户在前端页面中完成的请求。
|
||||
type BrowserTask struct {
|
||||
// URL 书源传入的原始地址,可能是 http(s),也可能是 data:text/html;base64,…。
|
||||
URL string
|
||||
// Title 页面标题(对应 legado 对话框标题)。
|
||||
Title string
|
||||
// HTML URL 为 data: 系地址时解析出的页面源码,空表示由服务层按 URL 抓取。
|
||||
HTML string
|
||||
// Refetch 对应 legado startBrowserAwait 的 refetchAfterSuccess:
|
||||
// 用户确认后重新抓取该 URL 作为返回内容,而不是回传页面 DOM。
|
||||
Refetch bool
|
||||
}
|
||||
|
||||
// BrowserResult 用户在页面上完成操作后的回传。
|
||||
type BrowserResult struct {
|
||||
// URL 最终地址(对应 StrResponse.url())。
|
||||
URL string
|
||||
// Body 页面内容:默认是用户操作后的 DOM 序列化,
|
||||
// Refetch 为真时是服务端重新抓取的响应体。
|
||||
Body string
|
||||
}
|
||||
|
||||
// BrowserHost 由服务层注入的宿主浏览器实现。
|
||||
type BrowserHost interface {
|
||||
// AwaitBrowser 展示页面并阻塞等待用户回传(java.startBrowserAwait)。
|
||||
AwaitBrowser(ctx context.Context, req BrowserTask) (BrowserResult, error)
|
||||
// OpenBrowser 展示页面但不等待(java.startBrowser / java.showBrowser)。
|
||||
OpenBrowser(ctx context.Context, req BrowserTask) error
|
||||
}
|
||||
|
||||
// ErrBrowserCancelled 用户在页面里主动取消。
|
||||
var ErrBrowserCancelled = errBrowserCancelled{}
|
||||
|
||||
type errBrowserCancelled struct{}
|
||||
|
||||
func (errBrowserCancelled) Error() string { return "用户取消了页面操作" }
|
||||
|
||||
// ─── 运行时侧:把页面交给宿主浏览器 ────────────────────────────────────────
|
||||
|
||||
// parseBrowserArgs 解析 startBrowser / startBrowserAwait 的位置参数。
|
||||
//
|
||||
// startBrowser(url, title[, html])
|
||||
// startBrowserAwait(url, title[, refetchAfterSuccess][, html])
|
||||
//
|
||||
// 两者第 3 参类型不同(html 字符串 vs boolean),按实际类型判定。
|
||||
func parseBrowserArgs(call goja.FunctionCall) BrowserTask {
|
||||
req := BrowserTask{
|
||||
URL: stringArg(call, 0),
|
||||
Title: stringArgOr(call, 1, ""),
|
||||
}
|
||||
req.HTML = ParseDataHTML(req.URL)
|
||||
// 第 3 / 第 4 参里的字符串按 html 处理(data: URL 之外显式传入的页面源码)。
|
||||
for _, i := range []int{2, 3} {
|
||||
if len(call.Arguments) <= i {
|
||||
continue
|
||||
}
|
||||
arg := call.Arguments[i]
|
||||
if goja.IsUndefined(arg) || goja.IsNull(arg) {
|
||||
continue
|
||||
}
|
||||
if s, ok := arg.Export().(string); ok && strings.TrimSpace(s) != "" {
|
||||
req.HTML = s
|
||||
}
|
||||
}
|
||||
return req
|
||||
}
|
||||
|
||||
// openBrowser 展示页面但不等待。
|
||||
//
|
||||
// 注入了宿主浏览器时交给宿主(前端可交互的页面面板);未注入时退化为
|
||||
// 记录地址供前端代开,不中断书源逻辑。
|
||||
func (r *JSRunner) openBrowser(req BrowserTask) error {
|
||||
if req.URL == "" && req.HTML == "" {
|
||||
return nil
|
||||
}
|
||||
if r.cfg.Browser != nil {
|
||||
// 宿主抓取待展示的页面同样是网络等待,不能吃掉 JS 执行预算
|
||||
// (与 awaitBrowser 同一套暂停机制,见 pauseTimeout 的说明)。
|
||||
resume := r.pauseTimeout()
|
||||
defer resume()
|
||||
return r.cfg.Browser.OpenBrowser(r.ctx(), req)
|
||||
}
|
||||
if r.state != nil {
|
||||
r.state.OpenBrowser(req.URL, req.Title)
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
// awaitBrowser 展示页面并阻塞等待用户完成。
|
||||
//
|
||||
// 等待期间暂停 JS 超时计时:用户可能要看很久,默认 10s 的执行预算不能
|
||||
// 用在「等人」上。
|
||||
func (r *JSRunner) awaitBrowser(req BrowserTask) (BrowserResult, error) {
|
||||
if req.URL == "" && req.HTML == "" {
|
||||
return BrowserResult{}, fmt.Errorf("startBrowserAwait: 缺少页面地址")
|
||||
}
|
||||
if r.cfg.Browser == nil {
|
||||
if r.state != nil {
|
||||
r.state.OpenBrowser(req.URL, req.Title)
|
||||
}
|
||||
return BrowserResult{}, fmt.Errorf("服务端未提供浏览器能力,需要人工操作的页面请手动打开")
|
||||
}
|
||||
resume := r.pauseTimeout()
|
||||
defer resume()
|
||||
return r.cfg.Browser.AwaitBrowser(r.ctx(), req)
|
||||
}
|
||||
|
||||
// ctx 返回本次执行的上下文(未注入时用 Background)。
|
||||
func (r *JSRunner) ctx() context.Context {
|
||||
if r.cfg.Ctx != nil {
|
||||
return r.cfg.Ctx
|
||||
}
|
||||
return context.Background()
|
||||
}
|
||||
|
||||
// requestUIRefresh 请求宿主重新渲染登录表单(java.reLoginView / refreshExplore)。
|
||||
func (r *JSRunner) requestUIRefresh() {
|
||||
if ui, ok := r.state.(UIState); ok {
|
||||
ui.RequestUIRefresh()
|
||||
}
|
||||
}
|
||||
|
||||
// applyLoginData 把 java.upLoginData 传入的值合并进登录表单。
|
||||
func (r *JSRunner) applyLoginData(args []goja.Value) {
|
||||
if len(args) == 0 || goja.IsUndefined(args[0]) || goja.IsNull(args[0]) {
|
||||
return
|
||||
}
|
||||
ui, ok := r.state.(UIState)
|
||||
if !ok {
|
||||
return
|
||||
}
|
||||
exported, ok := args[0].Export().(map[string]any)
|
||||
if !ok {
|
||||
return
|
||||
}
|
||||
data := make(map[string]string, len(exported))
|
||||
for k, v := range exported {
|
||||
data[k] = anyToStringish(v)
|
||||
}
|
||||
ui.ApplyLoginData(data)
|
||||
}
|
||||
|
||||
func anyToStringish(v any) string {
|
||||
switch t := v.(type) {
|
||||
case nil:
|
||||
return ""
|
||||
case string:
|
||||
return t
|
||||
case bool:
|
||||
if t {
|
||||
return "true"
|
||||
}
|
||||
return "false"
|
||||
default:
|
||||
return fmt.Sprintf("%v", v)
|
||||
}
|
||||
}
|
||||
|
||||
// ParseDataHTML 解析 data:text/html 系地址,返回内嵌的页面源码。
|
||||
// 支持 base64 与百分号编码两种形式;非 data 地址或解码失败返回空串。
|
||||
func ParseDataHTML(rawURL string) string {
|
||||
s := strings.TrimSpace(rawURL)
|
||||
if !strings.HasPrefix(s, "data:") {
|
||||
return ""
|
||||
}
|
||||
comma := strings.Index(s, ",")
|
||||
if comma < 0 {
|
||||
return ""
|
||||
}
|
||||
meta := s[len("data:"):comma]
|
||||
payload := s[comma+1:]
|
||||
// MIME 限定在 text/*(书源也可能用 text/plain 塞 HTML)
|
||||
mime := meta
|
||||
if i := strings.Index(mime, ";"); i >= 0 {
|
||||
mime = mime[:i]
|
||||
}
|
||||
if mime != "" && !strings.HasPrefix(mime, "text/") {
|
||||
return ""
|
||||
}
|
||||
if strings.Contains(strings.ToLower(meta), ";base64") {
|
||||
// base64 可能带换行;同时兼容 URL-safe 变体
|
||||
cleaned := strings.Map(func(r rune) rune {
|
||||
switch r {
|
||||
case '\n', '\r', ' ', '\t':
|
||||
return -1
|
||||
}
|
||||
return r
|
||||
}, payload)
|
||||
if b, err := base64.StdEncoding.DecodeString(cleaned); err == nil {
|
||||
return string(b)
|
||||
}
|
||||
if b, err := base64.RawStdEncoding.DecodeString(cleaned); err == nil {
|
||||
return string(b)
|
||||
}
|
||||
if b, err := base64.URLEncoding.DecodeString(cleaned); err == nil {
|
||||
return string(b)
|
||||
}
|
||||
if b, err := base64.RawURLEncoding.DecodeString(cleaned); err == nil {
|
||||
return string(b)
|
||||
}
|
||||
return ""
|
||||
}
|
||||
if dec, err := url.PathUnescape(payload); err == nil {
|
||||
return dec
|
||||
}
|
||||
return ""
|
||||
}
|
||||
@@ -0,0 +1,305 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"context"
|
||||
"encoding/base64"
|
||||
"encoding/json"
|
||||
"strings"
|
||||
"sync"
|
||||
"testing"
|
||||
"time"
|
||||
)
|
||||
|
||||
// 本文件:书源宿主浏览器(java.startBrowser / startBrowserAwait)的运行时测试。
|
||||
//
|
||||
// 这些用例对应真实聚合类书源的「切换线路」「用户后台」路径:
|
||||
// 书源把页面交给宿主,宿主回传用户操作后的页面源码,书源再从 DOM 里解析结果。
|
||||
|
||||
// fakeBrowserHost 记录宿主调用并按用例给定的函数回传结果。
|
||||
type fakeBrowserHost struct {
|
||||
mu sync.Mutex
|
||||
awaitCalls []BrowserTask
|
||||
openCalls []BrowserTask
|
||||
respond func(req BrowserTask) (BrowserResult, error)
|
||||
openErr error
|
||||
}
|
||||
|
||||
func (f *fakeBrowserHost) AwaitBrowser(_ context.Context, req BrowserTask) (BrowserResult, error) {
|
||||
f.mu.Lock()
|
||||
f.awaitCalls = append(f.awaitCalls, req)
|
||||
respond := f.respond
|
||||
f.mu.Unlock()
|
||||
if respond == nil {
|
||||
return BrowserResult{URL: req.URL, Body: ""}, nil
|
||||
}
|
||||
return respond(req)
|
||||
}
|
||||
|
||||
func (f *fakeBrowserHost) OpenBrowser(_ context.Context, req BrowserTask) error {
|
||||
f.mu.Lock()
|
||||
defer f.mu.Unlock()
|
||||
f.openCalls = append(f.openCalls, req)
|
||||
return f.openErr
|
||||
}
|
||||
|
||||
// newBrowserRunner 用聚合源固件构建运行时,并注入宿主浏览器。
|
||||
func newBrowserRunner(t *testing.T, state SourceState, host BrowserHost, timeout time.Duration) *JSRunner {
|
||||
t.Helper()
|
||||
props := sampleSourceProps(t)
|
||||
r := NewJSRunner(JSConfig{
|
||||
SourceProps: props,
|
||||
JSLib: props["jsLib"].(string),
|
||||
State: state,
|
||||
BaseURL: "https://v1.example-aggregate.com",
|
||||
Browser: host,
|
||||
Timeout: timeout,
|
||||
Fetch: func(req *Request) (string, string, int, error) {
|
||||
return `{}`, req.URL, 200, nil
|
||||
},
|
||||
})
|
||||
if err := r.JSLibErr(); err != nil {
|
||||
t.Fatalf("jsLib 执行失败: %v", err)
|
||||
}
|
||||
return r
|
||||
}
|
||||
|
||||
// TestParseDataHTML 内嵌页面地址应被解码(base64 与百分号编码两种形式)。
|
||||
func TestParseDataHTML(t *testing.T) {
|
||||
page := "<html><body>线路</body></html>"
|
||||
cases := map[string]string{
|
||||
"base64": "data:text/html;base64," + base64.StdEncoding.EncodeToString([]byte(page)),
|
||||
"charset+base64": "data:text/html;charset=utf-8;base64," +
|
||||
base64.StdEncoding.EncodeToString([]byte(page)),
|
||||
"urlencoded": "data:text/html," + "%3Chtml%3E%3Cbody%3E%23%3C%2Fbody%3E%3C%2Fhtml%3E",
|
||||
}
|
||||
for name, raw := range cases {
|
||||
got := ParseDataHTML(raw)
|
||||
if got == "" {
|
||||
t.Fatalf("%s: 未解码出页面内容", name)
|
||||
}
|
||||
}
|
||||
if got := ParseDataHTML("https://example.com/a"); got != "" {
|
||||
t.Fatalf("非 data 地址不应解码,实际 %q", got)
|
||||
}
|
||||
if got := ParseDataHTML("data:image/png;base64,AAAA"); got != "" {
|
||||
t.Fatalf("非文本 MIME 不应解码,实际 %q", got)
|
||||
}
|
||||
// base64 带换行(书源拼接长 HTML 时常见)
|
||||
wrapped := "data:text/html;base64," + wrapBase64(base64.StdEncoding.EncodeToString([]byte(page)))
|
||||
if got := ParseDataHTML(wrapped); got != page {
|
||||
t.Fatalf("带换行的 base64 解码失败: %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
func wrapBase64(s string) string {
|
||||
var b strings.Builder
|
||||
for i, r := range s {
|
||||
if i > 0 && i%40 == 0 {
|
||||
b.WriteByte('\n')
|
||||
}
|
||||
b.WriteRune(r)
|
||||
}
|
||||
return b.String()
|
||||
}
|
||||
|
||||
// TestSwitchLineViaStartBrowserAwait 核心回归:书源用 startBrowserAwait 打开
|
||||
// 内嵌页面,宿主回传用户操作后的 DOM,书源据此切换线路。
|
||||
func TestSwitchLineViaStartBrowserAwait(t *testing.T) {
|
||||
state := NewMemoryState()
|
||||
host := &fakeBrowserHost{}
|
||||
host.respond = func(req BrowserTask) (BrowserResult, error) {
|
||||
// 宿主应当已经把 data: 地址解码成可承载的页面源码
|
||||
if !strings.Contains(req.HTML, `id="serverValue"`) {
|
||||
t.Errorf("宿主未收到解码后的页面: %q", req.HTML)
|
||||
}
|
||||
if req.Title != "线路设置" {
|
||||
t.Errorf("标题 = %q", req.Title)
|
||||
}
|
||||
if req.Refetch {
|
||||
t.Error("第 3 个参数为 false 时不应要求重新抓取")
|
||||
}
|
||||
// 模拟用户在页面里选了 v2 线路后回传的 DOM
|
||||
return BrowserResult{
|
||||
URL: req.URL,
|
||||
Body: `<html><body><span id="serverValue">https://v2.example-aggregate.com</span></body></html>`,
|
||||
}, nil
|
||||
}
|
||||
r := newBrowserRunner(t, state, host, 5*time.Second)
|
||||
loginJS := loadTestdata(t, "sample_loginUrl.js")
|
||||
|
||||
v, err := r.EvalAction(loginJS+"\nswitchLine()", nil)
|
||||
if err != nil {
|
||||
t.Fatalf("switchLine 执行失败: %v", err)
|
||||
}
|
||||
if got := anyToString(v); got != "https://v2.example-aggregate.com" {
|
||||
t.Fatalf("switchLine 返回 %q", got)
|
||||
}
|
||||
// 书源应把解析出的线路写进源变量,并让 BaseUrl 跟随
|
||||
if got := state.GetVariable(); !strings.Contains(got, "v2.example-aggregate.com") {
|
||||
t.Fatalf("线路未写入源变量: %q", got)
|
||||
}
|
||||
base, err := r.EvalAction(`BaseUrl()`, nil)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got := anyToString(base); got != "https://v2.example-aggregate.com" {
|
||||
t.Fatalf("BaseUrl 未跟随线路: %q", got)
|
||||
}
|
||||
if len(host.awaitCalls) != 1 {
|
||||
t.Fatalf("宿主 await 调用次数 = %d", len(host.awaitCalls))
|
||||
}
|
||||
}
|
||||
|
||||
// TestStartBrowserAwaitPausesJSTimeout 等待人工操作期间不应被 JS 执行超时打断。
|
||||
// 宿主故意比 Timeout 慢,若看门狗没有暂停,goja 会在恢复执行时抛出超时异常。
|
||||
func TestStartBrowserAwaitPausesJSTimeout(t *testing.T) {
|
||||
state := NewMemoryState()
|
||||
host := &fakeBrowserHost{}
|
||||
host.respond = func(req BrowserTask) (BrowserResult, error) {
|
||||
// 模拟用户慢慢点选:远超 Timeout
|
||||
time.Sleep(300 * time.Millisecond)
|
||||
return BrowserResult{URL: req.URL, Body: "ok"}, nil
|
||||
}
|
||||
r := newBrowserRunner(t, state, host, 80*time.Millisecond)
|
||||
loginJS := loadTestdata(t, "sample_loginUrl.js")
|
||||
|
||||
if _, err := r.EvalAction(loginJS+"\nswitchLine()", nil); err != nil {
|
||||
t.Fatalf("等待人工操作期间被误判为 JS 超时: %v", err)
|
||||
}
|
||||
}
|
||||
|
||||
// TestStartBrowserAwaitCancel 用户取消页面时应把异常抛回书源,
|
||||
// 书源走自己的降级分支(提示语),而不是拿到空 body 当成成功。
|
||||
func TestStartBrowserAwaitCancel(t *testing.T) {
|
||||
state := NewMemoryState()
|
||||
host := &fakeBrowserHost{}
|
||||
host.respond = func(BrowserTask) (BrowserResult, error) {
|
||||
return BrowserResult{}, ErrBrowserCancelled
|
||||
}
|
||||
r := newBrowserRunner(t, state, host, 5*time.Second)
|
||||
loginJS := loadTestdata(t, "sample_loginUrl.js")
|
||||
|
||||
_, err := r.EvalAction(loginJS+"\nswitchLine()", nil)
|
||||
if err == nil {
|
||||
t.Fatal("取消后应把异常抛回书源")
|
||||
}
|
||||
if !strings.Contains(err.Error(), "取消") {
|
||||
t.Fatalf("错误信息未说明取消: %v", err)
|
||||
}
|
||||
}
|
||||
|
||||
// TestUserBackendUsesAwait 用户后台按钮同样走 startBrowserAwait(普通 http 地址)。
|
||||
func TestUserBackendUsesAwait(t *testing.T) {
|
||||
state := NewMemoryState()
|
||||
state.SetCookie("https://v1.example-aggregate.com", "qttoken=TOKEN_abcdefghijklmn")
|
||||
host := &fakeBrowserHost{}
|
||||
host.respond = func(req BrowserTask) (BrowserResult, error) {
|
||||
if !strings.HasSuffix(req.URL, "/user") {
|
||||
t.Errorf("用户后台地址 = %q", req.URL)
|
||||
}
|
||||
return BrowserResult{URL: req.URL, Body: "<html>用户后台</html>"}, nil
|
||||
}
|
||||
r := newBrowserRunner(t, state, host, 5*time.Second)
|
||||
loginJS := loadTestdata(t, "sample_loginUrl.js")
|
||||
|
||||
if _, err := r.EvalAction(loginJS+"\nuser()", nil); err != nil {
|
||||
t.Fatalf("user() 执行失败: %v", err)
|
||||
}
|
||||
if len(host.awaitCalls) != 1 {
|
||||
t.Fatalf("应为用户后台打开页面,await 调用 = %d", len(host.awaitCalls))
|
||||
}
|
||||
}
|
||||
|
||||
// TestStartBrowserDoesNotWait startBrowser 只展示,不阻塞也不回传。
|
||||
func TestStartBrowserDoesNotWait(t *testing.T) {
|
||||
state := NewMemoryState()
|
||||
host := &fakeBrowserHost{}
|
||||
r := newBrowserRunner(t, state, host, 5*time.Second)
|
||||
|
||||
if _, err := r.EvalAction(`java.startBrowser('https://vip.example.com', '光遇看书')`, nil); err != nil {
|
||||
t.Fatalf("startBrowser 失败: %v", err)
|
||||
}
|
||||
if len(host.openCalls) != 1 || len(host.awaitCalls) != 0 {
|
||||
t.Fatalf("startBrowser 应只展示不等待: open=%d await=%d", len(host.openCalls), len(host.awaitCalls))
|
||||
}
|
||||
if host.openCalls[0].URL != "https://vip.example.com" {
|
||||
t.Fatalf("展示地址 = %q", host.openCalls[0].URL)
|
||||
}
|
||||
}
|
||||
|
||||
// TestShowBrowserOpensDialog showBrowser 与 startBrowser 一样是「打开即返回」。
|
||||
func TestShowBrowserOpensDialog(t *testing.T) {
|
||||
state := NewMemoryState()
|
||||
host := &fakeBrowserHost{}
|
||||
r := newBrowserRunner(t, state, host, 5*time.Second)
|
||||
|
||||
if _, err := r.EvalAction(`java.showBrowser('https://example.com/settings', '', '光遇书源设置', '')`, nil); err != nil {
|
||||
t.Fatalf("showBrowser 失败: %v", err)
|
||||
}
|
||||
if len(host.openCalls) != 1 {
|
||||
t.Fatalf("showBrowser 应展示页面: %d", len(host.openCalls))
|
||||
}
|
||||
}
|
||||
|
||||
// TestEnvProbeAPIsThrow 服务端不具备的 UI 函数必须抛异常,而不是静默成功。
|
||||
//
|
||||
// 这条对真实书源很关键:聚合源用 checkEnv() 探测运行环境,其中
|
||||
// try { java.qread(); return "轻阅"; } catch (e) {}
|
||||
// 旧实现把 qread 做成「返回 null 的空操作」,于是探测结果变成「轻阅」,
|
||||
// 书源据此跳过自己的降级分支,表现为「按钮点了没反应」。
|
||||
func TestEnvProbeAPIsThrow(t *testing.T) {
|
||||
r := NewJSRunner(JSConfig{State: NewMemoryState()})
|
||||
for _, name := range []string{"qread", "showReadingBrowser", "startBrowserDp", "open", "searchBook"} {
|
||||
if _, err := r.EvalAction("java."+name+"()", nil); err == nil {
|
||||
t.Fatalf("java.%s 应当抛异常(服务端无此能力)", name)
|
||||
}
|
||||
}
|
||||
// checkEnv 结构:qread 抛异常后应落到最后的 "改版",而不是 "轻阅"
|
||||
v, err := r.EvalAction(`(function () {
|
||||
try { java.qread(); return '轻阅'; } catch (e) {}
|
||||
try { java.deviceID(); return '苹果'; } catch (e) {}
|
||||
return '改版';
|
||||
})()`, nil)
|
||||
if err != nil {
|
||||
t.Fatalf("环境探测执行失败: %v", err)
|
||||
}
|
||||
if got := anyToString(v); got != "改版" {
|
||||
t.Fatalf("环境探测 = %q,期望 改版", got)
|
||||
}
|
||||
}
|
||||
|
||||
// TestUpLoginDataAndReLoginView 表单回填与重画信号应被显式记录。
|
||||
func TestUpLoginDataAndReLoginView(t *testing.T) {
|
||||
state := NewMemoryState()
|
||||
r := NewJSRunner(JSConfig{State: state})
|
||||
|
||||
if _, err := r.EvalAction(`java.upLoginData({邮箱:'user@example.com', 昵称:'tester'})`, nil); err != nil {
|
||||
t.Fatalf("upLoginData 失败: %v", err)
|
||||
}
|
||||
var m map[string]string
|
||||
if err := json.Unmarshal([]byte(state.GetLoginInfo()), &m); err != nil {
|
||||
t.Fatalf("登录信息不是 JSON: %q", state.GetLoginInfo())
|
||||
}
|
||||
if m["邮箱"] != "user@example.com" || m["昵称"] != "tester" {
|
||||
t.Fatalf("表单值未回填: %v", m)
|
||||
}
|
||||
if !state.UIRefreshRequested() {
|
||||
t.Fatal("upLoginData 应请求重画表单")
|
||||
}
|
||||
// 标记是「取走即清」
|
||||
if state.UIRefreshRequested() {
|
||||
t.Fatal("重画标记应被消费")
|
||||
}
|
||||
if _, err := r.EvalAction(`java.reLoginView()`, nil); err != nil {
|
||||
t.Fatalf("reLoginView 失败: %v", err)
|
||||
}
|
||||
if !state.UIRefreshRequested() {
|
||||
t.Fatal("reLoginView 应请求重画表单")
|
||||
}
|
||||
if _, err := r.EvalAction(`java.refreshExplore()`, nil); err != nil {
|
||||
t.Fatalf("refreshExplore 失败: %v", err)
|
||||
}
|
||||
if !state.UIRefreshRequested() {
|
||||
t.Fatal("refreshExplore 应请求重画表单")
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,285 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"crypto/aes"
|
||||
"crypto/cipher"
|
||||
"crypto/des"
|
||||
"encoding/hex"
|
||||
"errors"
|
||||
"fmt"
|
||||
"strings"
|
||||
|
||||
"github.com/dop251/goja"
|
||||
)
|
||||
|
||||
// 对称加密:对应 legado JsEncodeUtils.createSymmetricCrypto(hutool SymmetricCrypto 语义)。
|
||||
// transformation 形如 "AES/CBC/PKCS5Padding" / "DES/ECB/NoPadding" / "DESede/CBC/PKCS5Padding"。
|
||||
|
||||
type symmetricCipher struct {
|
||||
block cipher.Block
|
||||
mode string
|
||||
padding string
|
||||
iv []byte
|
||||
}
|
||||
|
||||
func newSymmetricCipher(transformation, key, iv string) (*symmetricCipher, error) {
|
||||
parts := strings.Split(strings.TrimSpace(transformation), "/")
|
||||
algo := strings.ToUpper(strings.TrimSpace(parts[0]))
|
||||
mode := "CBC"
|
||||
padding := "PKCS5Padding"
|
||||
if len(parts) > 1 && strings.TrimSpace(parts[1]) != "" {
|
||||
mode = strings.ToUpper(strings.TrimSpace(parts[1]))
|
||||
}
|
||||
if len(parts) > 2 && strings.TrimSpace(parts[2]) != "" {
|
||||
padding = strings.TrimSpace(parts[2])
|
||||
}
|
||||
|
||||
keyBytes := normalizeKey([]byte(key), algo)
|
||||
if len(keyBytes) == 0 {
|
||||
return nil, errors.New("加密密钥为空")
|
||||
}
|
||||
var block cipher.Block
|
||||
var err error
|
||||
switch algo {
|
||||
case "AES":
|
||||
block, err = aes.NewCipher(keyBytes)
|
||||
case "DES":
|
||||
block, err = des.NewCipher(keyBytes)
|
||||
case "DESEDE", "3DES", "TRIPLEDES":
|
||||
block, err = des.NewTripleDESCipher(keyBytes)
|
||||
default:
|
||||
return nil, fmt.Errorf("不支持的加密算法: %s", algo)
|
||||
}
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
ivBytes := []byte(iv)
|
||||
if mode == "ECB" {
|
||||
ivBytes = nil
|
||||
}
|
||||
return &symmetricCipher{block: block, mode: mode, padding: padding, iv: ivBytes}, nil
|
||||
}
|
||||
|
||||
// normalizeKey 对齐 hutool 的密钥处理:按算法要求补齐/截断密钥长度。
|
||||
func normalizeKey(key []byte, algo string) []byte {
|
||||
var sizes []int
|
||||
switch algo {
|
||||
case "AES":
|
||||
sizes = []int{16, 24, 32}
|
||||
case "DES":
|
||||
sizes = []int{8}
|
||||
case "DESEDE", "3DES", "TRIPLEDES":
|
||||
sizes = []int{24}
|
||||
default:
|
||||
return key
|
||||
}
|
||||
for _, n := range sizes {
|
||||
if len(key) == n {
|
||||
return key
|
||||
}
|
||||
}
|
||||
for _, n := range sizes {
|
||||
if len(key) < n {
|
||||
out := make([]byte, n)
|
||||
copy(out, key)
|
||||
return out
|
||||
}
|
||||
}
|
||||
return key[:sizes[len(sizes)-1]]
|
||||
}
|
||||
|
||||
func (c *symmetricCipher) blockSize() int { return c.block.BlockSize() }
|
||||
|
||||
func (c *symmetricCipher) crypt(dst, src []byte, decrypt bool) error {
|
||||
bs := c.blockSize()
|
||||
if c.mode == "ECB" {
|
||||
for i := 0; i < len(src); i += bs {
|
||||
if decrypt {
|
||||
c.block.Decrypt(dst[i:i+bs], src[i:i+bs])
|
||||
} else {
|
||||
c.block.Encrypt(dst[i:i+bs], src[i:i+bs])
|
||||
}
|
||||
}
|
||||
return nil
|
||||
}
|
||||
iv := make([]byte, bs)
|
||||
if len(c.iv) > 0 {
|
||||
copy(iv, c.iv[:min(len(c.iv), bs)])
|
||||
}
|
||||
switch c.mode {
|
||||
case "CBC":
|
||||
if decrypt {
|
||||
cipher.NewCBCDecrypter(c.block, iv).CryptBlocks(dst, src)
|
||||
} else {
|
||||
cipher.NewCBCEncrypter(c.block, iv).CryptBlocks(dst, src)
|
||||
}
|
||||
case "CTR":
|
||||
cipher.NewCTR(c.block, iv).XORKeyStream(dst, src)
|
||||
case "OFB":
|
||||
cipher.NewOFB(c.block, iv).XORKeyStream(dst, src)
|
||||
case "CFB", "CFB8":
|
||||
stream := cipher.NewCFBEncrypter(c.block, iv)
|
||||
if decrypt {
|
||||
stream = cipher.NewCFBDecrypter(c.block, iv)
|
||||
}
|
||||
stream.XORKeyStream(dst, src)
|
||||
default:
|
||||
return fmt.Errorf("不支持的加密模式: %s", c.mode)
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
func (c *symmetricCipher) encrypt(data []byte) ([]byte, error) {
|
||||
bs := c.blockSize()
|
||||
data = applyPadding(data, bs, c.padding)
|
||||
if len(data)%bs != 0 {
|
||||
return nil, errors.New("加密数据长度未按块对齐")
|
||||
}
|
||||
out := make([]byte, len(data))
|
||||
if err := c.crypt(out, data, false); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
return out, nil
|
||||
}
|
||||
|
||||
func (c *symmetricCipher) decrypt(data []byte) ([]byte, error) {
|
||||
bs := c.blockSize()
|
||||
if len(data) == 0 || len(data)%bs != 0 {
|
||||
return nil, errors.New("密文长度不是块大小的整数倍")
|
||||
}
|
||||
out := make([]byte, len(data))
|
||||
if err := c.crypt(out, data, true); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
return stripPadding(out, c.padding), nil
|
||||
}
|
||||
|
||||
// decryptAuto 对应 hutool decryptStr(String):输入自动识别 Base64 / Hex / 原文。
|
||||
func (c *symmetricCipher) decryptAuto(s string) ([]byte, error) {
|
||||
if b, err := base64DecodeBytes(s); err == nil && len(b) > 0 && len(b)%c.blockSize() == 0 {
|
||||
return c.decrypt(b)
|
||||
}
|
||||
if b, err := hex.DecodeString(strings.TrimSpace(s)); err == nil && len(b) > 0 && len(b)%c.blockSize() == 0 {
|
||||
return c.decrypt(b)
|
||||
}
|
||||
return c.decrypt([]byte(s))
|
||||
}
|
||||
|
||||
func applyPadding(data []byte, blockSize int, padding string) []byte {
|
||||
switch padding {
|
||||
case "NoPadding":
|
||||
return data
|
||||
case "ZeroPadding":
|
||||
n := (len(data)+blockSize-1)/blockSize*blockSize
|
||||
if n == 0 {
|
||||
n = blockSize
|
||||
}
|
||||
out := make([]byte, n)
|
||||
copy(out, data)
|
||||
return out
|
||||
default: // PKCS5Padding / PKCS7Padding
|
||||
pad := blockSize - len(data)%blockSize
|
||||
out := make([]byte, len(data)+pad)
|
||||
copy(out, data)
|
||||
for i := len(data); i < len(out); i++ {
|
||||
out[i] = byte(pad)
|
||||
}
|
||||
return out
|
||||
}
|
||||
}
|
||||
|
||||
func stripPadding(data []byte, padding string) []byte {
|
||||
switch padding {
|
||||
case "NoPadding":
|
||||
return data
|
||||
case "ZeroPadding":
|
||||
for len(data) > 0 && data[len(data)-1] == 0 {
|
||||
data = data[:len(data)-1]
|
||||
}
|
||||
return data
|
||||
default:
|
||||
if len(data) == 0 {
|
||||
return data
|
||||
}
|
||||
pad := int(data[len(data)-1])
|
||||
if pad <= 0 || pad > len(data) {
|
||||
return data
|
||||
}
|
||||
for _, b := range data[len(data)-pad:] {
|
||||
if int(b) != pad {
|
||||
return data
|
||||
}
|
||||
}
|
||||
return data[:len(data)-pad]
|
||||
}
|
||||
}
|
||||
|
||||
// newCipherObject 对应 hutool SymmetricCrypto 的 JS 方法面:
|
||||
// encrypt(data) / encryptBase64(data) / encryptHex(data) / decrypt(data) /
|
||||
// decryptStr(data) / decryptBase64(data) / decryptHex(data)。
|
||||
func newCipherObject(vm *goja.Runtime, c *symmetricCipher) *goja.Object {
|
||||
o := vm.NewObject()
|
||||
set := func(k string, v any) {
|
||||
if err := o.Set(k, v); err != nil {
|
||||
panic(vm.ToValue(err.Error()))
|
||||
}
|
||||
}
|
||||
set("encrypt", func(call goja.FunctionCall) goja.Value {
|
||||
out, err := c.encrypt([]byte(stringArg(call, 0)))
|
||||
if err != nil {
|
||||
panic(vm.ToValue(err.Error()))
|
||||
}
|
||||
return vm.ToValue(vm.NewArrayBuffer(out))
|
||||
})
|
||||
set("encryptBase64", func(call goja.FunctionCall) goja.Value {
|
||||
out, err := c.encrypt([]byte(stringArg(call, 0)))
|
||||
if err != nil {
|
||||
panic(vm.ToValue(err.Error()))
|
||||
}
|
||||
return vm.ToValue(base64StdEncodeBytes(out))
|
||||
})
|
||||
