diff --git a/README.md b/README.md index 0ae5bdfaf..efecfdbcb 100644 --- a/README.md +++ b/README.md @@ -52,6 +52,7 @@ A share of the revenue helps fund the development of Navidrome at no additional - **Multi-platform**, runs on macOS, Linux and Windows. **Docker** images are also provided - Ready to use binaries for all major platforms, including **Raspberry Pi** - Automatically **monitors your library** for changes, importing new files and reloading new metadata + - Supports lyrics from sidecar **.ttml**, **.elrc**, **.lrc**, **.srt**, **.txt** files and embedded **TTML**, **Enhanced LRC**, **LRC**, **SRT**, and plain-text tags (via `lyricspriority`) - **Themeable**, modern and responsive **Web interface** based on [Material UI](https://material-ui.com) - **Compatible** with all Subsonic/Madsonic/Airsonic [clients](https://www.navidrome.org/docs/overview/#apps) - **Transcoding** on the fly. Can be set per user/player. **Opus encoding is supported** diff --git a/conf/configuration.go b/conf/configuration.go index 08f12fc94..219b7b0bd 100644 --- a/conf/configuration.go +++ b/conf/configuration.go @@ -776,7 +776,7 @@ func setViperDefaults() { viper.SetDefault("artistartpriority", "artist.*, album/artist.*, external") viper.SetDefault("artistimagefolder", "") viper.SetDefault("discartpriority", "disc*.*, cd*.*, cover.*, folder.*, front.*, discsubtitle, embedded") - viper.SetDefault("lyricspriority", ".lrc,.txt,embedded") + viper.SetDefault("lyricspriority", ".ttml,.elrc,.lrc,.srt,.txt,embedded") viper.SetDefault("enablegravatar", false) viper.SetDefault("enablefavourites", true) viper.SetDefault("enablestarrating", true) diff --git a/core/lyrics/embedded.go b/core/lyrics/embedded.go new file mode 100644 index 000000000..4cdd898ee --- /dev/null +++ b/core/lyrics/embedded.go @@ -0,0 +1,64 @@ +package lyrics + +import ( + "encoding/xml" + "strings" + + "github.com/navidrome/navidrome/log" + "github.com/navidrome/navidrome/model" +) + +// ParseEmbedded parses lyrics read from media-file metadata tags. It detects rich +// payloads before falling back to the generic LRC/plain-text parser, because +// text sanitization would otherwise strip TTML XML markup. +func ParseEmbedded(language, text string) (model.LyricList, error) { + text = strings.TrimPrefix(text, "\ufeff") + + if isTTMLDocument(text) { + list, err := parseTTMLWithDefaultLang([]byte(text), language) + if err == nil && len(list) > 0 { + return list, nil + } + if err != nil { + log.Warn("Error parsing embedded TTML lyrics, falling back to plain lyrics", "error", err) + } + } + + list, err := parseSRTWithLanguage([]byte(text), language) + if err == nil && len(list) > 0 { + return list, nil + } + if err != nil && strings.Contains(text, "-->") { + log.Warn("Error parsing embedded SRT lyrics, falling back to plain lyrics", "error", err) + } + + lyric, err := model.ToLyrics(language, text) + if err != nil { + return nil, err + } + if lyric == nil || lyric.IsEmpty() { + return nil, nil + } + return model.LyricList{*lyric}, nil +} + +func isTTMLDocument(text string) bool { + decoder := xml.NewDecoder(strings.NewReader(strings.TrimSpace(text))) + for { + token, err := decoder.Token() + if err != nil { + return false + } + if start, ok := token.(xml.StartElement); ok { + return strings.EqualFold(start.Name.Local, "tt") + } + } +} + +func normalizeEmbeddedLanguage(language string) string { + language = strings.ToLower(strings.TrimSpace(language)) + if language == "" { + return "xxx" + } + return language +} diff --git a/core/lyrics/embedded_test.go b/core/lyrics/embedded_test.go new file mode 100644 index 000000000..d875bbdcf --- /dev/null +++ b/core/lyrics/embedded_test.go @@ -0,0 +1,169 @@ +package lyrics + +import ( + "strings" + + "github.com/navidrome/navidrome/model" + . "github.com/onsi/ginkgo/v2" + . "github.com/onsi/gomega" +) + +var _ = Describe("ParseEmbedded", func() { + It("should parse embedded TTML with the tag language as the default", func() { + content := ` + + + + Lead Vocal + + + + +
+

+ Hello world +

+
+ +` + + list, err := ParseEmbedded("ENG", content) + + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(1)) + Expect(list[0].Kind).To(Equal("main")) + Expect(list[0].Lang).To(Equal("eng")) + Expect(list[0].Synced).To(BeTrue()) + Expect(list[0].Agents).To(Equal([]model.Agent{{ID: "lead", Role: "main", Name: "Lead Vocal"}})) + Expect(list[0].Line).To(HaveLen(1)) + Expect(list[0].Line[0].Start).To(Equal(ptr(int64(1000)))) + Expect(list[0].Line[0].End).To(Equal(ptr(int64(3000)))) + Expect(list[0].Line[0].Value).To(Equal("Hello world")) + Expect(list[0].Line[0].Cue).To(HaveLen(2)) + Expect(list[0].Line[0].Cue[0].AgentID).To(Equal("lead")) + Expect(list[0].Line[0].Cue[0].ByteStart).To(Equal(0)) + Expect(list[0].Line[0].Cue[0].ByteEnd).To(Equal(5)) + Expect(list[0].Line[0].Cue[1].ByteStart).To(Equal(6)) + Expect(list[0].Line[0].Cue[1].ByteEnd).To(Equal(10)) + }) + + It("should preserve embedded TTML translation and pronunciation tracks", func() { + content := ` + + + + + + Hola + + + + + konni + + + + + + +
+

こんにちは

+
+ +` + + list, err := ParseEmbedded("eng", content) + + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(3)) + Expect(list[0].Kind).To(Equal("main")) + Expect(list[0].Lang).To(Equal("ja")) + Expect(list[0].Line[0].Value).To(Equal("こんにちは")) + Expect(list[1].Kind).To(Equal("translation")) + Expect(list[1].Lang).To(Equal("es")) + Expect(list[1].Line[0].Value).To(Equal("Hola")) + Expect(list[2].Kind).To(Equal("pronunciation")) + Expect(list[2].Lang).To(Equal("ja-latn")) + Expect(list[2].Line[0].Value).To(Equal("konni")) + Expect(list[2].Line[0].Cue).To(HaveLen(2)) + }) + + It("should parse embedded SRT with the tag language", func() { + content := `1 +00:00:18,800 --> 00:00:22,800 +We're from subtitles + +2 +00:00:22,801 --> 00:00:26,000 +Another subtitle line` + + list, err := ParseEmbedded("POR", content) + + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(Equal(model.LyricList{ + { + Lang: "por", + Line: []model.Line{ + { + Start: ptr(int64(18800)), + End: ptr(int64(22800)), + Value: "We're from subtitles", + }, + { + Start: ptr(int64(22801)), + End: ptr(int64(26000)), + Value: "Another subtitle line", + }, + }, + Synced: true, + }, + })) + }) + + It("should parse embedded SRT blocks separated by whitespace-only blank lines", func() { + content := "1\n00:00:01,000 --> 00:00:02,000\nFirst subtitle\n \n2\n00:00:03,000 --> 00:00:04,000\nSecond subtitle" + + list, err := ParseEmbedded("eng", content) + + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(1)) + Expect(list[0].Line).To(Equal([]model.Line{ + {Start: ptr(int64(1000)), End: ptr(int64(2000)), Value: "First subtitle"}, + {Start: ptr(int64(3000)), End: ptr(int64(4000)), Value: "Second subtitle"}, + })) + }) + + It("should keep embedded enhanced LRC cues", func() { + content := "[00:01.00]<00:01.00>Lead <00:01.50>words" + + list, err := ParseEmbedded("eng", content) + + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(1)) + Expect(list[0].Lang).To(Equal("eng")) + Expect(list[0].Synced).To(BeTrue()) + Expect(list[0].Line[0].Value).To(Equal("Lead words")) + Expect(list[0].Line[0].Cue).To(HaveLen(2)) + }) + + It("should fall back to plain lyrics when embedded TTML is invalid", func() { + content := ` + +

