优化在线字幕

2026-09-28 03:06:38 +08:00 · 2026-09-17 23:53:22 +08:00
parent ae8db26c1e
commit f4abdf78e2
3 changed files with 374 additions and 9 deletions
@@ -410,14 +410,10 @@ func (s *SubtitleService) Serve(ctx context.Context, mediaID, sub string, w io.W
 		}
 		return s.serveEmbedded(ctx, m, index, w)
 	}
-	abs, err := filepath.Abs(sub)
+	abs, err := readExternalSubtitlePath(m, sub)
 	if err != nil {
 		return err
 	}
-	mediaDir, _ := filepath.Abs(filepath.Dir(m.Path))
-	if !pathWithin(abs, mediaDir) {
-		return fmt.Errorf("path escape")
-	}

 	f, err := os.Open(abs) // #nosec G304 -- abs is constrained to the media file directory with pathWithin.
 	if err != nil {
@@ -428,14 +424,17 @@ func (s *SubtitleService) Serve(ctx context.Context, mediaID, sub string, w io.W
 	if err != nil {
 		return err
 	}
+	// 非 UTF-8 的外挂字幕（UTF-16、GBK/Big5 等）必须先归一化：浏览器只能按
+	// UTF-8 解析 <track> 内容，否则整篇都会变成替换字符。
+	text := decodeSubtitleText(body)

 	switch strings.ToLower(filepath.Ext(abs)) {
 	case ".vtt":
-		_, err = w.Write(body)
+		_, err = io.WriteString(w, text)
 	case ".srt":
-		_, err = w.Write([]byte(srtToVTT(string(body))))
+		_, err = io.WriteString(w, srtToVTT(text))
 	case ".ass", ".ssa":
-		_, err = w.Write([]byte(assToVTT(string(body))))
+		_, err = io.WriteString(w, assToVTT(text))
 	default:
 		return errors.New("unsupported subtitle format")
 	}
@@ -517,12 +516,42 @@ func (s *SubtitleService) ServeASS(ctx context.Context, mediaID, sub string, w i
 	}
 	switch strings.ToLower(filepath.Ext(sub)) {
 	case ".ass", ".ssa":
-		return s.ServeRaw(ctx, mediaID, sub, w)
+		// 外挂 ASS 同样要归一化成 UTF-8：libass 只认 UTF-8，UTF-16/GBK 的
+		// 字幕交给它会解析不到任何事件，表现为「字幕选中了却不显示」。
+		abs, err := readExternalSubtitlePath(m, sub)
+		if err != nil {
+			return err
+		}
+		f, err := os.Open(abs) // #nosec G304 -- abs is constrained to the media file directory with pathWithin.
+		if err != nil {
+			return err
+		}
+		defer f.Close()
+		body, err := io.ReadAll(f)
+		if err != nil {
+			return err
+		}
+		_, err = io.WriteString(w, decodeSubtitleText(body))
+		return err
 	default:
 		return errors.New("subtitle is not ASS/SSA")
 	}
 }

+// readExternalSubtitlePath 校验外挂字幕路径（必须落在媒体文件所在目录内）并返回
+// 绝对路径。Serve 与 ServeASS 共用，避免两处各自实现出现安全口径不一致。
+func readExternalSubtitlePath(m *model.Media, sub string) (string, error) {
+	abs, err := filepath.Abs(sub)
+	if err != nil {
+		return "", err
+	}
+	mediaDir, _ := filepath.Abs(filepath.Dir(m.Path))
+	if !pathWithin(abs, mediaDir) {
+		return "", fmt.Errorf("path escape")
+	}
+	return abs, nil
+}
+
 func (s *SubtitleService) serveEmbeddedASS(ctx context.Context, media *model.Media, streamIndex int, w io.Writer) error {
 	input, err := s.resolveInput(ctx, media)
 	if err != nil {
@@ -0,0 +1,167 @@
+package service
+
+import (
+	"bytes"
+	"encoding/binary"
+	"math"