set("encryptHex", func(call goja.FunctionCall) goja.Value {
|
||||
out, err := c.encrypt([]byte(stringArg(call, 0)))
|
||||
if err != nil {
|
||||
panic(vm.ToValue(err.Error()))
|
||||
}
|
||||
return vm.ToValue(hexEncodeBytes(out))
|
||||
})
|
||||
set("decrypt", func(call goja.FunctionCall) goja.Value {
|
||||
out, err := c.decryptAuto(stringArg(call, 0))
|
||||
if err != nil {
|
||||
panic(vm.ToValue(err.Error()))
|
||||
}
|
||||
return vm.ToValue(vm.NewArrayBuffer(out))
|
||||
})
|
||||
set("decryptStr", func(call goja.FunctionCall) goja.Value {
|
||||
out, err := c.decryptAuto(stringArg(call, 0))
|
||||
if err != nil {
|
||||
panic(vm.ToValue(err.Error()))
|
||||
}
|
||||
return vm.ToValue(string(out))
|
||||
})
|
||||
set("decryptBase64", func(call goja.FunctionCall) goja.Value {
|
||||
b, err := base64DecodeBytes(stringArg(call, 0))
|
||||
if err != nil {
|
||||
panic(vm.ToValue(err.Error()))
|
||||
}
|
||||
out, err := c.decrypt(b)
|
||||
if err != nil {
|
||||
panic(vm.ToValue(err.Error()))
|
||||
}
|
||||
return vm.ToValue(string(out))
|
||||
})
|
||||
set("decryptHex", func(call goja.FunctionCall) goja.Value {
|
||||
b, err := hexDecodeBytes(strings.TrimSpace(stringArg(call, 0)))
|
||||
if err != nil {
|
||||
panic(vm.ToValue(err.Error()))
|
||||
}
|
||||
out, err := c.decrypt(b)
|
||||
if err != nil {
|
||||
panic(vm.ToValue(err.Error()))
|
||||
}
|
||||
return vm.ToValue(string(out))
|
||||
})
|
||||
return o
|
||||
}
|
||||
@@ -0,0 +1,81 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"encoding/base64"
|
||||
"encoding/hex"
|
||||
"regexp"
|
||||
"strings"
|
||||
)
|
||||
|
||||
// 本文件:data: 地址与「type 声明」——书源用它当参数信封。
|
||||
//
|
||||
// 聚合类书源(光遇聚合等)会把上一阶段的结果打包进一个 data: 地址当 URL 用:
|
||||
//
|
||||
// searchUrl → `data:;base64,<base64(参数JSON)>,{"type":"gysearch"}`
|
||||
// bookUrl → `data:;base64,<base64({book_id,sources,…})>,{"type":"gydetail"}`
|
||||
// tocUrl / chapterUrl 同理
|
||||
//
|
||||
// legado 对这类地址的处理(AnalyzeUrl):
|
||||
// 1. getByteArrayIfDataUri():地址以 data: 开头时本地 base64 解码取字节,不发网络请求;
|
||||
// 2. getStrResponseAwait():只要选项里声明了 type(值本身不参与判断),
|
||||
// 就直接返回 `HexUtil.encodeHexStr(字节)`——即原始字节的十六进制串。
|
||||
//
|
||||
// 书源随后用 `java.hexDecodeToString(result)` 把 hex 还原成 JSON 取回参数,
|
||||
// 再自行发出真正的请求。所以「type」不是内容类型,而是一个「请以 hex 返回
|
||||
// 原始字节」的开关;MeBox 早期把它当成不支持的能力直接拒绝了,导致这类书源
|
||||
// 在搜索第一步就报「书源 URL 声明了不支持的 type」。
|
||||
|
||||
// dataURIPayloadRe 对应 legado AppPattern.dataUriRegex:`^data:.*?;base64,(.*)`。
|
||||
// 只认带 `;base64,` 的数据地址(legado 也只处理这一种)。
|
||||
var dataURIPayloadRe = regexp.MustCompile(`(?s)^data:.*?;base64,(.*)$`)
|
||||
|
||||
// DecodeDataURI 解析 `data:…;base64,…` 地址,返回解码后的原始字节。
|
||||
// ok 为 false 表示这不是一个可解析的 base64 数据地址。
|
||||
// 对应 legado AnalyzeUrl.getByteArrayIfDataUri()。
|
||||
func DecodeDataURI(raw string) ([]byte, bool) {
|
||||
m := dataURIPayloadRe.FindStringSubmatch(strings.TrimSpace(raw))
|
||||
if m == nil {
|
||||
return nil, false
|
||||
}
|
||||
// base64 允许换行与空白(书源拼接长载荷时常见),逐种变体尝试。
|
||||
cleaned := strings.Map(func(r rune) rune {
|
||||
switch r {
|
||||
case '\n', '\r', ' ', '\t':
|
||||
return -1
|
||||
}
|
||||
return r
|
||||
}, m[1])
|
||||
if cleaned == "" {
|
||||
return []byte{}, true
|
||||
}
|
||||
for _, enc := range []*base64.Encoding{
|
||||
base64.StdEncoding, base64.RawStdEncoding,
|
||||
base64.URLEncoding, base64.RawURLEncoding,
|
||||
} {
|
||||
if b, err := enc.DecodeString(cleaned); err == nil {
|
||||
return b, true
|
||||
}
|
||||
}
|
||||
return nil, false
|
||||
}
|
||||
|
||||
// IsDataURI 判断地址是否是 data: 地址。
|
||||
func IsDataURI(raw string) bool {
|
||||
return strings.HasPrefix(strings.TrimSpace(raw), "data:")
|
||||
}
|
||||
|
||||
// EncodeRuleBody 按请求声明的响应形态把原始字节转成规则层看到的 body。
|
||||
//
|
||||
// HexBody(选项里声明了 type)→ 原始字节的十六进制串
|
||||
// 其他 → 按 charset 解码的文本
|
||||
//
|
||||
// 对应 legado getStrResponseAwait 里的两个分支。
|
||||
func EncodeRuleBody(req *Request, raw []byte, charset string) string {
|
||||
if req != nil && req.HexBody {
|
||||
return hex.EncodeToString(raw)
|
||||
}
|
||||
if body, err := DecodeBytes(raw, charset); err == nil {
|
||||
return body
|
||||
}
|
||||
return string(raw)
|
||||
}
|
||||
@@ -0,0 +1,108 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"encoding/base64"
|
||||
"encoding/hex"
|
||||
"strings"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// 本文件:data: 参数信封 + URL 选项 type 的回归测试。
|
||||
//
|
||||
// 背景:聚合类书源把上一阶段的结果打包成
|
||||
// `data:;base64,<base64(参数JSON)>,{"type":"gysearch"}`
|
||||
// 当 URL 用。legado 对它的处理是:① 本地 base64 解码取字节;② 因为声明了 type,
|
||||
// 把字节按 hex 返回。MeBox 早期把带 type 的 URL 直接判为「不支持的能力」,
|
||||
// 导致这类书源第一步就报「书源 URL 声明了不支持的 type」。
|
||||
|
||||
func TestDecodeDataURI(t *testing.T) {
|
||||
payload := `{"key":"宠魅","page":1}`
|
||||
enc := base64.StdEncoding.EncodeToString([]byte(payload))
|
||||
|
||||
cases := map[string]string{
|
||||
"标准 base64": "data:;base64," + enc,
|
||||
"带 MIME": "data:application/json;base64," + enc,
|
||||
"无 padding": "data:;base64," + base64.RawStdEncoding.EncodeToString([]byte(payload)),
|
||||
"URL-safe": "data:;base64," + base64.URLEncoding.EncodeToString([]byte(payload)),
|
||||
"含空白/换行": "data:;base64," + wrapBase64(enc),
|
||||
}
|
||||
for name, raw := range cases {
|
||||
got, ok := DecodeDataURI(raw)
|
||||
if !ok {
|
||||
t.Fatalf("%s: 应能解码", name)
|
||||
}
|
||||
if string(got) != payload {
|
||||
t.Fatalf("%s: 解码结果 %q", name, string(got))
|
||||
}
|
||||
}
|
||||
|
||||
for _, raw := range []string{
|
||||
"https://example.com/a",
|
||||
"data:text/plain,hello", // 非 base64 形式
|
||||
"",
|
||||
} {
|
||||
if _, ok := DecodeDataURI(raw); ok {
|
||||
t.Fatalf("不应识别为 base64 数据地址: %q", raw)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestParseAnalyzeUrlDataTypeDeclaration 带 type 的 data: 地址应被解析为
|
||||
// 「hex 返回」的请求,而不是不支持的能力。
|
||||
func TestParseAnalyzeUrlDataTypeDeclaration(t *testing.T) {
|
||||
payload := base64.StdEncoding.EncodeToString([]byte(`{"key":"宠魅"}`))
|
||||
raw := "data:;base64," + payload + `,{"type":"gysearch"}`
|
||||
|
||||
req, err := ParseAnalyzeUrlWithJS(raw, "", 0, "https://example.com", NewJSRunner(JSConfig{}))
|
||||
if err != nil {
|
||||
t.Fatalf("解析失败: %v", err)
|
||||
}
|
||||
if req.Unsupported != nil {
|
||||
t.Fatalf("带 type 的地址不应被判为不支持: %v", req.Unsupported)
|
||||
}
|
||||
if !req.HexBody {
|
||||
t.Fatal("声明了 type 时应按 hex 返回响应")
|
||||
}
|
||||
// 选项里的 JSON 不能被当成数据载荷的一部分
|
||||
if strings.Contains(req.URL, `"type"`) {
|
||||
t.Fatalf("选项未与地址分离: %q", req.URL)
|
||||
}
|
||||
if _, ok := DecodeDataURI(req.URL); !ok {
|
||||
t.Fatalf("URL 不是可解码的数据地址: %q", req.URL)
|
||||
}
|
||||
}
|
||||
|
||||
// TestParseAnalyzeUrlWithoutType 没有 type 时不应进入 hex 模式。
|
||||
func TestParseAnalyzeUrlWithoutType(t *testing.T) {
|
||||
req, err := ParseAnalyzeUrlWithJS("https://example.com/a?x=1", "", 0, "https://example.com", NewJSRunner(JSConfig{}))
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if req.HexBody {
|
||||
t.Fatal("未声明 type 时不应按 hex 返回")
|
||||
}
|
||||
if req.URL != "https://example.com/a?x=1" {
|
||||
t.Fatalf("URL = %q", req.URL)
|
||||
}
|
||||
}
|
||||
|
||||
// TestEncodeRuleBody 声明 type 时返回原始字节的 hex,否则返回解码后的文本。
|
||||
func TestEncodeRuleBody(t *testing.T) {
|
||||
raw := []byte(`{"key":"宠魅"}`)
|
||||
|
||||
hexReq := &Request{HexBody: true}
|
||||
got := EncodeRuleBody(hexReq, raw, "")
|
||||
if got != hex.EncodeToString(raw) {
|
||||
t.Fatalf("hex 模式 = %q", got)
|
||||
}
|
||||
// 书源会用它还原
|
||||
back, err := hex.DecodeString(got)
|
||||
if err != nil || string(back) != string(raw) {
|
||||
t.Fatalf("hex 往返失败: %v %q", err, string(back))
|
||||
}
|
||||
|
||||
plainReq := &Request{}
|
||||
if got := EncodeRuleBody(plainReq, raw, ""); got != string(raw) {
|
||||
t.Fatalf("非 hex 模式 = %q", got)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,15 @@
|
||||
// Package rule 移植自 legado(refgd/legado,GPL-3.0)的规则引擎:
|
||||
// io.legado.app.model.analyzeRule 包下的 AnalyzeRule / AnalyzeByJSoup /
|
||||
// AnalyzeByJSonPath / AnalyzeByXPath / AnalyzeByRegex / AnalyzeUrl / RuleAnalyzer。
|
||||
// 语义以 legado 源码为基准逐项对齐,注释中标注了对应的 Kotlin 方法名。
|
||||
package rule
|
||||
|
||||
import "errors"
|
||||
|
||||
var (
|
||||
// ErrJsUnsupported 书源规则中包含 JS(<js>/@js:/{{}}/js 选项)。
|
||||
// JS 引擎(goja + java.* 桥)在 P2 阶段接入,届时移除本错误路径。
|
||||
ErrJsUnsupported = errors.New("书源使用了 JS 规则,当前阶段暂不支持")
|
||||
// ErrWebJSUnsupported 书源依赖 webView/webJs 抓取,服务端无头浏览器不在支持范围。
|
||||
ErrWebJSUnsupported = errors.New("书源依赖 webView/webJs(需要无头浏览器),服务端不支持;该源请在阅读 App 内使用")
|
||||
)
|
||||
@@ -0,0 +1,528 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"context"
|
||||
"fmt"
|
||||
"strings"
|
||||
"sync"
|
||||
"time"
|
||||
|
||||
"github.com/dop251/goja"
|
||||
)
|
||||
|
||||
// 本文件对应 legado 的 Rhino JS 执行层(RhinoScriptEngine / evalJS),
|
||||
// 用纯 Go 的 goja 实现。`java` 对象的函数在 bridge.go 中以 Go 原生实现,
|
||||
// 函数名与 legado JsExtensions 保持一致(存量书源 JS 硬编码了这些名字)。
|
||||
|
||||
const (
|
||||
defaultJSTimeout = 10 * time.Second
|
||||
// programCacheSize 编译产物缓存上限(超出后整体清空,防无限增长)。
|
||||
programCacheSize = 256
|
||||
)
|
||||
|
||||
var (
|
||||
programCacheMu sync.Mutex
|
||||
programCache = map[string]*goja.Program{}
|
||||
)
|
||||
|
||||
func compileCached(js string) (*goja.Program, error) {
|
||||
programCacheMu.Lock()
|
||||
defer programCacheMu.Unlock()
|
||||
if p, ok := programCache[js]; ok {
|
||||
return p, nil
|
||||
}
|
||||
p, err := goja.Compile(fmt.Sprintf("<js:%d>", len(js)), js, false)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
if len(programCache) >= programCacheSize {
|
||||
programCache = map[string]*goja.Program{}
|
||||
}
|
||||
programCache[js] = p
|
||||
return p, nil
|
||||
}
|
||||
|
||||
// JSFetcher 由服务层注入:桥接函数发起网络请求用(走 MeBox 的 HTTP 客户端,
|
||||
// 自动携带书源级请求头、UA、代理与重定向)。
|
||||
type JSFetcher func(req *Request) (body string, finalURL string, code int, err error)
|
||||
|
||||
// JSConfig 构造 JS 运行时的配置。
|
||||
type JSConfig struct {
|
||||
Fetch JSFetcher
|
||||
// SourceProps 注入为 JS 的 `source` 对象(书源 JSON 原样);
|
||||
// 同时供 source.getLoginInfoMap 从 loginUi 的 default 初始化登录信息。
|
||||
SourceProps map[string]any
|
||||
// Log 对应 java.log。
|
||||
Log func(msg string)
|
||||
// Timeout 单次 JS 执行超时,默认 10s。
|
||||
Timeout time.Duration
|
||||
// BaseURL 对应 evalJS 的 baseUrl 绑定。
|
||||
BaseURL string
|
||||
// Key / Page 搜索上下文绑定({{key}}/{{page}} 在 JS 里的取值)。
|
||||
Key string
|
||||
Page int
|
||||
// State 书源会话状态(变量/登录信息/Cookie)。nil 时用进程内 MemoryState。
|
||||
State SourceState
|
||||
// JSLib 书源 jsLib:在运行时创建后立即执行一次,
|
||||
// 其顶层函数与 lexical 绑定对该源后续所有 JS 可见(对应 legado SharedJsScope)。
|
||||
JSLib string
|
||||
// StateOnly 只构建会话状态与 jsLib 环境(登录交互用),
|
||||
// 不注入 book/result 等规则上下文。
|
||||
StateOnly bool
|
||||
// Browser 宿主浏览器实现(java.startBrowser / startBrowserAwait)。
|
||||
// nil 时这两个函数抛出不支持错误。
|
||||
Browser BrowserHost
|
||||
// Ctx 本次执行的可取消上下文,透传给 BrowserHost 的等待。
|
||||
Ctx context.Context
|
||||
}
|
||||
|
||||
// interruptGuard 是 JS 执行超时的看门狗:到期后中断虚拟机。
|
||||
//
|
||||
// 单独抽出来的原因是 java.startBrowserAwait 会阻塞等待用户在网页上操作
|
||||
// (可达数分钟),这段时间必须暂停计时,否则默认 10s 的超时会在用户还没
|
||||
// 点完 √ 时就把脚本打断。Pause/Resume 之间不计时,Resume 后重新起算完整
|
||||
// 的一段预算——语义即「每一段自动执行各有一次预算,等人不算」。
|
||||
type interruptGuard struct {
|
||||
vm *goja.Runtime
|
||||
timeout time.Duration
|
||||
reason string
|
||||
mu sync.Mutex
|
||||
paused int
|
||||
stopped bool
|
||||
timer *time.Timer
|
||||
}
|
||||
|
||||
func newInterruptGuard(vm *goja.Runtime, timeout time.Duration, reason string) *interruptGuard {
|
||||
g := &interruptGuard{vm: vm, timeout: timeout, reason: reason}
|
||||
g.start()
|
||||
return g
|
||||
}
|
||||
|
||||
// start 起一个新的超时计时(调用方需持锁或处于初始化阶段)。
|
||||
func (g *interruptGuard) start() {
|
||||
g.timer = time.AfterFunc(g.timeout, func() {
|
||||
g.mu.Lock()
|
||||
skip := g.stopped || g.paused > 0
|
||||
g.mu.Unlock()
|
||||
if !skip {
|
||||
g.vm.Interrupt(g.reason)
|
||||
}
|
||||
})
|
||||
}
|
||||
|
||||
// Pause 暂停计时(等待人工操作),返回恢复函数。
|
||||
func (g *interruptGuard) Pause() func() {
|
||||
g.mu.Lock()
|
||||
g.paused++
|
||||
if g.timer != nil {
|
||||
g.timer.Stop()
|
||||
g.timer = nil
|
||||
}
|
||||
g.mu.Unlock()
|
||||
return g.Resume
|
||||
}
|
||||
|
||||
// Resume 恢复计时。
|
||||
func (g *interruptGuard) Resume() {
|
||||
g.mu.Lock()
|
||||
defer g.mu.Unlock()
|
||||
if g.paused > 0 {
|
||||
g.paused--
|
||||
}
|
||||
if g.paused == 0 && !g.stopped && g.timer == nil {
|
||||
g.start()
|
||||
}
|
||||
}
|
||||
|
||||
// Stop 永久停止计时(本次 JS 执行结束)。
|
||||
func (g *interruptGuard) Stop() {
|
||||
g.mu.Lock()
|
||||
defer g.mu.Unlock()
|
||||
g.stopped = true
|
||||
if g.timer != nil {
|
||||
g.timer.Stop()
|
||||
g.timer = nil
|
||||
}
|
||||
}
|
||||
|
||||
// JSRunner 是一个单协程使用的 JS 运行时(每个 AnalyzeRule 一个实例)。
|
||||
type JSRunner struct {
|
||||
vm *goja.Runtime
|
||||
cfg JSConfig
|
||||
vars map[string]string // runner 级变量(URL 上下文 java.put/get)
|
||||
state SourceState
|
||||
// jsLibErr 记录 jsLib 执行失败原因(登录接口需要如实回报)。
|
||||
jsLibErr error
|
||||
// guard 当前执行的超时看门狗;java.startBrowserAwait 阻塞期间置为 nil。
|
||||
guardMu sync.Mutex
|
||||
guard *interruptGuard
|
||||
}
|
||||
|
||||
// setGuard 记录/清除当前执行的看门狗。
|
||||
func (r *JSRunner) setGuard(g *interruptGuard) {
|
||||
r.guardMu.Lock()
|
||||
r.guard = g
|
||||
r.guardMu.Unlock()
|
||||
}
|
||||
|
||||
// pauseTimeout 暂停当前 JS 执行的超时计时,返回恢复函数。
|
||||
// 供 java.startBrowserAwait 在等待人工操作期间调用。
|
||||
func (r *JSRunner) pauseTimeout() func() {
|
||||
r.guardMu.Lock()
|
||||
g := r.guard
|
||||
r.guardMu.Unlock()
|
||||
if g == nil {
|
||||
return func() {}
|
||||
}
|
||||
return g.Pause()
|
||||
}
|
||||
|
||||
// fetch 执行一次桥接网络请求,等待期间暂停 JS 超时看门狗。
|
||||
//
|
||||
// 书源会把「线路重试」写进规则 JS:光遇聚合的 request() 会串行试 7 条线路,单条
|
||||
// 线路最长可能等到客户端的 30s 超时。不暂停的话整条规则会被 10s 的 JS 超时打断,
|
||||
// 而这个中断是 goja 的 Go panic,书源自己写的 try/catch 接不住——表现就是「线路
|
||||
// 还在重试,接口已经 400」。java.startBrowserAwait 等待人工操作时用的是同一套暂停
|
||||
// 机制。暂停只覆盖网络等待,纯 CPU 死循环仍然受 Timeout 约束。
|
||||
func (r *JSRunner) fetch(req *Request) (string, string, int, error) {
|
||||
if r.cfg.Fetch == nil {
|
||||
return "", "", 0, ErrJsUnsupported
|
||||
}
|
||||
resume := r.pauseTimeout()
|
||||
defer resume()
|
||||
return r.cfg.Fetch(req)
|
||||
}
|
||||
|
||||
// NewJSRunner 创建运行时:注入全局对象 cookie / cache / source,并执行 jsLib。
|
||||
func NewJSRunner(cfg JSConfig) *JSRunner {
|
||||
vm := goja.New()
|
||||
if cfg.Timeout == 0 {
|
||||
cfg.Timeout = defaultJSTimeout
|
||||
}
|
||||
state := cfg.State
|
||||
if state == nil {
|
||||
state = NewMemoryState()
|
||||
}
|
||||
r := &JSRunner{vm: vm, cfg: cfg, vars: map[string]string{}, state: state}
|
||||
vm.Set("cookie", newCookieObject(vm, state))
|
||||
vm.Set("cache", newCacheObject(vm))
|
||||
// source 必须在 jsLib 之前注入:jsLib 的 getVariable/BaseUrl 依赖它。
|
||||
srcObj := newSourceObject(vm, state, cfg.SourceProps)
|
||||
vm.Set("source", srcObj)
|
||||
// java 也要在 jsLib 之前就位(jsLib 顶层可能引用 java.*)。
|
||||
r.installJava(nil)
|
||||
// 部分源把 source 的方法也当 java 成员用(同一 Kotlin 对象暴露两份)。
|
||||
r.loadJSLib()
|
||||
return r
|
||||
}
|
||||
|
||||
// installJava 安装/刷新本次执行可见的 java 对象(桥回指定解析器)。
|
||||
func (r *JSRunner) installJava(a *AnalyzeRule) {
|
||||
r.vm.Set("java", newJavaObject(r.vm, r, a))
|
||||
}
|
||||
|
||||
// loadJSLib 执行书源 jsLib(一次),使其中定义的函数与 lexical 绑定
|
||||
// 对后续 Run 可见(goja 的顶层 let/const 与函数声明在同一 Runtime 内保持)。
|
||||
func (r *JSRunner) loadJSLib() {
|
||||
if strings.TrimSpace(r.cfg.JSLib) == "" {
|
||||
return
|
||||
}
|
||||
prog, err := compileCached(r.cfg.JSLib)
|
||||
if err != nil {
|
||||
r.jsLibErr = fmt.Errorf("jsLib 编译失败: %w", err)
|
||||
return
|
||||
}
|
||||
g := newInterruptGuard(r.vm, r.cfg.Timeout, "jsLib 执行超时")
|
||||
r.setGuard(g)
|
||||
defer func() {
|
||||
g.Stop()
|
||||
r.setGuard(nil)
|
||||
}()
|
||||
if _, err := r.vm.RunProgram(prog); err != nil {
|
||||
r.jsLibErr = fmt.Errorf("jsLib 执行失败: %v", err)
|
||||
}
|
||||
}
|
||||
|
||||
// JSLibErr 返回 jsLib 的执行错误(nil 表示正常)。
|
||||
func (r *JSRunner) JSLibErr() error { return r.jsLibErr }
|
||||
|
||||
// State 返回运行时使用的会话状态。
|
||||
func (r *JSRunner) State() SourceState { return r.state }
|
||||
|
||||
// EvalAction 在已装载 jsLib 的环境中执行一段交互 JS(登录按钮 action 等)。
|
||||
// bindings 为附加的 JS 全局绑定(如 result)。返回脚本的返回值。
|
||||
//
|
||||
// 绑定集合与 Run 对齐(baseUrl/key/page),因为登录交互 JS 同样会用到它们。
|
||||
func (r *JSRunner) EvalAction(js string, bindings map[string]any) (any, error) {
|
||||
if r.jsLibErr != nil {
|
||||
return nil, r.jsLibErr
|
||||
}
|
||||
vm := r.vm
|
||||
r.installJava(nil)
|
||||
vm.Set("baseUrl", r.cfg.BaseURL)
|
||||
if r.cfg.Key != "" {
|
||||
vm.Set("key", r.cfg.Key)
|
||||
} else {
|
||||
vm.Set("key", nil)
|
||||
}
|
||||
if r.cfg.Page > 0 {
|
||||
vm.Set("page", r.cfg.Page)
|
||||
} else {
|
||||
vm.Set("page", nil)
|
||||
}
|
||||
for k, v := range bindings {
|
||||
vm.Set(k, toJSValue(vm, v))
|
||||
}
|
||||
prog, err := compileCached(js)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("JS 编译失败: %w", err)
|
||||
}
|
||||
g := newInterruptGuard(vm, r.cfg.Timeout, "JS 执行超时")
|
||||
r.setGuard(g)
|
||||
defer func() {
|
||||
g.Stop()
|
||||
r.setGuard(nil)
|
||||
}()
|
||||
v, err := vm.RunProgram(prog)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("JS 执行失败: %v", err)
|
||||
}
|
||||
return exportValue(v), nil
|
||||
}
|
||||
|
||||
// EvalLoginCheck 执行书源的 loginCheckJs(对应 legado WebBook 的 checkJs 钩子)。
|
||||
//
|
||||
// 语义对齐 legado:作用域里的 `result` 是一个响应对象(body()/code()/url()),
|
||||
// 脚本须返回响应对象或字符串;返回新 body 时调用方以之替换原响应体,
|
||||
// 从而支持「检测到会话失效 → 重新登录 → 重取页面」。
|
||||
func (r *JSRunner) EvalLoginCheck(js, body string, code int, finalURL string) (string, bool, error) {
|
||||
if r.jsLibErr != nil {
|
||||
return "", false, r.jsLibErr
|
||||
}
|
||||
vm := r.vm
|
||||
r.installJava(nil)
|
||||
vm.Set("baseUrl", r.cfg.BaseURL)
|
||||
vm.Set("result", newResponseObject(vm, body, code, finalURL, nil))
|
||||
|
||||
prog, err := compileCached(js)
|
||||
if err != nil {
|
||||
return "", false, fmt.Errorf("loginCheckJs 编译失败: %w", err)
|
||||
}
|
||||
g := newInterruptGuard(vm, r.cfg.Timeout, "loginCheckJs 执行超时")
|
||||
r.setGuard(g)
|
||||
defer func() {
|
||||
g.Stop()
|
||||
r.setGuard(nil)
|
||||
}()
|
||||
v, err := vm.RunProgram(prog)
|
||||
if err != nil {
|
||||
return "", false, fmt.Errorf("loginCheckJs 执行失败: %v", err)
|
||||
}
|
||||
if v == nil || goja.IsUndefined(v) || goja.IsNull(v) {
|
||||
return "", false, nil
|
||||
}
|
||||
// 返回响应对象时取 body()(对应 legado 的 StrResponse)
|
||||
if obj, ok := v.(*goja.Object); ok {
|
||||
if fn, ok := goja.AssertFunction(obj.Get("body")); ok {
|
||||
res, err := fn(obj)
|
||||
if err != nil {
|
||||
return "", false, fmt.Errorf("loginCheckJs 执行失败: %v", err)
|
||||
}
|
||||
return res.String(), true, nil
|
||||
}
|
||||
}
|
||||
return v.String(), true, nil
|
||||
}
|
||||
|
||||
|
||||
// ForAnalyzer 返回绑定到指定解析器的执行函数(java.getString 等规则回调
|
||||
// 会桥回该解析器,对应 legado 中 AnalyzeRule 自身实现 JsExtensions)。
|
||||
func (r *JSRunner) ForAnalyzer(a *AnalyzeRule) func(js string, result any) (any, error) {
|
||||
return func(js string, result any) (any, error) {
|
||||
return r.Run(a, js, result, "")
|
||||
}
|
||||
}
|
||||
|
||||
// RunWithBaseURL 与 Run 相同,但覆盖本次执行的 baseUrl 绑定(AnalyzeUrl 用)。
|
||||
func (r *JSRunner) RunWithBaseURL(a *AnalyzeRule, js string, result any, baseURL string) (any, error) {
|
||||
return r.Run(a, js, result, baseURL)
|
||||
}
|
||||
|
||||
// Run 执行一段书源 JS。绑定集合对应 legado evalJS:
|
||||
// java / cookie / cache / source / book / result / baseUrl / chapter / title /
|
||||
// src / page / key / nextChapterUrl。
|
||||
func (r *JSRunner) Run(a *AnalyzeRule, js string, result any, baseURL string) (any, error) {
|
||||
vm := r.vm
|
||||
// java 对象:每次执行重建(桥回当前解析器)
|
||||
r.installJava(a)
|
||||
// 上下文绑定
|
||||
if a != nil {
|
||||
vm.Set("book", newBookObject(vm, a))
|
||||
vm.Set("chapter", newChapterObject(vm, a))
|
||||
vm.Set("title", a.chapterTitle)
|
||||
if a.content != nil {
|
||||
vm.Set("src", resultString(a.content))
|
||||
} else {
|
||||
vm.Set("src", nil)
|
||||
}
|
||||
} else {
|
||||
vm.Set("book", nil)
|
||||
vm.Set("chapter", nil)
|
||||
vm.Set("title", nil)
|
||||
vm.Set("src", nil)
|
||||
}
|
||||
base := baseURL
|
||||
// baseUrl 绑定优先用「解析器当前处理的页面地址」再退回书源地址。
|
||||
//
|
||||
// 对应 legado:evalJS 里 bindings["baseUrl"] = analyzeRule.baseUrl,
|
||||
// 而 baseUrl 由各阶段 setBaseUrl(bookUrl / tocUrl / chapterUrl) 设定。
|
||||
// 这一点很关键:聚合类书源会用 String(baseUrl).startsWith("data:")
|
||||
// 判断「当前这一层是不是书源自搭的参数信封」,若把 baseUrl 固定成书源地址,
|
||||
// 书源会走 else 分支直接把 hex 原文当结果返回,详情/目录随之全空。
|
||||
if base == "" && a != nil {
|
||||
base = a.baseUrl
|
||||
}
|
||||
if base == "" {
|
||||
base = r.cfg.BaseURL
|
||||
}
|
||||
vm.Set("baseUrl", base)
|
||||
vm.Set("result", toJSValue(vm, result))
|
||||
if r.cfg.Key != "" {
|
||||
vm.Set("key", r.cfg.Key)
|
||||
} else {
|
||||
vm.Set("key", nil)
|
||||
}
|
||||
if r.cfg.Page > 0 {
|
||||
vm.Set("page", r.cfg.Page)
|
||||
} else {
|
||||
vm.Set("page", nil)
|
||||
}
|
||||
vm.Set("nextChapterUrl", nil)
|
||||
|
||||
prog, err := compileCached(scopedRuleJS(js))
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("JS 编译失败: %w", err)
|
||||
}
|
||||
g := newInterruptGuard(vm, r.cfg.Timeout, "JS 执行超时")
|
||||
r.setGuard(g)
|
||||
defer func() {
|
||||
g.Stop()
|
||||
r.setGuard(nil)
|
||||
}()
|
||||
v, err := vm.RunProgram(prog)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("JS 执行失败: %v", err)
|
||||
}
|
||||
return exportValue(v), nil
|
||||
}
|
||||
|
||||
// scopedRuleJS 把一段规则 JS 包进块作用域后编译。
|
||||
//
|
||||
// 同一个 goja Runtime 会被一个书源的所有规则 JS 复用,而顶层 let/const 会留在
|
||||
// 全局词法环境里,于是「前一个脚本声明过的名字,后一个脚本再声明」会直接报
|
||||
// `SyntaxError: Identifier 'x' has already been declared`。
|
||||
//
|
||||
// 典型触发(光遇聚合):搜索列表规则是
|
||||
//
|
||||
// <js>const { key, tab, sourcesKey, page, ... } = res; ...</js>$.data
|
||||
//
|
||||
// 而单本书的 bookUrl 规则是
|
||||
//
|
||||
// <js>let book_id = ...; let tab = result.tab || '小说'; ...</js>
|
||||
//
|
||||
// 两者在同一轮解析里先后执行,第二个必然编译失败;因为失败发生在「逐条取字段」
|
||||
// 阶段且被 continue 跳过,表现出来就是「搜索有结果但一条都读不出来」。
|
||||
// legado 用的 Rhino 对顶层 let 更宽松,所以同一书源在阅读 App 里是正常的。
|
||||
//
|
||||
// 包一层块即可隔离词法声明,同时保留块最后表达式的值(JS 规范中块的完成值就是
|
||||
// 最后一条语句的值),也不改变 this(仍是全局对象,书源的 this.getVariable /
|
||||
// this.BaseUrl 照常可用)。
|
||||
func scopedRuleJS(js string) string {
|
||||
return "{\n" + js + "\n}"
|
||||
}
|
||||
|
||||
// exportValue 把 JS 返回值转为 Go 值(字符串/数值/映射/切片)。
|
||||
func exportValue(v goja.Value) any {
|
||||
if v == nil || goja.IsUndefined(v) || goja.IsNull(v) {
|
||||
return nil
|
||||
}
|
||||
switch v.ExportType() {
|
||||
case nil:
|
||||
return nil
|
||||
default:
|
||||
return v.Export()
|
||||
}
|
||||
}
|
||||
|
||||
// toJSValue 把引擎内部结果转为可注入 JS 的值。
|
||||
// Element 列表等 DOM 结果以序列化字符串传入(对应 Rhino 的 Java 对象字符串化)。
|
||||
func toJSValue(vm *goja.Runtime, v any) any {
|
||||
switch t := v.(type) {
|
||||
case nil:
|
||||
return nil
|
||||
case string, bool, int, int32, int64, float64, map[string]any, []any:
|
||||
return t
|
||||
// 登录表单等以 map[string]string 传入,交给 goja 直接反射转换
|
||||
case map[string]string:
|
||||
m := make(map[string]any, len(t))
|
||||
for k, v := range t {
|
||||
m[k] = v
|
||||
}
|
||||
return m
|
||||
case []string:
|
||||
arr := make([]any, len(t))
|
||||
for i, s := range t {
|
||||
arr[i] = s
|
||||
}
|
||||
return arr
|
||||
case *goja.Object:
|
||||
return t
|
||||
case *goja.Runtime:
|
||||
return nil
|
||||
default:
|
||||
return resultString(v)
|
||||
}
|
||||
}
|
||||
|
||||
// newResponseObject 构造 Connection.Response / StrResponse 的 JS 等价物:
|
||||
// body()/code()/url()/header(name)。
|
||||
func newResponseObject(vm *goja.Runtime, body string, code int, finalURL string, headers map[string][]string) *goja.Object {
|
||||
o := vm.NewObject()
|
||||
mustSet := func(k string, v any) {
|
||||
if err := o.Set(k, v); err != nil {
|
||||
panic(vm.ToValue(err.Error()))
|
||||
}
|
||||
}
|
||||
mustSet("body", func(call goja.FunctionCall) goja.Value { return vm.ToValue(body) })
|
||||
mustSet("bodyStr", body)
|
||||
mustSet("code", func(call goja.FunctionCall) goja.Value { return vm.ToValue(code) })
|
||||
mustSet("url", func(call goja.FunctionCall) goja.Value { return vm.ToValue(finalURL) })
|
||||
mustSet("header", func(call goja.FunctionCall) goja.Value {
|
||||
name := strings.ToLower(strings.TrimSpace(toStringArg(call, 0)))
|
||||
for k, vs := range headers {
|
||||
if strings.ToLower(k) == name && len(vs) > 0 {
|
||||
return vm.ToValue(vs[0])
|
||||
}
|
||||
}
|
||||
return goja.Null()
|
||||
})
|
||||
return o
|
||||
}
|
||||
|
||||
func toStringArg(call goja.FunctionCall, i int) string {
|
||||