Broken

+ +` + + list, err := ParseEmbedded("eng", content) + + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(1)) + Expect(list[0].Lang).To(Equal("eng")) + Expect(list[0].Synced).To(BeFalse()) + Expect(list[0].Line).ToNot(BeEmpty()) + values := make([]string, 0, len(list[0].Line)) + for _, line := range list[0].Line { + values = append(values, line.Value) + } + Expect(strings.Join(values, "\n")).To(ContainSubstring("Broken")) + }) +}) diff --git a/core/lyrics/external_test_helpers_test.go b/core/lyrics/external_test_helpers_test.go new file mode 100644 index 000000000..6fd41aba4 --- /dev/null +++ b/core/lyrics/external_test_helpers_test.go @@ -0,0 +1,5 @@ +package lyrics_test + +func ptr[T any](v T) *T { + return &v +} diff --git a/core/lyrics/lyrics.go b/core/lyrics/lyrics.go index 758053042..73a0479f3 100644 --- a/core/lyrics/lyrics.go +++ b/core/lyrics/lyrics.go @@ -14,6 +14,12 @@ type Lyrics interface { GetLyrics(ctx context.Context, mf *model.MediaFile) (model.LyricList, error) } +// BatchLyrics can resolve lyrics across multiple candidate media files while +// still honoring the configured source priority globally. +type BatchLyrics interface { + GetLyricsForMediaFiles(ctx context.Context, mediaFiles []model.MediaFile) (model.LyricList, error) +} + // PluginLoader discovers and loads lyrics provider plugins. type PluginLoader interface { LoadLyricsProvider(name string) (Lyrics, bool) @@ -32,28 +38,53 @@ func NewLyrics(pluginLoader PluginLoader) Lyrics { // GetLyrics returns lyrics for the given media file, trying sources in the // order specified by conf.Server.LyricsPriority. func (l *lyricsService) GetLyrics(ctx context.Context, mf *model.MediaFile) (model.LyricList, error) { - var lyricsList model.LyricList - var err error + return l.getLyricsForCandidates(ctx, []*model.MediaFile{mf}) +} +// GetLyricsForMediaFiles resolves lyrics across duplicate media files while +// preserving the configured source priority across the full candidate set. +func (l *lyricsService) GetLyricsForMediaFiles(ctx context.Context, mediaFiles []model.MediaFile) (model.LyricList, error) { + candidates := make([]*model.MediaFile, 0, len(mediaFiles)) + for i := range mediaFiles { + candidates = append(candidates, &mediaFiles[i]) + } + return l.getLyricsForCandidates(ctx, candidates) +} + +func (l *lyricsService) getLyricsForCandidates(ctx context.Context, mediaFiles []*model.MediaFile) (model.LyricList, error) { for pattern := range strings.SplitSeq(conf.Server.LyricsPriority, ",") { pattern = strings.TrimSpace(pattern) - switch { - case strings.EqualFold(pattern, "embedded"): - lyricsList, err = fromEmbedded(ctx, mf) - case strings.HasPrefix(pattern, "."): - lyricsList, err = fromExternalFile(ctx, mf, strings.ToLower(pattern)) - default: - lyricsList, err = l.fromPlugin(ctx, mf, pattern) + if pattern == "" { + continue } - if err != nil { - log.Error(ctx, "error getting lyrics", "source", pattern, err) - } + for _, mf := range mediaFiles { + if mf == nil { + continue + } - if len(lyricsList) > 0 { - return lyricsList, nil + lyricsList, err := l.getLyricsFromSource(ctx, mf, pattern) + if err != nil { + log.Error(ctx, "error getting lyrics", "source", pattern, err) + continue + } + + if len(lyricsList) > 0 { + return lyricsList, nil + } } } return nil, nil } + +func (l *lyricsService) getLyricsFromSource(ctx context.Context, mf *model.MediaFile, pattern string) (model.LyricList, error) { + switch { + case strings.EqualFold(pattern, "embedded"): + return fromEmbedded(ctx, mf) + case strings.HasPrefix(pattern, "."): + return fromExternalFile(ctx, mf, pattern) + default: + return l.fromPlugin(ctx, mf, pattern) + } +} diff --git a/core/lyrics/lyrics_test.go b/core/lyrics/lyrics_test.go index 9ab732ad1..68a7f844d 100644 --- a/core/lyrics/lyrics_test.go +++ b/core/lyrics/lyrics_test.go @@ -5,6 +5,7 @@ import ( "encoding/json" "fmt" "os" + "path/filepath" "github.com/navidrome/navidrome/conf" "github.com/navidrome/navidrome/conf/configtest" @@ -44,6 +45,71 @@ var _ = Describe("sources", func() { }, } + elrcLyrics := model.LyricList{ + model.Lyrics{ + DisplayArtist: "ELRC Artist", + DisplayTitle: "ELRC Song", + Lang: "eng", + Line: []model.Line{ + { + Start: ptr(int64(1000)), + End: ptr(int64(3000)), + Value: "Lead words", + Cue: []model.Cue{ + { + Start: ptr(int64(1000)), + End: ptr(int64(1500)), + Value: "Lead ", + ByteStart: 0, + ByteEnd: 4, + }, + { + Start: ptr(int64(1500)), + End: ptr(int64(3000)), + Value: "words", + ByteStart: 5, + ByteEnd: 9, + }, + }, + }, + { + Start: ptr(int64(3000)), + Value: "Fallback line", + }, + }, + Synced: true, + }, + } + + ttmlLyrics := model.LyricList{ + model.Lyrics{ + Kind: "main", + Lang: "eng", + Line: []model.Line{ + { + Start: ptr(int64(18800)), + Value: "We're no strangers to love", + }, + { + Start: ptr(int64(22800)), + Value: "You know the rules and so do I", + }, + }, + Synced: true, + }, + model.Lyrics{ + Kind: "main", + Lang: "por", + Line: []model.Line{ + { + Start: ptr(int64(18800)), + Value: "Nao somos estranhos ao amor", + }, + }, + Synced: true, + }, + } + unsyncedLyrics := model.LyricList{ model.Lyrics{ Lang: "xxx", @@ -59,6 +125,25 @@ var _ = Describe("sources", func() { }, } + srtLyrics := model.LyricList{ + model.Lyrics{ + Lang: "xxx", + Line: []model.Line{ + { + Start: ptr(int64(18800)), + End: ptr(int64(22800)), + Value: "We're from subtitles", + }, + { + Start: ptr(int64(22801)), + End: ptr(int64(26000)), + Value: "Another subtitle line", + }, + }, + Synced: true, + }, + } + BeforeEach(func() { DeferCleanup(configtest.SetupConfig()) @@ -80,7 +165,64 @@ var _ = Describe("sources", func() { }, Entry("embedded > lrc > txt", "embedded,.lrc,.txt", embeddedLyrics), Entry("lrc > embedded > txt", ".lrc,embedded,.txt", syncedLyrics), - Entry("txt > lrc > embedded", ".txt,.lrc,embedded", unsyncedLyrics)) + Entry("elrc > lrc > embedded", ".elrc,.lrc,embedded", elrcLyrics), + Entry("srt > txt > embedded", ".srt,.txt,embedded", srtLyrics), + Entry("txt > lrc > embedded", ".txt,.lrc,embedded", unsyncedLyrics), + Entry("ttml > elrc > lrc > srt > embedded", ".ttml,.elrc,.lrc,.srt,embedded", ttmlLyrics)) + + It("resolves source priority across duplicate media files", func() { + conf.Server.LyricsPriority = ".ttml,embedded" + embeddedJSON, err := json.Marshal(embeddedLyrics) + Expect(err).To(BeNil()) + + svc := lyrics.NewLyrics(nil) + batchSvc, ok := svc.(lyrics.BatchLyrics) + Expect(ok).To(BeTrue()) + + list, err := batchSvc.GetLyricsForMediaFiles(ctx, []model.MediaFile{ + { + Lyrics: string(embeddedJSON), + Path: "tests/fixtures/01 Invisible (RED) Edit Version.mp3", + }, + { + Lyrics: "[]", + Path: "tests/fixtures/test.mp3", + }, + }) + Expect(err).To(BeNil()) + Expect(list).To(Equal(ttmlLyrics)) + }) + + It("preserves configured sidecar suffix casing on case-sensitive filesystems", func() { + dir, err := os.MkdirTemp("", "lyrics-case-*") + Expect(err).ToNot(HaveOccurred()) + DeferCleanup(func() { + Expect(os.RemoveAll(dir)).To(Succeed()) + }) + + probe := filepath.Join(dir, "CASECHECK") + Expect(os.WriteFile(probe, []byte("probe"), 0644)).To(Succeed()) + _, err = os.Stat(filepath.Join(dir, "casecheck")) + if err == nil { + Skip("filesystem is case-insensitive") + } + Expect(os.IsNotExist(err)).To(BeTrue()) + + conf.Server.LyricsPriority = ".LRC" + Expect(os.WriteFile(filepath.Join(dir, "song.LRC"), []byte("[00:01.00]Upper suffix"), 0644)).To(Succeed()) + + svc := lyrics.NewLyrics(nil) + list, err := svc.GetLyrics(ctx, &model.MediaFile{ + LibraryPath: dir, + Path: "song.mp3", + }) + + Expect(err).To(BeNil()) + Expect(list).To(HaveLen(1)) + Expect(list[0].Line).To(Equal([]model.Line{ + {Start: ptr(int64(1000)), Value: "Upper suffix"}, + })) + }) Context("Errors", func() { var RegularUserContext = XContext diff --git a/core/lyrics/sources.go b/core/lyrics/sources.go index 82a10ca41..7586c944f 100644 --- a/core/lyrics/sources.go +++ b/core/lyrics/sources.go @@ -5,6 +5,7 @@ import ( "errors" "os" "path" + "strings" "github.com/navidrome/navidrome/log" "github.com/navidrome/navidrome/model" @@ -36,18 +37,38 @@ func fromExternalFile(ctx context.Context, mf *model.MediaFile, suffix string) ( return nil, err } - lyrics, err := model.ToLyrics("xxx", string(contents)) - if err != nil { - log.Error(ctx, "error parsing lyric external file", "path", externalLyric, err) - return nil, err - } else if lyrics == nil { + var list model.LyricList + switch { + case strings.EqualFold(suffix, ".ttml"): + list, err = parseTTML(contents) + if err != nil { + log.Error(ctx, "error parsing ttml external file", "path", externalLyric, err) + return nil, err + } + case strings.EqualFold(suffix, ".srt"): + list, err = parseSRT(contents) + if err != nil { + log.Error(ctx, "error parsing srt external file", "path", externalLyric, err) + return nil, err + } + default: + lyrics, err := model.ToLyrics("xxx", string(contents)) + if err != nil { + log.Error(ctx, "error parsing lyric external file", "path", externalLyric, err) + return nil, err + } + if lyrics != nil { + list = model.LyricList{*lyrics} + } + } + + if len(list) == 0 { log.Trace(ctx, "empty lyrics from external file", "path", externalLyric) return nil, nil } log.Trace(ctx, "retrieved lyrics from external file", "path", externalLyric) - - return model.LyricList{*lyrics}, nil + return list, nil } // fromPlugin attempts to load lyrics from a plugin with the given name. diff --git a/core/lyrics/sources_test.go b/core/lyrics/sources_test.go index d1aefcb5d..7ca4a0a45 100644 --- a/core/lyrics/sources_test.go +++ b/core/lyrics/sources_test.go @@ -87,6 +87,89 @@ var _ = Describe("sources", func() { })) }) + It("should return Enhanced LRC lyrics with word-level cues from a file", func() { + mf := model.MediaFile{Path: "tests/fixtures/test-enhanced.mp3"} + lyrics, err := fromExternalFile(ctx, &mf, ".lrc") + + Expect(err).To(BeNil()) + Expect(lyrics).To(HaveLen(1)) + Expect(lyrics[0].DisplayArtist).To(Equal("Test Artist")) + Expect(lyrics[0].DisplayTitle).To(Equal("Enhanced Test")) + Expect(lyrics[0].Lang).To(Equal("eng")) + Expect(lyrics[0].Synced).To(BeTrue()) + Expect(lyrics[0].Line).To(HaveLen(3)) + + // Line 1: has inline markers → Cue array populated + Expect(lyrics[0].Line[0].Start).To(Equal(ptr(int64(1000)))) + Expect(lyrics[0].Line[0].End).To(Equal(ptr(int64(3000)))) + Expect(lyrics[0].Line[0].Value).To(Equal("Some lyrics here")) + Expect(lyrics[0].Line[0].Cue).To(HaveLen(3)) + Expect(*lyrics[0].Line[0].Cue[0].Start).To(Equal(int64(1000))) + Expect(lyrics[0].Line[0].Cue[0].Value).To(Equal("Some ")) + Expect(lyrics[0].Line[0].Cue[0].End).To(Equal(ptr(int64(1500)))) + Expect(lyrics[0].Line[0].Cue[0].ByteStart).To(Equal(0)) + Expect(lyrics[0].Line[0].Cue[0].ByteEnd).To(Equal(4)) + Expect(*lyrics[0].Line[0].Cue[1].Start).To(Equal(int64(1500))) + Expect(lyrics[0].Line[0].Cue[1].Value).To(Equal("lyrics ")) + Expect(lyrics[0].Line[0].Cue[1].End).To(Equal(ptr(int64(2000)))) + Expect(lyrics[0].Line[0].Cue[1].ByteStart).To(Equal(5)) + Expect(lyrics[0].Line[0].Cue[1].ByteEnd).To(Equal(11)) + Expect(*lyrics[0].Line[0].Cue[2].Start).To(Equal(int64(2000))) + Expect(lyrics[0].Line[0].Cue[2].Value).To(Equal("here")) + Expect(lyrics[0].Line[0].Cue[2].End).To(Equal(ptr(int64(3000)))) + Expect(lyrics[0].Line[0].Cue[2].ByteStart).To(Equal(12)) + Expect(lyrics[0].Line[0].Cue[2].ByteEnd).To(Equal(15)) + + // Line 2: has inline markers + Expect(lyrics[0].Line[1].Start).To(Equal(ptr(int64(3000)))) + Expect(lyrics[0].Line[1].End).To(Equal(ptr(int64(5000)))) + Expect(lyrics[0].Line[1].Value).To(Equal("More words")) + Expect(lyrics[0].Line[1].Cue).To(HaveLen(2)) + Expect(lyrics[0].Line[1].Cue[0].End).To(Equal(ptr(int64(3500)))) + Expect(lyrics[0].Line[1].Cue[1].End).To(Equal(ptr(int64(5000)))) + Expect(lyrics[0].Line[1].Cue[0].ByteStart).To(Equal(0)) + Expect(lyrics[0].Line[1].Cue[0].ByteEnd).To(Equal(4)) + Expect(lyrics[0].Line[1].Cue[1].ByteStart).To(Equal(5)) + Expect(lyrics[0].Line[1].Cue[1].ByteEnd).To(Equal(9)) + + // Line 3: plain line, no cues + Expect(lyrics[0].Line[2].Start).To(Equal(ptr(int64(5000)))) + Expect(lyrics[0].Line[2].Value).To(Equal("Plain line without inline markers")) + Expect(lyrics[0].Line[2].Cue).To(BeNil()) + }) + + It("should return Enhanced LRC lyrics from an ELRC file", func() { + mf := model.MediaFile{Path: "tests/fixtures/test.mp3"} + lyrics, err := fromExternalFile(ctx, &mf, ".elrc") + + Expect(err).To(BeNil()) + Expect(lyrics).To(HaveLen(1)) + Expect(lyrics[0].DisplayArtist).To(Equal("ELRC Artist")) + Expect(lyrics[0].DisplayTitle).To(Equal("ELRC Song")) + Expect(lyrics[0].Lang).To(Equal("eng")) + Expect(lyrics[0].Synced).To(BeTrue()) + Expect(lyrics[0].Line).To(HaveLen(2)) + + Expect(lyrics[0].Line[0].Start).To(Equal(ptr(int64(1000)))) + Expect(lyrics[0].Line[0].End).To(Equal(ptr(int64(3000)))) + Expect(lyrics[0].Line[0].Value).To(Equal("Lead words")) + Expect(lyrics[0].Line[0].Cue).To(HaveLen(2)) + Expect(*lyrics[0].Line[0].Cue[0].Start).To(Equal(int64(1000))) + Expect(lyrics[0].Line[0].Cue[0].Value).To(Equal("Lead ")) + Expect(lyrics[0].Line[0].Cue[0].End).To(Equal(ptr(int64(1500)))) + Expect(lyrics[0].Line[0].Cue[0].ByteStart).To(Equal(0)) + Expect(lyrics[0].Line[0].Cue[0].ByteEnd).To(Equal(4)) + Expect(*lyrics[0].Line[0].Cue[1].Start).To(Equal(int64(1500))) + Expect(lyrics[0].Line[0].Cue[1].Value).To(Equal("words")) + Expect(lyrics[0].Line[0].Cue[1].End).To(Equal(ptr(int64(3000)))) + Expect(lyrics[0].Line[0].Cue[1].ByteStart).To(Equal(5)) + Expect(lyrics[0].Line[0].Cue[1].ByteEnd).To(Equal(9)) + + Expect(lyrics[0].Line[1].Start).To(Equal(ptr(int64(3000)))) + Expect(lyrics[0].Line[1].Value).To(Equal("Fallback line")) + Expect(lyrics[0].Line[1].Cue).To(BeNil()) + }) + It("should return unsynchronized lyrics from a file", func() { mf := model.MediaFile{Path: "tests/fixtures/test.mp3"} lyrics, err := fromExternalFile(ctx, &mf, ".txt") @@ -108,6 +191,66 @@ var _ = Describe("sources", func() { })) }) + It("should return synchronized lyrics from an SRT file", func() { + mf := model.MediaFile{Path: "tests/fixtures/test.mp3"} + lyrics, err := fromExternalFile(ctx, &mf, ".srt") + + Expect(err).To(BeNil()) + Expect(lyrics).To(Equal(model.LyricList{ + model.Lyrics{ + Lang: "xxx", + Line: []model.Line{ + { + Start: ptr(int64(18800)), + End: ptr(int64(22800)), + Value: "We're from subtitles", + }, + { + Start: ptr(int64(22801)), + End: ptr(int64(26000)), + Value: "Another subtitle line", + }, + }, + Synced: true, + }, + })) + }) + + It("should return synchronized multilingual lyrics from a TTML file", func() { + mf := model.MediaFile{Path: "tests/fixtures/test.mp3"} + lyrics, err := fromExternalFile(ctx, &mf, ".ttml") + + Expect(err).To(BeNil()) + Expect(lyrics).To(Equal(model.LyricList{ + { + Kind: "main", + Lang: "eng", + Line: []model.Line{ + { + Start: ptr(int64(18800)), + Value: "We're no strangers to love", + }, + { + Start: ptr(int64(22800)), + Value: "You know the rules and so do I", + }, + }, + Synced: true, + }, + { + Kind: "main", + Lang: "por", + Line: []model.Line{ + { + Start: ptr(int64(18800)), + Value: "Nao somos estranhos ao amor", + }, + }, + Synced: true, + }, + })) + }) + It("should handle LRC files with UTF-8 BOM marker (issue #4631)", func() { // The function looks for , so we need to pass // a MediaFile with .mp3 path and look for .lrc suffix @@ -141,5 +284,33 @@ var _ = Describe("sources", func() { Expect(lyrics[0].Line[1].Start).To(Equal(new(int64(22801)))) Expect(lyrics[0].Line[1].Value).To(Equal("You know the rules and so do I")) }) + + It("should handle TTML files with UTF-8 BOM marker", func() { + mf := model.MediaFile{Path: "tests/fixtures/bom-test.mp3"} + lyrics, err := fromExternalFile(ctx, &mf, ".ttml") + + Expect(err).To(BeNil()) + Expect(lyrics).To(HaveLen(1)) + Expect(lyrics[0].Kind).To(Equal("main")) + Expect(lyrics[0].Synced).To(BeTrue()) + Expect(lyrics[0].Line).To(HaveLen(1)) + Expect(lyrics[0].Line[0].Start).To(Equal(ptr(int64(0)))) + Expect(lyrics[0].Line[0].Value).To(Equal("BOM test line")) + }) + + It("should handle UTF-16 BE encoded TTML files", func() { + mf := model.MediaFile{Path: "tests/fixtures/bom-utf16-test.mp3"} + lyrics, err := fromExternalFile(ctx, &mf, ".ttml") + + Expect(err).To(BeNil()) + Expect(lyrics).To(HaveLen(1)) + Expect(lyrics[0].Kind).To(Equal("main")) + Expect(lyrics[0].Synced).To(BeTrue()) + Expect(lyrics[0].Line).To(HaveLen(2)) + Expect(lyrics[0].Line[0].Start).To(Equal(ptr(int64(18800)))) + Expect(lyrics[0].Line[0].Value).To(Equal("UTF16 line one")) + Expect(lyrics[0].Line[1].Start).To(Equal(ptr(int64(22801)))) + Expect(lyrics[0].Line[1].Value).To(Equal("UTF16 line two")) + }) }) }) diff --git a/core/lyrics/srt.go b/core/lyrics/srt.go new file mode 100644 index 000000000..05a43ca57 --- /dev/null +++ b/core/lyrics/srt.go @@ -0,0 +1,168 @@ +package lyrics + +import ( + "bytes" + "regexp" + "strconv" + "strings" + + "github.com/navidrome/navidrome/model" + "github.com/navidrome/navidrome/utils/str" +) + +var ( + srtTimeRegex = regexp.MustCompile(`^\s*(\d{1,2}):(\d{2}):(\d{2})[,.](\d{1,3})\s*$`) + srtBlockSeparatorRegex = regexp.MustCompile(`\n\s*\n`) +) + +func parseSRT(contents []byte) (model.LyricList, error) { + return parseSRTWithLanguage(contents, "xxx") +} + +func parseSRTWithLanguage(contents []byte, language string) (model.LyricList, error) { + raw := strings.ReplaceAll(string(contents), "\r\n", "\n") + raw = strings.ReplaceAll(raw, "\r", "\n") + + blocks := splitSRTBlocks(raw) + lines := make([]model.Line, 0, len(blocks)) + + for _, block := range blocks { + line, ok, err := parseSRTBlock(block) + if err != nil { + return nil, err + } + if ok { + lines = append(lines, line) + } + } + + if len(lines) == 0 { + return nil, nil + } + + lyrics := model.NormalizeLyrics(model.Lyrics{ + Lang: normalizeEmbeddedLanguage(language), + Line: lines, + Synced: true, + }) + return model.LyricList{lyrics}, nil +} + +func splitSRTBlocks(raw string) []string { + raw = strings.TrimSpace(raw) + if raw == "" { + return nil + } + + parts := srtBlockSeparatorRegex.Split(raw, -1) + blocks := make([]string, 0, len(parts)) + for _, part := range parts { + part = strings.TrimSpace(part) + if part != "" { + blocks = append(blocks, part) + } + } + return blocks +} + +func parseSRTBlock(block string) (model.Line, bool, error) { + scanner := bytes.Split([]byte(block), []byte("\n")) + if len(scanner) == 0 { + return model.Line{}, false, nil + } + + lines := make([]string, 0, len(scanner)) + for _, line := range scanner { + lines = append(lines, strings.TrimSpace(string(line))) + } + + if len(lines) == 0 { + return model.Line{}, false, nil + } + + startIdx := 0 + if digitsOnly(lines[0]) { + startIdx = 1 + } + if startIdx >= len(lines) { + return model.Line{}, false, nil + } + + timing := strings.Split(lines[startIdx], "-->") + if len(timing) != 2 { + return model.Line{}, false, nil + } + + startMs, err := parseSRTTime(timing[0]) + if err != nil { + return model.Line{}, false, err + } + endMs, err := parseSRTTime(timing[1]) + if err != nil { + return model.Line{}, false, err + } + + textLines := make([]string, 0, len(lines)-startIdx-1) + for _, line := range lines[startIdx+1:] { + if line == "" { + continue + } + textLines = append(textLines, line) + } + + value := str.SanitizeText(strings.Join(textLines, "\n")) + if value == "" { + return model.Line{}, false, nil + } + + return model.Line{ + Start: &startMs, + End: &endMs, + Value: value, + }, true, nil +} + +func parseSRTTime(value string) (int64, error) { + match := srtTimeRegex.FindStringSubmatch(strings.TrimSpace(value)) + if match == nil { + return 0, strconv.ErrSyntax + } + + hours, err := strconv.ParseInt(match[1], 10, 64) + if err != nil { + return 0, err + } + minutes, err := strconv.ParseInt(match[2], 10, 64) + if err != nil { + return 0, err + } + seconds, err := strconv.ParseInt(match[3], 10, 64) + if err != nil { + return 0, err + } + millis, err := strconv.ParseInt(match[4], 10, 64) + if err != nil { + return 0, err + } + + switch len(match[4]) { + case 1: + millis *= 100 + case 2: + millis *= 10 + } + + return (((hours*60)+minutes)*60+seconds)*1000 + millis, nil +} + +func digitsOnly(value string) bool { + if value == "" { + return false + } + for _, ch := range value { + if ch < '0' || ch > '9' { + return false + } + } + return true +} diff --git a/core/lyrics/test_helpers_test.go b/core/lyrics/test_helpers_test.go new file mode 100644 index 000000000..a4a03f0fe --- /dev/null +++ b/core/lyrics/test_helpers_test.go @@ -0,0 +1,5 @@ +package lyrics + +func ptr[T any](v T) *T { + return &v +} diff --git a/core/lyrics/ttml.go b/core/lyrics/ttml.go new file mode 100644 index 000000000..163bd657c --- /dev/null +++ b/core/lyrics/ttml.go @@ -0,0 +1,1278 @@ +package lyrics + +import ( + "bytes" + "encoding/xml" + "errors" + "io" + "math" + "regexp" + "sort" + "strconv" + "strings" + "unicode" + + "github.com/navidrome/navidrome/log" + "github.com/navidrome/navidrome/model" + "github.com/navidrome/navidrome/utils/str" +) + +const ( + defaultTTMLFrameRate = 30.0 + defaultTTMLSubFrameRate = 1.0 + defaultTTMLTickRate = 1.0 + + ttmlLyricKindMain = "main" + ttmlLyricKindTranslation = "translation" + ttmlLyricKindPronunciation = "pronunciation" + ttmlBackgroundAgentPrefix = "__nd_bg__|" +) + +var offsetTimeRegex = regexp.MustCompile(`^([0-9]+(?:\.[0-9]+)?)(h|m|s|ms|f|t)$`) +var xmlEncodingRegex = regexp.MustCompile(`(?i)<\?xml([^>]*?)encoding\s*=\s*["'][^"']+["']([^>]*)\?>`) + +type ttmlTimeKind int + +const ( + ttmlTimeAbsolute ttmlTimeKind = iota + ttmlTimeOffset + ttmlTimeAmbiguous +) + +type ttmlTimingParams struct { + frameRate float64 + subFrameRate float64 + tickRate float64 +} + +type ttmlTimingContext struct { + lang string + role string + agentID string + begin int64 + hasBegin bool + end int64 + hasEnd bool + invalid bool +} + +type ttmlLineRef struct { + order int + line model.Line +} + +type ttmlMetadataEntry struct { + key string + line model.Line + seq int +} + +type ttmlResolvedMetadataLine struct { + order int + seq int + line model.Line +} + +type ttmlDefinedAgent struct { + ID string + Type string + Name string +} + +type ttmlPiece struct { + raw string + cue *model.Cue +} + +type ttmlParser struct { + decoder *xml.Decoder + params ttmlTimingParams + + mainLangOrder []string + mainLinesByLang map[string][]model.Line + + mainLineRefsByKey map[string]ttmlLineRef + mainLineOrder int + + translationLangOrder []string + translationEntriesByLg map[string][]ttmlMetadataEntry + + pronunciationLangOrder []string + pronunciationEntriesByLg map[string][]ttmlMetadataEntry + + definedAgents map[string]ttmlDefinedAgent + + metadataSeq int +} + +func parseTTML(contents []byte) (model.LyricList, error) { + return parseTTMLWithDefaultLang(contents, "xxx") +} + +func parseTTMLWithDefaultLang(contents []byte, defaultLang string) (model.LyricList, error) { + contents = xmlEncodingRegex.ReplaceAll(contents, []byte(``)) + + p := ttmlParser{ + decoder: xml.NewDecoder(bytes.NewReader(contents)), + params: ttmlTimingParams{ + frameRate: defaultTTMLFrameRate, + subFrameRate: defaultTTMLSubFrameRate, + tickRate: defaultTTMLTickRate, + }, + mainLinesByLang: make(map[string][]model.Line), + mainLineRefsByKey: make(map[string]ttmlLineRef), + translationEntriesByLg: make(map[string][]ttmlMetadataEntry), + pronunciationEntriesByLg: make(map[string][]ttmlMetadataEntry), + definedAgents: make(map[string]ttmlDefinedAgent), + } + + root := ttmlTimingContext{lang: normalizeTTMLLang(defaultLang)} + + for { + token, err := p.decoder.Token() + if errors.Is(err, io.EOF) { + break + } + if err != nil { + return nil, err + } + + start, ok := token.(xml.StartElement) + if !ok { + continue + } + + if err := p.parseElement(start, root); err != nil { + return nil, err + } + } + + return p.toLyricList(), nil +} + +func (p *ttmlParser) parseElement(start xml.StartElement, parent ttmlTimingContext) error { + local := strings.ToLower(start.Name.Local) + if local == "tt" { + p.updateTimingParams(start.Attr) + } + + switch local { + case "translation": + return p.parseMetadataTrack(start, parent, ttmlLyricKindTranslation) + case "transliteration": + return p.parseMetadataTrack(start, parent, ttmlLyricKindPronunciation) + case "agent": + return p.parseAgentDefinition(start) + } + + ctx := p.childContext(start.Attr, parent) + if local == "p" { + lineText, tokens, err := p.parseParagraph(ctx) + if err != nil { + return err + } + if ctx.invalid || lineText == "" { + return nil + } + + parsedLine := model.Line{Value: lineText} + if ctx.hasBegin { + startMs := ctx.begin + parsedLine.Start = &startMs + } + if ctx.hasEnd { + endMs := ctx.end + parsedLine.End = &endMs + } + if len(tokens) > 0 { + parsedLine.Cue = tokens + } + parsedLine = hydrateLineTimingFromTokens(parsedLine) + + lineKey, _ := attrValue(start.Attr, "key") + p.addMainLine(ctx.lang, lineKey, parsedLine) + return nil + } + + for { + token, err := p.decoder.Token() + if err != nil { + return err + } + + switch t := token.(type) { + case xml.StartElement: + nextParent := ctx + if ctx.invalid { + // Best effort: ignore invalid timing in container elements, and + // continue traversing descendants with parent context. + nextParent = parent + } + if err := p.parseElement(t, nextParent); err != nil { + return err + } + case xml.EndElement: + if strings.EqualFold(t.Name.Local, start.Name.Local) { + return nil + } + } + } +} + +func (p *ttmlParser) parseMetadataTrack(start xml.StartElement, parent ttmlTimingContext, kind string) error { + ctx := p.childContext(start.Attr, parent) + lang := normalizeTTMLLang(ctx.lang) + + for { + token, err := p.decoder.Token() + if err != nil { + return err + } + + switch t := token.(type) { + case xml.StartElement: + if strings.EqualFold(t.Name.Local, "text") { + entry, ok, err := p.parseMetadataText(t, ctx) + if err != nil { + return err + } + if ok { + p.addMetadataEntry(kind, lang, entry) + } + continue + } + + nextParent := ctx + if ctx.invalid { + nextParent = parent + } + if err := p.parseElement(t, nextParent); err != nil { + return err + } + case xml.EndElement: + if strings.EqualFold(t.Name.Local, start.Name.Local) { + return nil + } + } + } +} + +func (p *ttmlParser) parseAgentDefinition(start xml.StartElement) error { + id, ok := attrValue(start.Attr, "id") + id = strings.TrimSpace(id) + if !ok || id == "" { + return p.skipElement(start) + } + + agent := ttmlDefinedAgent{ + ID: id, + Type: strings.ToLower(strings.TrimSpace(attrOrEmpty(start.Attr, "type"))), + } + + for { + token, err := p.decoder.Token() + if err != nil { + return err + } + + switch t := token.(type) { + case xml.StartElement: + if strings.EqualFold(t.Name.Local, "name") { + name, err := p.collectElementText(t) + if err != nil { + return err + } + name = sanitizeTTMLText(name) + if name != "" && agent.Name == "" { + agent.Name = name + } + continue + } + if err := p.skipElement(t); err != nil { + return err + } + case xml.EndElement: + if strings.EqualFold(t.Name.Local, start.Name.Local) { + p.definedAgents[agent.ID] = agent + return nil + } + } + } +} + +func (p *ttmlParser) parseMetadataText(start xml.StartElement, parent ttmlTimingContext) (ttmlMetadataEntry, bool, error) { + forKey, hasFor := attrValue(start.Attr, "for") + forKey = strings.TrimSpace(forKey) + + pieces, err := p.parseInlineElement(start, parent) + if err != nil { + return ttmlMetadataEntry{}, false, err + } + if !hasFor || forKey == "" { + return ttmlMetadataEntry{}, false, nil + } + + ctx := p.childContext(start.Attr, parent) + if ctx.invalid { + return ttmlMetadataEntry{}, false, nil + } + + value, tokens := buildTTMLLineFromPieces(pieces) + line := model.Line{Value: value} + if ctx.hasBegin { + startMs := ctx.begin + line.Start = &startMs + } + if ctx.hasEnd { + endMs := ctx.end + line.End = &endMs + } + if len(tokens) > 0 { + line.Cue = tokens + } + line = hydrateLineTimingFromTokens(line) + + if line.Value == "" && len(line.Cue) == 0 { + return ttmlMetadataEntry{}, false, nil + } + + return ttmlMetadataEntry{key: forKey, line: line}, true, nil +} + +func (p *ttmlParser) parseParagraph(parent ttmlTimingContext) (string, []model.Cue, error) { + var pieces []ttmlPiece + + for { + token, err := p.decoder.Token() + if err != nil { + return "", nil, err + } + + switch t := token.(type) { + case xml.StartElement: + inlinePieces, err := p.parseInlineElement(t, parent) + if err != nil { + return "", nil, err + } + pieces = append(pieces, inlinePieces...) + case xml.EndElement: + if strings.EqualFold(t.Name.Local, "p") { + value, tokens := buildTTMLLineFromPieces(pieces) + return value, tokens, nil + } + case xml.CharData: + pieces = append(pieces, ttmlPiece{raw: string(t)}) + } + } +} + +func (p *ttmlParser) parseInlineElement(start xml.StartElement, parent ttmlTimingContext) ([]ttmlPiece, error) { + local := strings.ToLower(start.Name.Local) + if local == "br" { + return []ttmlPiece{{raw: "\n"}}, nil + } + + ctx := p.childContext(start.Attr, parent) + _, hasBegin := attrValue(start.Attr, "begin") + _, hasEnd := attrValue(start.Attr, "end") + _, hasDur := attrValue(start.Attr, "dur") + hasOwnTiming := hasBegin || hasEnd || hasDur + + var pieces []ttmlPiece + + for { + token, err := p.decoder.Token() + if err != nil { + return nil, err + } + + switch t := token.(type) { + case xml.StartElement: + inlinePieces, err := p.parseInlineElement(t, ctx) + if err != nil { + return nil, err + } + pieces = append(pieces, inlinePieces...) + case xml.EndElement: + if !strings.EqualFold(t.Name.Local, start.Name.Local) { + continue + } + + if local == "span" && hasOwnTiming && !ctx.invalid && !ttmlPiecesContainCue(pieces) { + rawValue := concatTTMLPieceRaw(pieces) + tokenText := sanitizeTTMLText(rawValue) + if tokenText != "" { + parsedToken := model.Cue{ + AgentID: p.resolveCueAgentID(ctx), + } + if ctx.hasBegin { + startMs := ctx.begin + parsedToken.Start = &startMs + } + if ctx.hasEnd { + endMs := ctx.end + parsedToken.End = &endMs + } + + return []ttmlPiece{{ + raw: rawValue, + cue: &parsedToken, + }}, nil + } + } + + return pieces, nil + case xml.CharData: + pieces = append(pieces, ttmlPiece{raw: string(t)}) + } + } +} + +func buildTTMLLineFromPieces(pieces []ttmlPiece) (string, []model.Cue) { + finalized := finalizeTTMLLines(splitTTMLPiecesByNewline(pieces)) + for len(finalized) > 0 && finalized[0].text == "" && len(finalized[0].cues) == 0 { + finalized = finalized[1:] + } + for len(finalized) > 0 { + last := finalized[len(finalized)-1] + if last.text != "" || len(last.cues) > 0 { + break + } + finalized = finalized[:len(finalized)-1] + } + + var value strings.Builder + cues := make([]model.Cue, 0, 8) + byteOffset := 0 + for i, line := range finalized { + if i > 0 { + value.WriteByte('\n') + byteOffset++ + } + value.WriteString(line.text) + for _, cue := range line.cues { + cue.ByteStart += byteOffset + cue.ByteEnd += byteOffset + cues = append(cues, cue) + } + byteOffset += len(line.text) + } + + return value.String(), cues +} + +type ttmlFinalLine struct { + text string + cues []model.Cue +} + +func finalizeTTMLLines(lines [][]ttmlPiece) []ttmlFinalLine { + finalized := make([]ttmlFinalLine, 0, len(lines)) + for _, line := range lines { + text, cues := finalizeTTMLLogicalLine(line) + finalized = append(finalized, ttmlFinalLine{text: text, cues: cues}) + } + return finalized +} + +func splitTTMLPiecesByNewline(pieces []ttmlPiece) [][]ttmlPiece { + lines := [][]ttmlPiece{{}} + for _, piece := range pieces { + raw := normalizeTTMLPieceRaw(piece.raw) + if raw == "" { + continue + } + + start := 0 + for i := 0; i < len(raw); i++ { + if raw[i] != '\n' { + continue + } + if start < i { + lines[len(lines)-1] = append(lines[len(lines)-1], ttmlPiece{ + raw: raw[start:i], + cue: cloneTTMLCue(piece.cue), + }) + } + lines = append(lines, []ttmlPiece{}) + start = i + 1 + } + if start < len(raw) { + lines[len(lines)-1] = append(lines[len(lines)-1], ttmlPiece{ + raw: raw[start:], + cue: cloneTTMLCue(piece.cue), + }) + } + } + return lines +} + +func finalizeTTMLLogicalLine(line []ttmlPiece) (string, []model.Cue) { + rawLine := concatTTMLPieceRaw(line) + if rawLine == "" { + return "", nil + } + + leftTrimBytes := len(rawLine) - len(strings.TrimLeftFunc(rawLine, unicode.IsSpace)) + rightTrimBytes := len(rawLine) - len(strings.TrimRightFunc(rawLine, unicode.IsSpace)) + trimmedEnd := len(rawLine) - rightTrimBytes + if trimmedEnd < leftTrimBytes { + trimmedEnd = leftTrimBytes + } + + trimmed := strings.TrimSpace(rawLine) + cues := make([]model.Cue, 0, len(line)) + cursor := 0 + for _, piece := range line { + pieceEnd := cursor + len(piece.raw) + if piece.cue != nil { + byteStart := max(cursor, leftTrimBytes) + byteEnd := min(pieceEnd, trimmedEnd) + if byteStart < byteEnd { + cue := *piece.cue + cue.Value = rawLine[byteStart:byteEnd] + cue.ByteStart = byteStart - leftTrimBytes + cue.ByteEnd = byteEnd - leftTrimBytes - 1 + cues = append(cues, cue) + } + } + cursor = pieceEnd + } + + return trimmed, cues +} + +func normalizeTTMLPieceRaw(raw string) string { + raw = str.SanitizeText(raw) + raw = strings.ReplaceAll(raw, "\r\n", "\n") + raw = strings.ReplaceAll(raw, "\r", "\n") + return raw +} + +func concatTTMLPieceRaw(pieces []ttmlPiece) string { + var raw strings.Builder + for _, piece := range pieces { + raw.WriteString(normalizeTTMLPieceRaw(piece.raw)) + } + return raw.String() +} + +func ttmlPiecesContainCue(pieces []ttmlPiece) bool { + for _, piece := range pieces { + if piece.cue != nil { + return true + } + } + return false +} + +func cloneTTMLCue(cue *model.Cue) *model.Cue { + if cue == nil { + return nil + } + + cloned := *cue + return &cloned +} + +func (p *ttmlParser) toLyricList() model.LyricList { + res := make(model.LyricList, 0, len(p.mainLangOrder)+len(p.translationLangOrder)+len(p.pronunciationLangOrder)) + for _, lang := range p.mainLangOrder { + lines := p.mainLinesByLang[lang] + if len(lines) == 0 { + continue + } + res = append(res, p.finalizeLyrics(model.Lyrics{ + Kind: ttmlLyricKindMain, + Lang: lang, + Line: lines, + Synced: linesAreSynced(lines), + })) + } + + res = append(res, p.buildMetadataLyrics(ttmlLyricKindTranslation, p.translationLangOrder, p.translationEntriesByLg)...) + res = append(res, p.buildMetadataLyrics(ttmlLyricKindPronunciation, p.pronunciationLangOrder, p.pronunciationEntriesByLg)...) + return res +} + +func (p *ttmlParser) buildMetadataLyrics(kind string, langOrder []string, entriesByLang map[string][]ttmlMetadataEntry) model.LyricList { + res := make(model.LyricList, 0, len(langOrder)) + + for _, lang := range langOrder { + entries := entriesByLang[lang] + if len(entries) == 0 { + continue + } + + seenKeys := make(map[string]struct{}, len(entries)) + resolved := make([]ttmlResolvedMetadataLine, 0, len(entries)) + for _, entry := range entries { + if _, exists := seenKeys[entry.key]; exists { + continue + } + seenKeys[entry.key] = struct{}{} + + ref, ok := p.mainLineRefsByKey[entry.key] + if !ok { + log.Warn("Skipping TTML metadata line without matching key", "kind", kind, "lang", lang, "key", entry.key) + continue + } + + line := entry.line + if line.Start == nil && ref.line.Start != nil { + startMs := *ref.line.Start + line.Start = &startMs + } + if line.End == nil && ref.line.End != nil { + endMs := *ref.line.End + line.End = &endMs + } + line = hydrateLineTimingFromTokens(line) + + if line.Value == "" && len(line.Cue) == 0 { + continue + } + + resolved = append(resolved, ttmlResolvedMetadataLine{ + order: ref.order, + seq: entry.seq, + line: line, + }) + } + + if len(resolved) == 0 { + continue + } + + sort.SliceStable(resolved, func(i, j int) bool { + if resolved[i].order != resolved[j].order { + return resolved[i].order < resolved[j].order + } + return resolved[i].seq < resolved[j].seq + }) + + lines := make([]model.Line, len(resolved)) + for i := range resolved { + lines[i] = resolved[i].line + } + + res = append(res, p.finalizeLyrics(model.Lyrics{ + Kind: kind, + Lang: lang, + Line: lines, + Synced: linesAreSynced(lines), + })) + } + + return res +} + +func (p *ttmlParser) finalizeLyrics(lyrics model.Lyrics) model.Lyrics { + lyrics.Line, lyrics.Agents = p.resolveAgents(lyrics.Line) + return model.NormalizeLyrics(lyrics) +} + +func (p *ttmlParser) resolveAgents(lines []model.Line) ([]model.Line, []model.Agent) { + if len(lines) == 0 { + return lines, nil + } + + usedOrder := make([]string, 0, 4) + usedSet := make(map[string]struct{}, 4) + sawEmptyCue := false + + for i := range lines { + for j := range lines[i].Cue { + agentID := strings.TrimSpace(lines[i].Cue[j].AgentID) + if agentID == "" { + sawEmptyCue = true + continue + } + if _, exists := usedSet[agentID]; !exists { + usedSet[agentID] = struct{}{} + usedOrder = append(usedOrder, agentID) + } + } + } + + if len(usedOrder) == 0 { + return lines, nil + } + + mainID := "" + for _, agentID := range usedOrder { + role := p.baseRoleForAgent(agentID) + if role != "bg" && role != "group" { + mainID = agentID + break + } + } + if mainID == "" && sawEmptyCue { + mainID = "main" + } + if mainID == "" { + for _, agentID := range usedOrder { + if p.baseRoleForAgent(agentID) != "bg" { + mainID = agentID + break + } + } + } + if mainID == "" { + mainID = usedOrder[0] + } + + if _, exists := usedSet[mainID]; !exists { + usedSet[mainID] = struct{}{} + usedOrder = append([]string{mainID}, usedOrder...) + } + + for i := range lines { + for j := range lines[i].Cue { + if strings.TrimSpace(lines[i].Cue[j].AgentID) == "" { + lines[i].Cue[j].AgentID = mainID + } + } + } + + agents := make([]model.Agent, 0, len(usedOrder)) + for _, agentID := range usedOrder { + role := p.baseRoleForAgent(agentID) + if agentID == mainID { + role = "main" + } + agent := model.Agent{ + ID: agentID, + Role: role, + Name: p.agentNameForID(agentID), + } + agents = append(agents, agent) + } + + return lines, agents +} + +func (p *ttmlParser) resolveCueAgentID(ctx ttmlTimingContext) string { + agentID := strings.TrimSpace(ctx.agentID) + if contextHasRole(ctx.role, "x-bg") { + if agentID == "" { + agentID = "main" + } + return backgroundAgentID(agentID) + } + return agentID +} + +func (p *ttmlParser) baseRoleForAgent(agentID string) string { + if isBackgroundAgentID(agentID) { + return "bg" + } + + if agent, ok := p.definedAgents[agentID]; ok { + switch agent.Type { + case "group": + return "group" + default: + return "voice" + } + } + + return "voice" +} + +func (p *ttmlParser) agentNameForID(agentID string) string { + if isBackgroundAgentID(agentID) { + baseID := strings.TrimPrefix(agentID, ttmlBackgroundAgentPrefix) + if baseID == "main" { + return "" + } + if agent, ok := p.definedAgents[baseID]; ok { + return agent.Name + } + return "" + } + + if agent, ok := p.definedAgents[agentID]; ok { + return agent.Name + } + + return "" +} + +func backgroundAgentID(agentID string) string { + return ttmlBackgroundAgentPrefix + agentID +} + +func isBackgroundAgentID(agentID string) bool { + return strings.HasPrefix(agentID, ttmlBackgroundAgentPrefix) +} + +func contextHasRole(roles string, role string) bool { + for _, candidate := range strings.Fields(strings.ToLower(roles)) { + if candidate == strings.ToLower(role) { + return true + } + } + return false +} + +func appendTTMLRoles(existing string, roles string) string { + for _, role := range strings.Fields(roles) { + if contextHasRole(existing, role) { + continue + } + if existing == "" { + existing = role + } else { + existing += " " + role + } + } + return existing +} + +func (p *ttmlParser) addMainLine(lang string, lineKey string, line model.Line) { + lang = normalizeTTMLLang(lang) + if _, ok := p.mainLinesByLang[lang]; !ok { + p.mainLangOrder = append(p.mainLangOrder, lang) + } + p.mainLinesByLang[lang] = append(p.mainLinesByLang[lang], line) + + lineKey = strings.TrimSpace(lineKey) + if lineKey != "" { + if _, exists := p.mainLineRefsByKey[lineKey]; !exists { + p.mainLineRefsByKey[lineKey] = ttmlLineRef{ + order: p.mainLineOrder, + line: line, + } + } + } + p.mainLineOrder++ +} + +func (p *ttmlParser) addMetadataEntry(kind string, lang string, entry ttmlMetadataEntry) { + lang = normalizeTTMLLang(lang) + entry.seq = p.metadataSeq + p.metadataSeq++ + + switch kind { + case ttmlLyricKindTranslation: + if _, ok := p.translationEntriesByLg[lang]; !ok { + p.translationLangOrder = append(p.translationLangOrder, lang) + } + p.translationEntriesByLg[lang] = append(p.translationEntriesByLg[lang], entry) + case ttmlLyricKindPronunciation: + if _, ok := p.pronunciationEntriesByLg[lang]; !ok { + p.pronunciationLangOrder = append(p.pronunciationLangOrder, lang) + } + p.pronunciationEntriesByLg[lang] = append(p.pronunciationEntriesByLg[lang], entry) + } +} + +func (p *ttmlParser) childContext(attrs []xml.Attr, parent ttmlTimingContext) ttmlTimingContext { + ctx := parent + + if lang, ok := attrValue(attrs, "lang"); ok { + ctx.lang = normalizeTTMLLang(lang) + } + if agentID, ok := attrValue(attrs, "agent"); ok { + ctx.agentID = strings.TrimSpace(agentID) + } + if role, ok := attrValue(attrs, "role"); ok { + role = strings.TrimSpace(role) + if role != "" { + ctx.role = appendTTMLRoles(ctx.role, role) + } + } + + beginExpr, hasBegin := attrValue(attrs, "begin") + endExpr, hasEnd := attrValue(attrs, "end") + durExpr, hasDur := attrValue(attrs, "dur") + + if hasBegin { + begin, kind, ok := parseTTMLTimeExpression(beginExpr, p.params) + if !ok { + ctx.invalid = true + return ctx + } + + base := int64(0) + if parent.hasBegin { + base = parent.begin + } + ctx.begin = resolveTTMLTime(begin, kind, base, parent) + ctx.hasBegin = true + } else { + ctx.begin = parent.begin + ctx.hasBegin = parent.hasBegin + } + + var calculatedEnd int64 + calculatedHasEnd := false + + if hasEnd { + end, kind, ok := parseTTMLTimeExpression(endExpr, p.params) + if !ok { + ctx.invalid = true + return ctx + } + + base := ctx.begin + if !ctx.hasBegin { + base = parent.begin + } + calculatedEnd = resolveTTMLTime(end, kind, base, parent) + calculatedHasEnd = true + } + + if hasDur { + dur, ok := parseTTMLDurationExpression(durExpr, p.params) + if !ok { + ctx.invalid = true + return ctx + } + if ctx.hasBegin { + durEnd := ctx.begin + dur + if !calculatedHasEnd || durEnd < calculatedEnd { + calculatedEnd = durEnd + calculatedHasEnd = true + } + } + } + + if !calculatedHasEnd && parent.hasEnd { + calculatedEnd = parent.end + calculatedHasEnd = true + } + + ctx.end = calculatedEnd + ctx.hasEnd = calculatedHasEnd + return ctx +} + +func (p *ttmlParser) updateTimingParams(attrs []xml.Attr) { + frameRate := p.params.frameRate + if value, ok := attrValue(attrs, "frameRate"); ok { + if parsed, err := strconv.ParseFloat(value, 64); err == nil && parsed > 0 { + frameRate = parsed + } + } + + if value, ok := attrValue(attrs, "frameRateMultiplier"); ok { + parts := strings.Fields(value) + if len(parts) == 2 { + numerator, errA := strconv.ParseFloat(parts[0], 64) + denominator, errB := strconv.ParseFloat(parts[1], 64) + if errA == nil && errB == nil && denominator > 0 { + frameRate = frameRate * (numerator / denominator) + } + } + } + + subFrameRate := p.params.subFrameRate + if value, ok := attrValue(attrs, "subFrameRate"); ok { + if parsed, err := strconv.ParseFloat(value, 64); err == nil && parsed > 0 { + subFrameRate = parsed + } + } + + tickRate := p.params.tickRate + if value, ok := attrValue(attrs, "tickRate"); ok { + if parsed, err := strconv.ParseFloat(value, 64); err == nil && parsed > 0 { + tickRate = parsed + } + } + + p.params.frameRate = positiveOrDefault(frameRate, defaultTTMLFrameRate) + p.params.subFrameRate = positiveOrDefault(subFrameRate, defaultTTMLSubFrameRate) + p.params.tickRate = positiveOrDefault(tickRate, defaultTTMLTickRate) +} + +func parseTTMLDurationExpression(expr string, params ttmlTimingParams) (int64, bool) { + value, _, ok := parseTTMLTimeExpression(expr, params) + return value, ok +} + +func resolveTTMLTime(value int64, kind ttmlTimeKind, base int64, parent ttmlTimingContext) int64 { + switch kind { + case ttmlTimeAbsolute: + return value + case ttmlTimeOffset: + return base + value + case ttmlTimeAmbiguous: + absolute := value + offset := base + value + + // No parent timing context → no reference frame for offsets. + // Prefer absolute when offset differs (i.e., base > 0). + if !parent.hasBegin && !parent.hasEnd && base != 0 { + return absolute + } + + if parent.hasBegin && parent.hasEnd { + absoluteInParent := absolute >= parent.begin && absolute <= parent.end + offsetInParent := offset >= parent.begin && offset <= parent.end + if absoluteInParent && !offsetInParent { + return absolute + } + if offsetInParent && !absoluteInParent { + return offset + } + } + + if parent.hasBegin { + if absolute < parent.begin && offset >= parent.begin { + return offset + } + if absolute >= parent.begin && offset > absolute { + return absolute + } + } + return offset + default: + return base + value + } +} + +func parseTTMLTimeExpression(expr string, params ttmlTimingParams) (int64, ttmlTimeKind, bool) { + expr = strings.TrimSpace(expr) + if expr == "" { + return 0, ttmlTimeOffset, false + } + + lower := strings.ToLower(expr) + if strings.Contains(lower, "wallclock(") || + strings.Contains(lower, ".begin") || + strings.Contains(lower, ".end") { + log.Warn("Unsupported TTML time expression", "value", expr) + return 0, ttmlTimeOffset, false + } + + // Best-effort support for non-standard TTML seen in the wild where a + // bare decimal value is used (implicitly seconds), e.g. "0.170". + if value, err := strconv.ParseFloat(lower, 64); err == nil && value >= 0 { + return int64(math.Round(value * 1000)), ttmlTimeAmbiguous, true + } + + if matches := offsetTimeRegex.FindStringSubmatch(lower); len(matches) == 3 { + value, err := strconv.ParseFloat(matches[1], 64) + if err != nil { + return 0, ttmlTimeOffset, false + } + + unit := matches[2] + seconds := 0.0 + switch unit { + case "h": + seconds = value * 60 * 60 + case "m": + seconds = value * 60 + case "s": + seconds = value + case "ms": + seconds = value / 1000 + case "f": + seconds = value / params.frameRate + case "t": + seconds = value / params.tickRate + default: + return 0, ttmlTimeOffset, false + } + + return int64(math.Round(seconds * 1000)), ttmlTimeOffset, true + } + + colonCount := strings.Count(expr, ":") + switch colonCount { + case 1, 2: + clockMs, ok := parseTTMLClockTime(expr) + if !ok { + return 0, ttmlTimeAbsolute, false + } + return clockMs, ttmlTimeAbsolute, true + case 3: + framesMs, ok := parseTTMLFrameTime(expr, params) + if !ok { + return 0, ttmlTimeAbsolute, false + } + return framesMs, ttmlTimeAbsolute, true + default: + log.Warn("Unsupported TTML time expression", "value", expr) + return 0, ttmlTimeOffset, false + } +} + +func parseTTMLClockTime(value string) (int64, bool) { + parts := strings.Split(value, ":") + if len(parts) != 2 && len(parts) != 3 { + return 0, false + } + + hours := int64(0) + minutesIdx := 0 + if len(parts) == 3 { + h, err := strconv.ParseInt(parts[0], 10, 64) + if err != nil { + return 0, false + } + hours = h + minutesIdx = 1 + } + + minutes, err := strconv.ParseInt(parts[minutesIdx], 10, 64) + if err != nil { + return 0, false + } + + seconds, err := strconv.ParseFloat(parts[minutesIdx+1], 64) + if err != nil { + return 0, false + } + + totalSeconds := float64(hours*60*60+minutes*60) + seconds + return int64(math.Round(totalSeconds * 1000)), true +} + +func parseTTMLFrameTime(value string, params ttmlTimingParams) (int64, bool) { + parts := strings.Split(value, ":") + if len(parts) != 4 { + return 0, false + } + + hours, err := strconv.ParseInt(parts[0], 10, 64) + if err != nil { + return 0, false + } + + minutes, err := strconv.ParseInt(parts[1], 10, 64) + if err != nil { + return 0, false + } + + seconds, err := strconv.ParseInt(parts[2], 10, 64) + if err != nil { + return 0, false + } + + frameParts := strings.SplitN(parts[3], ".", 2) + frames, err := strconv.ParseFloat(frameParts[0], 64) + if err != nil { + return 0, false + } + + subFrames := 0.0 + if len(frameParts) == 2 { + subFrames, err = strconv.ParseFloat(frameParts[1], 64) + if err != nil { + return 0, false + } + } + + totalSeconds := float64(hours*60*60 + minutes*60 + seconds) + totalSeconds += frames / params.frameRate + totalSeconds += subFrames / (params.subFrameRate * params.frameRate) + + return int64(math.Round(totalSeconds * 1000)), true +} + +func attrValue(attrs []xml.Attr, key string) (string, bool) { + for _, attr := range attrs { + if strings.EqualFold(attr.Name.Local, key) { + return strings.TrimSpace(attr.Value), true + } + } + return "", false +} + +func attrOrEmpty(attrs []xml.Attr, key string) string { + value, _ := attrValue(attrs, key) + return value +} + +func (p *ttmlParser) collectElementText(start xml.StartElement) (string, error) { + var text strings.Builder + + for { + token, err := p.decoder.Token() + if err != nil { + return "", err + } + + switch t := token.(type) { + case xml.StartElement: + value, err := p.collectElementText(t) + if err != nil { + return "", err + } + text.WriteString(value) + case xml.EndElement: + if strings.EqualFold(t.Name.Local, start.Name.Local) { + return text.String(), nil + } + case xml.CharData: + text.WriteString(string(t)) + } + } +} + +func (p *ttmlParser) skipElement(_ xml.StartElement) error { + depth := 1 + for depth > 0 { + token, err := p.decoder.Token() + if err != nil { + return err + } + + switch token.(type) { + case xml.StartElement: + depth++ + case xml.EndElement: + depth-- + } + } + return nil +} + +func normalizeTTMLLang(lang string) string { + lang = strings.ToLower(strings.TrimSpace(lang)) + if lang == "" { + return "xxx" + } + return lang +} + +func sanitizeTTMLText(raw string) string { + raw = str.SanitizeText(raw) + raw = strings.ReplaceAll(raw, "\r\n", "\n") + raw = strings.ReplaceAll(raw, "\r", "\n") + + lines := strings.Split(raw, "\n") + for i := range lines { + lines[i] = strings.TrimSpace(lines[i]) + } + return strings.TrimSpace(strings.Join(lines, "\n")) +} + +func linesAreSynced(lines []model.Line) bool { + for i := range lines { + if lines[i].Start != nil { + return true + } + for j := range lines[i].Cue { + if lines[i].Cue[j].Start != nil { + return true + } + } + } + return false +} + +func hydrateLineTimingFromTokens(line model.Line) model.Line { + return model.NormalizeLineTiming(line) +} + +func positiveOrDefault(v float64, fallback float64) float64 { + if v <= 0 { + return fallback + } + return v +} diff --git a/core/lyrics/ttml_test.go b/core/lyrics/ttml_test.go new file mode 100644 index 000000000..07f41a080 --- /dev/null +++ b/core/lyrics/ttml_test.go @@ -0,0 +1,430 @@ +package lyrics + +import ( + "github.com/navidrome/navidrome/model" + . "github.com/onsi/ginkgo/v2" + . "github.com/onsi/gomega" +) + +var _ = Describe("parseTTML", func() { + Describe("Multi-language and timing", func() { + It("should parse multiple language divs with inherited offsets and frame/tick timing", func() { + content := []byte(` + + +
+