+	"regexp"
+	"strings"
+	"unicode"
+	"unicode/utf16"
+	"unicode/utf8"
+
+	"golang.org/x/text/encoding"
+	"golang.org/x/text/encoding/japanese"
+	"golang.org/x/text/encoding/korean"
+	"golang.org/x/text/encoding/simplifiedchinese"
+	"golang.org/x/text/encoding/traditionalchinese"
+)
+
+// 外挂字幕文件的编码归一化。
+//
+// 字幕文件由字幕组/工具生成，编码并不统一：除 UTF-8 外，中文圈常见 GBK/GB18030，
+// 港台常见 Big5，部分 Windows 工具（Aegisub 早期版本、Subtitle Edit 的某些导出）
+// 会写成带 BOM 的 UTF-16LE。浏览器 <track> 与 libass（JASSUB）都只按 UTF-8 解析，
+// 后端如果原样下发字节，非 UTF-8 的字幕会整篇变成替换字符——表现就是「字幕轨已经
+// 选中，但画面上一个字都不显示」。所以外挂字幕在服务端就要统一解码成 UTF-8。
+//
+// 注意：只用于下发给浏览器/播放器的文本路径。Emby 兼容层的 /subtitles 原始字节
+// 接口（ServeRaw）必须保持字节一致，不能经过这里。
+
+// 各编码的 BOM 前缀。
+var (
+	utf8BOM    = []byte{0xEF, 0xBB, 0xBF}
+	utf16BEBOM = []byte{0xFE, 0xFF}
+	utf16LEBOM = []byte{0xFF, 0xFE}
+)
+
+// subtitleLegacyEncodings 是无 BOM 时的候选编码，按优先级排列。
+// GB18030 是 GBK/GB2312 的超集，中文外挂字幕绝大多数是它，因此排在最前——
+// 多个候选都能解出合法字幕时，靠前的优先。
+var subtitleLegacyEncodings = []struct {
+	name string
+	enc  encoding.Encoding
+}{
+	{"gb18030", simplifiedchinese.GB18030},
+	{"big5", traditionalchinese.Big5},
+	{"shift_jis", japanese.ShiftJIS},
+	{"euc-kr", korean.EUCKR},
+}
+
+// subtitleTimecodePattern 匹配 SRT（00:00:01,000 --> ...）与 WebVTT
+// （00:00:01.000 --> ...）的时间行。
+var subtitleTimecodePattern = regexp.MustCompile(`\d{1,2}:\d{2}:\d{2}[,.]\d{1,3}\s*-->`)
+
+// decodeSubtitleText 把字幕文件字节归一化成 UTF-8 文本。识别不出编码时原样返回，
+// 保证不会比「直接透传」更糟。
+func decodeSubtitleText(raw []byte) string {
+	switch {
+	case bytes.HasPrefix(raw, utf8BOM):
+		return string(raw[len(utf8BOM):])
+	case bytes.HasPrefix(raw, utf16LEBOM):
+		return decodeUTF16(raw[len(utf16LEBOM):], false)
+	case bytes.HasPrefix(raw, utf16BEBOM):
+		return decodeUTF16(raw[len(utf16BEBOM):], true)
+	}
+
+	// 无 BOM 且是合法 UTF-8：这是绝大多数情况，直接返回。
+	if utf8.Valid(raw) {
+		return string(raw)
+	}
+
+	best := ""
+	bestScore := math.MinInt
+	consider := func(text string) {
+		if !utf8.ValidString(text) || !looksLikeSubtitle(text) {
+			return
+		}
+		score := subtitlePlausibility(text)
+		if score > bestScore {
+			best, bestScore = text, score
+		}
+	}
+
+	// 无 BOM 的 UTF-16 少见（多来自手工编辑），但一旦是它，所有单字节编码都会
+	// 解出带大量 NUL 的垃圾，因此先用「格式特征」筛掉这些候选。
+	consider(decodeUTF16(raw, false))
+	consider(decodeUTF16(raw, true))
+
+	for _, candidate := range subtitleLegacyEncodings {
+		text, err := candidate.enc.NewDecoder().Bytes(raw)
+		if err != nil {
+			continue
+		}
+		consider(string(text))
+	}
+	if best != "" {
+		return best
+	}
+	return string(raw)
+}
+
+// decodeUTF16 按给定字节序把 UTF-16 字节解码为字符串（含代理对）。
+func decodeUTF16(raw []byte, bigEndian bool) string {
+	units := make([]uint16, 0, len(raw)/2)
+	for i := 0; i+1 < len(raw); i += 2 {
+		if bigEndian {
+			units = append(units, binary.BigEndian.Uint16(raw[i:i+2]))
+		} else {