if i >= len(call.Arguments) {
|
||||
return ""
|
||||
}
|
||||
return call.Arguments[i].String()
|
||||
}
|
||||
|
||||
func stringArg(call goja.FunctionCall, i int) string {
|
||||
return strings.TrimSpace(toStringArg(call, i))
|
||||
}
|
||||
|
||||
func stringArgOr(call goja.FunctionCall, i int, def string) string {
|
||||
if i >= len(call.Arguments) || goja.IsUndefined(call.Arguments[i]) || goja.IsNull(call.Arguments[i]) {
|
||||
return def
|
||||
}
|
||||
return call.Arguments[i].String()
|
||||
}
|
||||
@@ -0,0 +1,263 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"bytes"
|
||||
"strings"
|
||||
|
||||
"github.com/andybalholm/cascadia"
|
||||
"golang.org/x/net/html"
|
||||
)
|
||||
|
||||
// 本文件在 golang.org/x/net/html 的 *html.Node 上实现 jsoup 的元素语义,
|
||||
// 供 jsoup 风格分析器使用。语义对齐 org.jsoup.nodes.Element。
|
||||
|
||||
func isElement(n *html.Node) bool {
|
||||
return n != nil && n.Type == html.ElementNode
|
||||
}
|
||||
|
||||
// childrenElements 对应 Element.children():直接子元素。
|
||||
func childrenElements(n *html.Node) []*html.Node {
|
||||
var out []*html.Node
|
||||
if n == nil {
|
||||
return out
|
||||
}
|
||||
for c := n.FirstChild; c != nil; c = c.NextSibling {
|
||||
if isElement(c) {
|
||||
out = append(out, c)
|
||||
}
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// collectElements 对应 jsoup Collector.collect(evaluator, root):
|
||||
// 前序遍历,包含 root 自身。
|
||||
func collectElements(root *html.Node, pred func(*html.Node) bool) []*html.Node {
|
||||
var out []*html.Node
|
||||
if root == nil {
|
||||
return out
|
||||
}
|
||||
var walk func(n *html.Node)
|
||||
walk = func(n *html.Node) {
|
||||
if n.Type == html.ElementNode && pred(n) {
|
||||
out = append(out, n)
|
||||
}
|
||||
for c := n.FirstChild; c != nil; c = c.NextSibling {
|
||||
walk(c)
|
||||
}
|
||||
}
|
||||
// root 自身参与匹配(jsoup select/getElementsByXxx 均包含自身)
|
||||
if root.Type == html.ElementNode && pred(root) {
|
||||
out = append(out, root)
|
||||
}
|
||||
for c := root.FirstChild; c != nil; c = c.NextSibling {
|
||||
walk(c)
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
func hasClassToken(n *html.Node, class string) bool {
|
||||
for _, a := range n.Attr {
|
||||
if a.Key == "class" {
|
||||
for _, tok := range strings.Fields(a.Val) {
|
||||
if tok == class {
|
||||
return true
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
// getElementsByClass 对应 Element.getElementsByClass(含自身)。
|
||||
func getElementsByClass(root *html.Node, class string) []*html.Node {
|
||||
return collectElements(root, func(n *html.Node) bool { return hasClassToken(n, class) })
|
||||
}
|
||||
|
||||
// getElementsByTag 对应 Element.getElementsByTag(含自身)。
|
||||
func getElementsByTag(root *html.Node, tag string) []*html.Node {
|
||||
return collectElements(root, func(n *html.Node) bool { return n.Data == tag })
|
||||
}
|
||||
|
||||
// getElementsById 对应 Collector.collect(Evaluator.Id(id), root)(含自身)。
|
||||
func getElementsById(root *html.Node, id string) []*html.Node {
|
||||
return collectElements(root, func(n *html.Node) bool {
|
||||
for _, a := range n.Attr {
|
||||
if a.Key == "id" && a.Val == id {
|
||||
return true
|
||||
}
|
||||
}
|
||||
return false
|
||||
})
|
||||
}
|
||||
|
||||
// getElementsContainingOwnText 对应 Evaluator.ContentsOwnText 语义:
|
||||
// ownText 包含目标串的元素。
|
||||
func getElementsContainingOwnText(root *html.Node, text string) []*html.Node {
|
||||
return collectElements(root, func(n *html.Node) bool {
|
||||
return strings.Contains(nodeOwnText(n), text)
|
||||
})
|
||||
}
|
||||
|
||||
// selectCSS 对应 Element.select(css):以 root 为起点(含自身)执行 CSS 选择。
|
||||
func selectCSS(root *html.Node, sel string) []*html.Node {
|
||||
compiled, err := cascadia.Compile(sel)
|
||||
if err != nil {
|
||||
return nil
|
||||
}
|
||||
return selectWithCompiled(root, compiled)
|
||||
}
|
||||
|
||||
func selectWithCompiled(root *html.Node, sel cascadia.Selector) []*html.Node {
|
||||
var out []*html.Node
|
||||
if root.Type == html.ElementNode && sel(root) {
|
||||
out = append(out, root)
|
||||
}
|
||||
var walk func(n *html.Node)
|
||||
walk = func(n *html.Node) {
|
||||
for c := n.FirstChild; c != nil; c = c.NextSibling {
|
||||
if isElement(c) && sel(c) {
|
||||
out = append(out, c)
|
||||
}
|
||||
walk(c)
|
||||
}
|
||||
}
|
||||
walk(root)
|
||||
return out
|
||||
}
|
||||
|
||||
// nodeOwnText 对应 Element.ownText():直接子文本节点,规整空白后空格连接。
|
||||
func nodeOwnText(n *html.Node) string {
|
||||
var parts []string
|
||||
for c := n.FirstChild; c != nil; c = c.NextSibling {
|
||||
if c.Type == html.TextNode {
|
||||
t := normalizeWhitespace(c.Data)
|
||||
if t != "" {
|
||||
parts = append(parts, t)
|
||||
}
|
||||
}
|
||||
}
|
||||
return strings.Join(parts, " ")
|
||||
}
|
||||
|
||||
// nodeText 对应 Element.text():全部后代文本规整空白(<br> 记空格,
|
||||
// 跳过 script/style),多段空白折叠为单个空格。
|
||||
func nodeText(n *html.Node) string {
|
||||
var sb bytes.Buffer
|
||||
var walk func(n *html.Node)
|
||||
walk = func(n *html.Node) {
|
||||
if n.Type == html.TextNode {
|
||||
sb.WriteString(n.Data)
|
||||
return
|
||||
}
|
||||
if n.Type == html.ElementNode && (n.Data == "script" || n.Data == "style") {
|
||||
return
|
||||
}
|
||||
if n.Type == html.ElementNode && n.Data == "br" {
|
||||
sb.WriteString(" ")
|
||||
return
|
||||
}
|
||||
for c := n.FirstChild; c != nil; c = c.NextSibling {
|
||||
walk(c)
|
||||
}
|
||||
}
|
||||
walk(n)
|
||||
return normalizeWhitespace(sb.String())
|
||||
}
|
||||
|
||||
// normalizeWhitespace 对应 jsoup TextUtil 的空白规整。
|
||||
func normalizeWhitespace(s string) string {
|
||||
return strings.Join(strings.Fields(s), " ")
|
||||
}
|
||||
|
||||
// nodeTextNodes 对应 Element.textNodes():直接子文本节点(trim 非空)。
|
||||
func nodeTextNodes(n *html.Node) []string {
|
||||
var out []string
|
||||
for c := n.FirstChild; c != nil; c = c.NextSibling {
|
||||
if c.Type == html.TextNode {
|
||||
t := strings.TrimSpace(c.Data)
|
||||
if t != "" {
|
||||
out = append(out, t)
|
||||
}
|
||||
}
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// nodeData 对应 Element.data():script/style 的原始内容。
|
||||
func nodeData(n *html.Node) string {
|
||||
if n.Type != html.ElementNode || (n.Data != "script" && n.Data != "style") {
|
||||
return ""
|
||||
}
|
||||
var sb bytes.Buffer
|
||||
for c := n.FirstChild; c != nil; c = c.NextSibling {
|
||||
if c.Type == html.TextNode {
|
||||
sb.WriteString(c.Data)
|
||||
}
|
||||
}
|
||||
return sb.String()
|
||||
}
|
||||
|
||||
// outerHTML 对应 Element.outerHtml()。
|
||||
func outerHTML(n *html.Node) string {
|
||||
var buf bytes.Buffer
|
||||
if err := html.Render(&buf, n); err != nil {
|
||||
return ""
|
||||
}
|
||||
return buf.String()
|
||||
}
|
||||
|
||||
// outerHTMLNoScript 对应 getResultLast "html" 分支:移除 script/style 后的 outerHtml。
|
||||
func outerHTMLNoScript(n *html.Node) string {
|
||||
clone := cloneNodeShallowTree(n)
|
||||
removeTags(clone, "script")
|
||||
removeTags(clone, "style")
|
||||
return outerHTML(clone)
|
||||
}
|
||||
|
||||
func removeTags(n *html.Node, tag string) {
|
||||
var toRemove []*html.Node
|
||||
for c := n.FirstChild; c != nil; c = c.NextSibling {
|
||||
if isElement(c) && c.Data == tag {
|
||||
toRemove = append(toRemove, c)
|
||||
}
|
||||
removeTags(c, tag)
|
||||
}
|
||||
for _, r := range toRemove {
|
||||
n.RemoveChild(r)
|
||||
}
|
||||
}
|
||||
|
||||
// cloneNodeShallowTree 深拷贝节点树(html.Render 需要)。
|
||||
func cloneNodeShallowTree(n *html.Node) *html.Node {
|
||||
c := &html.Node{
|
||||
Type: n.Type,
|
||||
DataAtom: n.DataAtom,
|
||||
Data: n.Data,
|
||||
Attr: append([]html.Attribute(nil), n.Attr...),
|
||||
}
|
||||
for ch := n.FirstChild; ch != nil; ch = ch.NextSibling {
|
||||
cc := cloneNodeShallowTree(ch)
|
||||
c.AppendChild(cc)
|
||||
}
|
||||
return c
|
||||
}
|
||||
|
||||
func attrValue(n *html.Node, key string) string {
|
||||
for _, a := range n.Attr {
|
||||
if a.Key == key {
|
||||
return a.Val
|
||||
}
|
||||
}
|
||||
return ""
|
||||
}
|
||||
|
||||
// parseHTML 对应 AnalyzeByJSoup.parse:字符串转节点树。
|
||||
// x/net/html 会补全 <html><body> 结构,选择器从 document 根开始匹配,
|
||||
// 与 jsoup 以 Document 为根的选择行为一致。
|
||||
func parseHTML(s string) *html.Node {
|
||||
nodes, err := html.Parse(strings.NewReader(s))
|
||||
if err != nil {
|
||||
return nil
|
||||
}
|
||||
return nodes
|
||||
}
|
||||
@@ -0,0 +1,58 @@
|
||||
package rule
|
||||
|
||||
import "testing"
|
||||
|
||||
// 书源普遍用 legado 的内嵌模板拼字段(聚合源的 kind / 最新章节就是这种写法):
|
||||
//
|
||||
// kind = "{{$.status}},{{$.score}},{{$.tags}}"
|
||||
// lastChapter = "{{$.source}} {{$.last_chapter_title}}"
|
||||
//
|
||||
// 模板必须在**当前元素**上求值。修复前这里拿整份响应求值,模板一律取空,
|
||||
// 表现为 kind=",,,"、最新章节为空——搜索结果和换源列表里所有条目长得一模一样。
|
||||
func TestInnerRuleTemplateEvaluatesAgainstElement(t *testing.T) {
|
||||
const body = `{"data":[{"book_name":"全球高武","status":"连载","score":"9.2","tags":"都市","source":"svip_酷我","last_chapter_title":"第694章"}]}`
|
||||
a := NewAnalyzeRule()
|
||||
a.SetContent(body, "")
|
||||
els, err := a.GetElements("$.data")
|
||||
if err != nil {
|
||||
t.Fatalf("GetElements: %v", err)
|
||||
}
|
||||
if len(els) != 1 {
|
||||
t.Fatalf("want 1 element, got %d", len(els))
|
||||
}
|
||||
el := els[0]
|
||||
|
||||
got, err := a.GetString("{{$.status}},{{$.score}},{{$.tags}}", el, false)
|
||||
if err != nil {
|
||||
t.Fatalf("GetString: %v", err)
|
||||
}
|
||||
if got != "连载,9.2,都市" {
|
||||
t.Fatalf("kind 模板 = %q, want 连载,9.2,都市", got)
|
||||
}
|
||||
|
||||
got, err = a.GetString("{{$.source}} {{$.last_chapter_title}}", el, false)
|
||||
if err != nil {
|
||||
t.Fatalf("GetString: %v", err)
|
||||
}
|
||||
if got != "svip_酷我 第694章" {
|
||||
t.Fatalf("最新章节模板 = %q, want svip_酷我 第694章", got)
|
||||
}
|
||||
|
||||
// 列表规则(GetStringList)走的是同一套 MakeUpRule,必须同样生效
|
||||
list, err := a.GetStringList("{{$.status}},{{$.score}}", el, false)
|
||||
if err != nil {
|
||||
t.Fatalf("GetStringList: %v", err)
|
||||
}
|
||||
if len(list) != 1 || list[0] != "连载,9.2" {
|
||||
t.Fatalf("列表模板 = %v, want [连载,9.2]", list)
|
||||
}
|
||||
|
||||
// 模板里取不存在的字段应得到空串,而不是报错
|
||||
got, err = a.GetString("[{{$.not_exist}}]", el, false)
|
||||
if err != nil {
|
||||
t.Fatalf("GetString: %v", err)
|
||||
}
|
||||
if got != "[]" {
|
||||
t.Fatalf("缺失字段 = %q, want []", got)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,98 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"strings"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// 本文件:规则 JS 作用域隔离的回归测试。
|
||||
//
|
||||
// 同一个 goja Runtime 会被一个书源的所有规则 JS 复用,而顶层 let/const 会留在
|
||||
// 全局词法环境里。若不做隔离,「前一个脚本声明过的名字,后一个脚本再声明」就会
|
||||
// 报 `SyntaxError: Identifier 'x' has already been declared`。
|
||||
//
|
||||
// 光遇聚合正好踩中:搜索列表规则是 `const { key, tab, ... } = res`,
|
||||
// 单本书的 bookUrl 规则是 `let tab = result.tab || '小说'`。失败发生在「逐条取
|
||||
// 字段」阶段且被 continue 跳过,于是表现为「搜索有结果但一条都读不出来」。
|
||||
// legado 用的 Rhino 对顶层 let 更宽松,所以同一书源在阅读 App 里正常。
|
||||
|
||||
// TestRuleJSScopeIsIsolated 先后执行的规则 JS 不应因顶层 let/const 重名而失败。
|
||||
func TestRuleJSScopeIsIsolated(t *testing.T) {
|
||||
r := NewJSRunner(JSConfig{})
|
||||
|
||||
// 第一个脚本:解构声明(搜索列表规则的写法)
|
||||
v, err := r.Run(nil, `const { tab, key } = {tab:'A', key:'K'}; tab + key`, nil, "")
|
||||
if err != nil {
|
||||
t.Fatalf("第一个脚本失败: %v", err)
|
||||
}
|
||||
if got := anyToString(v); got != "AK" {
|
||||
t.Fatalf("第一个脚本返回 %q", got)
|
||||
}
|
||||
|
||||
// 第二个脚本重名声明(bookUrl 规则的写法):不得报「已声明」
|
||||
for i := 0; i < 3; i++ {
|
||||
v, err = r.Run(nil, `let tab = 'B'; let book_id = 'x' + tab; book_id`, nil, "")
|
||||
if err != nil {
|
||||
t.Fatalf("第 %d 次重名声明冲突: %v", i+2, err)
|
||||
}
|
||||
if got := anyToString(v); got != "xB" {
|
||||
t.Fatalf("第 %d 次返回 %q", i+2, got)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestRuleJSKeepsCompletionValue 包块不能吃掉「最后一条语句的值」——
|
||||
// 书源的 URL 规则几乎都靠这个完成值返回结果。
|
||||
func TestRuleJSKeepsCompletionValue(t *testing.T) {
|
||||
r := NewJSRunner(JSConfig{})
|
||||
|
||||
v, err := r.Run(nil, "`data:;base64,AAAA,{\"type\":\"gysearch\"}`", nil, "")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got := anyToString(v); !strings.HasPrefix(got, "data:;base64,") {
|
||||
t.Fatalf("模板字面量的完成值丢了: %q", got)
|
||||
}
|
||||
|
||||
// 末尾带行注释也不能把收尾的块注释掉
|
||||
v, err = r.Run(nil, "let a = 'ok'; a // 末尾注释", nil, "")
|
||||
if err != nil {
|
||||
t.Fatalf("带末尾注释的脚本失败: %v", err)
|
||||
}
|
||||
if got := anyToString(v); got != "ok" {
|
||||
t.Fatalf("返回 %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
// TestRuleJSThisStaysGlobal 包块后 this 仍是全局对象,
|
||||
// 书源的 this.getVariable / this.BaseUrl / this.request 才照常可用。
|
||||
func TestRuleJSThisStaysGlobal(t *testing.T) {
|
||||
r := NewJSRunner(JSConfig{})
|
||||
v, err := r.Run(nil, `String(this === globalThis) + '|' + typeof java.ajax`, nil, "")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got := anyToString(v); got != "true|function" {
|
||||
t.Fatalf("this 语义被破坏: %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
// TestRuleJSCanReadGlobalsFromJSLib 规则 JS 仍能读到 jsLib 定义的全局函数
|
||||
// 与顶层 lexical 绑定(hosts 这类)。
|
||||
func TestRuleJSCanReadGlobalsFromJSLib(t *testing.T) {
|
||||
r := NewJSRunner(JSConfig{
|
||||
JSLib: `let hosts = ['https://v1.example.com'];
|
||||
function BaseUrl(){ return hosts[0] }
|
||||
function pick(k){ return k + '@' + BaseUrl() }`,
|
||||
})
|
||||
if err := r.JSLibErr(); err != nil {
|
||||
t.Fatalf("jsLib 失败: %v", err)
|
||||
}
|
||||
v, err := r.Run(nil, `{ let x = 'A'; pick(x) }`, nil, "")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got := anyToString(v); got != "A@https://v1.example.com" {
|
||||
t.Fatalf("规则读不到 jsLib 的全局: %q", got)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,358 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"crypto/aes"
|
||||
"crypto/cipher"
|
||||
"encoding/base64"
|
||||
"encoding/hex"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"regexp"
|
||||
"strings"
|
||||
"testing"
|
||||
"time"
|
||||
)
|
||||
|
||||
// ─── 基础执行与绑定 ─────────────────────────────────────────────────────────
|
||||
|
||||
func newTestRunner() *JSRunner {
|
||||
return NewJSRunner(JSConfig{Key: "斗", Page: 2})
|
||||
}
|
||||
|
||||
func TestJSBasicEval(t *testing.T) {
|
||||
r := newTestRunner()
|
||||
a := NewAnalyzeRule()
|
||||
v, err := r.Run(a, "1 + 2", nil, "")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if anyToString(v) != "3" {
|
||||
t.Fatalf("1+2 = %v", v)
|
||||
}
|
||||
// key/page 绑定
|
||||
v, err = r.Run(a, "key + page", nil, "")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if anyToString(v) != "斗2" {
|
||||
t.Fatalf("key+page = %v", v)
|
||||
}
|
||||
}
|
||||
|
||||
func TestJSTimeoutInterrupt(t *testing.T) {
|
||||
r := NewJSRunner(JSConfig{Timeout: 200 * 1e6}) // 200ms
|
||||
_, err := r.Run(NewAnalyzeRule(), "while(true){}", nil, "")
|
||||
if err == nil || !strings.Contains(err.Error(), "超时") {
|
||||
t.Fatalf("expected timeout error, got %v", err)
|
||||
}
|
||||
}
|
||||
|
||||
// TestJSTimeoutPausedDuringFetch 网络等待期间不能被 JS 超时打断。
|
||||
//
|
||||
// 回归:书源把「线路重试」写在规则 JS 里(光遇聚合的 request() 串行试 7 条线路,
|
||||
// 单条最长等到 HTTP 客户端超时)。看门狗不暂停的话,整条规则会被 10s 的 JS 超时
|
||||
// 中断,而这个中断是 goja 的 Go panic,书源自己写的 try/catch 接不住——表现就是
|
||||
// 「线路还在重试,接口已经 400」。
|
||||
func TestJSTimeoutPausedDuringFetch(t *testing.T) {
|
||||
r := NewJSRunner(JSConfig{
|
||||
Timeout: 100 * time.Millisecond,
|
||||
Fetch: func(req *Request) (string, string, int, error) {
|
||||
time.Sleep(400 * time.Millisecond) // 远超过 JS 超时:模拟慢线路
|
||||
return `{"content":"正文"}`, req.URL, 200, nil
|
||||
},
|
||||
})
|
||||
ar := NewAnalyzeRule()
|
||||
v, err := r.Run(ar, `(function(){
|
||||
try { return java.ajax('https://slow.example.com/content'); }
|
||||
catch (e) { return 'caught:' + e; }
|
||||
})()`, nil, "")
|
||||
if err != nil {
|
||||
t.Fatalf("网络等待期间不应触发 JS 超时: %v", err)
|
||||
}
|
||||
if got := anyToString(v); !strings.Contains(got, "正文") {
|
||||
t.Fatalf("ajax 返回值 = %q,期望上游正文", got)
|
||||
}
|
||||
|
||||
// 暂停只覆盖网络等待:回到 JS 里的纯 CPU 死循环仍然要被超时打断。
|
||||
_, err = r.Run(ar, `java.ajax('https://slow.example.com/content'); while(true){}`, nil, "")
|
||||
if err == nil || !strings.Contains(err.Error(), "超时") {
|
||||
t.Fatalf("网络等待之后的 CPU 死循环应当仍然超时,实际: %v", err)
|
||||
}
|
||||
}
|
||||
|
||||
// ─── 规则引擎中的 JS({{}} / @js: / <js>) ──────────────────────────────────
|
||||
|
||||
func TestAnalyzeRuleJSEval(t *testing.T) {
|
||||
r := newTestRunner()
|
||||
a := NewAnalyzeRule()
|
||||
a.SetJSRunner(r.ForAnalyzer(a))
|
||||
a.SetContent("正文内容", "http://x.com")
|
||||
|
||||
// {{js}} 内嵌
|
||||
got, err := a.GetString(`{{baseUrl}}/next`, nil, false)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got != "http://x.com/next" {
|
||||
t.Fatalf("{{baseUrl}} = %q", got)
|
||||
}
|
||||
// @js: 前缀
|
||||
got, err = a.GetString(`@js:'hello ' + (40 + 2)`, nil, false)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got != "hello 42" {
|
||||
t.Fatalf("@js = %q", got)
|
||||
}
|
||||
// <js></js> 块
|
||||
got, err = a.GetString(`<js>"结果:" + result</js>`, nil, false)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got != "结果:正文内容" {
|
||||
t.Fatalf("<js> = %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
func TestJSGetStringBridge(t *testing.T) {
|
||||
r := newTestRunner()
|
||||
a := NewAnalyzeRule()
|
||||
a.SetJSRunner(r.ForAnalyzer(a))
|
||||
a.SetContent(testHTML, "http://x.com")
|
||||
got, err := a.GetString(`<js>java.getString("class.item.0@tag.a@text")</js>`, nil, false)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got != "斗破苍穹" {
|
||||
t.Fatalf("java.getString = %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
// ─── 编码与摘要 ─────────────────────────────────────────────────────────────
|
||||
|
||||
func TestJSEncodeFunctions(t *testing.T) {
|
||||
r := newTestRunner()
|
||||
a := NewAnalyzeRule()
|
||||
a.SetJSRunner(r.ForAnalyzer(a))
|
||||
a.SetContent("", "") // 规则循环要求 content 非空(与 legado 语义一致)
|
||||
|
||||
cases := []struct{ js, want string }{
|
||||
{`@js:java.base64Encode('你好')`, base64.StdEncoding.EncodeToString([]byte("你好"))},
|
||||
{`@js:java.base64Decode('` + base64.StdEncoding.EncodeToString([]byte("你好")) + `')`, "你好"},
|
||||
{`@js:java.hexEncodeToString('AB')`, "4142"},
|
||||
{`@js:java.hexDecodeToString('4142')`, "AB"},
|
||||
{`@js:java.md5Encode('abc')`, "900150983cd24fb0d6963f7d28e17f72"},
|
||||
{`@js:java.md5Encode16('abc')`, "3cd24fb0d6963f7d"},
|
||||
{`@js:java.digestHex('abc','SHA-256')`, "ba7816bf8f01cfea414140de5dae2223b00361a396177a9cb410ff61f20015ad"},
|
||||
}
|
||||
for _, c := range cases {
|
||||
got, err := a.GetString(c.js, nil, false)
|
||||
if err != nil {
|
||||
t.Fatalf("%s: %v", c.js, err)
|
||||
}
|
||||
if got != c.want {
|
||||
t.Errorf("%s = %q, want %q", c.js, got, c.want)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// ─── 对称加解密 ─────────────────────────────────────────────────────────────
|
||||
|
||||
func TestJSSymmetricCryptoRoundTrip(t *testing.T) {
|
||||
r := newTestRunner()
|
||||
a := NewAnalyzeRule()
|
||||
a.SetJSRunner(r.ForAnalyzer(a))
|
||||
a.SetContent("", "")
|
||||
|
||||
script := `
|
||||
var key = '0123456789abcdef';
|
||||
var iv = 'abcdef9876543210';
|
||||
var c = java.createSymmetricCrypto('AES/CBC/PKCS5Padding', key, iv);
|
||||
var enc = c.encryptBase64('测试明文内容');
|
||||
c.decryptStr(enc)
|
||||
`
|
||||
got, err := a.GetString(`@js:`+script, nil, false)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got != "测试明文内容" {
|
||||
t.Fatalf("roundtrip = %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
func TestAESVector(t *testing.T) {
|
||||
// 用 Go 标准库生成固定密文,验证 decryptAuto + PKCS7 对齐
|
||||
block, _ := aes.NewCipher([]byte("0123456789abcdef"))
|
||||
iv := []byte("abcdef9876543210")
|
||||
plain := []byte("hello legado")
|
||||
padded := applyPadding(plain, block.BlockSize(), "PKCS5Padding")
|
||||
out := make([]byte, len(padded))
|
||||
cipher.NewCBCEncrypter(block, iv).CryptBlocks(out, padded)
|
||||
ct := base64.StdEncoding.EncodeToString(out)
|
||||
|
||||
r := newTestRunner()
|
||||
a := NewAnalyzeRule()
|
||||
a.SetJSRunner(r.ForAnalyzer(a))
|
||||
a.SetContent("", "")
|
||||
got, err := a.GetString(`@js:java.aesBase64DecodeToString('`+ct+`', '0123456789abcdef', 'AES/CBC/PKCS5Padding', 'abcdef9876543210')`, nil, false)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got != "hello legado" {
|
||||
t.Fatalf("aes decode = %q", got)
|
||||
}
|
||||
if _, err := hex.DecodeString("00"); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
}
|
||||
|
||||
func TestECBCipher(t *testing.T) {
|
||||
r := newTestRunner()
|
||||
a := NewAnalyzeRule()
|
||||
a.SetJSRunner(r.ForAnalyzer(a))
|
||||
a.SetContent("", "")
|
||||
script := `
|
||||
var c = java.createSymmetricCrypto('AES/ECB/PKCS5Padding', '0123456789abcdef', '');
|
||||
var enc = c.encryptBase64('ECB模式测试');
|
||||
c.decryptStr(enc)
|
||||
`
|
||||
got, err := a.GetString(`@js:`+script, nil, false)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got != "ECB模式测试" {
|
||||
t.Fatalf("ecb roundtrip = %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
// ─── 网络桥(httptest) ─────────────────────────────────────────────────────
|
||||
|
||||
func TestJSAjaxBridge(t *testing.T) {
|
||||
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
if r.Header.Get("X-Token") != "tk" {
|
||||
http.Error(w, "no token", 401)
|
||||
return
|
||||
}
|
||||
_, _ = w.Write([]byte(`{"code":0,"data":"ok"}`))
|
||||
}))
|
||||
defer srv.Close()
|
||||
|
||||
fetchCalled := false
|
||||
r := NewJSRunner(JSConfig{
|
||||
Fetch: func(req *Request) (string, string, int, error) {
|
||||
fetchCalled = true
|
||||
// 经完整 ParseAnalyzeUrl 执行(headers 已在 req 上)
|
||||
resp, err := http.Get(req.URL)
|
||||
if err != nil {
|
||||
return "", "", 0, err
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
// 补上测试头
|
||||
req.Headers["X-Token"] = "tk"
|
||||
return `{"code":0,"data":"ok"}`, req.URL, 200, nil
|
||||
},
|
||||
})
|
||||
a := NewAnalyzeRule()
|
||||
a.SetJSRunner(r.ForAnalyzer(a))
|
||||
a.SetContent("", "")
|
||||
|
||||
got, err := a.GetString(`@js:java.get('`+srv.URL+`/api', {"X-Token":"tk"}).body()`, nil, false)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if !fetchCalled || !strings.Contains(got, `"ok"`) {
|
||||
t.Fatalf("ajax bridge = %q", got)
|
||||
}
|
||||
|
||||
// ajax 返回字符串
|
||||
got, err = a.GetString(`@js:java.ajax('`+srv.URL+`/api')`, nil, false)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if !strings.Contains(got, `"ok"`) {
|
||||
t.Fatalf("ajax = %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
// ─── URL 规则 JS ────────────────────────────────────────────────────────────
|
||||
|
||||
func TestParseAnalyzeUrlWithJSBlocks(t *testing.T) {
|
||||
r := NewJSRunner(JSConfig{Key: "斗罗", Page: 3})
|
||||
req, err := ParseAnalyzeUrlWithJS(`<js>'https://e.com/search?q=' + encodeURIComponent(key) + '&p=' + page</js>`, "斗罗", 3, "", r)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if req.Unsupported != nil {
|
||||
t.Fatalf("unsupported: %v", req.Unsupported)
|
||||
}
|
||||
if !strings.Contains(req.URL, "p=3") || !strings.Contains(req.URL, "q=") {
|
||||
t.Fatalf("url = %q", req.URL)
|
||||
}
|
||||
}
|
||||
|
||||
func TestParseAnalyzeUrlJSTemplate(t *testing.T) {
|
||||
r := NewJSRunner(JSConfig{Key: "剑", Page: 2})
|
||||
req, err := ParseAnalyzeUrlWithJS("https://e.com/api?page={{page + 1}}&kw={{key}}", "剑", 2, "", r)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if !strings.Contains(req.URL, "page=3") || !strings.Contains(req.URL, "kw=%E5%89%91") {
|
||||
t.Fatalf("url = %q", req.URL)
|
||||
}
|
||||
}
|
||||
|
||||
func TestParseAnalyzeUrlBodyJs(t *testing.T) {
|
||||
r := newTestRunner()
|
||||
// 与 legado 一致:option.js 里通过 result 引用当前 url
|
||||
req, err := ParseAnalyzeUrlWithJS(`https://e.com/x,{"bodyJs":"result + '!'","js":"result + '#anchor'"}`, "", 0, "", r)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if !strings.Contains(req.URL, "#anchor") {
|
||||
t.Fatalf("url = %q", req.URL)
|
||||
}
|
||||
if req.BodyJsFn == nil {
|
||||
t.Fatal("BodyJsFn should be set")
|
||||
}
|
||||
if got := req.BodyJsFn("body-x"); got != "body-x!" {
|
||||
t.Fatalf("bodyJs = %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
func TestJSUnsupportedStillWorks(t *testing.T) {
|
||||
// runner 为 nil 时,P0 行为保持:标记 Unsupported
|
||||
req, err := ParseAnalyzeUrl(`<js>'x'</js>`, "", 0, "")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if req.Unsupported == nil {
|
||||
t.Fatal("expected unsupported without runner")
|
||||
}
|
||||
}
|
||||
|
||||
// ─── 沙箱边界 ───────────────────────────────────────────────────────────────
|
||||
|
||||
func TestJSSandboxUnsupported(t *testing.T) {