Line one

+

Line two
with break

+
+
+

Linha

+
+ +`) + + list, err := parseTTML(content) + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(2)) + + By("parsing the English track") + eng := list[0] + Expect(eng.Lang).To(Equal("eng")) + Expect(eng.Synced).To(BeTrue()) + Expect(eng.Line[0].Start).To(Equal(ptr(int64(3000)))) + Expect(eng.Line[0].Value).To(Equal("Line one")) + Expect(eng.Line[1].Start).To(Equal(ptr(int64(4517)))) + Expect(eng.Line[1].Value).To(Equal("Line two\nwith break")) + + By("parsing the Portuguese track") + por := list[1] + Expect(por.Lang).To(Equal("por")) + Expect(por.Line[0].Start).To(Equal(ptr(int64(4500)))) + Expect(por.Line[0].Value).To(Equal("Linha")) + }) + }) + + Describe("Unsupported cue handling", func() { + It("should skip wallclock cues and keep valid ones", func() { + content := []byte(` + + +
+

Skip me

+

Keep me

+
+ +`) + + list, err := parseTTML(content) + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(1)) + Expect(list[0].Line).To(HaveLen(1)) + Expect(list[0].Line[0].Start).To(Equal(ptr(int64(1000)))) + Expect(list[0].Line[0].Value).To(Equal("Keep me")) + }) + }) + + Describe("Begin/End/Dur with inheritance", func() { + It("should correctly accumulate nested timing from body, div, and p elements", func() { + content := []byte(` + + +
+

First line

+

Second line

+
+ +`) + + list, err := parseTTML(content) + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(1)) + Expect(list[0].Lang).To(Equal("eng")) + Expect(list[0].Line).To(HaveLen(2)) + Expect(list[0].Line[0].Start).To(Equal(ptr(int64(16000)))) + Expect(list[0].Line[0].Value).To(Equal("First line")) + Expect(list[0].Line[1].Start).To(Equal(ptr(int64(18000)))) + Expect(list[0].Line[1].Value).To(Equal("Second line")) + }) + }) + + Describe("Non-standard bare second offsets", func() { + It("should parse bare decimal numbers as seconds", func() { + content := []byte(` + + +
+

First line

+

Second line

+
+ +`) + + list, err := parseTTML(content) + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(1)) + Expect(list[0].Line).To(HaveLen(2)) + Expect(list[0].Line[0].Start).To(Equal(ptr(int64(10170)))) + Expect(list[0].Line[0].Value).To(Equal("First line")) + Expect(list[0].Line[1].Start).To(Equal(ptr(int64(13710)))) + Expect(list[0].Line[1].Value).To(Equal("Second line")) + }) + }) + + Describe("Word timing tokens", func() { + It("should extract timed tokens from spans including background role", func() { + content := []byte(` + + +
+

+ Hello + echo +

+
+ +`) + + list, err := parseTTML(content) + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(1)) + Expect(list[0].Agents).To(Equal([]model.Agent{ + {ID: "main", Role: "main"}, + {ID: "__nd_bg__|main", Role: "bg"}, + })) + Expect(list[0].Line).To(HaveLen(1)) + + line := list[0].Line[0] + Expect(line.Start).To(Equal(ptr(int64(1000)))) + Expect(line.Value).To(Equal("Hello\necho")) + Expect(line.End).To(Equal(ptr(int64(3000)))) + Expect(line.Cue).To(HaveLen(3)) + + Expect(line.Cue[0]).To(Equal(model.Cue{Start: ptr(int64(1000)), End: ptr(int64(1400)), Value: "He", ByteStart: 0, ByteEnd: 1, AgentID: "main"})) + Expect(line.Cue[1]).To(Equal(model.Cue{Start: ptr(int64(1400)), End: ptr(int64(1800)), Value: "llo", ByteStart: 2, ByteEnd: 4, AgentID: "main"})) + Expect(line.Cue[2]).To(Equal(model.Cue{Start: ptr(int64(2000)), End: ptr(int64(2500)), Value: "echo", ByteStart: 6, ByteEnd: 9, AgentID: "__nd_bg__|main"})) + }) + + It("should append role tokens exactly instead of using substring matches", func() { + content := []byte(` + + +
+

LeadEcho

+
+ +`) + + list, err := parseTTML(content) + + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(1)) + Expect(list[0].Agents).To(Equal([]model.Agent{ + {ID: "main", Role: "main"}, + {ID: "__nd_bg__|main", Role: "bg"}, + })) + Expect(list[0].Line).To(HaveLen(1)) + Expect(list[0].Line[0].Cue).To(HaveLen(2)) + Expect(list[0].Line[0].Cue[0].AgentID).To(Equal("main")) + Expect(list[0].Line[0].Cue[1].AgentID).To(Equal("__nd_bg__|main")) + }) + + It("should parse named TTML agents into main, voice, and group roles", func() { + content := []byte(` + + + + Chris Martin + Jin + All + + + +
+

You

+

and

+

All

+
+ +`) + + list, err := parseTTML(content) + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(1)) + Expect(list[0].Agents).To(Equal([]model.Agent{ + {ID: "v1", Role: "main", Name: "Chris Martin"}, + {ID: "v2", Role: "voice", Name: "Jin"}, + {ID: "v1000", Role: "group", Name: "All"}, + })) + Expect(list[0].Line[0].Cue[0].AgentID).To(Equal("v1")) + Expect(list[0].Line[1].Cue[0].AgentID).To(Equal("v2")) + Expect(list[0].Line[2].Cue[0].AgentID).To(Equal("v1000")) + }) + + It("should avoid collisions between derived background agents and explicit TTML agent ids", func() { + content := []byte(` + + + + Lead + Existing Background Id + + + +
+

+ Lead + Echo +

+

+ Named +

+
+ +`) + + list, err := parseTTML(content) + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(1)) + Expect(list[0].Agents).To(Equal([]model.Agent{ + {ID: "lead", Role: "main", Name: "Lead"}, + {ID: "__nd_bg__|lead", Role: "bg", Name: "Lead"}, + {ID: "lead__bg", Role: "voice", Name: "Existing Background Id"}, + })) + Expect(list[0].Line).To(HaveLen(2)) + Expect(list[0].Line[0].Cue).To(HaveLen(2)) + Expect(list[0].Line[0].Cue[0].AgentID).To(Equal("lead")) + Expect(list[0].Line[0].Cue[1].AgentID).To(Equal("__nd_bg__|lead")) + Expect(list[0].Line[1].Cue).To(HaveLen(1)) + Expect(list[0].Line[1].Cue[0].AgentID).To(Equal("lead__bg")) + }) + + It("should fill missing cue agent ids with the resolved main agent", func() { + content := []byte(` + + + + Guest Vocal + + + +
+

+ Lead + Guest +

+
+ +`) + + list, err := parseTTML(content) + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(1)) + Expect(list[0].Agents).To(Equal([]model.Agent{ + {ID: "guest", Role: "main", Name: "Guest Vocal"}, + })) + Expect(list[0].Line).To(HaveLen(1)) + Expect(list[0].Line[0].Cue).To(HaveLen(2)) + Expect(list[0].Line[0].Cue[0].AgentID).To(Equal("guest")) + Expect(list[0].Line[0].Cue[1].AgentID).To(Equal("guest")) + }) + }) + + Describe("Ambiguous decimal timing", func() { + It("should prefer absolute timing when values fall inside parent window", func() { + content := []byte(` + + +
+

+ go + go +

+
+ +`) + + list, err := parseTTML(content) + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(1)) + Expect(list[0].Line).To(HaveLen(1)) + + line := list[0].Line[0] + Expect(line.Start).To(Equal(ptr(int64(43444)))) + Expect(line.Value).To(Equal("go\ngo")) + Expect(line.End).To(Equal(ptr(int64(45570)))) + Expect(line.Cue).To(HaveLen(2)) + Expect(line.Cue[0]).To(Equal(model.Cue{Start: ptr(int64(43444)), End: ptr(int64(43716)), Value: "go", ByteStart: 0, ByteEnd: 1})) + Expect(line.Cue[1]).To(Equal(model.Cue{Start: ptr(int64(43716)), End: ptr(int64(43887)), Value: "go", ByteStart: 3, ByteEnd: 4})) + }) + }) + + Describe("Unsynced fallback", func() { + It("should return unsynced lyrics when no timing is present", func() { + content := []byte(` + + +
+

No timing here

+
+ +`) + + list, err := parseTTML(content) + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(1)) + Expect(list[0].Lang).To(Equal("xxx")) + Expect(list[0].Synced).To(BeFalse()) + Expect(list[0].Line).To(HaveLen(1)) + Expect(list[0].Line[0].Start).To(BeNil()) + Expect(list[0].Line[0].Value).To(Equal("No timing here")) + }) + }) + + Describe("Metadata tracks", func() { + It("should produce main, translation, and pronunciation tracks from iTunesMetadata", func() { + content := []byte(` + + + + + + + Hola + Skip me + + + + + konni + + + + + + +
+

こんにちは

+

こんばんは

+
+ +`) + + list, err := parseTTML(content) + Expect(err).ToNot(HaveOccurred()) + Expect(list).To(HaveLen(3)) + + By("checking the main track") + main := list[0] + Expect(main.Kind).To(Equal("main")) + Expect(main.Lang).To(Equal("ja")) + Expect(main.Line).To(HaveLen(2)) + + By("checking the translation track") + translation := list[1] + Expect(translation.Kind).To(Equal("translation")) + Expect(translation.Lang).To(Equal("es")) + Expect(translation.Line).To(HaveLen(1)) + Expect(translation.Line[0].Start).To(Equal(ptr(int64(1000)))) + Expect(translation.Line[0].Value).To(Equal("Hola")) + Expect(translation.Line[0].End).To(Equal(ptr(int64(1500)))) + + By("checking the pronunciation track") + pronunciation := list[2] + Expect(pronunciation.Kind).To(Equal("pronunciation")) + Expect(pronunciation.Lang).To(Equal("ja-latn")) + Expect(pronunciation.Line).To(HaveLen(1)) + Expect(pronunciation.Line[0].Start).To(Equal(ptr(int64(2000)))) + Expect(pronunciation.Line[0].Value).To(Equal("konni")) + Expect(pronunciation.Line[0].End).To(Equal(ptr(int64(2600)))) + Expect(pronunciation.Line[0].Cue).To(HaveLen(2)) + Expect(pronunciation.Line[0].Cue[0]).To(Equal(model.Cue{Start: ptr(int64(2000)), End: ptr(int64(2300)), Value: "ko", ByteStart: 0, ByteEnd: 1})) + Expect(pronunciation.Line[0].Cue[1]).To(Equal(model.Cue{Start: ptr(int64(2300)), End: ptr(int64(2600)), Value: "nni", ByteStart: 2, ByteEnd: 4})) + }) + }) + + Describe("Pronunciation with bare decimal end times", func() { + It("should correctly parse bare decimal times in transliteration spans", func() { + content := []byte(` + + + + + + + I woke up + + + + + + +
+

