fix(subscription): apply rules on RSS path, dedup movies via library, MoviePilot-style default excludes

- runOne (RSS) now applies matchesSubscriptionRules so ExcludeWords/Resolution/Quality/Effects/ReleaseGroups take effect (previously only the Filter regex ran, so editing 排除 had no effect on RSS subscriptions)
- movie candidate selection now consults library availability: non-wash subscriptions skip titles already in the library, stopping repeated downloads of 10bit/Dolby releases (aligns with MoviePilot 'download once = satisfied')
- add MoviePilot-style default exclude list (cam/ts/枪版/trailer/sample...) merged with user excludes; latin tokens matched on word boundaries to avoid substring false positives
- frontend: send rule fields as raw strings on edit so 排除词 can be cleared/changed (was dropped by '|| undefined')

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
This commit is contained in:
soldosluka857
2026-05-30 14:36:42 +00:00
committed by Shuke
parent de3f1f3bb4
commit f1d87638f3
3 changed files with 218 additions and 11 deletions
+120 -7
View File
@@ -205,6 +205,13 @@ func (s *SubscriptionService) runOne(ctx context.Context, sub *model.Subscriptio
seenSet[g] = struct{}{}
}
// 非洗版订阅:成功下载一次即满足,预先算一次媒体库可用性用于跳过已入库的电影/剧集(对齐 MoviePilot)。
washOff := !sub.WashEnabled
var avail LocalAvailability
if washOff {
avail = SubscriptionLocalAvailability(ctx, s.repo, sub)
}
queued := 0
for _, item := range feed.Channel.Items {
guid := item.GUID
@@ -217,6 +224,15 @@ func (s *SubscriptionService) runOne(ctx context.Context, sub *model.Subscriptio
if filter != nil && !filter.MatchString(item.Title) {
continue
}
// 应用订阅高级规则(排除词/分辨率/质量/特效/发布组)—— 此前 RSS 路径完全跳过,导致排除不生效。
if !matchesSubscriptionRules(sub, item.Title) {
continue
}
if washOff && subscriptionItemAlreadyAvailable(sub, avail, item.Title) {
seen = append(seen, guid)
seenSet[guid] = struct{}{}
continue
}
download := item.Enclosure.URL
if download == "" {
download = item.Link
@@ -402,15 +418,20 @@ func selectSiteSearchCandidates(results []SearchResult, sub *model.Subscription,
return candidates[i].Item.Size > candidates[j].Item.Size
})
mediaType := normalizeMediaType(sub.MediaType, sub.Name+" "+sub.Filter, "")
if !isSubscriptionSeriesType(mediaType) {
return candidates[:1]
}
var local LocalAvailability
if len(availability) > 0 {
local = availability[0]
}
mediaType := normalizeMediaType(sub.MediaType, sub.Name+" "+sub.Filter, "")
if !isSubscriptionSeriesType(mediaType) {
// 对齐 MoviePilot:非洗版订阅成功下载一次即满足,媒体库/下载中已存在则不再重复下载。
if (sub == nil || !sub.WashEnabled) && local.LocalMediaCount > 0 {
return nil
}
return candidates[:1]
}
if local.LocalMediaCount > 0 {
if local.TotalEpisodes > 0 && len(local.MissingEpisodes) == 0 {
return nil
@@ -451,12 +472,19 @@ func selectSiteSearchCandidates(results []SearchResult, sub *model.Subscription,
return selected
}
// defaultExcludeWords 是参考 MoviePilot 默认过滤的「垃圾版本」排除清单,对所有订阅生效,
// 与用户自定义排除词合并。拉丁词在 containsAnyExcludeToken 里按词边界匹配以避免子串误伤。
const defaultExcludeWords = "cam,ts,tc,telesync,telecine,hdcam,hdts,枪版,抢先,抢鲜,预告,trailer,sample"
func matchesSubscriptionRules(sub *model.Subscription, title string) bool {
titleFold := strings.ToLower(title)
if containsAnyExcludeToken(titleFold, defaultExcludeWords) {
return false
}
if sub == nil {
return true
}
titleFold := strings.ToLower(title)
if sub.ExcludeWords != "" && containsAnyToken(titleFold, sub.ExcludeWords) {
if sub.ExcludeWords != "" && containsAnyExcludeToken(titleFold, sub.ExcludeWords) {
return false
}