|
||||
r := newTestRunner()
|
||||
a := NewAnalyzeRule()
|
||||
a.SetJSRunner(r.ForAnalyzer(a))
|
||||
a.SetContent("", "")
|
||||
_, err := a.GetString(`@js:java.readTxtFile('/etc/passwd')`, nil, false)
|
||||
if err == nil || !strings.Contains(err.Error(), "不支持") {
|
||||
t.Fatalf("expected sandbox error, got %v", err)
|
||||
}
|
||||
}
|
||||
|
||||
func TestJSTimeFormatShape(t *testing.T) {
|
||||
r := newTestRunner()
|
||||
a := NewAnalyzeRule()
|
||||
a.SetJSRunner(r.ForAnalyzer(a))
|
||||
a.SetContent("", "")
|
||||
got, err := a.GetString(`@js:java.timeFormat(1700000000000, 'yyyy-MM-dd HH:mm')`, nil, false)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if !regexp.MustCompile(`^\d{4}-\d{2}-\d{2} \d{2}:\d{2}$`).MatchString(got) {
|
||||
t.Fatalf("timeFormat = %q", got)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,476 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"encoding/base64"
|
||||
"encoding/hex"
|
||||
"encoding/json"
|
||||
"fmt"
|
||||
"strings"
|
||||
"sync"
|
||||
"time"
|
||||
|
||||
"github.com/dop251/goja"
|
||||
)
|
||||
|
||||
// 本文件:桥接层的小工具(编码字节转换、cookie/cache 对象、时间格式化)。
|
||||
|
||||
var b64EncodingVariants = []*base64.Encoding{
|
||||
base64.StdEncoding,
|
||||
base64.RawStdEncoding,
|
||||
base64.URLEncoding,
|
||||
base64.RawURLEncoding,
|
||||
}
|
||||
|
||||
// base64DecodeBytes 宽松解码:自动补 padding、支持 URL-safe 变体
|
||||
// (对应 legado Base64 解码的宽容行为)。
|
||||
func base64DecodeBytes(s string) ([]byte, error) {
|
||||
s = strings.Map(func(r rune) rune {
|
||||
if r == '\n' || r == '\r' || r == ' ' || r == '\t' {
|
||||
return -1
|
||||
}
|
||||
return r
|
||||
}, s)
|
||||
var lastErr error
|
||||
for _, enc := range b64EncodingVariants {
|
||||
b, err := enc.DecodeString(s)
|
||||
if err == nil {
|
||||
return b, nil
|
||||
}
|
||||
lastErr = err
|
||||
}
|
||||
// 缺 padding 的 std 变体
|
||||
if padded := s + strings.Repeat("=", (4-len(s)%4)%4); padded != s {
|
||||
if b, err := base64.StdEncoding.DecodeString(padded); err == nil {
|
||||
return b, nil
|
||||
}
|
||||
}
|
||||
if lastErr == nil {
|
||||
lastErr = fmt.Errorf("无效 Base64")
|
||||
}
|
||||
return nil, lastErr
|
||||
}
|
||||
|
||||
func base64DecodeString(s string) (string, error) {
|
||||
b, err := base64DecodeBytes(s)
|
||||
return string(b), err
|
||||
}
|
||||
|
||||
func hexDecodeBytes(s string) ([]byte, error) { return hex.DecodeString(s) }
|
||||
func base64StdEncodeBytes(b []byte) string { return base64.StdEncoding.EncodeToString(b) }
|
||||
func hexEncodeBytes(b []byte) string { return hex.EncodeToString(b) }
|
||||
|
||||
// ─── cookie 对象(对应 legado CookieStore 注入的 `cookie`) ─────────────────
|
||||
|
||||
// 会话状态由服务层注入(落库);未注入时用进程内 MemoryState。
|
||||
// 见 state.go —— 旧的全局 cookieJar 因无法按书源隔离且不落库已废弃。
|
||||
|
||||
// newCookieObject 构造 JS 的 `cookie` 对象:
|
||||
// getCookie(url[,key]) / setCookie(url,cookie) / replaceCookie / removeCookie / getKey。
|
||||
func newCookieObject(vm *goja.Runtime, state SourceState) *goja.Object {
|
||||
o := vm.NewObject()
|
||||
set := func(k string, v any) {
|
||||
if err := o.Set(k, v); err != nil {
|
||||
panic(vm.ToValue(err.Error()))
|
||||
}
|
||||
}
|
||||
set("getCookie", func(call goja.FunctionCall) goja.Value {
|
||||
tag := stringArg(call, 0)
|
||||
if len(call.Arguments) > 1 {
|
||||
return vm.ToValue(state.GetCookieKey(tag, stringArg(call, 1)))
|
||||
}
|
||||
return vm.ToValue(state.GetCookie(tag))
|
||||
})
|
||||
set("setCookie", func(call goja.FunctionCall) goja.Value {
|
||||
rawURL := stringArg(call, 0)
|
||||
cookie := stringArgOr(call, 1, "")
|
||||
if cookie != "" {
|
||||
state.SetCookie(rawURL, cookie)
|
||||
}
|
||||
return vm.ToValue(cookie)
|
||||
})
|
||||
// replaceCookie 与 setCookie 在服务端实现中同为覆盖式合并(对应 legado 的语义)
|
||||
set("replaceCookie", func(call goja.FunctionCall) goja.Value {
|
||||
rawURL := stringArg(call, 0)
|
||||
cookie := stringArgOr(call, 1, "")
|
||||
if cookie != "" {
|
||||
state.SetCookie(rawURL, cookie)
|
||||
}
|
||||
return vm.ToValue(cookie)
|
||||
})
|
||||
set("removeCookie", func(call goja.FunctionCall) goja.Value {
|
||||
state.RemoveCookie(stringArg(call, 0))
|
||||
return goja.Null()
|
||||
})
|
||||
set("getKey", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(state.GetCookieKey(stringArg(call, 0), stringArg(call, 1)))
|
||||
})
|
||||
return o
|
||||
}
|
||||
|
||||
// ─── 书籍 / 章节对象 ───────────────────────────────────────────────────────
|
||||
|
||||
// newBookObject 构造规则 JS 里的 `book`(对应 legado 的 Book 实体)。
|
||||
//
|
||||
// 书源会读它的元数据(name / author / coverUrl / durChapterIndex / order / type…)、
|
||||
// 给它赋值(book.type = …、book.imageStyle = …)、调用 setUseReplaceRule(),
|
||||
// 以及用 getVariable / putVariable 读写书籍自定义变量。
|
||||
//
|
||||
// 早期这里只绑了 {"name": ...},书源一碰 `book.setUseReplaceRule(false)`
|
||||
// 就 TypeError,整段详情/目录规则 JS 直接失败(表现为「详情空白、目录 0 章」)。
|
||||
func newBookObject(vm *goja.Runtime, a *AnalyzeRule) *goja.Object {
|
||||
o := vm.NewObject()
|
||||
set := func(k string, v any) {
|
||||
_ = o.Set(k, v)
|
||||
}
|
||||
for k, v := range a.bookMeta {
|
||||
if k == "type" {
|
||||
continue // type 用访问器,见下
|
||||
}
|
||||
set(k, v)
|
||||
}
|
||||
// name 以 SetBookContext 的值为准(legado 里 book.name 就是这个)
|
||||
set("name", a.bookName)
|
||||
set("bookName", a.bookName)
|
||||
|
||||
// book.type:书源会赋值来声明书籍类型(听书=1 / 漫画=2 …),
|
||||
// legado 会把它写回 Book.type,服务层据此决定正文按文本/音频/图片返回。
|
||||
// 用访问器把写入记下来,否则赋值只活在本次 JS 里,读完仍是文本。
|
||||
_ = o.DefineAccessorProperty("type",
|
||||
vm.ToValue(func(call goja.FunctionCall) goja.Value { return vm.ToValue(a.bookTypeValue()) }),
|
||||
vm.ToValue(func(call goja.FunctionCall) goja.Value {
|
||||
if len(call.Arguments) > 0 && !goja.IsUndefined(call.Arguments[0]) && !goja.IsNull(call.Arguments[0]) {
|
||||
a.SetBookType(int(call.Arguments[0].ToInteger()))
|
||||
}
|
||||
return goja.Undefined()
|
||||
}),
|
||||
goja.FLAG_FALSE, goja.FLAG_TRUE)
|
||||
|
||||
// readConfig:书源读 book.readConfig.useReplaceRule,并可能回写
|
||||
rc := vm.NewObject()
|
||||
_ = rc.Set("useReplaceRule", false)
|
||||
set("readConfig", rc)
|
||||
set("setUseReplaceRule", func(call goja.FunctionCall) goja.Value {
|
||||
if len(call.Arguments) > 0 {
|
||||
_ = rc.Set("useReplaceRule", call.Arguments[0].ToBoolean())
|
||||
}
|
||||
return goja.Null()
|
||||
})
|
||||
|
||||
// 书籍自定义变量(legado Book.variableMap)。
|
||||
//
|
||||
// 注意「缺省返回空串」:legado 的 RuleDataInterface.getVariable 是
|
||||
// variableMap[key] ?: getBigVariable(key) ?: ""
|
||||
// 返回 "" 而不是 null。书源会直接写 `String(book.getVariable('custom')) || ''`,
|
||||
// 若这里返回 null,String(null) 得到字符串 "null"(真值),会被当成
|
||||
// tone_id 发给站点,站点直接返回空正文(表现为「正文 0 字」)。
|
||||
set("getVariable", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(a.bookCustom[stringArg(call, 0)])
|
||||
})
|
||||
set("putVariable", func(call goja.FunctionCall) goja.Value {
|
||||
if a.bookCustom == nil {
|
||||
a.bookCustom = map[string]string{}
|
||||
}
|
||||
a.bookCustom[stringArg(call, 0)] = stringArgOr(call, 1, "")
|
||||
return goja.Null()
|
||||
})
|
||||
return o
|
||||
}
|
||||
|
||||
// newChapterObject 构造规则 JS 里的 `chapter`(对应 legado 的 BookChapter)。
|
||||
func newChapterObject(vm *goja.Runtime, a *AnalyzeRule) *goja.Object {
|
||||
o := vm.NewObject()
|
||||
_ = o.Set("title", a.chapterTitle)
|
||||
_ = o.Set("index", a.chapterIndex)
|
||||
_ = o.Set("isVip", false)
|
||||
_ = o.Set("isPay", false)
|
||||
return o
|
||||
}
|
||||
|
||||
// newSourceObject 构造 JS 的 `source` 对象。
|
||||
// 对应 legado BaseSource 的变量与登录信息读写。
|
||||
func newSourceObject(vm *goja.Runtime, state SourceState, props map[string]any) *goja.Object {
|
||||
o := vm.NewObject()
|
||||
set := func(k string, v any) {
|
||||
if err := o.Set(k, v); err != nil {
|
||||
panic(vm.ToValue(err.Error()))
|
||||
}
|
||||
}
|
||||
bindSourceState(vm, set, state, props)
|
||||
// source.get/put:书源级键值缓存(legado 中即 source 的方法)。
|
||||
// 注意 java.get/put 另有语义(网络 / 解析器变量),在 bridge.go 中定义,
|
||||
// 不在这里覆盖。
|
||||
set("get", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(sourceKeyGet(state, stringArg(call, 0)))
|
||||
})
|
||||
set("put", func(call goja.FunctionCall) goja.Value {
|
||||
val := stringArgOr(call, 1, "")
|
||||
sourceKeyPut(state, stringArg(call, 0), val)
|
||||
return vm.ToValue(val)
|
||||
})
|
||||
// 书源自身属性(bookSourceUrl 等)原样可读,存量 JS 会读 source.loginUi 等做能力探测。
|
||||
for k, v := range props {
|
||||
if o.Get(k) != nil {
|
||||
continue
|
||||
}
|
||||
_ = o.Set(k, v)
|
||||
}
|
||||
return o
|
||||
}
|
||||
|
||||
// bindSourceState 把书源会话状态方法绑定到目标 JS 对象。
|
||||
// legado 中 book source 的 evalJS 把 `java` 绑成书源对象自身,因此 `java` 与
|
||||
// `source` 都能读到 getVariable/getLoginInfo 等方法——这里保持一致。
|
||||
func bindSourceState(vm *goja.Runtime, set func(k string, v any), state SourceState, props map[string]any) {
|
||||
set("getVariable", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(state.GetVariable())
|
||||
})
|
||||
setVariable := func(call goja.FunctionCall) goja.Value {
|
||||
if len(call.Arguments) == 0 || goja.IsUndefined(call.Arguments[0]) || goja.IsNull(call.Arguments[0]) {
|
||||
state.SetVariable("")
|
||||
return goja.Null()
|
||||
}
|
||||
state.SetVariable(call.Arguments[0].String())
|
||||
return goja.Null()
|
||||
}
|
||||
set("setVariable", setVariable)
|
||||
set("putVariable", setVariable)
|
||||
|
||||
set("getLoginInfo", func(call goja.FunctionCall) goja.Value {
|
||||
if v := state.GetLoginInfo(); v != "" {
|
||||
return vm.ToValue(v)
|
||||
}
|
||||
return goja.Null()
|
||||
})
|
||||
set("putLoginInfo", func(call goja.FunctionCall) goja.Value {
|
||||
v := ""
|
||||
if len(call.Arguments) > 0 && !goja.IsUndefined(call.Arguments[0]) && !goja.IsNull(call.Arguments[0]) {
|
||||
v = call.Arguments[0].String()
|
||||
}
|
||||
state.SetLoginInfo(v)
|
||||
return vm.ToValue(true)
|
||||
})
|
||||
set("removeLoginInfo", func(call goja.FunctionCall) goja.Value {
|
||||
state.SetLoginInfo("")
|
||||
return goja.Null()
|
||||
})
|
||||
// getLoginInfoMap 对应 Kotlin:解析登录信息 JSON;
|
||||
// 未设置且 loginUi 非空时,用 loginUi 里各字段的 default 初始化。
|
||||
set("getLoginInfoMap", func(call goja.FunctionCall) goja.Value {
|
||||
if m := parseLoginInfoMap(state.GetLoginInfo()); m != nil {
|
||||
return vm.ToValue(m)
|
||||
}
|
||||
return vm.ToValue(initLoginInfoFromUI(props))
|
||||
})
|
||||
|
||||
set("getLoginHeader", func(call goja.FunctionCall) goja.Value {
|
||||
if v := state.GetLoginHeader(); v != "" {
|
||||
return vm.ToValue(v)
|
||||
}
|
||||
return goja.Null()
|
||||
})
|
||||
set("putLoginHeader", func(call goja.FunctionCall) goja.Value {
|
||||
header := stringArgOr(call, 0, "")
|
||||
state.SetLoginHeader(header)
|
||||
// 请求头里的 Cookie 同步进 cookie 存储(对应 legado putLoginHeader)
|
||||
if header != "" {
|
||||
var m map[string]any
|
||||
if json.Unmarshal([]byte(header), &m) == nil {
|
||||
for k, v := range m {
|
||||
if strings.EqualFold(k, "cookie") {
|
||||
state.SetCookie(headerCookieURL(props), fmt.Sprintf("%v", v))
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
return goja.Null()
|
||||
})
|
||||
set("removeLoginHeader", func(call goja.FunctionCall) goja.Value {
|
||||
state.SetLoginHeader("")
|
||||
state.RemoveCookie(headerCookieURL(props))
|
||||
return goja.Null()
|
||||
})
|
||||
set("getKey", func(call goja.FunctionCall) goja.Value {
|
||||
return vm.ToValue(headerCookieURL(props))
|
||||
})
|
||||
}
|
||||
|
||||
// sourceKeyGet / sourceKeyPut:source.get/put 的独立键值槽,
|
||||
// 以 "__kv__" 前缀编码进同一份变量 JSON,从而与 getVariable 一起持久化。
|
||||
const sourceKVVar = "__kv__"
|
||||
|
||||
func sourceKeyGet(state SourceState, key string) string {
|
||||
if key == "" {
|
||||
return ""
|
||||
}
|
||||
var m map[string]string
|
||||
_ = json.Unmarshal([]byte(state.GetVariable()), &m)
|
||||
return m[sourceKVVar+key]
|
||||
}
|
||||
|
||||
func sourceKeyPut(state SourceState, key, val string) {
|
||||
if key == "" {
|
||||
return
|
||||
}
|
||||
m := map[string]string{}
|
||||
_ = json.Unmarshal([]byte(state.GetVariable()), &m)
|
||||
if m == nil {
|
||||
m = map[string]string{}
|
||||
}
|
||||
m[sourceKVVar+key] = val
|
||||
b, err := json.Marshal(m)
|
||||
if err != nil {
|
||||
return
|
||||
}
|
||||
state.SetVariable(string(b))
|
||||
}
|
||||
|
||||
// headerCookieURL 取 loginHeader 里 Cookie 归属的 URL。
|
||||
func headerCookieURL(props map[string]any) string {
|
||||
if props != nil {
|
||||
if u, ok := props["bookSourceUrl"].(string); ok && u != "" {
|
||||
return u
|
||||
}
|
||||
}
|
||||
return ""
|
||||
}
|
||||
|
||||
// parseLoginInfoMap 解析登录信息 JSON;空串或非法 JSON 返回 nil。
|
||||
func parseLoginInfoMap(s string) map[string]string {
|
||||
if strings.TrimSpace(s) == "" {
|
||||
return nil
|
||||
}
|
||||
var raw map[string]any
|
||||
if json.Unmarshal([]byte(s), &raw) != nil {
|
||||
return nil
|
||||
}
|
||||
out := make(map[string]string, len(raw))
|
||||
for k, v := range raw {
|
||||
out[k] = fmt.Sprintf("%v", v)
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// initLoginInfoFromUI 用 loginUi 各字段 default 生成初始登录信息
|
||||
// (对应 legado getLoginInfoMap:跳过 button 类型)。
|
||||
func initLoginInfoFromUI(props map[string]any) map[string]string {
|
||||
out := map[string]string{}
|
||||
raw, _ := props["loginUi"].(string)
|
||||
if strings.TrimSpace(raw) != "" {
|
||||
var rows []map[string]any
|
||||
if json.Unmarshal([]byte(raw), &rows) == nil {
|
||||
for _, row := range rows {
|
||||
if t, _ := row["type"].(string); t == "button" {
|
||||
continue
|
||||
}
|
||||
name, _ := row["name"].(string)
|
||||
if name == "" {
|
||||
continue
|
||||
}
|
||||
def, _ := row["default"].(string)
|
||||
out[name] = def
|
||||
}
|
||||
}
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// ─── cache 对象(对应 legado CacheManager 注入的 `cache`) ──────────────────
|
||||
//
|
||||
// legado 的 cache 有两套存储:put/get/delete 落持久缓存(ACache),
|
||||
// putMemory/getFromMemory 落进程内内存缓存。书源靠后者记录「这条段评点过几次」
|
||||
// 这类临时状态——光遇聚合的 paraForAndroid 每一段带段评的文字都会调
|
||||
// cache.putMemory(url, 0),缺了它整条正文规则会抛 TypeError 直接失败。
|
||||
// 两套存储分开,否则 getFromMemory 会读到 put 写进去的持久值。
|
||||
var jsCache = struct {
|
||||
mu sync.Mutex
|
||||
m map[string]string
|
||||
mem map[string]string
|
||||
}{m: map[string]string{}, mem: map[string]string{}}
|
||||
|
||||
// jsCacheMaxEntries 单套存储的条目上限:超了整体清空,避免书源把内存吃满。
|
||||
const jsCacheMaxEntries = 4096
|
||||
|
||||
func newCacheObject(vm *goja.Runtime) *goja.Object {
|
||||
o := vm.NewObject()
|
||||
set := func(k string, v any) {
|
||||
if err := o.Set(k, v); err != nil {
|
||||
panic(vm.ToValue(err.Error()))
|
||||
}
|
||||
}
|
||||
|
||||
// 两套存储共用同一份读写实现,只有落点不同。
|
||||
putTo := func(store *map[string]string) func(goja.FunctionCall) goja.Value {
|
||||
return func(call goja.FunctionCall) goja.Value {
|
||||
key := stringArg(call, 0)
|
||||
val := ""
|
||||
if len(call.Arguments) > 1 && !goja.IsUndefined(call.Arguments[1]) && !goja.IsNull(call.Arguments[1]) {
|
||||
val = call.Arguments[1].String()
|
||||
}
|
||||
jsCache.mu.Lock()
|
||||
if len(*store) >= jsCacheMaxEntries {
|
||||
*store = map[string]string{}
|
||||
}
|
||||
(*store)[key] = val
|
||||
jsCache.mu.Unlock()
|
||||
return vm.ToValue(val)
|
||||
}
|
||||
}
|
||||
getFrom := func(store *map[string]string) func(goja.FunctionCall) goja.Value {
|
||||
return func(call goja.FunctionCall) goja.Value {
|
||||
key := stringArg(call, 0)
|
||||
jsCache.mu.Lock()
|
||||
v, ok := (*store)[key]
|
||||
jsCache.mu.Unlock()
|
||||
if !ok {
|
||||
return goja.Null()
|
||||
}
|
||||
return vm.ToValue(v)
|
||||
}
|
||||
}
|
||||
deleteFrom := func(store *map[string]string) func(goja.FunctionCall) goja.Value {
|
||||
return func(call goja.FunctionCall) goja.Value {
|
||||
key := stringArg(call, 0)
|
||||
jsCache.mu.Lock()
|
||||
delete(*store, key)
|
||||
jsCache.mu.Unlock()
|
||||
return goja.Null()
|
||||
}
|
||||
}
|
||||
|
||||
set("put", putTo(&jsCache.m))
|
||||
set("get", getFrom(&jsCache.m))
|
||||
set("delete", deleteFrom(&jsCache.m))
|
||||
// 内存缓存(legado Cache.getFromMemory / putMemory)
|
||||
set("putMemory", putTo(&jsCache.mem))
|
||||
set("getFromMemory", getFrom(&jsCache.mem))
|
||||
return o
|
||||
}
|
||||
|
||||
// ─── 时间格式化(对应 JsExtensions.timeFormat/timeFormatUTC) ───────────────
|
||||
|
||||
// javaTimeFormat 将 Java SimpleDateFormat 常用 pattern 转 Go 布局。
|
||||
// sh 为时区偏移小时数(timeFormatUTC 语义,0 表示本地时区)。
|
||||
func javaTimeFormat(ts int64, format string, sh int64) string {
|
||||
loc := time.Local
|
||||
if sh != 0 {
|
||||
loc = time.FixedZone(fmt.Sprintf("UTC%+d", sh), int(sh)*3600)
|
||||
}
|
||||
t := time.UnixMilli(ts).In(loc)
|
||||
return t.Format(javaPatternToGo(format))
|
||||
}
|
||||
|
||||
func javaPatternToGo(p string) string {
|
||||
replacements := []struct{ java, goLayout string }{
|
||||
{"yyyy", "2006"}, {"yy", "06"},
|
||||
{"MM", "01"}, {"dd", "02"},
|
||||
{"HH", "15"}, {"hh", "03"},
|
||||
{"mm", "04"}, {"ss", "05"},
|
||||
{"SSS", "000"},
|
||||
{"a", "PM"},
|
||||
}
|
||||
// 优先替换长 token,避免 yyyy 被 yy+yy 拆坏
|
||||
for _, r := range replacements {
|
||||
p = strings.ReplaceAll(p, r.java, "\x00"+r.goLayout+"\x00")
|
||||
}
|
||||
return strings.NewReplacer("\x00", "").Replace(p)
|
||||
}
|
||||
@@ -0,0 +1,53 @@
|
||||
package rule
|
||||
|
||||
import "testing"
|
||||
|
||||
// TestCacheMemoryRoundTrip 内存缓存(cache.putMemory / cache.getFromMemory)要能读写。
|
||||
//
|
||||
// 回归:「cache 对象只有 put/get/delete」曾让光遇聚合的正文规则整条失败——
|
||||
// paraForAndroid 每一段带段评的文字都会调 cache.putMemory(url, 0),缺了它就抛
|
||||
// TypeError: Object has no member 'putMemory',正文接口直接 400。
|
||||
func TestCacheMemoryRoundTrip(t *testing.T) {
|
||||
r := NewJSRunner(JSConfig{})
|
||||
ar := NewAnalyzeRule()
|
||||
|
||||
if _, err := r.Run(ar, `cache.putMemory('__test_mem', 3)`, nil, ""); err != nil {
|
||||
t.Fatalf("cache.putMemory 失败: %v", err)
|
||||
}
|
||||
v, err := r.Run(ar, `String(cache.getFromMemory('__test_mem'))`, nil, "")
|
||||
if err != nil {
|
||||
t.Fatalf("cache.getFromMemory 失败: %v", err)
|
||||
}
|
||||
if got := anyToString(v); got != "3" {
|
||||
t.Fatalf("getFromMemory = %q,期望 3", got)
|
||||
}
|
||||
|
||||
// 缺省返回 null(书源会写 `cache.getFromMemory(k) || ''` 这类兜底)。
|
||||
v, err = r.Run(ar, `String(cache.getFromMemory('__test_mem_missing'))`, nil, "")
|
||||
if err != nil {
|
||||
t.Fatalf("cache.getFromMemory 失败: %v", err)
|
||||
}
|
||||
if got := anyToString(v); got != "null" {
|
||||
t.Fatalf("缺失键 = %q,期望 null", got)
|
||||
}
|
||||
|
||||
// 持久缓存与内存缓存是两套存储(legado 的 put/get 走 ACache,putMemory 走内存),
|
||||
// 互不串门,否则书源会读到本该过期的值。
|
||||
if _, err := r.Run(ar, `cache.put('__test_persist', 'p')`, nil, ""); err != nil {
|
||||
t.Fatalf("cache.put 失败: %v", err)
|
||||
}
|
||||
v, err = r.Run(ar, `String(cache.getFromMemory('__test_persist'))`, nil, "")
|
||||
if err != nil {
|
||||
t.Fatalf("cache.getFromMemory 失败: %v", err)
|
||||
}
|
||||
if got := anyToString(v); got != "null" {
|
||||
t.Fatalf("getFromMemory 读到了持久缓存的值: %q", got)
|
||||
}
|
||||
v, err = r.Run(ar, `String(cache.get('__test_mem'))`, nil, "")
|
||||
if err != nil {
|
||||
t.Fatalf("cache.get 失败: %v", err)
|
||||
}
|
||||
if got := anyToString(v); got != "null" {
|
||||
t.Fatalf("cache.get 读到了内存缓存的值: %q", got)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,253 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"sort"
|
||||
"strconv"
|
||||
"strings"
|
||||
)
|
||||
|
||||
// 本文件对应 AnalyzeByJSonPath.kt(Jayway JsonPath 语义,Go 侧用
|
||||
// PaesslerAG/jsonpath 实现)。
|
||||
|
||||
type jsonAnalyzer struct {
|
||||
root any
|
||||
}
|
||||
|
||||
func newJSONAnalyzer(doc any) *jsonAnalyzer {
|
||||
switch t := doc.(type) {
|
||||
case string:
|
||||
var v any
|
||||
if err := json.Unmarshal([]byte(t), &v); err != nil {
|
||||
return &jsonAnalyzer{root: nil}
|
||||
}
|
||||
return &jsonAnalyzer{root: v}
|
||||
default:
|
||||
return &jsonAnalyzer{root: doc}
|
||||
}
|
||||
}
|
||||
|
||||
// jsonRead 对应 ctx.read(rule):路径求值,失败返回 nil。
|
||||
func jsonRead(root any, path string) any {
|
||||
if root == nil || path == "" {
|
||||
return nil
|
||||
}
|
||||
v, err := jsonpathGet(path, root)
|
||||
if err != nil {
|
||||
return nil
|
||||
}
|
||||
return v
|
||||
}
|
||||
|
||||
// jsonValueToString 对应 Kotlin 的 ob.toString() / joinToString("\n")。
|
||||
// 列表按 "\n" 拼接(legado 的 getString 就是这么做的);对象走 Java 的 Map.toString
|
||||
// 形态(见 javaValueString),书源的正则大量依赖这个形态。
|
||||
func jsonValueToString(ob any) string {
|
||||
switch t := ob.(type) {
|
||||
case nil:
|
||||
return ""
|
||||
case string:
|
||||
return t
|
||||
case []any:
|
||||
parts := make([]string, len(t))
|
||||
for i, e := range t {
|
||||
parts[i] = jsonValueToString(e)
|
||||
}
|
||||
return strings.Join(parts, "\n")
|
||||
default:
|
||||
return javaValueString(t)
|
||||
}
|
||||
}
|
||||
|
||||
// javaValueString 按 Java 的 Map/List toString 形态输出 JSON 值。
|
||||
//
|
||||
// 书源规则里的正则就是照着这个形态写的,最典型的是拷贝漫画的作者规则
|
||||
// `$.author##.*name=(.*?)\,.*##$1`:它期望拿到 `{name=岸本斉史, alias=…}`。
|
||||
// 之前这里用 Go 的 fmt 输出 `map[alias:… name:…]`,`name=` 根本匹配不到,
|
||||
// 作者字段就退化成一整串 map 文本。
|
||||
//
|
||||
// 已知差异:Java 的 LinkedTreeMap 保留 JSON 里的键顺序,而 Go 的
|
||||
// map[string]interface{} 不保序(JSONPath 库也只认这个类型),这里按 **键排序**
|
||||
// 输出以保证确定性。绝大多数「取某个键」的正则不受影响,但依赖原始键序的正则
|
||||
// (如 `.*name=(.*?)\,` 且 name 是最后一个键)仍可能与阅读 App 的结果不同。
|
||||
func javaValueString(v any) string {
|
||||
switch t := v.(type) {
|
||||
case nil:
|
||||
return "null"
|
||||
case string:
|
||||
return t
|
||||
case bool:
|
||||
return strconv.FormatBool(t)
|
||||
case float64:
|
||||
return strconv.FormatFloat(t, 'f', -1, 64)
|
||||
case int:
|
||||
return strconv.Itoa(t)
|
||||
case int64:
|
||||
return strconv.FormatInt(t, 10)
|
||||
case json.Number:
|
||||
return t.String()
|
||||
case []any:
|
||||
parts := make([]string, len(t))
|
||||
for i, e := range t {
|
||||
parts[i] = javaValueString(e)
|
||||
}
|
||||
return "[" + strings.Join(parts, ", ") + "]"
|
||||
case map[string]any:
|
||||
keys := make([]string, 0, len(t))
|
||||
for k := range t {
|
||||
keys = append(keys, k)
|
||||
}
|
||||
sort.Strings(keys)
|
||||
var sb strings.Builder
|
||||
sb.WriteByte('{')
|
||||
for i, k := range keys {
|
||||
if i > 0 {
|
||||
sb.WriteString(", ")
|
||||
}
|
||||
sb.WriteString(k)
|
||||
sb.WriteByte('=')
|
||||
sb.WriteString(javaValueString(t[k]))
|
||||
}
|
||||
sb.WriteByte('}')
|
||||
return sb.String()
|
||||
default:
|
||||
return anyToString(v)
|
||||
}
|
||||
}
|
||||
|
||||
// getString 对应 AnalyzeByJSonPath.getString。
|
||||
func (a *jsonAnalyzer) getString(rule string) string {
|
||||
if rule == "" {
|
||||
return ""
|
||||
}
|
||||
ra := NewRuleAnalyzer(rule, true)
|
||||
rules := ra.SplitRule("&&", "||")
|
||||
if len(rules) == 1 {
|
||||
ra.ReSetPos()
|
||||
result := ra.InnerRule("{$.", 1, 1, func(inner string) string {
|
||||
return a.getString(inner)
|
||||
})
|
||||
if result == "" {
|
||||
result = jsonValueToString(jsonRead(a.root, rule))
|
||||
}
|
||||
return result
|
||||
}
|
||||
var textList []string
|
||||
for _, rl := range rules {
|
||||
temp := a.getString(rl)
|
||||
if temp != "" {
|
||||
textList = append(textList, temp)
|
||||
if ra.ElementsType() == "||" {
|
||||
break
|
||||
}
|
||||
}
|
||||
}
|
||||
return strings.Join(textList, "\n")
|
||||
}
|
||||
|
||||
// getStringList 对应 AnalyzeByJSonPath.getStringList。
|
||||
func (a *jsonAnalyzer) getStringList(rule string) []string {
|
||||
var result []string
|
||||
if rule == "" {
|
||||
return result
|
||||
}
|
||||
ra := NewRuleAnalyzer(rule, true)
|
||||
rules := ra.SplitRule("&&", "||", "%%")
|
||||
if len(rules) == 1 {
|
||||
ra.ReSetPos()
|
||||
st := ra.InnerRule("{$.", 1, 1, func(inner string) string {
|
||||
return a.getString(inner)
|
||||
})
|
||||
if st == "" {
|
||||
ob := jsonRead(a.root, rule)
|
||||
if ob == nil {
|
||||
return result
|
||||
}
|
||||
if lst, ok := ob.([]any); ok {
|
||||
for _, o := range lst {
|
||||
result = append(result, jsonValueToString(o))
|
||||
}
|
||||
} else {
|
||||
result = append(result, jsonValueToString(ob))
|
||||
}
|
||||
} else {
|
||||
result = append(result, st)
|
||||
}
|
||||
return result
|
||||
}
|
||||
var results [][]string
|
||||
for _, rl := range rules {
|
||||
temp := a.getStringList(rl)
|
||||
if len(temp) > 0 {
|
||||
results = append(results, temp)
|
||||
if ra.ElementsType() == "||" {
|
||||
break
|
||||
}
|
||||
}
|
||||
}
|
||||
if len(results) > 0 {
|
||||
if ra.ElementsType() == "%%" {
|
||||
for i := 0; i < len(results[0]); i++ {
|
||||
for _, temp := range results {
|
||||
if i < len(temp) {
|
||||
result = append(result, temp[i])
|
||||
}
|
||||
}
|
||||
}
|
||||
} else {
|
||||
for _, temp := range results {
|
||||
result = append(result, temp...)