起きた

+
+ +`) + + list, err := parseTTML(content) + Expect(err).ToNot(HaveOccurred()) + + var pronunciation *model.Lyrics + for i := range list { + if list[i].Kind == "pronunciation" { + pronunciation = &list[i] + break + } + } + Expect(pronunciation).ToNot(BeNil()) + Expect(pronunciation.Line).To(HaveLen(1)) + + line := pronunciation.Line[0] + Expect(line.Start).To(Equal(ptr(int64(2747)))) + Expect(line.Value).To(Equal("I woke up")) + Expect(line.Cue).To(HaveLen(3)) + Expect(line.Cue[0]).To(Equal(model.Cue{Start: ptr(int64(2747)), End: ptr(int64(3018)), Value: "I", ByteStart: 0, ByteEnd: 0})) + Expect(line.Cue[1]).To(Equal(model.Cue{Start: ptr(int64(3018)), End: ptr(int64(3179)), Value: "woke", ByteStart: 2, ByteEnd: 5})) + Expect(line.Cue[2]).To(Equal(model.Cue{Start: ptr(int64(3179)), End: ptr(int64(3582)), Value: "up", ByteStart: 7, ByteEnd: 8})) + }) + }) +}) diff --git a/model/lyrics.go b/model/lyrics.go index f75f3b11b..f9f21b873 100644 --- a/model/lyrics.go +++ b/model/lyrics.go @@ -6,33 +6,57 @@ import ( "slices" "strconv" "strings" + "unicode" "github.com/navidrome/navidrome/log" "github.com/navidrome/navidrome/utils/str" ) +type Cue struct { + Start *int64 `structs:"start,omitempty" json:"start,omitempty"` + End *int64 `structs:"end,omitempty" json:"end,omitempty"` + Value string `structs:"value" json:"value"` + ByteStart int `structs:"byteStart" json:"byteStart"` + ByteEnd int `structs:"byteEnd" json:"byteEnd"` + AgentID string `structs:"agentId,omitempty" json:"agentId,omitempty"` +} + +type Agent struct { + ID string `structs:"id" json:"id"` + Role string `structs:"role" json:"role"` + Name string `structs:"name,omitempty" json:"name,omitempty"` +} + type Line struct { Start *int64 `structs:"start,omitempty" json:"start,omitempty"` + End *int64 `structs:"end,omitempty" json:"end,omitempty"` Value string `structs:"value" json:"value"` + Cue []Cue `structs:"cue,omitempty" json:"cue,omitempty"` } type Lyrics struct { - DisplayArtist string `structs:"displayArtist,omitempty" json:"displayArtist,omitempty"` - DisplayTitle string `structs:"displayTitle,omitempty" json:"displayTitle,omitempty"` - Lang string `structs:"lang" json:"lang"` - Line []Line `structs:"line" json:"line"` - Offset *int64 `structs:"offset,omitempty" json:"offset,omitempty"` - Synced bool `structs:"synced" json:"synced"` + DisplayArtist string `structs:"displayArtist,omitempty" json:"displayArtist,omitempty"` + DisplayTitle string `structs:"displayTitle,omitempty" json:"displayTitle,omitempty"` + Kind string `structs:"kind,omitempty" json:"kind,omitempty"` + Lang string `structs:"lang" json:"lang"` + Agents []Agent `structs:"agents,omitempty" json:"agents,omitempty"` + Line []Line `structs:"line" json:"line"` + Offset *int64 `structs:"offset,omitempty" json:"offset,omitempty"` + Synced bool `structs:"synced" json:"synced"` } // support the standard [mm:ss.mm], as well as [hh:*] and [*.mmm] -const timeRegexString = `\[([0-9]{1,2}:)?([0-9]{1,2}):([0-9]{1,2})(.[0-9]{1,3})?\]` +const timeRegexString = `\[([0-9]{1,2}:)?([0-9]{1,2}):([0-9]{1,2})(\.[0-9]{1,3})?\]` var ( // Should either be at the beginning of file, or beginning of line syncRegex = regexp.MustCompile(`(^|\n)\s*` + timeRegexString) timeRegex = regexp.MustCompile(timeRegexString) lrcIdRegex = regexp.MustCompile(`\[(ar|ti|offset|lang):([^]]+)]`) + + // Enhanced LRC: inline word-level timing markers like <00:12.34> + enhancedLRCTimeString = `<([0-9]{1,2}:)?([0-9]{1,2}):([0-9]{1,2})(\.[0-9]{1,3})?>` + enhancedLRCRegex = regexp.MustCompile(enhancedLRCTimeString) ) func (l Lyrics) IsEmpty() bool { @@ -106,9 +130,11 @@ func ToLyrics(language, text string) (*Lyrics, error) { if validLine { for idx := range timestamps { + value, cues := parseEnhancedLine(priorLine) structuredLines = append(structuredLines, Line{ Start: ×tamps[idx], - Value: strings.TrimSpace(priorLine), + Value: value, + Cue: cues, }) } timestamps = nil @@ -154,9 +180,11 @@ func ToLyrics(language, text string) (*Lyrics, error) { if validLine { for idx := range timestamps { + value, cues := parseEnhancedLine(priorLine) structuredLines = append(structuredLines, Line{ Start: ×tamps[idx], - Value: strings.TrimSpace(priorLine), + Value: value, + Cue: cues, }) } } @@ -173,13 +201,118 @@ func ToLyrics(language, text string) (*Lyrics, error) { DisplayArtist: artist, DisplayTitle: title, Lang: language, - Line: structuredLines, + Line: NormalizeCueLines(structuredLines), Offset: offset, Synced: synced, } return &lyrics, nil } +// parseEnhancedLine extracts word-level timing cues from Enhanced LRC inline markers +// and computes UTF-8 byte offsets against the final stripped line value. +func parseEnhancedLine(text string) (string, []Cue) { + matches := enhancedLRCRegex.FindAllStringSubmatchIndex(text, -1) + if len(matches) == 0 { + return strings.TrimSpace(text), nil + } + + type segment struct { + start int64 + rawStart int + rawEnd int + } + + segments := make([]segment, 0, len(matches)) + var rawValue strings.Builder + for i, match := range matches { + timeMs, err := parseTime( + // Rewrite <...> as [...] so parseTime can handle it with the same logic + "["+text[match[0]+1:match[1]-1]+"]", + // Adjust match indices to point into our rewritten string (need start/end pairs for each group) + []int{ + 0, match[1] - match[0], + adjustGroup(match, 2), adjustGroup(match, 3), + adjustGroup(match, 4), adjustGroup(match, 5), + adjustGroup(match, 6), adjustGroup(match, 7), + adjustGroup(match, 8), adjustGroup(match, 9), + }, + ) + if err != nil { + continue + } + + // Text runs from after this marker to the start of the next marker (or end of string) + textStart := match[1] + var textEnd int + if i+1 < len(matches) { + textEnd = matches[i+1][0] + } else { + textEnd = len(text) + } + + word := text[textStart:textEnd] + if word == "" { + continue + } + + rawStart := rawValue.Len() + rawValue.WriteString(word) + segments = append(segments, segment{ + start: timeMs, + rawStart: rawStart, + rawEnd: rawValue.Len(), + }) + } + + if len(segments) == 0 { + return strings.TrimSpace(stripEnhancedMarkers(text)), nil + } + + finalRaw := rawValue.String() + leftTrimBytes := len(finalRaw) - len(strings.TrimLeftFunc(finalRaw, unicode.IsSpace)) + rightTrimBytes := len(finalRaw) - len(strings.TrimRightFunc(finalRaw, unicode.IsSpace)) + trimmedEnd := len(finalRaw) - rightTrimBytes + if trimmedEnd < leftTrimBytes { + trimmedEnd = leftTrimBytes + } + + cues := make([]Cue, 0, len(segments)) + for _, seg := range segments { + start := seg.start + byteStart := max(seg.rawStart, leftTrimBytes) + byteEnd := min(seg.rawEnd, trimmedEnd) + if byteStart >= byteEnd { + continue + } + + cues = append(cues, Cue{ + Start: &start, + Value: finalRaw[byteStart:byteEnd], + ByteStart: byteStart - leftTrimBytes, + ByteEnd: byteEnd - leftTrimBytes - 1, + }) + } + + return strings.TrimSpace(finalRaw), cues +} + +// adjustGroup remaps a capture group index from the original match to our rewritten "[...]" string. +// The rewrite shifts by -1 (removed '<', added '[') so positions within the brackets stay the same. +func adjustGroup(match []int, groupIdx int) int { + orig := match[groupIdx] + if orig == -1 { + return -1 + } + // Offset is: original position minus the position of '<' in the original, plus 1 for '[' + return orig - match[0] +} + +// stripEnhancedMarkers removes all inline markers from text, +// returning the plain lyric text. +func stripEnhancedMarkers(text string) string { + return enhancedLRCRegex.ReplaceAllString(text, "") +} + func parseTime(line string, match []int) (int64, error) { var hours, millis int64 var err error @@ -227,3 +360,119 @@ func parseTime(line string, match []int) (int64, error) { } type LyricList []Lyrics + +func NormalizeLyrics(lyrics Lyrics) Lyrics { + lyrics.Line = NormalizeCueLines(lyrics.Line) + if len(lyrics.Agents) == 0 { + lyrics.Agents = nil + } + return lyrics +} + +func NormalizeCueLines(lines []Line) []Line { + if len(lines) == 0 { + return lines + } + + normalized := make([]Line, len(lines)) + copy(normalized, lines) + + for i := range normalized { + if len(normalized[i].Cue) > 0 { + normalized[i].Cue = slices.Clone(normalized[i].Cue) + } + + var fallbackEnd *int64 + if normalized[i].End != nil { + v := *normalized[i].End + fallbackEnd = &v + } else if i+1 < len(normalized) && normalized[i+1].Start != nil { + v := *normalized[i+1].Start + fallbackEnd = &v + } + + normalized[i] = normalizeCueLine(normalized[i], fallbackEnd) + } + + return normalized +} + +func NormalizeLineTiming(line Line) Line { + if len(line.Cue) == 0 { + return line + } + + var earliestStart *int64 + var latestEnd *int64 + for i := range line.Cue { + token := line.Cue[i] + if token.Start != nil { + if earliestStart == nil || *token.Start < *earliestStart { + v := *token.Start + earliestStart = &v + } + } + + candidateEnd := token.End + if candidateEnd == nil { + candidateEnd = token.Start + } + if candidateEnd != nil { + if latestEnd == nil || *candidateEnd > *latestEnd { + v := *candidateEnd + latestEnd = &v + } + } + } + + if line.Start == nil && earliestStart != nil { + v := *earliestStart + line.Start = &v + } + if line.End == nil && latestEnd != nil { + v := *latestEnd + line.End = &v + } + return line +} + +func normalizeCueLine(line Line, fallbackEnd *int64) Line { + if len(line.Cue) == 0 { + return line + } + + for i := range line.Cue { + if line.Cue[i].End != nil { + continue + } + + if i+1 < len(line.Cue) && line.Cue[i+1].Start != nil { + v := *line.Cue[i+1].Start + line.Cue[i].End = &v + continue + } + + if fallbackEnd != nil { + v := *fallbackEnd + line.Cue[i].End = &v + } + } + + for i := range line.Cue { + if line.Cue[i].End == nil { + line.Cue = clearCueEnds(line.Cue) + return NormalizeLineTiming(line) + } + } + + return NormalizeLineTiming(line) +} + +func clearCueEnds(cues []Cue) []Cue { + normalized := make([]Cue, len(cues)) + copy(normalized, cues) + for i := range normalized { + normalized[i].End = nil + } + return normalized +} diff --git a/model/lyrics_test.go b/model/lyrics_test.go index 644b85ad2..21bdf0e3d 100644 --- a/model/lyrics_test.go +++ b/model/lyrics_test.go @@ -108,4 +108,121 @@ var _ = Describe("ToLyrics", func() { {Start: new(int64(1000 * 60 * 60 * 51)), Value: "Test"}, })) }) + + It("should parse Enhanced LRC with word-level timing", func() { + lyrics, err := ToLyrics("xxx", "[00:01.00]<00:01.00>Some <00:01.50>lyrics <00:02.00>here\n[00:03.00]<00:03.00>More <00:03.50>words") + Expect(err).ToNot(HaveOccurred()) + Expect(lyrics.Synced).To(BeTrue()) + Expect(lyrics.Line).To(HaveLen(2)) + + t1000, t1500, t2000, t3000, t3500 := int64(1000), int64(1500), int64(2000), int64(3000), int64(3500) + + line0 := lyrics.Line[0] + Expect(line0.Start).To(Equal(&t1000)) + Expect(line0.End).To(Equal(&t3000)) + Expect(line0.Value).To(Equal("Some lyrics here")) + Expect(line0.Cue).To(Equal([]Cue{ + {Start: &t1000, End: &t1500, Value: "Some ", ByteStart: 0, ByteEnd: 4}, + {Start: &t1500, End: &t2000, Value: "lyrics ", ByteStart: 5, ByteEnd: 11}, + {Start: &t2000, End: &t3000, Value: "here", ByteStart: 12, ByteEnd: 15}, + })) + + line1 := lyrics.Line[1] + Expect(line1.Start).To(Equal(&t3000)) + Expect(line1.End).To(Equal(&t3500)) + Expect(line1.Value).To(Equal("More words")) + Expect(line1.Cue).To(Equal([]Cue{ + {Start: &t3000, Value: "More ", ByteStart: 0, ByteEnd: 4}, + {Start: &t3500, Value: "words", ByteStart: 5, ByteEnd: 9}, + })) + + Expect(line1.Cue[1].End).To(BeNil()) + }) + + It("should not parse malformed Enhanced LRC timing markers", func() { + lyrics, err := ToLyrics("xxx", "[00:01.00]<00:01a50>Not a marker") + Expect(err).ToNot(HaveOccurred()) + Expect(lyrics.Synced).To(BeTrue()) + Expect(lyrics.Line).To(Equal([]Line{ + {Start: new(int64(1000)), Value: "<00:01a50>Not a marker"}, + })) + }) + + It("should ignore Enhanced LRC markers and return plain lines when no markers present", func() { + a, b := int64(1000), int64(3000) + lyrics, err := ToLyrics("xxx", "[00:01.00]Plain line\n[00:03.00]Another plain line") + Expect(err).ToNot(HaveOccurred()) + Expect(lyrics.Line).To(Equal([]Line{ + {Start: &a, Value: "Plain line"}, + {Start: &b, Value: "Another plain line"}, + })) + }) + + It("should handle mixed Enhanced and plain LRC lines", func() { + lyrics, err := ToLyrics("xxx", "[00:01.00]<00:01.00>Some <00:01.50>lyrics\n[00:03.00]Plain line\n[00:05.00]<00:05.00>More <00:05.50>words") + Expect(err).ToNot(HaveOccurred()) + Expect(lyrics.Line).To(HaveLen(3)) + + t1000, t1500, t5000, t5500 := int64(1000), int64(1500), int64(5000), int64(5500) + t3000 := int64(3000) + + Expect(lyrics.Line[0].Cue).To(Equal([]Cue{ + {Start: &t1000, End: &t1500, Value: "Some ", ByteStart: 0, ByteEnd: 4}, + {Start: &t1500, End: &t3000, Value: "lyrics", ByteStart: 5, ByteEnd: 10}, + })) + Expect(lyrics.Line[0].Value).To(Equal("Some lyrics")) + Expect(lyrics.Line[0].End).To(Equal(&t3000)) + + Expect(lyrics.Line[1].Cue).To(BeNil()) + Expect(lyrics.Line[1].Value).To(Equal("Plain line")) + + Expect(lyrics.Line[2].Cue).To(Equal([]Cue{ + {Start: &t5000, Value: "More ", ByteStart: 0, ByteEnd: 4}, + {Start: &t5500, Value: "words", ByteStart: 5, ByteEnd: 9}, + })) + Expect(lyrics.Line[2].Value).To(Equal("More words")) + }) + + It("should preserve byte offsets for Enhanced LRC cues", func() { + lyrics, err := ToLyrics("xxx", "[00:00.00]<00:00.00>Oh <00:00.90>love<00:01.30> me <00:01.60>tonight") + Expect(err).ToNot(HaveOccurred()) + Expect(lyrics.Line).To(HaveLen(1)) + + t0, t900, t1300, t1600 := int64(0), int64(900), int64(1300), int64(1600) + line := lyrics.Line[0] + Expect(line.Value).To(Equal("Oh love me tonight")) + Expect(line.Cue).To(Equal([]Cue{ + {Start: &t0, Value: "Oh ", ByteStart: 0, ByteEnd: 2}, + {Start: &t900, Value: "love", ByteStart: 3, ByteEnd: 6}, + {Start: &t1300, Value: " me ", ByteStart: 7, ByteEnd: 10}, + {Start: &t1600, Value: "tonight", ByteStart: 11, ByteEnd: 17}, + })) + }) +}) + +var _ = Describe("NormalizeCueLines", func() { + It("should not mutate caller cue slices when filling missing cue end times", func() { + start0, start1, nextLineStart := int64(1000), int64(1500), int64(3000) + lines := []Line{ + { + Start: &start0, + Value: "Some lyrics", + Cue: []Cue{ + {Start: &start0, Value: "Some ", ByteStart: 0, ByteEnd: 4}, + {Start: &start1, Value: "lyrics", ByteStart: 5, ByteEnd: 10}, + }, + }, + { + Start: &nextLineStart, + Value: "Next line", + }, + } + + normalized := NormalizeCueLines(lines) + + Expect(normalized[0].Cue[0].End).To(Equal(&start1)) + Expect(normalized[0].Cue[1].End).To(Equal(&nextLineStart)) + Expect(lines[0].Cue[0].End).To(BeNil()) + Expect(lines[0].Cue[1].End).To(BeNil()) + }) }) diff --git a/model/metadata/map_mediafile.go b/model/metadata/map_mediafile.go index 824cad7c2..ecb0a9195 100644 --- a/model/metadata/map_mediafile.go +++ b/model/metadata/map_mediafile.go @@ -8,6 +8,7 @@ import ( "strconv" "github.com/navidrome/navidrome/conf" + lyricssvc "github.com/navidrome/navidrome/core/lyrics" "github.com/navidrome/navidrome/log" "github.com/navidrome/navidrome/model" "github.com/navidrome/navidrome/utils/str" @@ -129,7 +130,7 @@ func (md Metadata) mapGain(rg, r128 model.TagName) *float64 { } func (md Metadata) mapLyrics() string { - rawLyrics := md.Pairs(model.TagLyrics) + rawLyrics := md.rawPairs(model.TagLyrics) lyricList := make(model.LyricList, 0, len(rawLyrics)) @@ -137,13 +138,15 @@ func (md Metadata) mapLyrics() string { lang := raw.Key() text := raw.Value() - lyrics, err := model.ToLyrics(lang, text) + lyrics, err := lyricssvc.ParseEmbedded(lang, text) if err != nil { log.Warn("Unexpected failure occurred when parsing lyrics", "file", md.filePath, err) continue } - if !lyrics.IsEmpty() { - lyricList = append(lyricList, *lyrics) + for _, lyric := range lyrics { + if !lyric.IsEmpty() { + lyricList = append(lyricList, lyric) + } } } diff --git a/model/metadata/map_mediafile_test.go b/model/metadata/map_mediafile_test.go index 16142f526..486386718 100644 --- a/model/metadata/map_mediafile_test.go +++ b/model/metadata/map_mediafile_test.go @@ -4,6 +4,7 @@ import ( "encoding/json" "os" "sort" + "strings" "github.com/navidrome/navidrome/model" "github.com/navidrome/navidrome/model/metadata" @@ -116,5 +117,86 @@ var _ = Describe("ToMediaFile", func() { sort.Slice(expected, func(i, j int) bool { return expected[i].Lang < expected[j].Lang }) Expect(actual).To(Equal(expected)) }) + + It("should parse embedded TTML lyrics before sanitizing XML tags", func() { + mf = toMediaFile(model.RawTags{ + "LYRICS:ENG": {` + +
+

Embedded TTML line

+
+ +`}, + }) + var actual model.LyricList + err := json.Unmarshal([]byte(mf.Lyrics), &actual) + Expect(err).ToNot(HaveOccurred()) + + Expect(actual).To(Equal(model.LyricList{ + { + Kind: "main", + Lang: "eng", + Line: []model.Line{{Start: ptr(int64(1000)), End: ptr(int64(2500)), Value: "Embedded TTML line"}}, + Synced: true, + }, + })) + }) + + It("should parse embedded TTML lyrics longer than the metadata tag max length", func() { + padding := strings.Repeat(`padding`, 1400) + content := ` + + + + + ` + padding + ` + + + + + +
+