+			units = append(units, binary.LittleEndian.Uint16(raw[i:i+2]))
+		}
+	}
+	return string(utf16.Decode(units))
+}
+
+// looksLikeSubtitle 判断解码结果是否真的像一份字幕。解码器「没报错」并不代表
+// 解对了——很多编码能把任意字节解成合法文本，只有格式特征才能确认。
+func looksLikeSubtitle(text string) bool {
+	head := text
+	if len(head) > 4096 {
+		head = head[:4096]
+	}
+	if strings.Contains(head, "[Script Info]") || strings.Contains(head, "Dialogue:") {
+		return true // ASS/SSA
+	}
+	if strings.HasPrefix(strings.TrimSpace(head), "WEBVTT") {
+		return true
+	}
+	return subtitleTimecodePattern.MatchString(head) // SRT / WebVTT 时间行
+}
+
+// subtitlePlausibility 给解码结果打分，返回 -100..100 的可信度：正常文本字符加分，
+// 替换字符（解码失败）、控制字符、未分配码位扣分。
+//
+// 刻意用比例而不是字符数量：单字节回退会把一个双字节字符拆成两个字符，字符多的
+// 候选看起来「更正常」，实际是解错了——GBK 中文字幕被当成 Shift_JIS 解，就会得到
+// 一串半角片假名，数量比正确结果还多。半角片假名单独扣分正是为了压住这种情况。
+func subtitlePlausibility(text string) int {
+	total := 0
+	weight := 0
+	for _, r := range text {
+		total++
+		switch {
+		case r == utf8.RuneError:
+			weight -= 8
+		case r == '\n' || r == '\r' || r == '\t':
+			// 正常换行不计分也不扣分。
+		case r < 0x20 || r == 0x7F:
+			weight -= 8
+		case unicode.Is(unicode.Co, r) || unicode.Is(unicode.Cn, r):
+			weight -= 4
+		case r >= 0xFF61 && r <= 0xFF9F:
+			// 半角片假名：GBK 汉字被按 Shift_JIS 解出来的典型产物。
+			weight -= 2
+		case unicode.In(r, unicode.Han, unicode.Hiragana, unicode.Katakana, unicode.Hangul),
+			unicode.Is(unicode.Latin, r), unicode.IsDigit(r), unicode.IsPunct(r):
+			weight++
+		case unicode.IsSpace(r):
+			// 空格（含全角空格）不参与打分。
+		default:
+			weight--
+		}
+	}
+	if total == 0 {
+		return 0
+	}
+	return weight * 100 / total
+}
@@ -0,0 +1,169 @@
+package service
+
+import (
+	"bytes"
+	"encoding/binary"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+	"unicode/utf16"
+
+	"github.com/glebarez/sqlite"
+	"go.uber.org/zap"
+	"golang.org/x/text/encoding/simplifiedchinese"
+	"gorm.io/gorm"
+
+	"github.com/truewhile/MeBox/internal/config"
+	"github.com/truewhile/MeBox/internal/model"
+	"github.com/truewhile/MeBox/internal/repository"
+)
+
+const assFixture = "[Script Info]\n" +
+	"Title:Railgun 01 BD\n" +
+	"ScriptType:v4.00+\n" +
+	"\n" +
+	"[Events]\n" +
+	"Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text\n" +
+	"Dialogue: 0,0:00:01.00,0:00:02.00,Default,,0,0,0,,放て！心に刻んだ夢を\n"
+
+const srtFixture = "1\n" +
+	"00:00:01,000 --> 00:00:02,000\n" +
+	"只有我的超电磁炮\n" +
+	"\n"
+
+func encodeUTF16(s string, bigEndian bool, withBOM bool) []byte {
+	units := utf16.Encode([]rune(s))
+	out := make([]byte, 0, len(units)*2+2)
+	if withBOM {
+		if bigEndian {
+			out = append(out, 0xFE, 0xFF)
+		} else {
+			out = append(out, 0xFF, 0xFE)
+		}
+	}
+	buf := make([]byte, 2)
+	for _, unit := range units {
+		if bigEndian {
+			binary.BigEndian.PutUint16(buf, unit)
+		} else {
+			binary.LittleEndian.PutUint16(buf, unit)
+		}
+		out = append(out, buf...)