|
||||
}
|
||||
}
|
||||
}
|
||||
return result
|
||||
}
|
||||
|
||||
// getObject 对应 getObject:直接返回求值结果。
|
||||
func (a *jsonAnalyzer) getObject(rule string) any {
|
||||
return jsonRead(a.root, rule)
|
||||
}
|
||||
|
||||
// getList 对应 getList:要求路径结果为数组(对应 jayway read<ArrayList>,
|
||||
// 非数组时 legado 侧捕获异常返回空列表)。
|
||||
func (a *jsonAnalyzer) getList(rule string) []any {
|
||||
var result []any
|
||||
if rule == "" {
|
||||
return result
|
||||
}
|
||||
ra := NewRuleAnalyzer(rule, true)
|
||||
rules := ra.SplitRule("&&", "||", "%%")
|
||||
if len(rules) == 1 {
|
||||
ob := jsonRead(a.root, rules[0])
|
||||
if lst, ok := ob.([]any); ok {
|
||||
return lst
|
||||
}
|
||||
return result
|
||||
}
|
||||
var results [][]any
|
||||
for _, rl := range rules {
|
||||
temp := a.getList(rl)
|
||||
if len(temp) > 0 {
|
||||
results = append(results, temp)
|
||||
if ra.ElementsType() == "||" {
|
||||
break
|
||||
}
|
||||
}
|
||||
}
|
||||
if len(results) > 0 {
|
||||
if ra.ElementsType() == "%%" {
|
||||
for i := 0; i < len(results[0]); i++ {
|
||||
for _, temp := range results {
|
||||
if i < len(temp) {
|
||||
result = append(result, temp[i])
|
||||
}
|
||||
}
|
||||
}
|
||||
} else {
|
||||
for _, temp := range results {
|
||||
result = append(result, temp...)
|
||||
}
|
||||
}
|
||||
}
|
||||
return result
|
||||
}
|
||||
@@ -0,0 +1,626 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"strings"
|
||||
|
||||
"golang.org/x/net/html"
|
||||
)
|
||||
|
||||
// 本文件对应 AnalyzeByJSoup.kt(含 ElementsSingle 索引语法)。
|
||||
|
||||
type jsoupAnalyzer struct {
|
||||
root *html.Node
|
||||
}
|
||||
|
||||
func newJsoupAnalyzer(doc any) *jsoupAnalyzer {
|
||||
return &jsoupAnalyzer{root: toHTMLNode(doc)}
|
||||
}
|
||||
|
||||
// toHTMLNode 对应 AnalyzeByJSoup.parse(doc):节点直通,字符串解析为 HTML。
|
||||
func toHTMLNode(doc any) *html.Node {
|
||||
switch t := doc.(type) {
|
||||
case nil:
|
||||
return nil
|
||||
case *html.Node:
|
||||
return t
|
||||
default:
|
||||
return parseHTML(anyToString(t))
|
||||
}
|
||||
}
|
||||
|
||||
// getStringList 对应 AnalyzeByJSoup.getStringList。
|
||||
func (a *jsoupAnalyzer) getStringList(ruleStr string) []string {
|
||||
var textS []string
|
||||
if ruleStr == "" {
|
||||
return textS
|
||||
}
|
||||
isCss := false
|
||||
elementsRule := ruleStr
|
||||
if hasPrefixFold(ruleStr, "@CSS:") {
|
||||
isCss = true
|
||||
elementsRule = strings.TrimSpace(ruleStr[5:])
|
||||
}
|
||||
|
||||
if elementsRule == "" {
|
||||
if d := nodeData(a.root); d != "" {
|
||||
textS = append(textS, d)
|
||||
} else {
|
||||
textS = append(textS, "")
|
||||
}
|
||||
return textS
|
||||
}
|
||||
|
||||
ra := NewRuleAnalyzer(elementsRule, false)
|
||||
ruleStrS := ra.SplitRule("&&", "||", "%%")
|
||||
|
||||
var results [][]string
|
||||
for _, ruleStrX := range ruleStrS {
|
||||
var temp []string
|
||||
if isCss {
|
||||
// 对应 isCss 分支:lastIndexOf('@') 分离选择器与提取规则
|
||||
lastIndex := strings.LastIndex(ruleStrX, "@")
|
||||
var elements []*html.Node
|
||||
var lastRule string
|
||||
if lastIndex == -1 {
|
||||
elements = selectCSS(a.root, ruleStrX)
|
||||
lastRule = "text"
|
||||
} else {
|
||||
selector := ruleStrX[:lastIndex]
|
||||
if strings.TrimSpace(selector) == "" {
|
||||
elements = []*html.Node{a.root}
|
||||
} else {
|
||||
elements = selectCSS(a.root, selector)
|
||||
}
|
||||
lastRule = ruleStrX[lastIndex+1:]
|
||||
}
|
||||
temp = getResultLast(elements, lastRule)
|
||||
} else {
|
||||
temp = a.getResultList(ruleStrX)
|
||||
}
|
||||
if len(temp) > 0 {
|
||||
results = append(results, temp)
|
||||
if ra.ElementsType() == "||" {
|
||||
break
|
||||
}
|
||||
}
|
||||
}
|
||||
if len(results) > 0 {
|
||||
if ra.ElementsType() == "%%" {
|
||||
for i := 0; i < len(results[0]); i++ {
|
||||
for _, temp := range results {
|
||||
if i < len(temp) {
|
||||
textS = append(textS, temp[i])
|
||||
}
|
||||
}
|
||||
}
|
||||
} else {
|
||||
for _, temp := range results {
|
||||
textS = append(textS, temp...)
|
||||
}
|
||||
}
|
||||
}
|
||||
return textS
|
||||
}
|
||||
|
||||
// getString 对应 getString:多结果以 \n 连接。
|
||||
func (a *jsoupAnalyzer) getString(ruleStr string) string {
|
||||
list := a.getStringList(ruleStr)
|
||||
if len(list) == 0 {
|
||||
return ""
|
||||
}
|
||||
if len(list) == 1 {
|
||||
return list[0]
|
||||
}
|
||||
return strings.Join(list, "\n")
|
||||
}
|
||||
|
||||
// getString0 对应 getString0:只取第一个。
|
||||
func (a *jsoupAnalyzer) getString0(ruleStr string) string {
|
||||
list := a.getStringList(ruleStr)
|
||||
if len(list) == 0 {
|
||||
return ""
|
||||
}
|
||||
return list[0]
|
||||
}
|
||||
|
||||
// getElements 对应 getElements。
|
||||
func (a *jsoupAnalyzer) getElements(rule string) []*html.Node {
|
||||
if rule == "" || a.root == nil {
|
||||
return nil
|
||||
}
|
||||
isCss := false
|
||||
elementsRule := rule
|
||||
if hasPrefixFold(rule, "@CSS:") {
|
||||
isCss = true
|
||||
elementsRule = strings.TrimSpace(rule[5:])
|
||||
}
|
||||
ra := NewRuleAnalyzer(elementsRule, false)
|
||||
ruleStrS := ra.SplitRule("&&", "||", "%%")
|
||||
|
||||
var elementsList [][]*html.Node
|
||||
if isCss {
|
||||
for _, ruleStr := range ruleStrS {
|
||||
tempS := selectCSS(a.root, ruleStr)
|
||||
elementsList = append(elementsList, tempS)
|
||||
if len(tempS) > 0 && ra.ElementsType() == "||" {
|
||||
break
|
||||
}
|
||||
}
|
||||
} else {
|
||||
for _, ruleStr := range ruleStrS {
|
||||
rsRule := NewRuleAnalyzer(ruleStr, false)
|
||||
rsRule.Trim()
|
||||
rs := rsRule.SplitRule("@")
|
||||
var el []*html.Node
|
||||
if len(rs) > 1 {
|
||||
el = []*html.Node{a.root}
|
||||
for _, rl := range rs {
|
||||
var es []*html.Node
|
||||
for _, et := range el {
|
||||
es = append(es, a.getElementsOf(et, rl)...)
|
||||
}
|
||||
el = es
|
||||
}
|
||||
} else {
|
||||
es := newElementsSingle().getElementsSingle(a.root, ruleStr)
|
||||
el = es
|
||||
}
|
||||
elementsList = append(elementsList, el)
|
||||
if len(el) > 0 && ra.ElementsType() == "||" {
|
||||
break
|
||||
}
|
||||
}
|
||||
}
|
||||
var elements []*html.Node
|
||||
if len(elementsList) > 0 {
|
||||
if ra.ElementsType() == "%%" {
|
||||
for i := 0; i < len(elementsList[0]); i++ {
|
||||
for _, es := range elementsList {
|
||||
if i < len(es) {
|
||||
elements = append(elements, es[i])
|
||||
}
|
||||
}
|
||||
}
|
||||
} else {
|
||||
for _, es := range elementsList {
|
||||
elements = append(elements, es...)
|
||||
}
|
||||
}
|
||||
}
|
||||
return elements
|
||||
}
|
||||
|
||||
// getElementsOf 对应私有 getElements(temp, rule) 递归。
|
||||
func (a *jsoupAnalyzer) getElementsOf(temp *html.Node, rule string) []*html.Node {
|
||||
if temp == nil || rule == "" {
|
||||
return nil
|
||||
}
|
||||
isCss := false
|
||||
elementsRule := rule
|
||||
if hasPrefixFold(rule, "@CSS:") {
|
||||
isCss = true
|
||||
elementsRule = strings.TrimSpace(rule[5:])
|
||||
}
|
||||
ra := NewRuleAnalyzer(elementsRule, false)
|
||||
ruleStrS := ra.SplitRule("&&", "||", "%%")
|
||||
|
||||
var elementsList [][]*html.Node
|
||||
if isCss {
|
||||
for _, ruleStr := range ruleStrS {
|
||||
tempS := selectCSS(temp, ruleStr)
|
||||
elementsList = append(elementsList, tempS)
|
||||
if len(tempS) > 0 && ra.ElementsType() == "||" {
|
||||
break
|
||||
}
|
||||
}
|
||||
} else {
|
||||
for _, ruleStr := range ruleStrS {
|
||||
rsRule := NewRuleAnalyzer(ruleStr, false)
|
||||
rsRule.Trim()
|
||||
rs := rsRule.SplitRule("@")
|
||||
var el []*html.Node
|
||||
if len(rs) > 1 {
|
||||
el = []*html.Node{temp}
|
||||
for _, rl := range rs {
|
||||
var es []*html.Node
|
||||
for _, et := range el {
|
||||
es = append(es, a.getElementsOf(et, rl)...)
|
||||
}
|
||||
el = es
|
||||
}
|
||||
} else {
|
||||
el = newElementsSingle().getElementsSingle(temp, ruleStr)
|
||||
}
|
||||
elementsList = append(elementsList, el)
|
||||
if len(el) > 0 && ra.ElementsType() == "||" {
|
||||
break
|
||||
}
|
||||
}
|
||||
}
|
||||
var elements []*html.Node
|
||||
if len(elementsList) > 0 {
|
||||
if ra.ElementsType() == "%%" {
|
||||
for i := 0; i < len(elementsList[0]); i++ {
|
||||
for _, es := range elementsList {
|
||||
if i < len(es) {
|
||||
elements = append(elements, es[i])
|
||||
}
|
||||
}
|
||||
}
|
||||
} else {
|
||||
for _, es := range elementsList {
|
||||
elements = append(elements, es...)
|
||||
}
|
||||
}
|
||||
}
|
||||
return elements
|
||||
}
|
||||
|
||||
// getResultList 对应 getResultList:按 "@" 步进,最后一段作为提取规则。
|
||||
func (a *jsoupAnalyzer) getResultList(ruleStr string) []string {
|
||||
if ruleStr == "" {
|
||||
return nil
|
||||
}
|
||||
elements := []*html.Node{a.root}
|
||||
rule := NewRuleAnalyzer(ruleStr, false)
|
||||
rule.Trim()
|
||||
rules := rule.SplitRule("@")
|
||||
last := len(rules) - 1
|
||||
for i := 0; i < last; i++ {
|
||||
var es []*html.Node
|
||||
for _, elt := range elements {
|
||||
es = append(es, newElementsSingle().getElementsSingle(elt, rules[i])...)
|
||||
}
|
||||
elements = es
|
||||
}
|
||||
if len(elements) == 0 {
|
||||
return nil
|
||||
}
|
||||
return getResultLast(elements, rules[last])
|
||||
}
|
||||
|
||||
// getResultLast 对应 getResultLast:按最后一个规则提取内容。
|
||||
func getResultLast(elements []*html.Node, lastRule string) []string {
|
||||
var textS []string
|
||||
switch lastRule {
|
||||
case "text":
|
||||
for _, element := range elements {
|
||||
if text := nodeText(element); text != "" {
|
||||
textS = append(textS, text)
|
||||
}
|
||||
}
|
||||
case "textNodes":
|
||||
for _, element := range elements {
|
||||
tn := nodeTextNodes(element)
|
||||
if len(tn) > 0 {
|
||||
textS = append(textS, strings.Join(tn, "\n"))
|
||||
}
|
||||
}
|
||||
case "ownText":
|
||||
for _, element := range elements {
|
||||
if text := nodeOwnText(element); text != "" {
|
||||
textS = append(textS, text)
|
||||
}
|
||||
}
|
||||
case "html":
|
||||
for _, element := range elements {
|
||||
if h := outerHTMLNoScript(element); h != "" {
|
||||
textS = append(textS, h)
|
||||
}
|
||||
}
|
||||
case "all":
|
||||
var sb strings.Builder
|
||||
for _, element := range elements {
|
||||
sb.WriteString(outerHTML(element))
|
||||
}
|
||||
textS = append(textS, sb.String())
|
||||
default:
|
||||
for _, element := range elements {
|
||||
url := attrValue(element, lastRule)
|
||||
if url == "" || containsStr(textS, url) {
|
||||
continue
|
||||
}
|
||||
textS = append(textS, url)
|
||||
}
|
||||
}
|
||||
return textS
|
||||
}
|
||||
|
||||
func containsStr(list []string, s string) bool {
|
||||
for _, v := range list {
|
||||
if v == s {
|
||||
return true
|
||||
}
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
// ─── ElementsSingle:索引语法(对应 data class ElementsSingle) ─────────────
|
||||
|
||||
type indexRange struct {
|
||||
start *int
|
||||
end *int
|
||||
step int
|
||||
}
|
||||
|
||||
type elementsSingle struct {
|
||||
split byte // '.' 选择 / '!' 排除 / ' ' 无索引
|
||||
beforeRule string
|
||||
indexDefault []int
|
||||
indexes []any // int 或 indexRange
|
||||
}
|
||||
|
||||
func newElementsSingle() *elementsSingle {
|
||||
return &elementsSingle{split: '.'}
|
||||
}
|
||||
|
||||
func (e *elementsSingle) getElementsSingle(temp *html.Node, rule string) []*html.Node {
|
||||
e.findIndexSet(rule)
|
||||
|
||||
var elements []*html.Node
|
||||
if e.beforeRule == "" {
|
||||
elements = childrenElements(temp)
|
||||
} else {
|
||||
rules := strings.Split(e.beforeRule, ".")
|
||||
arg := ""
|
||||
if len(rules) > 1 {
|
||||
arg = rules[1]
|
||||
}
|
||||
switch rules[0] {
|
||||
case "children":
|
||||
elements = childrenElements(temp)
|
||||
case "class":
|
||||
if arg != "" {
|
||||
elements = getElementsByClass(temp, arg)
|
||||
}
|
||||
case "tag":
|
||||
if arg != "" {
|
||||
elements = getElementsByTag(temp, arg)
|
||||
}
|
||||
case "id":
|
||||
if arg != "" {
|
||||
elements = getElementsById(temp, arg)
|
||||
}
|
||||
case "text":
|
||||
if arg != "" {
|
||||
elements = getElementsContainingOwnText(temp, arg)
|
||||
}
|
||||
default:
|
||||
elements = selectCSS(temp, e.beforeRule)
|
||||
}
|
||||
}
|
||||
|
||||
// 索引集合:slice+set 模拟 Kotlin LinkedHashSet 的插入顺序去重
|
||||
indexSet := make([]int, 0, len(elements))
|
||||
seen := make(map[int]bool)
|
||||
addIndex := func(ix int) {
|
||||
if !seen[ix] {
|
||||
seen[ix] = true
|
||||
indexSet = append(indexSet, ix)
|
||||
}
|
||||
}
|
||||
|
||||
lenn := len(elements)
|
||||
lastIndexes := -1
|
||||
if len(e.indexDefault) > 0 {
|
||||
lastIndexes = len(e.indexDefault) - 1
|
||||
} else if len(e.indexes) > 0 {
|
||||
lastIndexes = len(e.indexes) - 1
|
||||
}
|
||||
|
||||
if len(e.indexes) == 0 {
|
||||
// 旧式索引:逆向遍历还原顺序
|
||||
for ix := lastIndexes; ix >= 0; ix-- {
|
||||
it := e.indexDefault[ix]
|
||||
if it >= 0 && it < lenn {
|
||||
addIndex(it)
|
||||
} else if it < 0 && lenn >= -it {
|
||||
addIndex(it + lenn)
|
||||
}
|
||||
}
|
||||
} else {
|
||||
for ix := lastIndexes; ix >= 0; ix-- {
|
||||
if rg, ok := e.indexes[ix].(indexRange); ok {
|
||||
start := 0
|
||||
if rg.start != nil {
|
||||
start = *rg.start
|
||||
}
|
||||
if start < 0 {
|
||||
start += lenn
|
||||
}
|
||||
end := lenn - 1
|
||||
if rg.end != nil {
|
||||
end = *rg.end
|
||||
}
|
||||
if end < 0 {
|
||||
end += lenn
|
||||
}
|
||||
if (start < 0 && end < 0) || (start >= lenn && end >= lenn) {
|
||||
continue
|
||||
}
|
||||
if start >= lenn {
|
||||
start = lenn - 1
|
||||
} else if start < 0 {
|
||||
start = 0
|
||||
}
|
||||
if end >= lenn {
|
||||
end = lenn - 1
|
||||
} else if end < 0 {
|
||||
end = 0
|
||||
}
|
||||
if start == end || rg.step >= lenn {
|
||||
addIndex(start)
|
||||
continue
|
||||
}
|
||||
step := rg.step
|
||||
if step > 0 {
|
||||
// 正向步长原样使用
|
||||
} else if -step < lenn {
|
||||
step = step + lenn
|
||||
} else {
|
||||
step = 1
|
||||
}
|
||||
if end > start {
|
||||
for i := start; i <= end; i += step {
|
||||
addIndex(i)
|
||||
}
|
||||
} else {
|
||||
for i := start; i >= end; i -= step {
|
||||
addIndex(i)
|
||||
}
|
||||
}
|
||||
} else {
|
||||
it := e.indexes[ix].(int)
|
||||
if it >= 0 && it < lenn {
|
||||
addIndex(it)
|
||||
} else if it < 0 && lenn >= -it {
|
||||
addIndex(it + lenn)
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if e.split == '!' {
|
||||
exclude := make(map[int]bool)
|
||||
for _, ix := range indexSet {
|
||||
exclude[ix] = true
|
||||
}
|
||||
var es []*html.Node
|
||||
for i, el := range elements {
|
||||
if !exclude[i] {
|
||||
es = append(es, el)
|
||||
}
|
||||
}
|
||||
elements = es
|
||||
} else if e.split == '.' {
|
||||
var es []*html.Node
|
||||
for _, ix := range indexSet {
|
||||
if ix >= 0 && ix < lenn {
|
||||
es = append(es, elements[ix])
|
||||
}
|
||||
}
|
||||
elements = es
|
||||
}
|
||||
return elements
|
||||
}
|
||||
|
||||
// findIndexSet 对应 ElementsSingle.findIndexSet:从右向左解析索引。
|
||||
// 支持旧式 tag.div.-1:10:2 / tag.div!0:3 与新式 tag.div[!-1, 3:-2:-10, 2]。
|
||||
func (e *elementsSingle) findIndexSet(rule string) {
|
||||
rus := []rune(strings.TrimSpace(rule))
|
||||
n := len(rus)
|
||||
if n == 0 {
|
||||
e.split = ' '
|
||||
e.beforeRule = ""
|
||||
return
|
||||
}
|
||||
var curList []*int // 区间临时列表(逆向压入:右端、左端、间隔)
|
||||
l := "" // 暂存数字字符串
|
||||
curMinus := false
|
||||
|
||||
head := rus[n-1] == ']'
|
||||
length := n
|
||||
if head {
|
||||
length-- // 跳过尾部 ']'
|
||||
}
|
||||
findLoop:
|
||||
for length >= 0 {
|
||||
length--
|
||||
if length < 0 {
|
||||
break
|
||||
}
|
||||
rl := rus[length]
|
||||
if rl == ' ' {
|
||||
continue
|
||||
}
|
||||
if rl >= '0' && rl <= '9' {
|
||||
l = string(rl) + l
|
||||
continue
|
||||
}
|
||||
if rl == '-' {
|
||||
curMinus = true
|
||||
continue
|
||||
}
|
||||
var curInt *int
|
||||
if l != "" {
|
||||
v := parseIntSafe(l)
|
||||
if curMinus {
|
||||
v = -v
|
||||
}
|
||||
curInt = &v
|
||||
}
|
||||
if head {
|
||||
switch rl {
|
||||
case ':':
|
||||
curList = append(curList, curInt)
|
||||
default:
|
||||
if len(curList) == 0 {
|
||||
if curInt == nil {
|
||||
break findLoop // 是 jsoup 选择器而非索引列表
|
||||
}
|
||||
e.indexes = append(e.indexes, *curInt)
|
||||
} else {
|
||||
rg := indexRange{start: curInt, end: curList[len(curList)-1], step: 1}
|
||||
if len(curList) == 2 && curList[0] != nil {
|
||||
rg.step = *curList[0]
|
||||
}
|
||||
e.indexes = append(e.indexes, rg)
|
||||
curList = curList[:0]
|
||||
}
|
||||
if rl == '!' {
|
||||
e.split = '!'