if sub.ReleaseGroups != "" && !containsAnyToken(titleFold, sub.ReleaseGroups) {
@@ -521,6 +549,63 @@ func containsAnyToken(titleFold, csv string) bool {
return false
}
// containsAnyExcludeToken 用于排除词匹配:纯 ASCII 字母数字的词按词边界匹配(避免 "ts"
// 误伤 "tsukihime"、"cam" 误伤 "camp" 之类的子串误判),含 CJK/符号的词仍按子串匹配。
func containsAnyExcludeToken(titleFold, csv string) bool {
for _, token := range strings.FieldsFunc(strings.ToLower(csv), func(r rune) bool {
return r == ',' || r == '/' || r == '|' || r == ';' || r == ','
}) {
token = strings.TrimSpace(token)
if token == "" {
continue
}
if isASCIIWordToken(token) {
if matchesWordBoundary(titleFold, token) {
return true
}
continue
}
if strings.Contains(titleFold, token) {
return true
}
}
return false
}
func isASCIIWordToken(token string) bool {
for _, r := range token {
if r > unicode.MaxASCII || !(unicode.IsLetter(r) || unicode.IsDigit(r)) {
return false
}
}
return token != ""
}
// matchesWordBoundary 判断 token 是否作为独立词出现在 title 中,词边界为「非字母数字」。
func matchesWordBoundary(titleFold, token string) bool {
isWordRune := func(r rune) bool {
return unicode.IsLetter(r) || unicode.IsDigit(r)
}
from := 0
for {
idx := strings.Index(titleFold[from:], token)
if idx < 0 {
return false
}
start := from + idx
end := start + len(token)
leftOK := start == 0 || !isWordRune(rune(titleFold[start-1]))
rightOK := end >= len(titleFold) || !isWordRune(rune(titleFold[end]))
if leftOK && rightOK {
return true
}
from = start + 1
if from >= len(titleFold) {
return false
}
}
}
func containsAnyEffect(titleFold, csv string) bool {
for _, token := range strings.FieldsFunc(strings.ToLower(csv), func(r rune) bool {
return r == ',' || r == '/' || r == '|' || r == ';' || r == ','
@@ -742,6 +827,34 @@ func mergeLocalAvailability(values ...LocalAvailability) LocalAvailability {
return out
}
// subscriptionItemAlreadyAvailable 判断某个订阅条目(按其标题解析出的季/集)是否已在媒体库存在。
// 电影/无集号条目:媒体库已有该片即视为已存在;剧集条目:对应季集已入库即视为已存在。
func subscriptionItemAlreadyAvailable(sub *model.Subscription, avail LocalAvailability, title string) bool {
if avail.LocalMediaCount == 0 {
return false
}
if !isSubscriptionSeriesType(subscriptionMediaType(sub)) {
return true
}
wantSeason, wantEpisode := ParseEpisode(title)
if wantEpisode <= 0 {
// 整季合集 / 无法解析集号:库里已有内容时保守跳过,避免重复整季下载。
return true
}
if wantSeason <= 0 {
wantSeason = 1
}
_, exists := avail.ExistingEpisodeKeys[episodeKey(wantSeason, wantEpisode)]
return exists
}
func subscriptionMediaType(sub *model.Subscription) string {
if sub == nil {
return ""
}
return sub.MediaType
}
func (s *SubscriptionService) downloadPathHasCandidate(ctx context.Context, sub *model.Subscription, title, savePath string) bool {
savePath = strings.TrimSpace(savePath)
if savePath == "" {
+93
View File
@@ -192,3 +192,96 @@ func TestSubscriptionPendingDownloadAvailabilitySkipsUnorganizedEpisodes(t *test
t.Fatal("did not expect missing E03 to be detected")
}
}
func TestMatchesSubscriptionRulesUserExcludeWords(t *testing.T) {
sub := &model.Subscription{ExcludeWords: "10bit,dolby vision,杜比"}
cases := []struct {
title string
want bool
}{
{"Movie 2024 1080p WEB-DL", true},
{"Movie 2024 2160p 10bit HEVC", false},
{"Movie 2024 2160p Dolby Vision", false},
{"电影 2024 杜比全景声", false},
}
for _, c := range cases {
if got := matchesSubscriptionRules(sub, c.title); got != c.want {
t.Errorf("matchesSubscriptionRules(%q) = %v, want %v", c.title, got, c.want)
}
}
}
func TestMatchesSubscriptionRulesDefaultExcludesJunkReleases(t *testing.T) {
sub := &model.Subscription{}
for _, title := range []string{