+	}
+	return out
+}
+
+func encodeGBK(t *testing.T, s string) []byte {
+	t.Helper()
+	encoded, err := simplifiedchinese.GBK.NewEncoder().Bytes([]byte(s))
+	if err != nil {
+		t.Fatalf("encode GBK: %v", err)
+	}
+	return encoded
+}
+
+// 这是本次线上问题的核心场景：字幕组给的 ASS 是带 BOM 的 UTF-16LE，
+// 之前原样下发，浏览器按 UTF-8 解开后整篇是替换字符，一个字都渲染不出来。
+func TestDecodeSubtitleTextNormalisesLegacyEncodings(t *testing.T) {
+	cases := []struct {
+		name string
+		raw  []byte
+	}{
+		{"utf8", []byte(assFixture)},
+		{"utf8-bom", append([]byte{0xEF, 0xBB, 0xBF}, []byte(assFixture)...)},
+		{"utf16le-bom", encodeUTF16(assFixture, false, true)},
+		{"utf16be-bom", encodeUTF16(assFixture, true, true)},
+		{"utf16le-no-bom", encodeUTF16(assFixture, false, false)},
+		{"gb18030", encodeGBK(t, assFixture)},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			got := decodeSubtitleText(tc.raw)
+			if !strings.Contains(got, "[Script Info]") {
+				t.Fatalf("decoded text lost ASS header: %q", got)
+			}
+			if !strings.Contains(got, "Dialogue:") {
+				t.Fatalf("decoded text lost Dialogue line: %q", got)
+			}
+			if !strings.Contains(got, "放て！心に刻んだ夢を") {
+				t.Fatalf("decoded text lost non-ASCII content: %q", got)
+			}
+			if strings.ContainsRune(got, 0) || strings.ContainsRune(got, '\uFFFD') {
+				t.Fatalf("decoded text still contains NUL or replacement runes: %q", got)
+			}
+		})
+	}
+}
+
+func TestDecodeSubtitleTextKeepsUnknownBytes(t *testing.T) {
+	raw := []byte("not a subtitle at all: \x81\x82\x83\x84")
+	if got := decodeSubtitleText(raw); got != string(raw) {
+		t.Fatalf("decodeSubtitleText mangled non-subtitle bytes: %q", got)
+	}
+}
+
+// 接口层回归：外挂 UTF-16 字幕经 /subtitles/:id/ass（libass 路径）与
+// /subtitles/:id（WebVTT 降级路径）下发时都必须是可解析的 UTF-8。
+func TestSubtitleHandlersNormaliseUTF16Sidecar(t *testing.T) {
+	db, err := gorm.Open(sqlite.Open("file:subtitle-encoding?mode=memory&cache=shared"), &gorm.Config{})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if err := db.AutoMigrate(&model.Library{}, &model.Media{}); err != nil {
+		t.Fatal(err)
+	}
+
+	dir := t.TempDir()
+	videoPath := filepath.Join(dir, "Railgun S01E01.mkv")
+	assPath := filepath.Join(dir, "Railgun S01E01.ass")
+	srtPath := filepath.Join(dir, "Railgun S01E01.zh.srt")
+	if err := os.WriteFile(videoPath, []byte("video"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	if err := os.WriteFile(assPath, encodeUTF16(assFixture, false, true), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	if err := os.WriteFile(srtPath, encodeGBK(t, srtFixture), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	media := model.Media{Title: "Railgun", Path: videoPath}
+	if err := db.Create(&media).Error; err != nil {
+		t.Fatal(err)
+	}
+	svc := NewSubtitleService(&config.Config{}, zap.NewNop(), repository.New(db))
+
+	var buf bytes.Buffer
+	if err := svc.ServeASS(t.Context(), media.ID, assPath, &buf); err != nil {
+		t.Fatalf("ServeASS: %v", err)
+	}
+	ass := buf.String()
+	if !bytes.HasPrefix(buf.Bytes(), []byte("[Script Info]")) {
+		t.Fatalf("ServeASS output is not UTF-8 ASS: %q", ass)
+	}
+	if !strings.Contains(ass, "放て！心に刻んだ夢を") {
+		t.Fatalf("ServeASS output lost dialogue text: %q", ass)
+	}
+
+	buf.Reset()
+	if err := svc.Serve(t.Context(), media.ID, assPath, &buf); err != nil {
+		t.Fatalf("Serve(ass): %v", err)
+	}
+	vtt := buf.String()
+	if !strings.HasPrefix(vtt, "WEBVTT") || !strings.Contains(vtt, "-->") {
+		t.Fatalf("ASS->WebVTT fallback produced no cues: %q", vtt)
+	}
+	if !strings.Contains(vtt, "放て！心に刻んだ夢を") {
+		t.Fatalf("ASS->WebVTT fallback lost dialogue text: %q", vtt)
+	}
+
+	buf.Reset()
+	if err := svc.Serve(t.Context(), media.ID, srtPath, &buf); err != nil {
+		t.Fatalf("Serve(srt): %v", err)
+	}
+	vtt = buf.String()
+	if !strings.Contains(vtt, "只有我的超电磁炮") {
+		t.Fatalf("GBK SRT was not transcoded: %q", vtt)
+	}
+}