|
||||
for {
|
||||
length--
|
||||
if length < 0 {
|
||||
break
|
||||
}
|
||||
rl = rus[length]
|
||||
if rl != ' ' {
|
||||
break
|
||||
}
|
||||
}
|
||||
}
|
||||
if rl == '[' {
|
||||
if length < 0 {
|
||||
length = 0
|
||||
}
|
||||
e.beforeRule = string(rus[:length])
|
||||
return
|
||||
}
|
||||
if rl != ',' {
|
||||
break findLoop
|
||||
}
|
||||
}
|
||||
} else {
|
||||
if rl == '!' || rl == '.' || rl == ':' {
|
||||
v := 0
|
||||
if curInt != nil {
|
||||
v = *curInt
|
||||
}
|
||||
e.indexDefault = append(e.indexDefault, v)
|
||||
if rl != ':' {
|
||||
e.split = byte(rl)
|
||||
e.beforeRule = string(rus[:length])
|
||||
return
|
||||
}
|
||||
} else {
|
||||
break findLoop
|
||||
}
|
||||
}
|
||||
l = ""
|
||||
curMinus = false
|
||||
}
|
||||
e.split = ' '
|
||||
e.beforeRule = string(rus)
|
||||
}
|
||||
|
||||
func parseIntSafe(s string) int {
|
||||
n := 0
|
||||
for _, c := range s {
|
||||
if c < '0' || c > '9' {
|
||||
return 0
|
||||
}
|
||||
n = n*10 + int(c-'0')
|
||||
}
|
||||
return n
|
||||
}
|
||||
@@ -0,0 +1,89 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"net/url"
|
||||
"strings"
|
||||
|
||||
"github.com/dop251/goja"
|
||||
)
|
||||
|
||||
// newJSURLObject 构造 legado JsURL 的等效对象(java.toURL)。
|
||||
//
|
||||
// 对应 io.legado.app.utils.JsURL:它只暴露四个属性 ——
|
||||
//
|
||||
// host 主机名
|
||||
// origin "scheme://host[:port]"(默认端口不写)
|
||||
// pathname 路径(原样,未解码)
|
||||
// searchParams 查询参数(值已 URL 解码;同名只留最后一个,与 legado 一致)
|
||||
//
|
||||
// legado 返回的是 Java Map,Rhino 下书源会写 searchParams.get("x"),
|
||||
// 而 goja 里更自然的是 searchParams.x,所以两者都支持。
|
||||
func newJSURLObject(vm *goja.Runtime, raw, base string) (*goja.Object, error) {
|
||||
u, err := parseJSURL(raw, base)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
obj := vm.NewObject()
|
||||
origin := u.Scheme + "://" + u.Hostname()
|
||||
if port := u.Port(); port != "" {
|
||||
origin += ":" + port
|
||||
}
|
||||
_ = obj.Set("host", u.Hostname())
|
||||
_ = obj.Set("origin", origin)
|
||||
_ = obj.Set("pathname", u.EscapedPath())
|
||||
|
||||
params := vm.NewObject()
|
||||
for key, values := range u.Query() {
|
||||
if len(values) == 0 {
|
||||
continue
|
||||
}
|
||||
_ = params.Set(key, values[len(values)-1])
|
||||
}
|
||||
// searchParams.get(name):兼容 Java Map 的取法;不存在时给 null(对应 Java 的 null)。
|
||||
get := func(call goja.FunctionCall) goja.Value {
|
||||
if len(call.Arguments) == 0 {
|
||||
return goja.Null()
|
||||
}
|
||||
v := params.Get(call.Arguments[0].String())
|
||||
if v == nil || goja.IsUndefined(v) {
|
||||
return goja.Null()
|
||||
}
|
||||
return v
|
||||
}
|
||||
_ = params.Set("get", get)
|
||||
_ = obj.Set("searchParams", params)
|
||||
return obj, nil
|
||||
}
|
||||
|
||||
// parseJSURL 解析 java.toURL 的入参。base 非空时按基础地址解析相对路径
|
||||
// (对应 Java 的 URL(base, url));解析不出绝对地址时报错(Java 会抛 MalformedURLException)。
|
||||
func parseJSURL(raw, base string) (*url.URL, error) {
|
||||
raw = strings.TrimSpace(raw)
|
||||
if raw == "" {
|
||||
return nil, fmt.Errorf("toURL: 地址为空")
|
||||
}
|
||||
if base != "" {
|
||||
b, err := url.Parse(strings.TrimSpace(base))
|
||||
if err != nil || !b.IsAbs() {
|
||||
return nil, fmt.Errorf("toURL: baseUrl 不是绝对地址: %s", base)
|
||||
}
|
||||
rel, err := url.Parse(raw)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("toURL: %v", err)
|
||||
}
|
||||
u := b.ResolveReference(rel)
|
||||
if !u.IsAbs() || u.Host == "" {
|
||||
return nil, fmt.Errorf("toURL: 解析结果不是绝对地址: %s", raw)
|
||||
}
|
||||
return u, nil
|
||||
}
|
||||
u, err := url.Parse(raw)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("toURL: %v", err)
|
||||
}
|
||||
if u.Scheme == "" || u.Host == "" {
|
||||
return nil, fmt.Errorf("toURL: 不是绝对地址: %s", raw)
|
||||
}
|
||||
return u, nil
|
||||
}
|
||||
@@ -0,0 +1,92 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"strings"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// TestJavaValueStringMatchesJavaMapToString 书源正则依赖 Java 的 Map.toString 形态。
|
||||
// 真实案例:拷贝漫画的 author 规则 `$.author##.*name=(.*?)\,.*##$1`,
|
||||
// 期望拿到 `{name=岸本斉史, alias=…}`;Go 的 `map[...]` 形态匹配不到,作者会退化成一串 map 文本。
|
||||
func TestJavaValueStringMatchesJavaMapToString(t *testing.T) {
|
||||
ar := NewAnalyzeRule()
|
||||
ar.SetContent(`{"author":[{"name":"岸本斉史","alias":"岸本齐史,キシモトマサ","path_word":"anbenqishi"}]}`, "https://api.example.com")
|
||||
|
||||
got, err := ar.GetString(`$.author##.*name=(.*?)\,.*##$1`, nil, false)
|
||||
if err != nil {
|
||||
t.Fatalf("err=%v", err)
|
||||
}
|
||||
if got != "岸本斉史" {
|
||||
t.Fatalf("author = %q,期望 岸本斉史(Java Map 形态没对上)", got)
|
||||
}
|
||||
|
||||
// 对象/数组/标量的字符串形态
|
||||
str, err := ar.GetString(`$.author[0]`, nil, false)
|
||||
if err != nil {
|
||||
t.Fatalf("err=%v", err)
|
||||
}
|
||||
for _, want := range []string{"{", "name=岸本斉史", "alias=岸本齐史,キシモトマサ", "path_word=anbenqishi"} {
|
||||
if !strings.Contains(str, want) {
|
||||
t.Errorf("对象字符串 %q 缺少 %q", str, want)
|
||||
}
|
||||
}
|
||||
if strings.Contains(str, "map[") {
|
||||
t.Errorf("仍是 Go 的 map 形态: %q", str)
|
||||
}
|
||||
}
|
||||
|
||||
// TestJavaValueStringScalars 标量与嵌套结构按 Java 习惯输出。
|
||||
func TestJavaValueStringScalars(t *testing.T) {
|
||||
cases := []struct {
|
||||
in any
|
||||
want string
|
||||
}{
|
||||
{nil, "null"},
|
||||
{"x", "x"},
|
||||
{true, "true"},
|
||||
{float64(399), "399"},
|
||||
{float64(1.5), "1.5"},
|
||||
{[]any{float64(1), "a"}, "[1, a]"},
|
||||
{map[string]any{"b": float64(1), "a": "x"}, "{a=x, b=1}"},
|
||||
{map[string]any{"n": []any{"x", "y"}}, "{n=[x, y]}"},
|
||||
}
|
||||
for _, c := range cases {
|
||||
if got := javaValueString(c.in); got != c.want {
|
||||
t.Errorf("javaValueString(%#v) = %q,期望 %q", c.in, got, c.want)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestJavaToURL legado JsURL 的四个属性(并兼容 Java Map 的 searchParams.get)。
|
||||
func TestJavaToURL(t *testing.T) {
|
||||
r := NewJSRunner(JSConfig{})
|
||||
ar := NewAnalyzeRule()
|
||||
|
||||
check := func(expr, want string) {
|
||||
t.Helper()
|
||||
v, err := r.Run(ar, expr, nil, "")
|
||||
if err != nil {
|
||||
t.Fatalf("%s 失败: %v", expr, err)
|
||||
}
|
||||
if got := anyToString(v); got != want {
|
||||
t.Errorf("%s = %q,期望 %q", expr, got, want)
|
||||
}
|
||||
}
|
||||
|
||||
check(`java.toURL('https://api.example.com:8443/a/b?x=1&y=%E4%B8%AD%E6%96%87').host`, "api.example.com")
|
||||
check(`java.toURL('https://api.example.com:8443/a/b?x=1&y=%E4%B8%AD%E6%96%87').origin`, "https://api.example.com:8443")
|
||||
check(`java.toURL('https://api.example.com/a/b').origin`, "https://api.example.com")
|
||||
check(`java.toURL('https://api.example.com/a/b?x=1').pathname`, "/a/b")
|
||||
check(`java.toURL('https://api.example.com/a/b?x=1&y=%E4%B8%AD%E6%96%87').searchParams.x`, "1")
|
||||
check(`java.toURL('https://api.example.com/a/b?x=1&y=%E4%B8%AD%E6%96%87').searchParams.y`, "中文")
|
||||
check(`java.toURL('https://api.example.com/a/b?x=1').searchParams.get('x')`, "1")
|
||||
// 相对地址按 baseUrl 解析(对应 Java 的 URL(base, url))
|
||||
check(`java.toURL('/next?p=2', 'https://api.example.com/a/b').origin`, "https://api.example.com")
|
||||
check(`java.toURL('/next?p=2', 'https://api.example.com/a/b').pathname`, "/next")
|
||||
// 不存在的参数给 null(对应 Java Map.get 返回 null)
|
||||
check(`String(java.toURL('https://api.example.com/a').searchParams.get('nope'))`, "null")
|
||||
// 非绝对地址必须报错(Java 抛 MalformedURLException),不能静默给一个空对象
|
||||
if _, err := r.Run(ar, `java.toURL('not-a-url').host`, nil, ""); err == nil {
|
||||
t.Error("非绝对地址应当报错")
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,180 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"net/url"
|
||||
"strings"
|
||||
|
||||
"golang.org/x/text/encoding/htmlindex"
|
||||
"golang.org/x/text/encoding/unicode"
|
||||
)
|
||||
|
||||
// 本文件对应 legado 的 NetworkUtils.kt / EncoderUtils.kt 中与规则引擎相关的函数。
|
||||
|
||||
// GetAbsoluteURL 对应 NetworkUtils.getAbsoluteURL(baseURL: String?, relativePath)。
|
||||
// baseURL 会先截掉 ",{...}" 选项段(substringBefore(","))。
|
||||
func GetAbsoluteURL(baseURL, relativePath string) string {
|
||||
rel := strings.TrimSpace(relativePath)
|
||||
if isAbsURL(rel) || isDataURL(rel) || strings.HasPrefix(rel, "javascript") {
|
||||
if strings.HasPrefix(rel, "javascript") {
|
||||
return ""
|
||||
}
|
||||
return rel
|
||||
}
|
||||
if baseURL == "" || isDataURL(baseURL) {
|
||||
return rel
|
||||
}
|
||||
base := baseURL
|
||||
if i := strings.Index(base, ","); i >= 0 {
|
||||
base = base[:i]
|
||||
}
|
||||
baseURLParsed, err := url.Parse(strings.TrimSpace(base))
|
||||
if err != nil {
|
||||
return rel
|
||||
}
|
||||
return GetAbsoluteURLParsed(baseURLParsed, rel)
|
||||
}
|
||||
|
||||
// GetAbsoluteURLParsed 对应 NetworkUtils.getAbsoluteURL(baseURL: URL?, relativePath)。
|
||||
func GetAbsoluteURLParsed(base *url.URL, relativePath string) string {
|
||||
rel := strings.TrimSpace(relativePath)
|
||||
if base == nil {
|
||||
return rel
|
||||
}
|
||||
if isAbsURL(rel) || isDataURL(rel) {
|
||||
return rel
|
||||
}
|
||||
if strings.HasPrefix(rel, "javascript") {
|
||||
return ""
|
||||
}
|
||||
ref, err := url.Parse(rel)
|
||||
if err != nil {
|
||||
return rel
|
||||
}
|
||||
return base.ResolveReference(ref).String()
|
||||
}
|
||||
|
||||
// GetBaseUrl 对应 NetworkUtils.getBaseUrl:scheme://host[:port]。
|
||||
func GetBaseUrl(u string) string {
|
||||
if len(u) >= 8 && (strings.EqualFold(u[:7], "http://") || (len(u) >= 9 && strings.EqualFold(u[:8], "https://"))) {
|
||||
if idx := strings.Index(u[8:], "/"); idx >= 0 {
|
||||
return u[:8+idx]
|
||||
}
|
||||
return u
|
||||
}
|
||||
return ""
|
||||
}
|
||||
|
||||
// isAbsURL 对应 String.isAbsUrl()。
|
||||
func isAbsURL(s string) bool {
|
||||
return strings.HasPrefix(s, "https://") || strings.HasPrefix(s, "http://")
|
||||
}
|
||||
|
||||
func isDataURL(s string) bool {
|
||||
return strings.HasPrefix(s, "data:")
|
||||
}
|
||||
|
||||
// notNeedEncodingQuery / notNeedEncodingForm 对应 NetworkUtils 的两个 BitSet。
|
||||
var (
|
||||
notNeedEncodingQuery = buildEncodingSet("!$&()*+,-./:;=?@[\\]^_`{|}~")
|
||||
notNeedEncodingForm = buildEncodingSet("*-._")
|
||||
)
|
||||
|
||||
func buildEncodingSet(extra string) map[rune]bool {
|
||||
set := make(map[rune]bool, 128)
|
||||
for r := 'a'; r <= 'z'; r++ {
|
||||
set[r] = true
|
||||
}
|
||||
for r := 'A'; r <= 'Z'; r++ {
|
||||
set[r] = true
|
||||
}
|
||||
for r := '0'; r <= '9'; r++ {
|
||||
set[r] = true
|
||||
}
|
||||
for _, r := range extra {
|
||||
set[r] = true
|
||||
}
|
||||
return set
|
||||
}
|
||||
|
||||
func isDigit16(r byte) bool {
|
||||
return (r >= '0' && r <= '9') || (r >= 'a' && r <= 'f') || (r >= 'A' && r <= 'F')
|
||||
}
|
||||
|
||||
// encodedQuery 对应 NetworkUtils.encodedQuery(str):判断字符串是否已按
|
||||
// urlEncode 规范编码(无需再编码返回 true)。
|
||||
func encodedQuery(s string) bool {
|
||||
return encodedWith(s, notNeedEncodingQuery)
|
||||
}
|
||||
|
||||
// encodedForm 对应 NetworkUtils.encodedForm(str)。
|
||||
func encodedForm(s string) bool {
|
||||
return encodedWith(s, notNeedEncodingForm)
|
||||
}
|
||||
|
||||
func encodedWith(s string, allow map[rune]bool) bool {
|
||||
rs := []rune(s)
|
||||
for i := 0; i < len(rs); i++ {
|
||||
r := rs[i]
|
||||
if r < 128 && allow[r] {
|
||||
continue
|
||||
}
|
||||
if r == '%' && i+2 < len(rs) && isDigit16(byte(rs[i+1])) && isDigit16(byte(rs[i+2])) {
|
||||
i += 2
|
||||
continue
|
||||
}
|
||||
return false
|
||||
}
|
||||
return len(rs) > 0
|
||||
}
|
||||
|
||||
// JSEscape 对应 EncoderUtils.escape(JS escape() 语义)。
|
||||
func JSEscape(src string) string {
|
||||
var sb strings.Builder
|
||||
for _, r := range src {
|
||||
code := int(r)
|
||||
if (code >= 48 && code <= 57) || (code >= 65 && code <= 90) || (code >= 97 && code <= 122) {
|
||||
sb.WriteRune(r)
|
||||
continue
|
||||
}
|
||||
switch {
|
||||
case code < 16:
|
||||
fmt.Fprintf(&sb, "%%0%X", code)
|
||||
case code < 256:
|
||||
fmt.Fprintf(&sb, "%%%X", code)
|
||||
default:
|
||||
fmt.Fprintf(&sb, "%%u%04X", code)
|
||||
}
|
||||
}
|
||||
return sb.String()
|
||||
}
|
||||
|
||||
// DecodeBytes 按字符集名解码字节流,name 为空时按 UTF-8(带 BOM 处理)。
|
||||
func DecodeBytes(b []byte, name string) (string, error) {
|
||||
if len(b) >= 3 && b[0] == 0xEF && b[1] == 0xBB && b[2] == 0xBF {
|
||||
return string(b[3:]), nil
|
||||
}
|
||||
if name == "" || strings.EqualFold(name, "utf-8") || strings.EqualFold(name, "utf8") {
|
||||
dec := unicode.UTF8.NewDecoder()
|
||||
out, err := dec.Bytes(b)
|
||||
if err != nil {
|
||||
return string(b), nil
|
||||
}
|
||||
return string(out), nil
|
||||
}
|
||||
enc, err := htmlindex.Get(name)
|
||||
if err != nil {
|
||||
return string(b), nil
|
||||
}
|
||||
out, derr := enc.NewDecoder().Bytes(b)
|
||||
if derr != nil {
|
||||
return string(b), nil
|
||||
}
|
||||
return string(out), nil
|
||||
}
|
||||
|
||||
// LooksLikeJSON 对应 String.isJson()(宽松:trim 后以 { 或 [ 开头)。
|
||||
func LooksLikeJSON(s string) bool {
|
||||
s = strings.TrimSpace(s)
|
||||
return strings.HasPrefix(s, "{") || strings.HasPrefix(s, "[")
|
||||
}
|
||||
@@ -0,0 +1,176 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"strings"
|
||||
"time"
|
||||
|
||||
"github.com/dlclark/regexp2"
|
||||
)
|
||||
|
||||
// 本文件对应 AnalyzeByRegex.kt。Java 正则语义用 regexp2 对齐
|
||||
// (支持前向后向断言与反向引用),匹配循环对齐 Matcher.find()。
|
||||
|
||||
// splitNotBlankAndTrim 对应 String.splitNotBlank("&&"):切分并去空白项。
|
||||
func splitNotBlankAndTrim(s, sep string) []string {
|
||||
var out []string
|
||||
for _, p := range strings.Split(s, sep) {
|
||||
if t := strings.TrimSpace(p); t != "" {
|
||||
out = append(out, t)
|
||||
}
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// regexGetElement 对应 AnalyzeByRegex.getElement:多段正则串联,
|
||||
// 最终返回第一个匹配的全部分组(含 group 0)。
|
||||
func regexGetElement(res string, regs []string, index int) []string {
|
||||
if index >= len(regs) {
|
||||
return nil
|
||||
}
|
||||
re, err := regexp2.Compile(regs[index], regexp2.None)
|
||||
if err != nil {
|
||||
return nil
|
||||
}
|
||||
m, err := re.FindStringMatchStartingAt(res, 0)
|
||||
if err != nil || m == nil {
|
||||
return nil
|
||||
}
|
||||
if index+1 == len(regs) {
|
||||
info := make([]string, 0, len(m.Groups()))
|
||||
for _, g := range m.Groups() {
|
||||
if len(g.Captures) > 0 {
|
||||
info = append(info, g.Captures[0].String())
|
||||
} else {
|
||||
info = append(info, "")
|
||||
}
|
||||
}
|
||||
return info
|
||||
}
|
||||
var sb strings.Builder
|
||||
for m != nil {
|
||||
sb.WriteString(m.String())
|
||||
m, _ = re.FindNextMatch(m)
|
||||
}
|
||||
return regexGetElement(sb.String(), regs, index+1)
|
||||
}
|
||||
|
||||
// regexGetElements 对应 AnalyzeByRegex.getElements:多段正则串联,
|
||||
// 最终按每个匹配返回一组分组列表。
|
||||
func regexGetElements(res string, regs []string, index int) [][]string {
|
||||
if index >= len(regs) {
|
||||
return nil
|
||||
}
|
||||
re, err := regexp2.Compile(regs[index], regexp2.None)
|
||||
if err != nil {
|
||||
return nil
|
||||
}
|
||||
m, err := re.FindStringMatchStartingAt(res, 0)
|
||||
if err != nil || m == nil {
|
||||
return nil
|
||||
}
|
||||
if index+1 == len(regs) {
|
||||
var books [][]string
|
||||
for m != nil {
|
||||
info := make([]string, 0, len(m.Groups()))
|
||||
for _, g := range m.Groups() {
|
||||
if len(g.Captures) > 0 {
|
||||
info = append(info, g.Captures[0].String())
|
||||
} else {
|
||||
info = append(info, "")
|
||||
}
|
||||
}
|
||||
books = append(books, info)
|
||||
m, _ = re.FindNextMatch(m)
|
||||
}
|
||||
return books
|
||||
}
|
||||
var sb strings.Builder
|
||||
for m != nil {
|
||||
sb.WriteString(m.String())
|
||||
m, _ = re.FindNextMatch(m)
|
||||
}
|
||||
return regexGetElements(sb.String(), regs, index+1)
|
||||
}
|
||||
|
||||
// regexReplaceAll 对应 Kotlin Regex.replace(result, replacement)
|
||||
// (Java $N 分组替换语义,regexp2 的 Replace 原生支持)。
|
||||
func regexReplaceAll(pattern, result, replacement string) string {
|
||||
re, err := regexp2.Compile(pattern, regexp2.None)
|
||||
if err != nil {
|
||||
return strings.ReplaceAll(result, pattern, replacement)
|
||||
}
|
||||
out, err := re.Replace(result, replacement, 0, -1)
|
||||
if err != nil {
|
||||
return result
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// ApplyUserReplace 应用一条用户替换净化规则(对应 legado ReplaceRule)。
|
||||
// isRegex=false 按字面替换;isRegex=true 用 Java 正则语义并带匹配超时
|
||||
// (防灾难性回溯挂死服务),编译失败回退字面替换。
|
||||
func ApplyUserReplace(content, pattern, replacement string, isRegex bool, timeoutMS int64) string {
|
||||
if pattern == "" {
|
||||
return content
|
||||
}
|
||||
if !isRegex {
|
||||
return strings.ReplaceAll(content, pattern, replacement)
|
||||
}
|
||||
re, err := regexp2.Compile(pattern, regexp2.None)
|
||||
if err != nil {
|
||||
return strings.ReplaceAll(content, pattern, replacement)
|
||||
}
|
||||
if timeoutMS > 0 {
|
||||
re.MatchTimeout = time.Duration(timeoutMS) * time.Millisecond
|
||||
} else {
|
||||
re.MatchTimeout = 3 * time.Second
|
||||
}
|
||||
out, err := re.Replace(content, replacement, 0, -1)
|
||||
if err != nil {
|
||||
return content
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// ApplyReplaceRegexString 应用书源 replaceRegex 字符串("##pattern##replace[##x]" 格式),
|
||||
// 对应 ContentRule.replaceRegex 的处理。
|
||||
func ApplyReplaceRegexString(content, replaceRegex string) string {
|
||||
if replaceRegex == "" {
|
||||
return content
|
||||
}
|
||||
segs := strings.Split(replaceRegex, "##")
|
||||
if len(segs) < 2 {
|
||||
return content
|
||||
}
|
||||
pattern := segs[1]
|
||||
replacement := ""
|
||||
replaceFirst := false
|
||||
if len(segs) > 2 {
|
||||
replacement = segs[2]
|
||||
}
|
||||
if len(segs) > 3 {
|
||||
replaceFirst = true
|
||||
}
|
||||
if replaceFirst {
|
||||
return regexReplaceFirstOnFirstMatch(pattern, content, replacement)
|
||||
}
|
||||
return regexReplaceAll(pattern, content, replacement)
|
||||
}
|
||||
|
||||
// regexReplaceFirstOnFirstMatch 对应 replaceRegex 的 replaceFirst 分支:
|
||||
// 找到第一个匹配(无匹配返回 ""),在匹配文本上做首次替换。
|
||||
func regexReplaceFirstOnFirstMatch(pattern, result, replacement string) string {
|
||||
re, err := regexp2.Compile(pattern, regexp2.None)
|
||||
if err != nil {
|
||||
return replacement
|
||||
}
|
||||
m, err := re.FindStringMatch(result)
|
||||
if err != nil || m == nil {
|
||||
return ""
|
||||
}
|
||||
out, err := re.Replace(m.String(), replacement, 0, 1)
|
||||
if err != nil {
|
||||
return m.String()
|
||||
}
|
||||
return out
|
||||
}
|
||||
@@ -0,0 +1,341 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"strings"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// ─── RuleAnalyzer ──────────────────────────────────────────────────────────
|
||||
|
||||
func TestRuleAnalyzerSplitAndOr(t *testing.T) {
|
||||
ra := NewRuleAnalyzer("class.a&&tag.b&&id.c", false)
|
||||
rules := ra.SplitRule("&&", "||", "%%")
|
||||
if len(rules) != 3 || rules[0] != "class.a" || rules[1] != "tag.b" || rules[2] != "id.c" {
|
||||
t.Fatalf("unexpected split: %#v", rules)
|
||||
}
|
||||
if ra.ElementsType() != "&&" {
|
||||
t.Fatalf("elementsType = %q, want &&", ra.ElementsType())
|
||||
}
|
||||
// 与 Kotlin 一致:consumeToAny 取最左侧出现的分隔符,
|
||||
// 之后只按该分隔符切分(混合操作符时右侧保留原样,由上层递归处理)。
|
||||
ra2 := NewRuleAnalyzer("class.a&&tag.b||id.c", false)
|
||||
rules2 := ra2.SplitRule("&&", "||", "%%")
|
||||
if len(rules2) != 2 || rules2[0] != "class.a" || rules2[1] != "tag.b||id.c" {
|
||||
t.Fatalf("mixed split: %#v", rules2)
|
||||
}
|
||||
}
|
||||
|
||||
func TestRuleAnalyzerBalancedGroup(t *testing.T) {
|
||||
// && 在选择器平衡组内不应被切分
|
||||
ra := NewRuleAnalyzer(`tag.div[class="x&&y"]@text&&class.z`, false)
|
||||
rules := ra.SplitRule("&&", "||", "%%")
|
||||
if len(rules) != 2 {
|
||||
t.Fatalf("unexpected split: %#v", rules)
|
||||
}
|
||||
if rules[0] != `tag.div[class="x&&y"]@text` {
|
||||
t.Fatalf("rule[0] = %q", rules[0])
|
||||
}
|
||||
}
|
||||
|
||||
func TestRuleAnalyzerSplitByAt(t *testing.T) {
|
||||
ra := NewRuleAnalyzer("class.bookbox@h4@a@text", false)
|
||||
ra.Trim()
|
||||
rules := ra.SplitRule("@")
|
||||
if len(rules) != 4 {
|
||||
t.Fatalf("unexpected split: %#v", rules)
|
||||
}
|
||||
}
|
||||
|
||||
// ─── SourceRule 模式识别 ────────────────────────────────────────────────────
|
||||
|
||||
func TestSourceRuleModeDetection(t *testing.T) {
|
||||
cases := []struct {
|
||||
rule string
|
||||
contentIsJSON bool
|
||||
want Mode
|
||||
}{
|
||||
{"class.a@text", false, ModeDefault},
|
||||
{"$.data.name", false, ModeJson},
|
||||
{"$[0].name", false, ModeJson},
|
||||
{"//div[@class='a']/text()", false, ModeXPath},
|
||||
{"@XPath://div", false, ModeXPath},
|
||||
{"@Json:$.a", false, ModeJson},
|
||||
{"@CSS:.a@text", false, ModeDefault},
|
||||
{"title", true, ModeJson}, // 内容为 JSON 时默认走 Json 模式
|
||||
{"/html/body", false, ModeXPath},
|
||||
}
|
||||
for _, c := range cases {
|
||||
got := newSourceRule(c.rule, ModeDefault, c.contentIsJSON)
|
||||
if got.Mode != c.want {
|
||||
t.Errorf("rule %q mode = %v, want %v", c.rule, got.Mode, c.want)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
func TestSourceRuleSplitPut(t *testing.T) {
|
||||
sr := newSourceRule(`class.a@text@put:{"key1":"class.b@text"}`, ModeDefault, false)
|
||||
if sr.Rule != "class.a@text" {
|
||||
t.Fatalf("rule after put split = %q", sr.Rule)
|
||||
}
|
||||
if sr.putMap["key1"] != "class.b@text" {
|
||||
t.Fatalf("putMap = %#v", sr.putMap)
|
||||
}
|
||||
}
|
||||
|
||||
func TestMakeUpRuleGetVariable(t *testing.T) {
|
||||
sr := newSourceRule(`@get:{kw}`, ModeDefault, false)
|
||||
deps := &RuleDeps{Get: func(key string) string { return "搜索词" }}
|
||||
resolved, err := sr.MakeUpRule(nil, deps)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if resolved.Rule != "搜索词" {
|
||||
t.Fatalf("resolved = %q", resolved.Rule)
|
||||
}
|
||||
}
|
||||
|
||||
func TestSplitSourceRuleJSBlocks(t *testing.T) {
|
||||
rules := SplitSourceRule(`<js>1+1</js>class.a@text`, false, false, nil)
|
||||
if len(rules) != 2 || rules[0].Mode != ModeJs || rules[1].Mode != ModeDefault {
|
||||
t.Fatalf("rules = %#v", rules)
|
||||
}
|
||||
if rules[0].Rule != "1+1" {
|
||||
t.Fatalf("js body = %q", rules[0].Rule)
|
||||
}
|
||||
}
|
||||
|
||||
// ─── jsoup 分析器 ──────────────────────────────────────────────────────────
|
||||
|
||||
const testHTML = `<!DOCTYPE html>
|
||||
<html><body>
|
||||
<div class="box" id="main">
|
||||
<div class="item"><h3><a href="/book/1">斗破苍穹</a></h3><span class="author">天蚕土豆</span><p class="intro">测试<strong>简介</strong></p></div>
|
||||
<div class="item"><h3><a href="/book/2">凡人修仙传</a></h3><span class="author">忘语</span><p class="intro">凡人流</p></div>
|
||||
<div class="item"><h3><a href="/book/3">遮天</a></h3><span class="author">辰东</span><p class="intro">九龙拉棺</p></div>
|
||||
</div>
|
||||
</body></html>`
|
||||
|
||||
func TestJsoupGetElementsAndFields(t *testing.T) {
|
||||
a := newJsoupAnalyzer(testHTML)
|
||||
els := a.getElements("class.item")
|
||||
if len(els) != 3 {
|
||||
t.Fatalf("elements = %d, want 3", len(els))
|
||||
}
|
||||
name := a.getString("class.item.0@tag.h3@tag.a@text")
|
||||
if name != "斗破苍穹" {
|
||||
t.Fatalf("name = %q", name)
|
||||
}
|
||||
href := a.getString("class.item.0@tag.h3@tag.a@href")
|
||||
if href != "/book/1" {
|
||||
t.Fatalf("href = %q", href)
|
||||
}
|
||||
author := a.getString("class.item.1@class.author@text")
|
||||
if author != "忘语" {
|
||||
t.Fatalf("author = %q", author)
|
||||
}
|
||||
// all:拼接所有
|
||||
names := a.getStringList("class.item@tag.h3@tag.a@text")
|
||||
if len(names) != 3 || names[2] != "遮天" {
|
||||
t.Fatalf("names = %#v", names)
|
||||
}
|
||||
}
|
||||
|
||||
func TestJsoupTextNodesAndOwnText(t *testing.T) {
|
||||
a := newJsoupAnalyzer(testHTML)
|
||||
// ownText:不含子元素文本
|
||||
intro := a.getString("class.item.0@tag.p@ownText")
|
||||
if intro != "测试" {
|
||||
t.Fatalf("ownText = %q", intro)
|
||||
}
|
||||
// text:含子元素文本(jsoup 不在内联元素间补空格)
|
||||
full := a.getString("class.item.0@tag.p@text")
|
||||
if full != "测试简介" {
|
||||
t.Fatalf("text = %q", full)
|
||||
}
|
||||
}
|
||||
|
||||
func TestJsoupIndexSyntax(t *testing.T) {
|
||||
a := newJsoupAnalyzer(testHTML)
|
||||
// 负索引:最后一个
|
||||
last := a.getString("class.item.-1@tag.a@text")
|
||||
if last != "遮天" {
|
||||
t.Fatalf("last = %q", last)
|
||||
}
|
||||
// 新式区间索引
|
||||
firstTwo := a.getStringList("class.item[0:1]@tag.a@text")
|
||||
if len(firstTwo) != 2 || firstTwo[0] != "斗破苍穹" || firstTwo[1] != "凡人修仙传" {
|
||||
t.Fatalf("range = %#v", firstTwo)
|
||||
}
|
||||
}
|
||||
|
||||
func TestJsoupAndOrPercent(t *testing.T) {
|
||||
a := newJsoupAnalyzer(testHTML)
|
||||
// &&:合并两路结果
|
||||
merged := a.getStringList(`class.item.0@tag.a@text&&class.item.1@tag.a@text`)
|
||||
if len(merged) != 2 {
|
||||
t.Fatalf("&& merged = %#v", merged)
|
||||
}
|
||||
// ||:第一个非空即停
|
||||
or := a.getStringList(`id.notexist@text||class.item.0@tag.a@text`)
|
||||
if len(or) != 1 || or[0] != "斗破苍穹" {
|
||||
t.Fatalf("|| result = %#v", or)
|
||||
}
|
||||
}
|
||||
|
||||
func TestJsoupCSSMode(t *testing.T) {
|
||||
a := newJsoupAnalyzer(testHTML)
|
||||
// @CSS: 末段为提取规则(与 jsoup 分析器一致的 @ 分离)
|
||||
name := a.getString("@CSS:#main .item:nth-child(1) a@text")
|
||||
if name != "斗破苍穹" {
|
||||
t.Fatalf("@css name = %q", name)
|
||||
}
|
||||
href := a.getString("@CSS:.item:nth-child(2) a@href")
|
||||
if href != "/book/2" {
|
||||
t.Fatalf("@css href = %q", href)
|
||||
}
|
||||
}
|
||||
|
||||
// ─── JSONPath 分析器 ───────────────────────────────────────────────────────
|
||||
|
||||
const testJSON = `{"data":{"list":[{"title":"第一章","url":"/c/1"},{"title":"第二章","url":"/c/2"}],"name":"测试书","page":2}}`
|
||||
|
||||
func TestJSONPathGetString(t *testing.T) {
|
||||
a := newJSONAnalyzer(testJSON)
|
||||
if got := a.getString("$.data.name"); got != "测试书" {
|
||||
t.Fatalf("name = %q", got)
|
||||
}
|
||||
if got := a.getString("$.data.list[*].title"); got != "第一章\n第二章" {
|
||||
t.Fatalf("titles = %q", got)
|
||||
}
|
||||
// || 首个非空
|
||||
if got := a.getString("$.data.missing||$.data.name"); got != "测试书" {
|
||||
t.Fatalf("|| = %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
func TestJSONPathInnerRule(t *testing.T) {
|
||||
a := newJSONAnalyzer(testJSON)
|
||||
// {$.data.page} 内嵌规则替换
|
||||
got := a.getString("/api/list/{$.data.page}/next.json")
|
||||
if got != "/api/list/2/next.json" {
|
||||
t.Fatalf("inner = %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
func TestJSONPathListAndElementContext(t *testing.T) {
|
||||
a := newJSONAnalyzer(testJSON)
|
||||
list := a.getList("$.data.list[*]")
|
||||
if len(list) != 2 {
|
||||
t.Fatalf("list = %#v", list)
|
||||
}
|
||||
// 以列表元素为根继续求值(对应 getString(rule, element))
|
||||
sub := newJSONAnalyzer(list[0])
|
||||
if got := sub.getString("$.title"); got != "第一章" {
|
||||
t.Fatalf("element title = %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
// ─── XPath 分析器 ──────────────────────────────────────────────────────────
|
||||
|
||||
func TestXPathAnalyzer(t *testing.T) {
|
||||
a := newXPathAnalyzer(testHTML)
|
||||
if got := a.getString(`//div[@class="item"][1]//a/text()`); got != "斗破苍穹" {
|
||||
t.Fatalf("xpath = %q", got)
|
||||
}
|
||||
hrefs := a.getStringList(`//div[@class="item"]//a/@href`)
|
||||
if len(hrefs) != 3 {
|
||||
t.Fatalf("hrefs = %#v", hrefs)
|
||||
}
|
||||
els := a.getElements(`//div[@class="item"]`)
|
||||
if len(els) != 3 {
|
||||
t.Fatalf("elements = %d", len(els))
|
||||
}
|
||||
}
|
||||
|
||||
// ─── 正则分析器 ────────────────────────────────────────────────────────────
|
||||
|
||||
func TestRegexAnalyzer(t *testing.T) {
|
||||
content := "第1章 开始 第2章 继续 第3章 结束"
|
||||
els := regexGetElements(content, []string{`第(\d+)章 ([^ ]+)`}, 0)
|
||||
if len(els) != 3 || els[0][1] != "1" || els[2][2] != "结束" {
|
||||
t.Fatalf("regex elements = %#v", els)
|
||||
}
|
||||
}
|
||||
|
||||
func TestReplaceRegex(t *testing.T) {
|
||||
a := NewAnalyzeRule()
|
||||