"Some Movie 2024 CAM",
"Some Movie 2024 HDTS",
"某电影 2024 枪版",
"Some Movie 2024 TELESYNC",
"Some Show 预告",
} {
if matchesSubscriptionRules(sub, title) {
t.Errorf("expected default rules to exclude junk release %q", title)
}
}
}
func TestMatchesSubscriptionRulesWordBoundaryAvoidsFalsePositives(t *testing.T) {
sub := &model.Subscription{}
// "ts" / "cam" / "tc" 作为子串出现在合法标题里时不应被默认排除误伤。
for _, title := range []string{
"Tsukihime 2024 1080p WEB-DL",
"Camp Rock 2024 1080p BluRay",
"Catch Me 2024 1080p WEB-DL",
} {
if !matchesSubscriptionRules(sub, title) {
t.Errorf("word-boundary match wrongly excluded %q", title)
}
}
}
func TestSelectSiteSearchCandidatesSkipsExistingMovieWhenNotWashing(t *testing.T) {
sub := &model.Subscription{Name: "Inception 自动订阅", Filter: "Inception 2010", MediaType: "movie"}
results := []SearchResult{
{Title: "Inception 2010 2160p 10bit Dolby Vision Atmos", DownloadURL: "https://pt/download/dovi", Seeders: 500},
{Title: "Inception 2010 1080p WEB-DL", DownloadURL: "https://pt/download/web", Seeders: 90},
}
availability := LocalAvailability{LocalMediaCount: 1, InLibrary: true, DownloadedEpisodes: 1, TotalEpisodes: 1}
got := selectSiteSearchCandidates(results, sub, map[string]struct{}{}, availability)
if len(got) != 0 {
t.Fatalf("selected %#v, want none (movie already in library, wash disabled)", got)
}
}
func TestSelectSiteSearchCandidatesAllowsMovieWashUpgrade(t *testing.T) {
sub := &model.Subscription{Name: "Inception 自动订阅", Filter: "Inception 2010", MediaType: "movie", WashEnabled: true, WashPriority: "resolution"}
results := []SearchResult{
{Title: "Inception 2010 2160p REMUX", DownloadURL: "https://pt/download/2160", Seeders: 80},
{Title: "Inception 2010 1080p WEB-DL", DownloadURL: "https://pt/download/1080", Seeders: 200},
}
availability := LocalAvailability{LocalMediaCount: 1, InLibrary: true, DownloadedEpisodes: 1, TotalEpisodes: 1}
got := selectSiteSearchCandidates(results, sub, map[string]struct{}{}, availability)
if len(got) != 1 || got[0].Download != "https://pt/download/2160" {
t.Fatalf("selected %#v, want 2160p upgrade allowed when washing", got)
}
}
func TestSubscriptionItemAlreadyAvailable(t *testing.T) {
movieSub := &model.Subscription{MediaType: "movie"}
if !subscriptionItemAlreadyAvailable(movieSub, LocalAvailability{LocalMediaCount: 1}, "Inception 2010 2160p") {
t.Fatal("movie already in library should be reported available")
}
if subscriptionItemAlreadyAvailable(movieSub, LocalAvailability{}, "Inception 2010 2160p") {
t.Fatal("empty library should not be reported available")
}
tvSub := &model.Subscription{MediaType: "tv"}
avail := LocalAvailability{LocalMediaCount: 1, ExistingEpisodeKeys: map[string]struct{}{episodeKey(1, 2): {}}}
if !subscriptionItemAlreadyAvailable(tvSub, avail, "Show S01E02 1080p") {
t.Fatal("existing episode should be reported available")
}
if subscriptionItemAlreadyAvailable(tvSub, avail, "Show S01E03 1080p") {
t.Fatal("missing episode should not be reported available")
}
}
+5 -4
View File
@@ -50,10 +50,11 @@ export function SubscriptionsPage() {
search_mode: searchMode,
imdb_id: imdbID || undefined,
resolution,
quality: quality || undefined,
effects: effects || undefined,
release_groups: releaseGroups || undefined,
exclude_words: excludeWords || undefined,
// 规则类字段发送原始字符串(含空串),以便编辑时可清空/修改;create 存空串无害,update 才能真正生效。
quality,
effects,
release_groups: releaseGroups,
exclude_words: excludeWords,
wash_enabled: washEnabled,
wash_priority: washPriority,
priority: 50,