a.SetContent(testHTML, "http://x.com")
|
||||
// ## 目标##替换
|
||||
got, err := a.GetString(`class.item.0@tag.a@text##斗破##破斗`, nil, false)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got != "破斗苍穹" {
|
||||
t.Fatalf("replace = %q", got)
|
||||
}
|
||||
}
|
||||
|
||||
// ─── AnalyzeUrl ────────────────────────────────────────────────────────────
|
||||
|
||||
func TestParseAnalyzeUrlBasic(t *testing.T) {
|
||||
req, err := ParseAnalyzeUrl("https://example.com/search/{{key}}/1.html", "斗罗", 1, "")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
// 与 legado 一致:路径段不做百分号编码(执行时由 HTTP 客户端转义)
|
||||
if !strings.Contains(req.URL, "/search/斗罗/1.html") {
|
||||
t.Fatalf("url = %q", req.URL)
|
||||
}
|
||||
}
|
||||
|
||||
func TestParseAnalyzeUrlPageList(t *testing.T) {
|
||||
// <1,20,40>:page=1 → 1;page=2 → 20;page=5 → 40(取最后一档)
|
||||
for page, want := range map[int]string{1: "1", 2: "20", 5: "40"} {
|
||||
req, err := ParseAnalyzeUrl("https://e.com/list/<1,20,40>.html", "k", page, "")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if !strings.Contains(req.URL, want+".html") {
|
||||
t.Fatalf("page=%d url = %q", page, req.URL)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
func TestParseAnalyzeUrlGBKQueryEncoding(t *testing.T) {
|
||||
req, err := ParseAnalyzeUrl("https://e.com/search.php?keyword={{key}},{\"charset\":\"gbk\"}", "斗罗", 1, "")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
// "斗罗" 的 GBK 编码 = B6 B7 C2 DE
|
||||
if !strings.Contains(req.URL, "%B6%B7%C2%DE") {
|
||||
t.Fatalf("gbk url = %q", req.URL)
|
||||
}
|
||||
}
|
||||
|
||||
func TestParseAnalyzeUrlPostForm(t *testing.T) {
|
||||
req, err := ParseAnalyzeUrl(`https://e.com/search,{"method":"POST","body":"searchkey={{key}}&submit=go"}`, "斗罗", 1, "")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if req.Method != "POST" {
|
||||
t.Fatalf("method = %s", req.Method)
|
||||
}
|
||||
if !req.IsForm {
|
||||
t.Fatalf("body should be form, got %q", req.Body)
|
||||
}
|
||||
if !strings.Contains(req.Body, "searchkey=") {
|
||||
t.Fatalf("body = %q", req.Body)
|
||||
}
|
||||
}
|
||||
|
||||
func TestParseAnalyzeUrlWebViewUnsupported(t *testing.T) {
|
||||
req, err := ParseAnalyzeUrl(`https://e.com/x,{"webView":true}`, "k", 1, "")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if req.Unsupported == nil {
|
||||
t.Fatal("expected unsupported for webView")
|
||||
}
|
||||
}
|
||||
|
||||
// ─── httptest 端到端:书源 JSON → 搜索解析全链路(见 reader 包 reader_test.go) ──
|
||||
@@ -0,0 +1,416 @@
|
||||
package rule
|
||||
|
||||
import "strings"
|
||||
|
||||
// RuleAnalyzer 移植自 legado RuleAnalyzer.kt,逐方法对齐。
|
||||
// 用于按 "@"/"&&"/"||"/"%%"/"@" 切分规则字符串,切分时跳过引号内与
|
||||
// 平衡组([...]、(...))内的分隔符,避免与选择器或正则内容冲突。
|
||||
|
||||
// RuleAnalyzer 是无状态的切分器实例(一次使用)。
|
||||
type RuleAnalyzer struct {
|
||||
queue string
|
||||
pos int
|
||||
start int
|
||||
startX int
|
||||
rule []string
|
||||
step int
|
||||
elementsType string
|
||||
code bool
|
||||
}
|
||||
|
||||
// NewRuleAnalyzer 对应 RuleAnalyzer(data, code);code=true 时平衡组按
|
||||
// 代码语义(处理转义、区分 [] 与 ())处理,用于 jsonPath 规则。
|
||||
func NewRuleAnalyzer(data string, code bool) *RuleAnalyzer {
|
||||
return &RuleAnalyzer{queue: data, code: code}
|
||||
}
|
||||
|
||||
// ElementsType 返回上次 splitRule 使用的组合符("&&"/"||"/"%%")。
|
||||
func (a *RuleAnalyzer) ElementsType() string { return a.elementsType }
|
||||
|
||||
// Rules 返回切分结果。
|
||||
func (a *RuleAnalyzer) Rules() []string { return a.rule }
|
||||
|
||||
// Trim 对应 trim():修剪当前规则之前的 "@" 或不可见字符。
|
||||
func (a *RuleAnalyzer) Trim() {
|
||||
if a.pos >= len(a.queue) {
|
||||
return
|
||||
}
|
||||
if a.queue[a.pos] == '@' || a.queue[a.pos] < '!' {
|
||||
a.pos++
|
||||
for a.pos < len(a.queue) && (a.queue[a.pos] == '@' || a.queue[a.pos] < '!') {
|
||||
a.pos++
|
||||
}
|
||||
a.start = a.pos
|
||||
a.startX = a.pos
|
||||
}
|
||||
}
|
||||
|
||||
// ReSetPos 对应 reSetPos()。
|
||||
func (a *RuleAnalyzer) ReSetPos() {
|
||||
a.pos = 0
|
||||
a.startX = 0
|
||||
}
|
||||
|
||||
// consumeTo 对应 consumeTo(seq)。
|
||||
func (a *RuleAnalyzer) consumeTo(seq string) bool {
|
||||
a.start = a.pos
|
||||
offset := strings.Index(a.queue[a.pos:], seq)
|
||||
if offset == -1 {
|
||||
return false
|
||||
}
|
||||
a.pos += offset
|
||||
return true
|
||||
}
|
||||
|
||||
// consumeToAny 对应 consumeToAny(seq)。
|
||||
func (a *RuleAnalyzer) consumeToAny(seqs ...string) bool {
|
||||
pos := a.pos
|
||||
for pos != len(a.queue) {
|
||||
for _, s := range seqs {
|
||||
if strings.HasPrefix(a.queue[pos:], s) {
|
||||
a.step = len(s)
|
||||
a.pos = pos
|
||||
return true
|
||||
}
|
||||
}
|
||||
pos++
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
// findToAny 对应 findToAny(seq: Char)。
|
||||
func (a *RuleAnalyzer) findToAny(chars ...byte) int {
|
||||
pos := a.pos
|
||||
for pos != len(a.queue) {
|
||||
for _, c := range chars {
|
||||
if a.queue[pos] == c {
|
||||
return pos
|
||||
}
|
||||
}
|
||||
pos++
|
||||
}
|
||||
return -1
|
||||
}
|
||||
|
||||
// chompCodeBalanced 对应 chompCodeBalanced(open, close):拉出一个平衡组,
|
||||
// 存在转义文本,'[' 与参数对 (open/close) 分别计数。
|
||||
func (a *RuleAnalyzer) chompCodeBalanced(open, close byte) bool {
|
||||
pos := a.pos
|
||||
depth := 0
|
||||
otherDepth := 0
|
||||
inSingle := false
|
||||
inDouble := false
|
||||
for {
|
||||
if pos == len(a.queue) {
|
||||
break
|
||||
}
|
||||
c := a.queue[pos]
|
||||
pos++
|
||||
if c != '\\' {
|
||||
if c == '\'' && !inDouble {
|
||||
inSingle = !inSingle
|
||||
} else if c == '"' && !inSingle {
|
||||
inDouble = !inDouble
|
||||
}
|
||||
if inSingle || inDouble {
|
||||
continue
|
||||
}
|
||||
if c == '[' {
|
||||
depth++
|
||||
} else if c == ']' {
|
||||
depth--
|
||||
} else if depth == 0 {
|
||||
if c == open {
|
||||
otherDepth++
|
||||
} else if c == close {
|
||||
otherDepth--
|
||||
}
|
||||
}
|
||||
} else {
|
||||
pos++
|
||||
}
|
||||
if !(depth > 0 || otherDepth > 0) {
|
||||
break
|
||||
}
|
||||
}
|
||||
if depth > 0 || otherDepth > 0 {
|
||||
return false
|
||||
}
|
||||
a.pos = pos
|
||||
return true
|
||||
}
|
||||
|
||||
// chompRuleBalanced 对应 chompRuleBalanced(open, close):引号外才处理转义。
|
||||
func (a *RuleAnalyzer) chompRuleBalanced(open, close byte) bool {
|
||||
pos := a.pos
|
||||
depth := 0
|
||||
inSingle := false
|
||||
inDouble := false
|
||||
for {
|
||||
if pos == len(a.queue) {
|
||||
break
|
||||
}
|
||||
c := a.queue[pos]
|
||||
pos++
|
||||
if c == '\'' && !inDouble {
|
||||
inSingle = !inSingle
|
||||
} else if c == '"' && !inSingle {
|
||||
inDouble = !inDouble
|
||||
}
|
||||
if inSingle || inDouble {
|
||||
continue
|
||||
}
|
||||
if c == '\\' {
|
||||
pos++
|
||||
continue
|
||||
}
|
||||
if c == open {
|
||||
depth++
|
||||
} else if c == close {
|
||||
depth--
|
||||
}
|
||||
if !(depth > 0) {
|
||||
break
|
||||
}
|
||||
}
|
||||
if depth > 0 {
|
||||
return false
|
||||
}
|
||||
a.pos = pos
|
||||
return true
|
||||
}
|
||||
|
||||
func (a *RuleAnalyzer) chompBalanced(open, close byte) bool {
|
||||
if a.code {
|
||||
return a.chompCodeBalanced(open, close)
|
||||
}
|
||||
return a.chompRuleBalanced(open, close)
|
||||
}
|
||||
|
||||
// SplitRule 对应 splitRule(vararg split):把 queue 切分成规则列表。
|
||||
// 首段按 splits 中最先出现的分隔符确定组合类型,其余按该类型循环切分。
|
||||
func (a *RuleAnalyzer) SplitRule(splits ...string) []string {
|
||||
if len(splits) == 1 {
|
||||
a.elementsType = splits[0]
|
||||
if !a.consumeTo(a.elementsType) {
|
||||
a.rule = append(a.rule, a.queue[a.startX:])
|
||||
return a.rule
|
||||
}
|
||||
a.step = len(a.elementsType)
|
||||
return a.splitRuleNext()
|
||||
}
|
||||
if !a.consumeToAny(splits...) {
|
||||
a.rule = append(a.rule, a.queue[a.startX:])
|
||||
return a.rule
|
||||
}
|
||||
|
||||
end := a.pos
|
||||
a.pos = a.start
|
||||
for {
|
||||
st := a.findToAny('[', '(')
|
||||
if st == -1 {
|
||||
a.rule = []string{a.queue[a.startX:end]}
|
||||
a.elementsType = a.queue[end : end+a.step]
|
||||
a.pos = end + a.step
|
||||
for a.consumeTo(a.elementsType) {
|
||||
a.rule = append(a.rule, a.queue[a.start:a.pos])
|
||||
a.pos += a.step
|
||||
}
|
||||
a.rule = append(a.rule, a.queue[a.pos:])
|
||||
return a.rule
|
||||
}
|
||||
if st > end {
|
||||
a.rule = []string{a.queue[a.startX:end]}
|
||||
a.elementsType = a.queue[end : end+a.step]
|
||||
a.pos = end + a.step
|
||||
for a.consumeTo(a.elementsType) && a.pos < st {
|
||||
a.rule = append(a.rule, a.queue[a.start:a.pos])
|
||||
a.pos += a.step
|
||||
}
|
||||
if a.pos > st {
|
||||
a.startX = a.start
|
||||
return a.splitRuleNext()
|
||||
}
|
||||
a.rule = append(a.rule, a.queue[a.pos:])
|
||||
return a.rule
|
||||
}
|
||||
a.pos = st
|
||||
next := byte(')')
|
||||
if a.queue[a.pos] == '[' {
|
||||
next = ']'
|
||||
}
|
||||
if !a.chompBalanced(a.queue[a.pos], next) {
|
||||
return []string{a.queue}
|
||||
}
|
||||
if end <= a.pos {
|
||||
break
|
||||
}
|
||||
}
|
||||
a.start = a.pos
|
||||
return a.splitRuleFirst(splits...)
|
||||
}
|
||||
|
||||
// splitRuleFirst 对应首段匹配的 tailrec 递归(elementsType 尚未确定)。
|
||||
func (a *RuleAnalyzer) splitRuleFirst(splits ...string) []string {
|
||||
if len(splits) == 1 {
|
||||
a.elementsType = splits[0]
|
||||
if !a.consumeTo(a.elementsType) {
|
||||
a.rule = append(a.rule, a.queue[a.startX:])
|
||||
return a.rule
|
||||
}
|
||||
a.step = len(a.elementsType)
|
||||
return a.splitRuleNext()
|
||||
}
|
||||
if !a.consumeToAny(splits...) {
|
||||
a.rule = append(a.rule, a.queue[a.startX:])
|
||||
return a.rule
|
||||
}
|
||||
end := a.pos
|
||||
a.pos = a.start
|
||||
for {
|
||||
st := a.findToAny('[', '(')
|
||||
if st == -1 {
|
||||
a.rule = []string{a.queue[a.startX:end]}
|
||||
a.elementsType = a.queue[end : end+a.step]
|
||||
a.pos = end + a.step
|
||||
for a.consumeTo(a.elementsType) {
|
||||
a.rule = append(a.rule, a.queue[a.start:a.pos])
|
||||
a.pos += a.step
|
||||
}
|
||||
a.rule = append(a.rule, a.queue[a.pos:])
|
||||
return a.rule
|
||||
}
|
||||
if st > end {
|
||||
a.rule = []string{a.queue[a.startX:end]}
|
||||
a.elementsType = a.queue[end : end+a.step]
|
||||
a.pos = end + a.step
|
||||
for a.consumeTo(a.elementsType) && a.pos < st {
|
||||
a.rule = append(a.rule, a.queue[a.start:a.pos])
|
||||
a.pos += a.step
|
||||
}
|
||||
if a.pos > st {
|
||||
a.startX = a.start
|
||||
return a.splitRuleNext()
|
||||
}
|
||||
a.rule = append(a.rule, a.queue[a.pos:])
|
||||
return a.rule
|
||||
}
|
||||
a.pos = st
|
||||
next := byte(')')
|
||||
if a.queue[a.pos] == '[' {
|
||||
next = ']'
|
||||
}
|
||||
if !a.chompBalanced(a.queue[a.pos], next) {
|
||||
return []string{a.queue}
|
||||
}
|
||||
if end <= a.pos {
|
||||
break
|
||||
}
|
||||
}
|
||||
a.start = a.pos
|
||||
return a.splitRuleFirst(splits...)
|
||||
}
|
||||
|
||||
// splitRuleNext 对应二段匹配 splitRule()(elementsType 已确定)。
|
||||
func (a *RuleAnalyzer) splitRuleNext() []string {
|
||||
for {
|
||||
end := a.pos
|
||||
a.pos = a.start
|
||||
for {
|
||||
st := a.findToAny('[', '(')
|
||||
if st == -1 {
|
||||
a.rule = append(a.rule, a.queue[a.startX:end])
|
||||
a.pos = end + a.step
|
||||
for a.consumeTo(a.elementsType) {
|
||||
a.rule = append(a.rule, a.queue[a.start:a.pos])
|
||||
a.pos += a.step
|
||||
}
|
||||
a.rule = append(a.rule, a.queue[a.pos:])
|
||||
return a.rule
|
||||
}
|
||||
if st > end {
|
||||
a.rule = append(a.rule, a.queue[a.startX:end])
|
||||
a.pos = end + a.step
|
||||
for a.consumeTo(a.elementsType) && a.pos < st {
|
||||
a.rule = append(a.rule, a.queue[a.start:a.pos])
|
||||
a.pos += a.step
|
||||
}
|
||||
if a.pos > st {
|
||||
a.startX = a.start
|
||||
return a.splitRuleNext()
|
||||
}
|
||||
a.rule = append(a.rule, a.queue[a.pos:])
|
||||
return a.rule
|
||||
}
|
||||
a.pos = st
|
||||
next := byte(')')
|
||||
if a.queue[a.pos] == '[' {
|
||||
next = ']'
|
||||
}
|
||||
if !a.chompBalanced(a.queue[a.pos], next) {
|
||||
return []string{a.queue}
|
||||
}
|
||||
if end <= a.pos {
|
||||
break
|
||||
}
|
||||
}
|
||||
a.start = a.pos
|
||||
if !a.consumeTo(a.elementsType) {
|
||||
a.rule = append(a.rule, a.queue[a.startX:])
|
||||
return a.rule
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// InnerRule 对应 innerRule(inner, startStep, endStep, fr) 第一变体:
|
||||
// 替换所有内嵌规则(如 {$.xxx}),fr 返回空表示该处不是有效内嵌规则。
|
||||
// 返回替换后的完整字符串;无一替换成功时返回 ""。
|
||||
func (a *RuleAnalyzer) InnerRule(inner string, startStep, endStep int, fr func(string) string) string {
|
||||
var sb []byte
|
||||
for {
|
||||
if !a.consumeTo(inner) {
|
||||
break
|
||||
}
|
||||
posPre := a.pos
|
||||
if a.chompCodeBalanced('{', '}') {
|
||||
frv := fr(a.queue[posPre+startStep : a.pos-endStep])
|
||||
if frv != "" {
|
||||
sb = append(sb, a.queue[a.startX:posPre]...)
|
||||
sb = append(sb, frv...)
|
||||
a.startX = a.pos
|
||||
continue
|
||||
}
|
||||
}
|
||||
a.pos += len(inner)
|
||||
}
|
||||
if a.startX == 0 {
|
||||
return ""
|
||||
}
|
||||
sb = append(sb, a.queue[a.startX:]...)
|
||||
return string(sb)
|
||||
}
|
||||
|
||||
// InnerRule2 对应 innerRule(startStr, endStr, fr) 第二变体:无平衡组检查。
|
||||
// 无一替换成功时返回原串。
|
||||
func (a *RuleAnalyzer) InnerRule2(startStr, endStr string, fr func(string) string) string {
|
||||
var sb []byte
|
||||
for {
|
||||
if !a.consumeTo(startStr) {
|
||||
break
|
||||
}
|
||||
a.pos += len(startStr)
|
||||
posPre := a.pos
|
||||
if a.consumeTo(endStr) {
|
||||
frv := fr(a.queue[posPre:a.pos])
|
||||
sb = append(sb, a.queue[a.startX:posPre-len(startStr)]...)
|
||||
sb = append(sb, frv...)
|
||||
a.pos += len(endStr)
|
||||
a.startX = a.pos
|
||||
}
|
||||
}
|
||||
if a.startX == 0 {
|
||||
return a.queue
|
||||
}
|
||||
sb = append(sb, a.queue[a.startX:]...)
|
||||
return string(sb)
|
||||
}
|
||||
@@ -0,0 +1,317 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"os"
|
||||
"strings"
|
||||
"testing"
|
||||
)
|
||||
|
||||
// 本文件用「聚合类登录书源」的典型结构验证服务端 JS 运行时:
|
||||
//
|
||||
// jsLib(公共函数库)+ loginUrl(登录逻辑)+ loginUi(表单)
|
||||
//
|
||||
// 与真实书源(如光遇聚合,jsLib 约 15 万字符)同构,但精简为可维护的固件。
|
||||
// 固件放在 testdata/,因此这些是常驻回归测试而非一次性 spike。
|
||||
|
||||
func loadTestdata(t *testing.T, name string) string {
|
||||
t.Helper()
|
||||
b, err := os.ReadFile("testdata/" + name)
|
||||
if err != nil {
|
||||
t.Fatalf("读取固件 %s 失败: %v", name, err)
|
||||
}
|
||||
return string(b)
|
||||
}
|
||||
|
||||
// sampleSourceProps 构造注入了 jsLib/loginUrl/loginUi 的书源属性。
|
||||
func sampleSourceProps(t *testing.T) map[string]any {
|
||||
t.Helper()
|
||||
return map[string]any{
|
||||
"bookSourceUrl": "https://v1.example-aggregate.com",
|
||||
"bookSourceName": "样例聚合源",
|
||||
"jsLib": loadTestdata(t, "sample_jsLib.js"),
|
||||
"loginUrl": loadTestdata(t, "sample_loginUrl.js"),
|
||||
"loginUi": loadTestdata(t, "sample_loginUi.json"),
|
||||
}
|
||||
}
|
||||
|
||||
// newSampleRunner 用聚合源固件构建运行时,网络由测试桩接管。
|
||||
func newSampleRunner(t *testing.T, state SourceState) *JSRunner {
|
||||
t.Helper()
|
||||
props := sampleSourceProps(t)
|
||||
r := NewJSRunner(JSConfig{
|
||||
SourceProps: props,
|
||||
JSLib: props["jsLib"].(string),
|
||||
State: state,
|
||||
BaseURL: "https://v1.example-aggregate.com",
|
||||
Fetch: func(req *Request) (string, string, int, error) {
|
||||
if strings.Contains(req.URL, "/login_api") {
|
||||
return `{"code":0,"key":"TOKEN_abcdefghijklmn"}`, req.URL, 200, nil
|
||||
}
|
||||
if strings.Contains(req.URL, "/user_api") {
|
||||
return `{"id":1,"email":"user@example.com","nickname":"tester"}`, req.URL, 200, nil
|
||||
}
|
||||
return `{}`, req.URL, 200, nil
|
||||
},
|
||||
})
|
||||
if err := r.JSLibErr(); err != nil {
|
||||
t.Fatalf("jsLib 执行失败: %v", err)
|
||||
}
|
||||
return r
|
||||
}
|
||||
|
||||
// TestJSLibProvidesHelpers jsLib 里的函数与 lexical 绑定(let hosts)应跨执行可见。
|
||||
func TestJSLibProvidesHelpers(t *testing.T) {
|
||||
r := newSampleRunner(t, NewMemoryState())
|
||||
v, err := r.EvalAction(`typeof getVariable + '|' + typeof request + '|' + typeof hosts`, nil)
|
||||
if err != nil {
|
||||
t.Fatalf("EvalAction 失败: %v", err)
|
||||
}
|
||||
if got := anyToString(v); got != "function|function|object" {
|
||||
t.Fatalf("jsLib 符号不可见: %q", got)
|
||||
}
|
||||
// BaseUrl() 依赖 lexical hosts 与 source.getVariable
|
||||
v, err = r.EvalAction(`BaseUrl()`, nil)
|
||||
if err != nil {
|
||||
t.Fatalf("BaseUrl 失败: %v", err)
|
||||
}
|
||||
if got := anyToString(v); got != "https://v1.example-aggregate.com" {
|
||||
t.Fatalf("BaseUrl() = %q", got)
|
||||
}
|
||||
// login 由 loginUrl 提供,未拼 loginUrl 时不应存在
|
||||
v, _ = r.EvalAction(`typeof login`, nil)
|
||||
if anyToString(v) != "undefined" {
|
||||
t.Fatalf("login 应仅由 loginUrl 定义,当前 %q", anyToString(v))
|
||||
}
|
||||
}
|
||||
|
||||
// TestLoginFlowWritesCookies login() 应请求登录接口、写入 Cookie,且 getToken 能读回。
|
||||
func TestLoginFlowWritesCookies(t *testing.T) {
|
||||
state := NewMemoryState()
|
||||
r := newSampleRunner(t, state)
|
||||
loginJS := loadTestdata(t, "sample_loginUrl.js")
|
||||
|
||||
_, err := r.EvalAction(loginJS+"\nlogin(true)", map[string]any{
|
||||
"result": map[string]string{"邮箱": "user@example.com", "密码": "pw123456"},
|
||||
})
|
||||
if err != nil {
|
||||
t.Fatalf("login(true) 执行失败: %v", err)
|
||||
}
|
||||
v, err := r.EvalAction(`getToken()`, nil)
|
||||
if err != nil {
|
||||
t.Fatalf("getToken 失败: %v", err)
|
||||
}
|
||||
if got := anyToString(v); got != "TOKEN_abcdefghijklmn" {
|
||||
t.Fatalf("getToken() = %q,期望登录后能读到 token", got)
|
||||
}
|
||||
// setAllCookies 遍历 hosts;两条线路同属 example-aggregate.com,
|
||||
// 按 eTLD+1 归并成一条 Cookie,两个线路地址都应能读到。
|
||||
if len(state.Snapshot()) != 1 {
|
||||
t.Fatalf("同站线路应归并为一条 Cookie,实际 %v", state.Snapshot())
|
||||
}
|
||||
for _, host := range []string{"https://v1.example-aggregate.com", "https://v2.example-aggregate.com"} {
|
||||
if !strings.Contains(state.GetCookie(host), "qttoken=TOKEN_abcdefghijklmn") {
|
||||
t.Fatalf("%s 未读到 token: %q", host, state.GetCookie(host))
|
||||
}
|
||||
}
|
||||
if !strings.Contains(strings.Join(state.Toasts(), "\n"), "登录成功") {
|
||||
t.Fatalf("未收到登录成功提示: %v", state.Toasts())
|
||||
}
|
||||
}
|
||||
|
||||
// TestLoginMissingCredentials 缺少账号密码时应给出提示且不产生 Cookie。
|
||||
func TestLoginMissingCredentials(t *testing.T) {
|
||||
state := NewMemoryState()
|
||||
r := newSampleRunner(t, state)
|
||||
loginJS := loadTestdata(t, "sample_loginUrl.js")
|
||||
|
||||
if _, err := r.EvalAction(loginJS+"\nlogin(true)", map[string]any{"result": map[string]string{}}); err != nil {
|
||||
t.Fatalf("执行失败: %v", err)
|
||||
}
|
||||
if len(state.Snapshot()) != 0 {
|
||||
t.Fatalf("未填表单不应写入 Cookie: %v", state.Snapshot())
|
||||
}
|
||||
if !strings.Contains(strings.Join(state.Toasts(), "\n"), "请先输入账号密码") {
|
||||
t.Fatalf("缺少提示语: %v", state.Toasts())
|
||||
}
|
||||
}
|
||||
|
||||
// TestStartBrowserAwaitWithoutHost startBrowserAwait 在未注入宿主浏览器时应
|
||||
// 明确报错,并记录待打开地址,避免书源逻辑把空 body 当成校验成功。
|
||||
func TestStartBrowserAwaitWithoutHost(t *testing.T) {
|
||||
state := NewMemoryState()
|
||||
r := newSampleRunner(t, state)
|
||||
loginJS := loadTestdata(t, "sample_loginUrl.js")
|
||||
|
||||
// 先造出登录态,让 user() 走到 startBrowserAwait
|
||||
state.SetCookie("https://v1.example-aggregate.com", "qttoken=TOKEN_abcdefghijklmn")
|
||||
_, err := r.EvalAction(loginJS+"\nuser()", nil)
|
||||
if err == nil || !strings.Contains(err.Error(), "浏览器") {
|
||||
t.Fatalf("应明确报不支持,实际: %v", err)
|
||||
}
|
||||
browsers := state.Browsers()
|
||||
if len(browsers) != 1 || !strings.HasSuffix(browsers[0].URL, "/user") {
|
||||
t.Fatalf("未记录待打开地址: %+v", browsers)
|
||||
}
|
||||
}
|
||||
|
||||
// TestSourceLoginInfoMapInit 未保存登录信息时,getLoginInfoMap 用 loginUi 的
|
||||
// 非按钮字段初始化(对应 legado)。
|
||||
func TestSourceLoginInfoMapInit(t *testing.T) {
|
||||
r := newSampleRunner(t, NewMemoryState())
|
||||
v, err := r.EvalAction(`JSON.stringify(source.getLoginInfoMap())`, nil)
|
||||
if err != nil {
|
||||
t.Fatalf("getLoginInfoMap 失败: %v", err)
|
||||
}
|
||||
var m map[string]string
|
||||
if err := json.Unmarshal([]byte(anyToString(v)), &m); err != nil {
|
||||
t.Fatalf("返回不是 JSON 对象: %q", anyToString(v))
|
||||
}
|
||||
if _, ok := m["邮箱"]; !ok {
|
||||
t.Fatalf("loginUi 的「邮箱」未出现在登录信息: %v", m)
|
||||
}
|
||||
if _, ok := m["密码"]; !ok {
|
||||
t.Fatalf("loginUi 的「密码」未出现在登录信息: %v", m)
|
||||
}
|
||||
for k := range m {
|
||||
if strings.Contains(k, "登录") || strings.Contains(k, "注册") || strings.Contains(k, "后台") {
|
||||
t.Fatalf("button 字段不应出现在登录信息: %v", m)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// TestSourceVariablesRoundTrip 变量经 source.setVariable 写入后可由 getVariable 读出。
|
||||
func TestSourceVariablesRoundTrip(t *testing.T) {
|
||||
state := NewMemoryState()
|
||||
r := newSampleRunner(t, state)
|
||||
|
||||
if _, err := r.EvalAction(`setVariable('线路','https://v2.example-aggregate.com',false)`, nil); err != nil {
|
||||
t.Fatalf("setVariable 失败: %v", err)
|
||||
}
|
||||
v, err := r.EvalAction(`getVariable('线路')`, nil)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got := anyToString(v); got != "https://v2.example-aggregate.com" {
|
||||
t.Fatalf("getVariable('线路') = %q", got)
|
||||
}
|
||||
// 变量变更应反映到状态存储(持久化的前提)
|
||||
if !strings.Contains(state.GetVariable(), "v2.example-aggregate.com") {
|
||||
t.Fatalf("变量未写回状态存储: %q", state.GetVariable())
|
||||
}
|
||||
// BaseUrl 也应跟随变量变化
|
||||
v, _ = r.EvalAction(`BaseUrl()`, nil)
|
||||
if anyToString(v) != "https://v2.example-aggregate.com" {
|
||||
t.Fatalf("BaseUrl 未跟随变量: %q", anyToString(v))
|
||||
}
|
||||
}
|
||||
|
||||
// TestCookieObjectIsolation Cookie 按站点隔离,同站子域共享,removeCookie 可清除。
|
||||
func TestCookieObjectIsolation(t *testing.T) {
|
||||
state := NewMemoryState()
|
||||
r := NewJSRunner(JSConfig{State: state})
|
||||
|
||||
if _, err := r.EvalAction(`cookie.setCookie('https://a.example.com/x','t=1; u=2')`, nil); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
v, err := r.EvalAction(`cookie.getCookie('https://a.example.com/y','t')`, nil)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if anyToString(v) != "1" {
|
||||
t.Fatalf("getCookie key = %q", anyToString(v))
|
||||
}
|
||||
// 同站子域共享 Cookie(对应 legado getSubDomain 取 eTLD+1)
|
||||
v, _ = r.EvalAction(`cookie.getCookie('https://www.example.com/')`, nil)
|
||||
if anyToString(v) == "" {
|
||||
t.Fatal("同站子域应共享 Cookie")
|
||||
}
|
||||
// 不同站点互不污染
|
||||
v, _ = r.EvalAction(`cookie.getCookie('https://other-site.net/')`, nil)
|
||||
if anyToString(v) != "" {
|
||||
t.Fatalf("跨站点读到了 Cookie: %q", anyToString(v))
|
||||
}
|
||||
// removeCookie 后读不到
|
||||
if _, err := r.EvalAction(`cookie.removeCookie('https://a.example.com')`, nil); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
v, _ = r.EvalAction(`cookie.getCookie('https://a.example.com/')`, nil)
|
||||
if anyToString(v) != "" {
|
||||
t.Fatalf("removeCookie 后仍读到: %q", anyToString(v))
|
||||
}
|
||||
}
|
||||
|
||||
// TestSourceStatePersistsAcrossRunners 状态存储是跨运行时共享的:
|
||||
// 模拟服务端重启(新建 runner)后登录态仍在。
|
||||
func TestSourceStatePersistsAcrossRunners(t *testing.T) {
|
||||
state := NewMemoryState()
|
||||
loginJS := loadTestdata(t, "sample_loginUrl.js")
|
||||
|
||||
r1 := newSampleRunner(t, state)
|
||||
if _, err := r1.EvalAction(loginJS+"\nlogin(true)", map[string]any{
|
||||
"result": map[string]string{"邮箱": "u@e.com", "密码": "pw"},
|
||||
}); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
// 新运行时复用同一 state(对应从 DB 重新载入)
|
||||
r2 := newSampleRunner(t, state)
|
||||
v, err := r2.EvalAction(`getToken()`, nil)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if anyToString(v) != "TOKEN_abcdefghijklmn" {
|
||||
t.Fatalf("新运行时未读到登录态: %q", anyToString(v))
|
||||
}
|
||||
// 已登录时再次 login 应提示已登录(走 getToken 短路分支)
|
||||
if _, err := r2.EvalAction(loginJS+"\nlogin(true)", map[string]any{
|
||||
"result": map[string]string{"邮箱": "u@e.com", "密码": "pw"},
|
||||
}); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if !strings.Contains(strings.Join(state.Toasts(), "\n"), "已登录") {
|
||||
t.Fatalf("已登录时应有提示: %v", state.Toasts())
|
||||
}
|
||||
}
|
||||
|
||||
// TestLoginCheckJsReplacesBody loginCheckJs 返回新响应体时应替换原 body
|
||||
// (书源借此检测会话失效并重取页面)。
|
||||
func TestLoginCheckJsReplacesBody(t *testing.T) {
|
||||
r := NewJSRunner(JSConfig{State: NewMemoryState()})
|
||||
// 返回响应对象(对应 legado 要求 loginCheckJs 返回 StrResponse)
|
||||
js := `if (result.body().indexOf('未登录') >= 0) { java.toast('会话失效,重新登录'); }
|
||||
var ok = JSON.stringify({code: 0, body: result.body()});
|
||||
result.body()`
|
||||
body, changed, err := r.EvalLoginCheck(js, "未登录", 200, "https://x.com/a")
|
||||
if err != nil {
|
||||
t.Fatalf("EvalLoginCheck 失败: %v", err)
|
||||
}
|
||||
if !changed || body != "未登录" {
|
||||
t.Fatalf("body=%q changed=%v", body, changed)
|
||||
}
|
||||
}
|
||||
|
||||
// TestLoginCheckJsReturnsString loginCheckJs 直接返回字符串时也应生效。
|
||||
func TestLoginCheckJsReturnsString(t *testing.T) {
|
||||
r := NewJSRunner(JSConfig{State: NewMemoryState()})
|
||||
body, changed, err := r.EvalLoginCheck(`'已修复内容'`, "原始", 200, "https://x.com/a")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if !changed || body != "已修复内容" {
|
||||
t.Fatalf("body=%q changed=%v", body, changed)
|
||||
}
|
||||
}
|
||||
|
||||
// TestLoginCheckJsResponseAccessors 验证 result 暴露 code()/url()/header()。
|
||||
func TestLoginCheckJsResponseAccessors(t *testing.T) {
|
||||
r := NewJSRunner(JSConfig{State: NewMemoryState()})
|
||||
body, changed, err := r.EvalLoginCheck(
|
||||
`result.code() + '|' + result.url()`, "x", 403, "https://x.com/y")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if !changed || body != "403|https://x.com/y" {
|
||||
t.Fatalf("body=%q", body)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,337 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"fmt"
|
||||
"regexp"
|
||||
"sort"
|
||||
"strings"
|
||||
)
|
||||
|
||||
// 本文件对应 AnalyzeRule.kt 中的 SourceRule 内部类与 splitSourceRule。
|
||||
|
||||
// Mode 规则模式(对应 AnalyzeRule.Mode)。
|
||||
type Mode int
|
||||
|
||||
const (
|
||||
ModeXPath Mode = iota
|
||||
ModeJson
|
||||
ModeDefault
|
||||
ModeJs
|
||||
ModeRegex
|
||||
ModeWebJs
|
||||
)
|
||||
|
||||
var (
|
||||
jsPatternRe = regexp.MustCompile(`(?i)<js>([\w\W]*?)</js>|@js:([\w\W]*)`)
|
||||
webJsPatternRe = regexp.MustCompile(`(?i)@webjs:([\w\W]{5,})`)
|
||||
putPatternRe = regexp.MustCompile(`(?i)@put:(\{[^}]+?\})`)
|
||||
evalPatternRe = regexp.MustCompile(`(?i)@get:\{[^}]+?\}|\{\{[\w\W]*?\}\}`)
|
||||
regexGroupRe = regexp.MustCompile(`\$\d{1,2}`)
|
||||
)
|
||||
|
||||
// makeUpRule 参数类型(对应 SourceRule 的 ruleType 常量)。
|
||||
const (
|
||||
paramGet = -2 // @get:{name}
|
||||
paramJs = -1 // {{js}}
|
||||
paramText = 0 // 字面文本
|
||||
paramGroupN = 1 // >0 为 $N 正则分组引用,typ 即分组序号
|
||||
)
|
||||
|
||||
type ruleParam struct {
|
||||
typ int
|
||||
val string
|
||||
}
|
||||
|
||||
// SourceRule 是拆分后的单条规则(对应 AnalyzeRule.SourceRule)。
|
||||
type SourceRule struct {
|
||||
Rule string
|
||||
Mode Mode
|
||||
|
||||
putMap map[string]string
|
||||
ruleParams []ruleParam
|
||||
}
|
||||
|
||||
// SplitSourceRule 对应 AnalyzeRule.splitSourceRule(ruleStr, allInOne)。
|
||||
// contentIsJSON 对应 Kotlin 成员 isJSON(当前内容是否为 JSON)。
|
||||
// isRegex 对应 AnalyzeRule.isRegex 持久标记,可为 nil。
|
||||
func SplitSourceRule(ruleStr string, allInOne, contentIsJSON bool, isRegex *bool) []*SourceRule {
|
||||
if ruleStr == "" {
|
||||
return nil
|
||||
}
|
||||
mode := ModeDefault
|
||||
start := 0
|
||||
if allInOne && strings.HasPrefix(ruleStr, ":") {
|
||||
mode = ModeRegex
|
||||
if isRegex != nil {
|
||||
*isRegex = true
|
||||
}
|
||||
start = 1
|
||||
} else if isRegex != nil && *isRegex {
|
||||
mode = ModeRegex
|
||||
}
|
||||
|
||||
type rulePart struct {
|
||||
start int
|
||||
end int
|
||||
rule string
|
||||
mode Mode
|
||||
}
|
||||
var parts []rulePart
|
||||
for _, g := range jsPatternRe.FindAllStringSubmatchIndex(ruleStr, -1) {
|
||||
jsBody := ""
|
||||
if g[2] >= 0 { // group(1): <js>...</js>
|
||||
jsBody = ruleStr[g[2]:g[3]]
|
||||
} else if g[4] >= 0 { // group(2): @js:...
|
||||
jsBody = ruleStr[g[4]:g[5]]
|
||||
}
|
||||
parts = append(parts, rulePart{g[0], g[1], jsBody, ModeJs})
|
||||
}
|
||||
for _, g := range webJsPatternRe.FindAllStringSubmatchIndex(ruleStr, -1) {
|
||||
parts = append(parts, rulePart{g[0], g[1], ruleStr[g[2]:g[3]], ModeWebJs})
|
||||
}
|
||||
sort.Slice(parts, func(i, j int) bool { return parts[i].start < parts[j].start })
|
||||
|
||||
var out []*SourceRule
|
||||
for _, p := range parts {
|
||||
if p.start < start {
|
||||
continue
|
||||
}
|
||||
if p.start > start {
|
||||
if tmp := strings.TrimSpace(ruleStr[start:p.start]); tmp != "" {
|
||||
out = append(out, newSourceRule(tmp, mode, contentIsJSON))
|
||||
}
|
||||
}
|
||||
out = append(out, newSourceRule(p.rule, p.mode, contentIsJSON))
|
||||
start = p.end
|
||||
}
|
||||
if len(ruleStr) > start {
|
||||
if tmp := strings.TrimSpace(ruleStr[start:]); tmp != "" {
|
||||
out = append(out, newSourceRule(tmp, mode, contentIsJSON))
|
||||
}
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// newSourceRule 对应 SourceRule.init:模式识别 + 分离 put + 拆分 @get/{{}}/$N。
|
||||
func newSourceRule(ruleStr string, mode Mode, contentIsJSON bool) *SourceRule {
|
||||
sr := &SourceRule{Mode: mode, putMap: map[string]string{}}
|
||||
rule := ruleStr
|
||||
switch {
|
||||
case mode == ModeJs || mode == ModeRegex:
|
||||
// 保持原样
|
||||
case hasPrefixFold(ruleStr, "@CSS:"):
|
||||
sr.Mode = ModeDefault
|
||||
case strings.HasPrefix(ruleStr, "@@"):
|
||||
sr.Mode = ModeDefault
|
||||
rule = ruleStr[2:]
|
||||
case hasPrefixFold(ruleStr, "@XPath:"):
|
||||
sr.Mode = ModeXPath
|
||||
rule = ruleStr[7:]
|
||||
case hasPrefixFold(ruleStr, "@Json:"):
|
||||
sr.Mode = ModeJson
|
||||
rule = ruleStr[6:]
|
||||
case contentIsJSON || strings.HasPrefix(ruleStr, "$.") || strings.HasPrefix(ruleStr, "$["):
|
||||
sr.Mode = ModeJson
|
||||
case strings.HasPrefix(ruleStr, "/"):
|
||||
sr.Mode = ModeXPath
|
||||
}
|
||||
// 分离 @put:{...}
|
||||
rule = splitPutRule(rule, sr.putMap)
|
||||
sr.Rule = rule
|
||||
sr.splitEvalParams()
|
||||
return sr
|
||||
}
|
||||
|
||||
func hasPrefixFold(s, prefix string) bool {
|
||||
return len(s) >= len(prefix) && strings.EqualFold(s[:len(prefix)], prefix)
|
||||
}
|
||||
|
||||
// splitPutRule 对应 splitPutRule:提取并移除 @put:{...} 段。
|
||||
func splitPutRule(ruleStr string, putMap map[string]string) string {
|
||||
out := ruleStr
|
||||
for _, m := range putPatternRe.FindAllStringSubmatch(ruleStr, -1) {
|
||||
out = strings.Replace(out, m[0], "", 1)
|
||||
parsed := map[string]string{}
|
||||
if err := json.Unmarshal([]byte(m[1]), &parsed); err == nil {
|
||||
for k, v := range parsed {
|
||||
putMap[k] = v
|
||||
}
|
||||
continue
|
||||
}
|
||||
// 宽松解析(对应 GSON lenient):key:val 键值对
|
||||
pairRe := regexp.MustCompile(`["']?(\w+)["']?\s*:\s*(["']?)([^,{}]*?)\2`)
|
||||
for _, pm := range pairRe.FindAllStringSubmatch(m[1], -1) {
|
||||
putMap[pm[1]] = pm[3]
|
||||
}
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// splitEvalParams 对应 init 中 @get/{{ }} 的拆分循环。
|
||||
func (sr *SourceRule) splitEvalParams() {
|
||||
rule := sr.Rule
|
||||
start := 0
|
||||
locs := evalPatternRe.FindAllStringIndex(rule, -1)
|
||||
if len(locs) > 0 {
|
||||
firstStart := locs[0][0]
|
||||
prefix := rule[:firstStart]
|
||||
if sr.Mode != ModeJs && sr.Mode != ModeRegex &&
|
||||
(firstStart == 0 || !strings.Contains(prefix, "##")) {
|
||||
sr.Mode = ModeRegex
|
||||
}
|
||||
for _, loc := range locs {
|
||||
if loc[0] > start {
|
||||
sr.splitRegex(rule[start:loc[0]])
|
||||
}
|
||||
tmp := rule[loc[0]:loc[1]]
|
||||
switch {
|
||||
case hasPrefixFold(tmp, "@get:"):
|
||||
sr.ruleParams = append(sr.ruleParams, ruleParam{paramGet, tmp[6 : len(tmp)-1]})
|
||||
case strings.HasPrefix(tmp, "{{"):
|
||||
sr.ruleParams = append(sr.ruleParams, ruleParam{paramJs, tmp[2 : len(tmp)-2]})
|
||||
default:
|
||||
sr.splitRegex(tmp)
|
||||
}
|
||||
start = loc[1]
|
||||
}
|
||||
}
|
||||
if len(rule) > start {
|
||||
sr.splitRegex(rule[start:])
|
||||
}
|
||||
}
|
||||
|
||||
// splitRegex 对应 SourceRule.splitRegex:拆分 $N 分组引用。
|
||||
// $N 匹配只作用于 "##" 前的第一段,"##..." 尾部作为字面参数保留,
|
||||
// 由 MakeUpRule 重新按 "##" 切分。
|
||||
func (sr *SourceRule) splitRegex(ruleStr string) {
|
||||
start := 0
|
||||
first := strings.Split(ruleStr, "##")[0]
|
||||
locs := regexGroupRe.FindAllStringIndex(first, -1)
|
||||
if len(locs) > 0 {
|
||||
if sr.Mode != ModeJs && sr.Mode != ModeRegex {
|
||||
sr.Mode = ModeRegex
|
||||
}
|
||||
for _, loc := range locs {
|
||||
if loc[0] > start {
|
||||
sr.ruleParams = append(sr.ruleParams, ruleParam{paramText, ruleStr[start:loc[0]]})
|
||||
}
|
||||
n := 0
|
||||
fmt.Sscanf(ruleStr[loc[0]+1:loc[1]], "%d", &n)
|
||||
sr.ruleParams = append(sr.ruleParams, ruleParam{n, ruleStr[loc[0]:loc[1]]})
|
||||
start = loc[1]
|
||||
}
|
||||
}
|
||||
if len(ruleStr) > start {
|
||||
sr.ruleParams = append(sr.ruleParams, ruleParam{paramText, ruleStr[start:]})
|
||||
}
|
||||
}
|
||||
|
||||
// ResolvedSourceRule 对应 AnalyzeRule.ResolvedSourceRule。
|
||||
type ResolvedSourceRule struct {
|
||||
Rule string
|
||||
ReplaceRegex string
|
||||
Replacement string
|
||||
ReplaceFirst bool
|
||||
ParamSize int
|
||||
}
|
||||
|
||||
// RuleDeps 是 MakeUpRule 解析内嵌 {{...}} 时需要的执行环境。
|
||||
type RuleDeps struct {
|
||||
// JS 执行 {{js}}(P0 未接入 goja 时返回 ErrJsUnsupported)
|
||||
JS func(js string, result any) (any, error)
|
||||
// Rule 执行 {{@rule}} / {{$.rule}} 形式的规则引用
|
||||
Rule func(rule string) (string, error)
|
||||
// Get 对应 AnalyzeRule.get(key)
|
||||
Get func(key string) string
|
||||
}
|
||||
|
||||
// MakeUpRule 对应 SourceRule.makeUpRule(result):替换 @get/{{}}/$N,
|
||||
// 再按 "##" 切分出正则段。
|
||||
func (sr *SourceRule) MakeUpRule(result any, deps *RuleDeps) (ResolvedSourceRule, error) {
|
||||
resolved := sr.Rule
|
||||
if len(sr.ruleParams) > 0 {
|
||||
var infoVal []byte
|
||||
for i := len(sr.ruleParams) - 1; i >= 0; i-- {
|
||||
p := sr.ruleParams[i]
|
||||
switch {
|
||||
case p.typ >= paramGroupN:
|
||||
// 对应 Kotlin:(result as? List<String?>) 成功但越界时不插入任何内容
|
||||
switch lst := result.(type) {
|
||||
case []string:
|
||||
if len(lst) > p.typ {
|
||||
infoVal = prepend(infoVal, lst[p.typ])
|
||||
}
|
||||
case []any:
|
||||
if len(lst) > p.typ {
|
||||
if s, ok := lst[p.typ].(string); ok {
|
||||
infoVal = prepend(infoVal, s)
|
||||
}
|
||||
}
|
||||
default:
|
||||
infoVal = prepend(infoVal, p.val)
|
||||
}
|
||||
case p.typ == paramJs:
|
||||
if isRuleString(p.val) {
|
||||
s, err := deps.Rule(p.val)
|
||||
if err != nil {
|
||||
return ResolvedSourceRule{}, err
|
||||
}
|
||||
infoVal = prepend(infoVal, s)
|
||||
} else {
|
||||
v, err := deps.JS(p.val, result)
|
||||
if err != nil {
|
||||
return ResolvedSourceRule{}, err
|
||||
}
|
||||
infoVal = prepend(infoVal, anyToString(v))
|
||||
}
|
||||
case p.typ == paramGet:
|
||||
infoVal = prepend(infoVal, deps.Get(p.val))
|
||||
default:
|
||||
infoVal = prepend(infoVal, p.val)
|
||||
}
|
||||
}
|
||||
resolved = string(infoVal)
|
||||
}
|
||||
segs := strings.Split(resolved, "##")
|
||||
r := ResolvedSourceRule{
|
||||
Rule: strings.TrimSpace(segs[0]),
|
||||
ReplaceFirst: len(segs) > 3,
|
||||
ParamSize: len(sr.ruleParams),
|
||||
}
|
||||
if len(segs) > 1 {
|
||||
r.ReplaceRegex = segs[1]
|
||||
}
|
||||
if len(segs) > 2 {
|
||||
r.Replacement = segs[2]
|
||||
}
|
||||
return r, nil
|
||||
}
|
||||
|
||||
func prepend(dst []byte, s string) []byte {
|
||||
return append([]byte(s), dst...)
|
||||
}
|
||||
|
||||
func isRuleString(s string) bool {
|
||||
return strings.HasPrefix(s, "@") ||
|
||||
strings.HasPrefix(s, "$.") ||
|
||||
strings.HasPrefix(s, "$[") ||
|
||||
strings.HasPrefix(s, "//")
|
||||
}
|
||||
|
||||
// anyToString 对应 Kotlin 值字符串化(整值 Double 去小数,对应 %.0f)。
|
||||
func anyToString(v any) string {
|
||||
switch t := v.(type) {
|
||||
case nil:
|
||||
return ""
|
||||
case string:
|
||||
return t
|
||||
case float64:
|
||||
if t == float64(int64(t)) {
|
||||
return fmt.Sprintf("%.0f", t)
|
||||
}
|
||||
return fmt.Sprintf("%v", t)
|
||||
default:
|
||||
return fmt.Sprintf("%v", t)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,292 @@
|
||||
package rule
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"net"
|
||||
"net/url"
|
||||
"strings"
|
||||
"sync"
|
||||
|
||||
"golang.org/x/net/publicsuffix"
|
||||
)
|
||||
|
||||
// 本文件:书源会话状态(变量 / 登录信息 / 登录请求头 / Cookie)的抽象。
|
||||
// 对应 legado BaseSource 的 getVariable/putVariable/getLoginInfo/putLoginInfo/
|
||||
// getLoginHeader/putLoginHeader 与 CookieStore。实现由服务层落库提供;
|
||||
// 未注入时退回进程内 MemoryState(单测、冒烟 CLI 用)。
|
||||
|
||||
// SourceState 书源会话状态读写接口。
|
||||
type SourceState interface {
|
||||
// GetVariable / SetVariable 对应 source.getVariable / source.setVariable。
|
||||
GetVariable() string
|
||||
SetVariable(v string)
|
||||
|
||||
// GetLoginInfo / SetLoginInfo 对应 source.getLoginInfo / source.putLoginInfo。
|
||||
// 内容为登录表单的 JSON(键为字段名)。
|
||||
GetLoginInfo() string
|
||||
SetLoginInfo(v string)
|
||||
|
||||
// GetLoginHeader / SetLoginHeader 对应 source 的登录请求头(JSON),
|
||||
// 其中 Cookie 键会在请求时合并进 Cookie 头。
|
||||
GetLoginHeader() string
|
||||
SetLoginHeader(v string)
|
||||
|
||||
// Cookie 读写:rawURL 可为完整 URL 或裸域名,按有效顶级域+1 归并。
|
||||
GetCookie(rawURL string) string
|
||||
GetCookieKey(rawURL, key string) string
|
||||
SetCookie(rawURL, cookie string)
|
||||
RemoveCookie(rawURL string)
|
||||
|
||||
// Toast 收集宿主提示(java.toast / java.longToast),登录反馈靠它回传前端。
|
||||
Toast(msg string)
|
||||
// OpenBrowser 记录需要浏览器完成的地址(java.startBrowser)——
|
||||
// 服务端无法弹窗,前端据此提供「在新标签打开」。
|
||||
OpenBrowser(url, title string)
|
||||
}
|
||||
|
||||
// UIState 是可选实现的登录界面信号接口。
|
||||
//
|
||||
// java.reLoginView / java.refreshExplore / java.upLoginData 在 legado 中直接
|
||||
// 操作登录对话框的控件。服务端不能碰 DOM,因此把意图显式交给宿主处理,
|
||||
// 而不是静默丢弃——静默丢弃会让书源以为表单已经按它的预期更新了。
|
||||
type UIState interface {
|
||||
// RequestUIRefresh 请求宿主重新渲染登录表单
|
||||
// (对应 legado SourceLoginJsExtensions.reLoginView / refreshExplore)。
|
||||
RequestUIRefresh()
|
||||
// ApplyLoginData 把书源给出的值合并进登录表单
|
||||
// (对应 legado SourceLoginJsExtensions.upLoginData)。
|
||||
ApplyLoginData(data map[string]string)
|
||||
}
|
||||
|
||||
// CookieDomain 取 URL 的有效顶级域 +1(对应 legado NetworkUtils.getSubDomain)。
|
||||
// 裸域名(无 scheme)按 http 处理;IP 原样返回;解析失败回退 host 本身。
|
||||
func CookieDomain(rawURL string) string {
|
||||
s := strings.TrimSpace(rawURL)
|
||||
if s == "" {
|
||||
return ""
|
||||
}
|
||||
if !strings.Contains(s, "://") {
|
||||
if strings.HasPrefix(s, "//") {
|
||||
s = "http:" + s
|
||||
} else {
|
||||
s = "http://" + s
|
||||
}
|
||||
}
|
||||
u, err := url.Parse(s)
|
||||
if err != nil {
|
||||
return ""
|
||||
}
|
||||
host := u.Hostname()
|
||||
if host == "" {
|
||||
return ""
|
||||
}
|
||||
if net.ParseIP(host) != nil {
|
||||
return host
|
||||
}
|
||||
etld1, err := publicsuffix.EffectiveTLDPlusOne(host)
|
||||
if err != nil {
|
||||
return host
|
||||
}
|
||||
return etld1
|
||||
}
|
||||
|
||||
// ParseCookie 把 "a=1; b=2" 拆为键值对(值不 Trim 内部空白,仅去首尾)。
|
||||
func ParseCookie(cookie string) map[string]string {
|
||||
out := map[string]string{}
|
||||
for _, part := range strings.Split(cookie, ";") {
|
||||
kv := strings.SplitN(strings.TrimSpace(part), "=", 2)
|
||||
if len(kv) != 2 || kv[0] == "" {
|
||||
continue
|
||||
}
|
||||
out[kv[0]] = kv[1]
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// MergeCookie 将新 cookie 覆盖式合并进旧 cookie(对应 CookieStore.replaceCookie)。
|
||||
func MergeCookie(old, newCookie string) string {
|
||||
merged := ParseCookie(old)
|
||||
for k, v := range ParseCookie(newCookie) {
|
||||
merged[k] = v
|
||||
}
|
||||
return formatCookie(merged)
|
||||
}
|
||||
|
||||
func formatCookie(kv map[string]string) string {
|
||||
parts := make([]string, 0, len(kv))
|
||||
for k, v := range kv {
|
||||
parts = append(parts, k+"="+v)
|
||||
}
|
||||
return strings.Join(parts, "; ")
|
||||
}
|
||||
|
||||
// MemoryState 进程内会话状态(未注入持久化实现时的回退)。
|
||||
type MemoryState struct {
|
||||
mu sync.Mutex
|
||||
variable string
|
||||
loginInfo string
|
||||
loginHeader string
|
||||
cookies map[string]map[string]string // domain → name → value
|
||||
toasts []string
|
||||
browsers []BrowserRequest
|
||||
uiRefresh bool
|
||||
}
|
||||
|
||||
// BrowserRequest 前端可代为打开的浏览器地址(java.startBrowser 收集)。
|
||||
type BrowserRequest struct {
|
||||
URL string `json:"url"`
|
||||
Title string `json:"title"`
|
||||
}
|
||||
|
||||
// NewMemoryState 创建进程内状态。
|
||||
func NewMemoryState() *MemoryState {
|
||||
return &MemoryState{cookies: map[string]map[string]string{}}
|
||||
}
|
||||
|
||||
func (m *MemoryState) GetVariable() string { m.mu.Lock(); defer m.mu.Unlock(); return m.variable }
|
||||
func (m *MemoryState) SetVariable(v string) {
|
||||
m.mu.Lock()
|
||||
defer m.mu.Unlock()
|
||||
m.variable = v
|
||||
}
|
||||
|
||||
func (m *MemoryState) GetLoginInfo() string { m.mu.Lock(); defer m.mu.Unlock(); return m.loginInfo }
|
||||
func (m *MemoryState) SetLoginInfo(v string) {
|
||||
m.mu.Lock()
|
||||
defer m.mu.Unlock()
|
||||
m.loginInfo = v
|
||||
}
|
||||
|
||||
func (m *MemoryState) GetLoginHeader() string { m.mu.Lock(); defer m.mu.Unlock(); return m.loginHeader }
|
||||
func (m *MemoryState) SetLoginHeader(v string) {
|
||||
m.mu.Lock()
|
||||
defer m.mu.Unlock()
|
||||
m.loginHeader = v
|
||||
}
|
||||
|
||||
func (m *MemoryState) GetCookie(rawURL string) string {
|
||||
return m.GetCookieKey(rawURL, "")
|
||||
}
|
||||
|
||||
func (m *MemoryState) GetCookieKey(rawURL, key string) string {
|
||||
domain := CookieDomain(rawURL)
|
||||
if domain == "" {
|
||||
return ""
|
||||
}
|
||||
m.mu.Lock()
|
||||
defer m.mu.Unlock()
|
||||
jar, ok := m.cookies[domain]
|
||||
if !ok {
|
||||
return ""
|
||||
}
|
||||
if key != "" {
|
||||
return jar[key]
|
||||
}
|
||||
return formatCookie(jar)
|
||||
}
|
||||
|
||||
func (m *MemoryState) SetCookie(rawURL, cookie string) {
|
||||
domain := CookieDomain(rawURL)
|
||||
if domain == "" || strings.TrimSpace(cookie) == "" {
|
||||
return
|
||||
}
|
||||
m.mu.Lock()
|
||||
defer m.mu.Unlock()
|
||||
jar, ok := m.cookies[domain]
|
||||
if !ok {
|
||||
jar = map[string]string{}
|
||||
m.cookies[domain] = jar
|
||||
}
|
||||
for k, v := range ParseCookie(cookie) {
|
||||
jar[k] = v
|
||||
}
|
||||
}
|
||||
|
||||
func (m *MemoryState) RemoveCookie(rawURL string) {
|
||||
domain := CookieDomain(rawURL)
|
||||
if domain == "" {
|
||||
return
|
||||
}
|
||||
m.mu.Lock()
|
||||
defer m.mu.Unlock()
|
||||
delete(m.cookies, domain)
|
||||
}
|
||||
|
||||
func (m *MemoryState) Toast(msg string) {
|
||||
m.mu.Lock()
|
||||
defer m.mu.Unlock()
|
||||
m.toasts = append(m.toasts, msg)
|
||||
}
|
||||
|
||||
func (m *MemoryState) OpenBrowser(url, title string) {
|
||||
m.mu.Lock()
|
||||
defer m.mu.Unlock()
|
||||
m.browsers = append(m.browsers, BrowserRequest{URL: url, Title: title})
|
||||
}
|
||||
|
||||
// RequestUIRefresh 实现 UIState(进程内实现,供单测)。
|
||||
func (m *MemoryState) RequestUIRefresh() {
|
||||
m.mu.Lock()
|
||||
defer m.mu.Unlock()
|
||||
m.uiRefresh = true
|
||||
}
|
||||
|
||||
// ApplyLoginData 实现 UIState:合并进 loginInfo。
|
||||
func (m *MemoryState) ApplyLoginData(data map[string]string) {
|
||||
if len(data) == 0 {
|
||||
return
|
||||
}
|
||||
m.mu.Lock()
|
||||
defer m.mu.Unlock()
|
||||
cur := map[string]string{}
|
||||
if m.loginInfo != "" {
|
||||
_ = json.Unmarshal([]byte(m.loginInfo), &cur)
|
||||
}
|
||||
for k, v := range data {
|
||||
cur[k] = v
|
||||
}
|
||||
if b, err := json.Marshal(cur); err == nil {
|
||||
m.loginInfo = string(b)
|
||||
}
|
||||
m.uiRefresh = true
|
||||
}
|
||||
|
||||
// UIRefreshRequested 返回并清空「重画登录表单」标记。
|
||||
func (m *MemoryState) UIRefreshRequested() bool {
|
||||
m.mu.Lock()
|
||||
defer m.mu.Unlock()
|
||||
out := m.uiRefresh
|
||||
m.uiRefresh = false
|
||||
return out
|
||||
}
|
||||
|
||||
// Toasts 返回并清空已收集的宿主提示。
|
||||
func (m *MemoryState) Toasts() []string {
|
||||
m.mu.Lock()
|
||||
defer m.mu.Unlock()
|
||||
out := m.toasts
|
||||
m.toasts = nil
|
||||
return out
|
||||
}
|
||||
|
||||
// Browsers 返回并清空已收集的浏览器地址。
|
||||
func (m *MemoryState) Browsers() []BrowserRequest {
|
||||
m.mu.Lock()
|
||||
defer m.mu.Unlock()
|
||||
out := m.browsers
|
||||
m.browsers = nil
|
||||
return out
|
||||
}
|
||||
|
||||
// Snapshot 返回当前 Cookie(domain → cookie 串)副本。
|
||||
func (m *MemoryState) Snapshot() map[string]string {
|
||||
m.mu.Lock()
|
||||
defer m.mu.Unlock()
|
||||
out := make(map[string]string, len(m.cookies))
|
||||
for d, jar := range m.cookies {
|
||||
out[d] = formatCookie(jar)
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
var _ SourceState = (*MemoryState)(nil)
|
||||
@@ -0,0 +1,117 @@
|
||||
// 测试用 jsLib:模拟聚合类书源的公共函数库结构。
|
||||
//
|
||||
// 真实书源(如光遇聚合)的 jsLib 有十几万字符,但关键结构就是这些:
|
||||
// 用 source.getVariable 做设置读写、用 cookie 存取凭证、用 BaseUrl() 拼接口地址。
|
||||
// 该文件由 JSRunner 在创建时执行一次,其函数对后续所有规则 JS 可见。
|
||||
// 同时保留了源码里常见的 lexical 声明(let hosts),验证跨执行可见性。
|
||||
|
||||
// 当前书源版本号
|
||||
let localVersion = '26.9.29.1';
|
||||
|
||||
// 初始服务器列表
|
||||
let hosts = [
|
||||
'https://v1.example-aggregate.com',
|
||||
'https://v2.example-aggregate.com'
|
||||
];
|
||||
|
||||
// 源变量默认值
|
||||
const defaultConfig = {
|
||||
线路: hosts[0],
|
||||
发现页来源: '番茄'
|
||||
};
|
||||
|
||||
// 获取源变量
|
||||
function getVariable(k) {
|
||||
if (k == undefined) k = "";
|
||||
let parsed = {};
|
||||
try {
|
||||
parsed = JSON.parse(source.getVariable());
|
||||
} catch (e) {}
|
||||
if (k == "") {
|
||||
return parsed;
|
||||
}
|
||||
let value = parsed[k];
|
||||
if (value == undefined) {
|
||||
value = defaultConfig[k];
|
||||
}
|
||||
return value != undefined ? value : "";
|
||||
}
|
||||
|
||||
// 设置源变量
|
||||
function setVariable(k, v, t) {
|
||||
if (t == undefined) t = true;
|
||||
const vs = getVariable();
|
||||
vs[k] = v;
|
||||
source.setVariable(JSON.stringify(vs, null, 4));
|
||||
if (k != '云端配置' && t) {
|
||||
java.toast('设置 ' + k + ' 为 ' + v);
|
||||
}
|
||||
}
|
||||
|
||||
// 获取正在使用的线路
|
||||
function BaseUrl() {
|
||||
let h = getVariable("线路");
|
||||
if (!h || String(h) == "undefined") {
|
||||
h = hosts[0];
|
||||
}
|
||||
return h;
|
||||
}
|
||||
|
||||
// 获取登陆 token
|
||||
function getToken() {
|
||||
let hostsbk = getVariable('云端配置') && getVariable('云端配置')['hosts'] || hosts;
|
||||
for (let i = 0; i < hostsbk.length; i++) {
|
||||
let cookieValue = String(cookie.getCookie(hostsbk[i]));
|
||||
let parts = cookieValue.split(";");
|
||||
for (let j = 0; j < parts.length; j++) {
|
||||
if (parts[j].indexOf("qttoken") != -1) {
|
||||
return parts[j].split("=")[1];
|
||||
}
|
||||
}
|
||||
}
|
||||
return "";
|
||||
}
|
||||
|
||||
// 设置 ck
|
||||
function setAllCookies(ck) {
|
||||
let hostsbk = hosts;
|
||||
for (let i = 0; i < hostsbk.length; i++) {
|
||||
cookie.setCookie(hostsbk[i], ck);
|
||||
}
|
||||
}
|
||||
|
||||
// 移除 ck
|
||||
function removeAllCookies() {
|
||||
for (let i = 0; i < hosts.length; i++) {
|
||||
cookie.removeCookie(hosts[i]);
|
||||
}
|
||||
java.toast('已退出登陆');
|
||||
}
|
||||
|
||||
// 请求封装(自动带 token)
|
||||
function request(url, method, body) {
|
||||
if (method == undefined) method = 'GET';
|
||||
if (body == undefined) body = {};
|
||||
let urla = url;
|
||||
if (url.indexOf('http') != 0) {
|
||||
urla = BaseUrl() + url;
|
||||
}
|
||||
let qttoken = getToken();
|
||||
let options = {
|
||||
method: method,
|
||||
headers: {
|
||||
'cookie': 'qttoken=' + qttoken,
|
||||
'Content-Type': 'application/json'
|
||||
},
|
||||
body: JSON.stringify(body)
|
||||
};
|
||||
return java.ajax(urla + ',' + JSON.stringify(options));
|
||||
}
|
||||
|
||||
function parseJsonSafely(str) {
|
||||
try {
|
||||
return JSON.parse(str);
|
||||
} catch (e) {
|
||||
return null;
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,46 @@
|
||||
[
|
||||
{
|
||||
"name": "线路设置",
|
||||
"type": "button",
|
||||
"action": "switchLine()",
|
||||
"style": { "layout_flexGrow": 1, "layout_flexBasisPercent": 0.4 }
|
||||
},
|
||||
{
|
||||
"name": "邮箱",
|
||||
"type": "text"
|
||||
},
|
||||
{
|
||||
"name": "密码",
|
||||
"type": "password"
|
||||
},
|
||||
{
|
||||
"name": "🔅登录账号",
|
||||
"type": "button",
|
||||
"action": "login(true)",
|
||||
"style": { "layout_flexGrow": 1, "layout_flexBasisPercent": 0.4 }
|
||||
},
|
||||
{
|
||||
"name": "🔐注册账号",
|
||||
"type": "button",
|
||||
"action": "register()",
|
||||
"style": { "layout_flexGrow": 1, "layout_flexBasisPercent": 0.4 }
|
||||
},
|
||||
{
|
||||
"name": " 🔚 退出登录",
|
||||
"type": "button",
|
||||
"action": "logout()",
|
||||
"style": { "layout_flexGrow": 1, "layout_flexBasisPercent": 0.4 }
|
||||
},
|
||||
{
|
||||
"name": "🪪 查看信息",
|
||||
"type": "button",
|
||||
"action": "checkStatus()",
|
||||
"style": { "layout_flexGrow": 1, "layout_flexBasisPercent": 0.4 }
|
||||
},
|
||||
{
|
||||
"name": "🏝用户后台",
|
||||
"type": "button",
|
||||
"action": "user()",
|
||||
"style": { "layout_flexGrow": 1, "layout_flexBasisPercent": 0.4 }
|
||||
}
|
||||
]
|
||||
Some files were not shown because too many files have changed in this diff Show More
Reference in New Issue
Block a user