Long embedded TTML line

+
+ +` + + Expect(len(content)).To(BeNumerically(">", 32768)) + + mf = toMediaFile(model.RawTags{ + "LYRICS:ENG": {content}, + }) + var actual model.LyricList + err := json.Unmarshal([]byte(mf.Lyrics), &actual) + Expect(err).ToNot(HaveOccurred()) + + Expect(actual).To(HaveLen(1)) + Expect(actual[0].Kind).To(Equal("main")) + Expect(actual[0].Lang).To(Equal("en")) + Expect(actual[0].Line).To(Equal([]model.Line{ + {Start: ptr(int64(1000)), End: ptr(int64(2500)), Value: "Long embedded TTML line"}, + })) + }) + + It("should parse embedded SRT lyrics with the tag language", func() { + mf = toMediaFile(model.RawTags{ + "LYRICS:POR": {`1 +00:00:18,800 --> 00:00:22,800 +Estamos nas legendas`}, + }) + var actual model.LyricList + err := json.Unmarshal([]byte(mf.Lyrics), &actual) + Expect(err).ToNot(HaveOccurred()) + + Expect(actual).To(Equal(model.LyricList{ + { + Lang: "por", + Line: []model.Line{ + {Start: ptr(int64(18800)), End: ptr(int64(22800)), Value: "Estamos nas legendas"}, + }, + Synced: true, + }, + })) + }) }) }) diff --git a/model/metadata/metadata.go b/model/metadata/metadata.go index 48928f989..c62f33776 100644 --- a/model/metadata/metadata.go +++ b/model/metadata/metadata.go @@ -70,6 +70,7 @@ func New(filePath string, info Info) Metadata { return Metadata{ filePath: filePath, fileInfo: info.FileInfo, + rawTags: lowerTags(info.Tags), tags: clean(filePath, info.Tags), audioProps: info.AudioProperties, hasPicture: info.HasPicture, @@ -79,6 +80,7 @@ func New(filePath string, info Info) Metadata { type Metadata struct { filePath string fileInfo FileInfo + rawTags model.Tags tags model.Tags audioProps AudioProperties hasPicture bool @@ -114,6 +116,14 @@ func (md Metadata) Pairs(key model.TagName) []Pair { values := md.tags[key] return slice.Map(values, func(v string) Pair { return Pair(v) }) } +func (md Metadata) rawPairs(key model.TagName) []Pair { + mapping, ok := model.TagMappings()[key] + if !ok { + return nil + } + values := filterDuplicatedOrEmptyValues(processPairMapping(key, mapping, md.rawTags)) + return slice.Map(values, func(v string) Pair { return Pair(v) }) +} func (md Metadata) first(key model.TagName) string { if v, ok := md.tags[key]; ok && len(v) > 0 { return v[0] diff --git a/model/metadata/test_helpers_test.go b/model/metadata/test_helpers_test.go new file mode 100644 index 000000000..abb65dfdb --- /dev/null +++ b/model/metadata/test_helpers_test.go @@ -0,0 +1,5 @@ +package metadata_test + +func ptr[T any](v T) *T { + return &v +} diff --git a/server/subsonic/helpers.go b/server/subsonic/helpers.go index e4c39e373..565b064a2 100644 --- a/server/subsonic/helpers.go +++ b/server/subsonic/helpers.go @@ -494,14 +494,79 @@ func mapExplicitStatus(explicitStatus string) string { return "" } -func buildStructuredLyric(mf *model.MediaFile, lyrics model.Lyrics) responses.StructuredLyric { +func buildStructuredLyric(mf *model.MediaFile, lyrics model.Lyrics, enhanced bool) responses.StructuredLyric { lines := make([]responses.Line, len(lyrics.Line)) + var cueLines []responses.CueLine + agentOrderByID := make(map[string]int, len(lyrics.Agents)) + agentRoleByID := make(map[string]string, len(lyrics.Agents)) + responseAgents := make([]responses.Agent, 0, len(lyrics.Agents)) + + for i, agent := range lyrics.Agents { + agentOrderByID[agent.ID] = i + agentRoleByID[agent.ID] = agent.Role + responseAgents = append(responseAgents, responses.Agent{ + ID: agent.ID, + Role: agent.Role, + Name: agent.Name, + }) + } for i, line := range lyrics.Line { lines[i] = responses.Line{ Start: line.Start, Value: line.Value, } + if !enhanced || len(line.Cue) == 0 { + continue + } + + agentOrder := make([]string, 0, 2) + cuesByAgent := make(map[string][]model.Cue) + for _, cue := range line.Cue { + if cue.Start == nil { + continue + } + agentID := strings.TrimSpace(cue.AgentID) + if _, exists := cuesByAgent[agentID]; !exists { + agentOrder = append(agentOrder, agentID) + } + cuesByAgent[agentID] = append(cuesByAgent[agentID], cue) + } + + sort.SliceStable(agentOrder, func(i, j int) bool { + leftRole := agentRoleByID[agentOrder[i]] + rightRole := agentRoleByID[agentOrder[j]] + if leftRole == "main" && rightRole != "main" { + return true + } + if rightRole == "main" && leftRole != "main" { + return false + } + + leftOrder, leftOK := agentOrderByID[agentOrder[i]] + rightOrder, rightOK := agentOrderByID[agentOrder[j]] + if leftOK && rightOK && leftOrder != rightOrder { + return leftOrder < rightOrder + } + if leftOK != rightOK { + return leftOK + } + return i < j + }) + + for _, agentID := range agentOrder { + cueLine := responses.CueLine{ + Index: int32(i), + Start: line.Start, + End: line.End, + Value: line.Value, + Cue: buildLyricCues(cuesByAgent[agentID], line.End), + } + if agentID != "" { + cueLine.AgentID = agentID + } + cueLines = append(cueLines, cueLine) + } } structured := responses.StructuredLyric{ @@ -509,10 +574,22 @@ func buildStructuredLyric(mf *model.MediaFile, lyrics model.Lyrics) responses.St DisplayTitle: lyrics.DisplayTitle, Lang: lyrics.Lang, Line: lines, + CueLine: cueLines, Offset: lyrics.Offset, Synced: lyrics.Synced, } + if enhanced { + kind := strings.TrimSpace(lyrics.Kind) + if kind == "" { + kind = "main" + } + structured.Kind = kind + if len(cueLines) > 0 && len(responseAgents) > 0 { + structured.Agents = responseAgents + } + } + if structured.DisplayArtist == "" { structured.DisplayArtist = mf.Artist } @@ -523,11 +600,86 @@ func buildStructuredLyric(mf *model.MediaFile, lyrics model.Lyrics) responses.St return structured } -func buildLyricsList(mf *model.MediaFile, lyricsList model.LyricList) *responses.LyricsList { - lyricList := make(responses.StructuredLyrics, len(lyricsList)) +func buildLyricCues(cues []model.Cue, lineEnd *int64) []responses.LyricCue { + if len(cues) == 0 { + return nil + } - for i, lyrics := range lyricsList { - lyricList[i] = buildStructuredLyric(mf, lyrics) + hasAnyEnd := false + for i := range cues { + if cues[i].End != nil { + hasAnyEnd = true + break + } + } + + normalized := make([]responses.LyricCue, 0, len(cues)) + for i := range cues { + if cues[i].Start == nil { + continue + } + + cue := responses.LyricCue{ + Start: *cues[i].Start, + Value: cues[i].Value, + ByteStart: cues[i].ByteStart, + ByteEnd: cues[i].ByteEnd, + } + if hasAnyEnd { + end := cues[i].End + if end == nil { + if i+1 < len(cues) && cues[i+1].Start != nil { + v := *cues[i+1].Start + end = &v + } else if lineEnd != nil { + v := *lineEnd + end = &v + } + } + if end != nil && i+1 < len(cues) && cues[i+1].Start != nil && *end > *cues[i+1].Start { + v := *cues[i+1].Start + end = &v + } + if end != nil && *end < cue.Start { + v := cue.Start + end = &v + } + cue.End = end + } + normalized = append(normalized, cue) + } + + if hasAnyEnd { + for i := range normalized { + if normalized[i].End == nil { + for j := range normalized { + normalized[j].End = nil + } + break + } + } + } + + return normalized +} + +func buildLyricsList(mf *model.MediaFile, lyricsList model.LyricList, enhanced bool) *responses.LyricsList { + var filtered model.LyricList + if enhanced { + filtered = lyricsList + } else { + // Without enhanced, only return "main" kind entries + for _, l := range lyricsList { + kind := strings.TrimSpace(l.Kind) + if kind == "" || kind == "main" { + filtered = append(filtered, l) + } + } + } + + lyricList := make(responses.StructuredLyrics, len(filtered)) + for i, lyrics := range filtered { + lyricList[i] = buildStructuredLyric(mf, lyrics, enhanced) } res := &responses.LyricsList{ diff --git a/server/subsonic/media_retrieval.go b/server/subsonic/media_retrieval.go index 9ab3a20b0..8461723ed 100644 --- a/server/subsonic/media_retrieval.go +++ b/server/subsonic/media_retrieval.go @@ -10,6 +10,7 @@ import ( "github.com/navidrome/navidrome/conf" "github.com/navidrome/navidrome/consts" + lyricssvc "github.com/navidrome/navidrome/core/lyrics" "github.com/navidrome/navidrome/log" "github.com/navidrome/navidrome/model" "github.com/navidrome/navidrome/resources" @@ -19,6 +20,8 @@ import ( "github.com/navidrome/navidrome/utils/req" ) +const maxLegacyLyricsCandidates = 10 + func (api *Router) GetAvatar(w http.ResponseWriter, r *http.Request) (*responses.Subsonic, error) { if !conf.Server.EnableGravatar { return api.getPlaceHolderAvatar(w, r) @@ -98,7 +101,11 @@ func (api *Router) GetLyrics(r *http.Request) (*responses.Subsonic, error) { response := newResponse() lyricsResponse := responses.Lyrics{} response.Lyrics = &lyricsResponse - mediaFiles, err := api.ds.MediaFile(r.Context()).GetAll(filter.SongsByArtistTitleWithLyricsFirst(artist, title)) + opts := filter.SongsByArtistTitleWithLyricsFirst(artist, title) + // Search a bounded duplicate window so source-priority fallback can still + // reach older matches without turning legacy getLyrics into an unbounded scan. + opts.Max = maxLegacyLyricsCandidates + mediaFiles, err := api.ds.MediaFile(r.Context()).GetAll(opts) if err != nil { return nil, err @@ -108,9 +115,22 @@ func (api *Router) GetLyrics(r *http.Request) (*responses.Subsonic, error) { return response, nil } - structuredLyrics, err := api.lyrics.GetLyrics(r.Context(), &mediaFiles[0]) - if err != nil { - return nil, err + var structuredLyrics model.LyricList + if batchLyrics, ok := api.lyrics.(lyricssvc.BatchLyrics); ok { + structuredLyrics, err = batchLyrics.GetLyricsForMediaFiles(r.Context(), mediaFiles) + if err != nil { + return nil, err + } + } else { + for i := range mediaFiles { + structuredLyrics, err = api.lyrics.GetLyrics(r.Context(), &mediaFiles[i]) + if err != nil { + return nil, err + } + if len(structuredLyrics) > 0 { + break + } + } } if len(structuredLyrics) == 0 { @@ -124,7 +144,6 @@ func (api *Router) GetLyrics(r *http.Request) (*responses.Subsonic, error) { for _, line := range structuredLyrics[0].Line { lyricsText.WriteString(line.Value + "\n") } - lyricsResponse.Value = lyricsText.String() return response, nil @@ -146,8 +165,10 @@ func (api *Router) GetLyricsBySongId(r *http.Request) (*responses.Subsonic, erro return nil, err } + enhanced, _ := req.Params(r).Bool("enhanced") + response := newResponse() - response.LyricsList = buildLyricsList(mediaFile, structuredLyrics) + response.LyricsList = buildLyricsList(mediaFile, structuredLyrics, enhanced) return response, nil } diff --git a/server/subsonic/media_retrieval_test.go b/server/subsonic/media_retrieval_test.go index 12c0dff56..68c9cf1e6 100644 --- a/server/subsonic/media_retrieval_test.go +++ b/server/subsonic/media_retrieval_test.go @@ -180,6 +180,41 @@ var _ = Describe("MediaRetrievalController", func() { Expect(response.Lyrics.Title).To(Equal("Never Gonna Give You Up")) Expect(response.Lyrics.Value).To(Equal("We're no strangers to love\nYou know the rules and so do I\n")) }) + + It("should prefer higher-priority sidecar lyrics across duplicate candidates", func() { + conf.Server.LyricsPriority = ".ttml,embedded" + r := newGetRequest("artist=Rick+Astley", "title=Never+Gonna+Give+You+Up") + baseTime := time.Date(2025, 1, 1, 0, 0, 0, 0, time.UTC) + embedded, err := model.ToLyrics("eng", "Newest duplicate embedded lyrics") + Expect(err).ToNot(HaveOccurred()) + embeddedJSON, err := json.Marshal(model.LyricList{*embedded}) + Expect(err).ToNot(HaveOccurred()) + mockRepo.SetData(model.MediaFiles{ + { + ID: "1", + Path: "tests/fixtures/01 Invisible (RED) Edit Version.mp3", + Artist: "Rick Astley", + Title: "Never Gonna Give You Up", + Lyrics: string(embeddedJSON), + UpdatedAt: baseTime.Add(2 * time.Hour), // Newer duplicate with embedded lyrics only + }, + { + ID: "2", + Path: "tests/fixtures/test.mp3", + Artist: "Rick Astley", + Title: "Never Gonna Give You Up", + Lyrics: "[]", + UpdatedAt: baseTime.Add(1 * time.Hour), // Older, but has TTML sidecar + }, + }) + + response, err := router.GetLyrics(r) + Expect(err).ToNot(HaveOccurred()) + Expect(response.Lyrics.Artist).To(Equal("Rick Astley")) + Expect(response.Lyrics.Title).To(Equal("Never Gonna Give You Up")) + Expect(response.Lyrics.Value).To(Equal("We're no strangers to love\nYou know the rules and so do I\n")) + Expect(mockRepo.Options.Max).To(Equal(maxLegacyLyricsCandidates)) + }) }) Describe("GetLyricsBySongId", func() { @@ -196,8 +231,10 @@ var _ = Describe("MediaRetrievalController", func() { Expect(realLyric.DisplayArtist).To(Equal(expectedLyric.DisplayArtist)) Expect(realLyric.DisplayTitle).To(Equal(expectedLyric.DisplayTitle)) + Expect(realLyric.Kind).To(Equal(expectedLyric.Kind)) Expect(realLyric.Lang).To(Equal(expectedLyric.Lang)) Expect(realLyric.Synced).To(Equal(expectedLyric.Synced)) + Expect(realLyric.Agents).To(Equal(expectedLyric.Agents)) if expectedLyric.Offset == nil { Expect(realLyric.Offset).To(BeNil()) @@ -216,6 +253,38 @@ var _ = Describe("MediaRetrievalController", func() { Expect(*realLine.Start).To(Equal(*expectedLine.Start)) } } + + Expect(realLyric.CueLine).To(HaveLen(len(expectedLyric.CueLine))) + for j, realCueLine := range realLyric.CueLine { + expectedCueLine := expectedLyric.CueLine[j] + Expect(realCueLine.Index).To(Equal(expectedCueLine.Index)) + Expect(realCueLine.Value).To(Equal(expectedCueLine.Value)) + Expect(realCueLine.AgentID).To(Equal(expectedCueLine.AgentID)) + if expectedCueLine.Start == nil { + Expect(realCueLine.Start).To(BeNil()) + } else { + Expect(*realCueLine.Start).To(Equal(*expectedCueLine.Start)) + } + if expectedCueLine.End == nil { + Expect(realCueLine.End).To(BeNil()) + } else { + Expect(*realCueLine.End).To(Equal(*expectedCueLine.End)) + } + + Expect(realCueLine.Cue).To(HaveLen(len(expectedCueLine.Cue))) + for k, realCue := range realCueLine.Cue { + expectedCue := expectedCueLine.Cue[k] + Expect(realCue.Value).To(Equal(expectedCue.Value)) + Expect(realCue.Start).To(Equal(expectedCue.Start)) + Expect(realCue.ByteStart).To(Equal(expectedCue.ByteStart)) + Expect(realCue.ByteEnd).To(Equal(expectedCue.ByteEnd)) + if expectedCue.End == nil { + Expect(realCue.End).To(BeNil()) + } else { + Expect(*realCue.End).To(Equal(*expectedCue.End)) + } + } + } } } @@ -316,6 +385,427 @@ var _ = Describe("MediaRetrievalController", func() { }, }) }) + + It("should return multilingual TTML sidecar lyrics", func() { + conf.Server.LyricsPriority = ".ttml,embedded" + r := newGetRequest("id=1") + + mockRepo.SetData(model.MediaFiles{ + { + ID: "1", + Path: "tests/fixtures/test.mp3", + Artist: "Rick Astley", + Title: "Never Gonna Give You Up", + Lyrics: "[]", + }, + }) + + response, err := router.GetLyricsBySongId(r) + Expect(err).ToNot(HaveOccurred()) + + porTime := int64(18800) + ttmlTime := int64(22800) + compareResponses(response.LyricsList, responses.LyricsList{ + StructuredLyrics: responses.StructuredLyrics{ + { + DisplayArtist: "Rick Astley", + DisplayTitle: "Never Gonna Give You Up", + Lang: "eng", + Synced: true, + Line: []responses.Line{ + { + Start: ×[0], + Value: "We're no strangers to love", + }, + { + Start: &ttmlTime, + Value: "You know the rules and so do I", + }, + }, + }, + { + DisplayArtist: "Rick Astley", + DisplayTitle: "Never Gonna Give You Up", + Lang: "por", + Synced: true, + Line: []responses.Line{ + { + Start: &porTime, + Value: "Nao somos estranhos ao amor", + }, + }, + }, + }, + }) + }) + + It("should return metadata-linked translation and pronunciation tracks from TTML", func() { + conf.Server.LyricsPriority = ".ttml,embedded" + r := newGetRequest("id=1&enhanced=true") + + mockRepo.SetData(model.MediaFiles{ + { + ID: "1", + Path: "tests/fixtures/test-metadata.mp3", + Artist: "Rick Astley", + Title: "Never Gonna Give You Up", + Lyrics: "[]", + }, + }) + + response, err := router.GetLyricsBySongId(r) + Expect(err).ToNot(HaveOccurred()) + + mainStartA := int64(1000) + mainStartB := int64(2000) + tokenStartA := int64(2000) + tokenEndA := int64(2300) + tokenStartB := int64(2300) + tokenEndB := int64(2600) + compareResponses(response.LyricsList, responses.LyricsList{ + StructuredLyrics: responses.StructuredLyrics{ + { + DisplayArtist: "Rick Astley", + DisplayTitle: "Never Gonna Give You Up", + Kind: "main", + Lang: "ja", + Synced: true, + Line: []responses.Line{ + { + Start: &mainStartA, + Value: "こんにちは", + }, + { + Start: &mainStartB, + Value: "こんばんは", + }, + }, + }, + { + DisplayArtist: "Rick Astley", + DisplayTitle: "Never Gonna Give You Up", + Kind: "translation", + Lang: "es", + Synced: true, + Line: []responses.Line{ + { + Start: &mainStartA, + Value: "Hola", + }, + }, + }, + { + DisplayArtist: "Rick Astley", + DisplayTitle: "Never Gonna Give You Up", + Kind: "pronunciation", + Lang: "ja-latn", + Synced: true, + Line: []responses.Line{ + { + Start: &mainStartB, + Value: "konni", + }, + }, + CueLine: []responses.CueLine{ + { + Index: 0, + Start: &mainStartB, + End: &tokenEndB, + Value: "konni", + Cue: []responses.LyricCue{ + { + Start: tokenStartA, + End: &tokenEndA, + ByteStart: 0, + ByteEnd: 1, + Value: "ko", + }, + { + Start: tokenStartB, + End: &tokenEndB, + ByteStart: 2, + ByteEnd: 4, + Value: "nni", + }, + }, + }, + }, + }, + }, + }) + }) + + It("should return cue lines for songLyrics v2 clients with enhanced=true", func() { + r := newGetRequest("id=1&enhanced=true") + + lineStart := int64(1000) + lineEnd := int64(3000) + tokenStartA := int64(1000) + tokenEndA := int64(1400) + tokenStartB := int64(2000) + tokenEndB := int64(2500) + lyricsJson, err := json.Marshal(model.LyricList{ + { + Lang: "eng", + Agents: []model.Agent{{ID: "lead", Role: "main"}, {ID: "__nd_bg__|lead", Role: "bg"}}, + Synced: true, + Line: []model.Line{ + { + Start: &lineStart, + End: &lineEnd, + Value: "Hello echo", + Cue: []model.Cue{ + { + Start: &tokenStartA, + End: &tokenEndA, + Value: "Hello", + ByteStart: 0, + ByteEnd: 4, + AgentID: "lead", + }, + { + Start: &tokenStartB, + End: &tokenEndB, + Value: "echo", + ByteStart: 6, + ByteEnd: 9, + AgentID: "__nd_bg__|lead", + }, + }, + }, + }, + }, + }) + Expect(err).ToNot(HaveOccurred()) + + mockRepo.SetData(model.MediaFiles{ + { + ID: "1", + Artist: "Rick Astley", + Title: "Never Gonna Give You Up", + Lyrics: string(lyricsJson), + }, + }) + + response, err := router.GetLyricsBySongId(r) + Expect(err).ToNot(HaveOccurred()) + compareResponses(response.LyricsList, responses.LyricsList{ + StructuredLyrics: responses.StructuredLyrics{ + { + DisplayArtist: "Rick Astley", + DisplayTitle: "Never Gonna Give You Up", + Kind: "main", + Lang: "eng", + Synced: true, + Agents: []responses.Agent{ + {ID: "lead", Role: "main"}, + {ID: "__nd_bg__|lead", Role: "bg"}, + }, + Line: []responses.Line{ + { + Start: &lineStart, + Value: "Hello echo", + }, + }, + CueLine: []responses.CueLine{ + { + Index: 0, + Start: &lineStart, + End: &lineEnd, + Value: "Hello echo", + AgentID: "lead", + Cue: []responses.LyricCue{ + { + Start: tokenStartA, + End: &tokenEndA, + ByteStart: 0, + ByteEnd: 4, + Value: "Hello", + }, + }, + }, + { + Index: 0, + Start: &lineStart, + End: &lineEnd, + Value: "Hello echo", + AgentID: "__nd_bg__|lead", + Cue: []responses.LyricCue{ + { + Start: tokenStartB, + End: &tokenEndB, + ByteStart: 6, + ByteEnd: 9, + Value: "echo", + }, + }, + }, + }, + }, + }, + }) + }) + + It("should keep enhanced line-level lyrics when no cue data is available", func() { + r := newGetRequest("id=1&enhanced=true") + + lineStart := int64(1000) + lineEnd := int64(3000) + lyricsJSON, err := json.Marshal(model.LyricList{ + { + Kind: "main", + Lang: "eng", + Synced: true, + Line: []model.Line{ + { + Start: &lineStart, + End: &lineEnd, + Value: "Line without word timing", + }, + }, + }, + }) + Expect(err).ToNot(HaveOccurred()) + + mockRepo.SetData(model.MediaFiles{ + { + ID: "1", + Artist: "Rick Astley", + Title: "Never Gonna Give You Up", + Lyrics: string(lyricsJSON), + }, + }) + + response, err := router.GetLyricsBySongId(r) + Expect(err).ToNot(HaveOccurred()) + compareResponses(response.LyricsList, responses.LyricsList{ + StructuredLyrics: responses.StructuredLyrics{ + { + DisplayArtist: "Rick Astley", + DisplayTitle: "Never Gonna Give You Up", + Kind: "main", + Lang: "eng", + Synced: true, + Line: []responses.Line{ + { + Start: &lineStart, + Value: "Line without word timing", + }, + }, + }, + }, + }) + }) + + It("should return required cue byte offsets for ambiguous and multibyte cue lines", func() { + r := newGetRequest("id=1&enhanced=true") + + asciiLineStart := int64(0) + asciiLineEnd := int64(2400) + asciiCueStartA := int64(0) + asciiCueEndA := int64(300) + asciiCueStartB := int64(900) + asciiCueEndB := int64(1300) + asciiCueStartC := int64(1300) + asciiCueEndC := int64(1600) + asciiCueStartD := int64(1600) + + utfLineStart := int64(2747) + utfLineEnd := int64(6214) + utfCueStartA := int64(2747) + utfCueEndA := int64(3018) + utfCueStartB := int64(3018) + utfCueEndB := int64(3179) + utfCueStartC := int64(3582) + utfCueEndC := int64(4100) + utfCueStartD := int64(4500) + utfCueEndD := int64(6214) + + lyricsJSON, err := json.Marshal(model.LyricList{ + { + Lang: "eng", + Synced: true, + Line: []model.Line{ + { + Start: &asciiLineStart, + End: &asciiLineEnd, + Value: "Oh love love me tonight", + Cue: []model.Cue{ + {Start: &asciiCueStartA, End: &asciiCueEndA, Value: "Oh", ByteStart: 0, ByteEnd: 1}, + {Start: &asciiCueStartB, End: &asciiCueEndB, Value: "love", ByteStart: 8, ByteEnd: 11}, + {Start: &asciiCueStartC, End: &asciiCueEndC, Value: "me", ByteStart: 13, ByteEnd: 14}, + {Start: &asciiCueStartD, Value: "tonight", ByteStart: 16, ByteEnd: 22}, + }, + }, + { + Start: &utfLineStart, + End: &utfLineEnd, + Value: "눈을 뜬 순간", + Cue: []model.Cue{ + {Start: &utfCueStartA, End: &utfCueEndA, Value: "눈", ByteStart: 0, ByteEnd: 2}, + {Start: &utfCueStartB, End: &utfCueEndB, Value: "을", ByteStart: 3, ByteEnd: 5}, + {Start: &utfCueStartC, End: &utfCueEndC, Value: "뜬", ByteStart: 7, ByteEnd: 9}, + {Start: &utfCueStartD, End: &utfCueEndD, Value: "순간", ByteStart: 11, ByteEnd: 16}, + }, + }, + }, + }, + }) + Expect(err).ToNot(HaveOccurred()) + + mockRepo.SetData(model.MediaFiles{ + { + ID: "1", + Artist: "Rick Astley", + Title: "Never Gonna Give You Up", + Lyrics: string(lyricsJSON), + }, + }) + + response, err := router.GetLyricsBySongId(r) + Expect(err).ToNot(HaveOccurred()) + compareResponses(response.LyricsList, responses.LyricsList{ + StructuredLyrics: responses.StructuredLyrics{ + { + DisplayArtist: "Rick Astley", + DisplayTitle: "Never Gonna Give You Up", + Kind: "main", + Lang: "eng", + Synced: true, + Line: []responses.Line{ + {Start: &asciiLineStart, Value: "Oh love love me tonight"}, + {Start: &utfLineStart, Value: "눈을 뜬 순간"}, + }, + CueLine: []responses.CueLine{ + { + Index: 0, + Start: &asciiLineStart, + End: &asciiLineEnd, + Value: "Oh love love me tonight", + Cue: []responses.LyricCue{ + {Start: asciiCueStartA, End: &asciiCueEndA, Value: "Oh", ByteStart: 0, ByteEnd: 1}, + {Start: asciiCueStartB, End: &asciiCueEndB, Value: "love", ByteStart: 8, ByteEnd: 11}, + {Start: asciiCueStartC, End: &asciiCueEndC, Value: "me", ByteStart: 13, ByteEnd: 14}, + {Start: asciiCueStartD, End: &asciiLineEnd, Value: "tonight", ByteStart: 16, ByteEnd: 22}, + }, + }, + { + Index: 1, + Start: &utfLineStart, + End: &utfLineEnd, + Value: "눈을 뜬 순간", + Cue: []responses.LyricCue{ + {Start: utfCueStartA, End: &utfCueEndA, Value: "눈", ByteStart: 0, ByteEnd: 2}, + {Start: utfCueStartB, End: &utfCueEndB, Value: "을", ByteStart: 3, ByteEnd: 5}, + {Start: utfCueStartC, End: &utfCueEndC, Value: "뜬", ByteStart: 7, ByteEnd: 9}, + {Start: utfCueStartD, End: &utfCueEndD, Value: "순간", ByteStart: 11, ByteEnd: 16}, + }, + }, + }, + }, + }, + }) + }) }) }) diff --git a/server/subsonic/opensubsonic.go b/server/subsonic/opensubsonic.go index 85edb1012..97b3cafcc 100644 --- a/server/subsonic/opensubsonic.go +++ b/server/subsonic/opensubsonic.go @@ -11,7 +11,7 @@ func (api *Router) GetOpenSubsonicExtensions(_ *http.Request) (*responses.Subson extensions := responses.OpenSubsonicExtensions{ {Name: "transcodeOffset", Versions: []int32{1}}, {Name: "formPost", Versions: []int32{1}}, - {Name: "songLyrics", Versions: []int32{1}}, + {Name: "songLyrics", Versions: []int32{1, 2}}, {Name: "indexBasedQueue", Versions: []int32{1}}, {Name: "transcoding", Versions: []int32{1}}, {Name: "playbackReport", Versions: []int32{1}}, diff --git a/server/subsonic/opensubsonic_test.go b/server/subsonic/opensubsonic_test.go index 3ccbf232e..e4217303f 100644 --- a/server/subsonic/opensubsonic_test.go +++ b/server/subsonic/opensubsonic_test.go @@ -58,7 +58,7 @@ var _ = Describe("GetOpenSubsonicExtensions", func() { HaveLen(6), ContainElement(responses.OpenSubsonicExtension{Name: "transcodeOffset", Versions: []int32{1}}), ContainElement(responses.OpenSubsonicExtension{Name: "formPost", Versions: []int32{1}}), - ContainElement(responses.OpenSubsonicExtension{Name: "songLyrics", Versions: []int32{1}}), + ContainElement(responses.OpenSubsonicExtension{Name: "songLyrics", Versions: []int32{1, 2}}), ContainElement(responses.OpenSubsonicExtension{Name: "indexBasedQueue", Versions: []int32{1}}), ContainElement(responses.OpenSubsonicExtension{Name: "transcoding", Versions: []int32{1}}), ContainElement(responses.OpenSubsonicExtension{Name: "playbackReport", Versions: []int32{1}}), @@ -88,7 +88,7 @@ var _ = Describe("GetOpenSubsonicExtensions", func() { HaveLen(7), ContainElement(responses.OpenSubsonicExtension{Name: "transcodeOffset", Versions: []int32{1}}), ContainElement(responses.OpenSubsonicExtension{Name: "formPost", Versions: []int32{1}}), - ContainElement(responses.OpenSubsonicExtension{Name: "songLyrics", Versions: []int32{1}}), + ContainElement(responses.OpenSubsonicExtension{Name: "songLyrics", Versions: []int32{1, 2}}), ContainElement(responses.OpenSubsonicExtension{Name: "indexBasedQueue", Versions: []int32{1}}), ContainElement(responses.OpenSubsonicExtension{Name: "transcoding", Versions: []int32{1}}), ContainElement(responses.OpenSubsonicExtension{Name: "playbackReport", Versions: []int32{1}}), diff --git a/server/subsonic/responses/responses.go b/server/subsonic/responses/responses.go index dcb458932..7e41a1daa 100644 --- a/server/subsonic/responses/responses.go +++ b/server/subsonic/responses/responses.go @@ -547,13 +547,39 @@ type Line struct { Value string `xml:",chardata" json:"value"` } +type LyricCue struct { + Start int64 `xml:"start,attr" json:"start"` + End *int64 `xml:"end,attr,omitempty" json:"end,omitempty"` + ByteStart int `xml:"byteStart,attr" json:"byteStart"` + ByteEnd int `xml:"byteEnd,attr" json:"byteEnd"` + Value string `xml:",chardata" json:"value"` +} + +type Agent struct { + ID string `xml:"id,attr" json:"id"` + Role string `xml:"role,attr" json:"role"` + Name string `xml:"name,attr,omitempty" json:"name,omitempty"` +} + +type CueLine struct { + Index int32 `xml:"index,attr" json:"index"` + Start *int64 `xml:"start,attr,omitempty" json:"start,omitempty"` + End *int64 `xml:"end,attr,omitempty" json:"end,omitempty"` + Value string `xml:"value,attr" json:"value"` + AgentID string `xml:"agentId,attr,omitempty" json:"agentId,omitempty"` + Cue []LyricCue `xml:"cue,omitempty" json:"cue,omitempty"` +} + type StructuredLyric struct { - DisplayArtist string `xml:"displayArtist,attr,omitempty" json:"displayArtist,omitempty"` - DisplayTitle string `xml:"displayTitle,attr,omitempty" json:"displayTitle,omitempty"` - Lang string `xml:"lang,attr" json:"lang"` - Line []Line `xml:"line" json:"line"` - Offset *int64 `xml:"offset,attr,omitempty" json:"offset,omitempty"` - Synced bool `xml:"synced,attr" json:"synced"` + DisplayArtist string `xml:"displayArtist,attr,omitempty" json:"displayArtist,omitempty"` + DisplayTitle string `xml:"displayTitle,attr,omitempty" json:"displayTitle,omitempty"` + Kind string `xml:"kind,attr,omitempty" json:"kind,omitempty"` + Lang string `xml:"lang,attr" json:"lang"` + Line []Line `xml:"line" json:"line"` + Agents []Agent `xml:"agent,omitempty" json:"agents,omitempty"` + CueLine []CueLine `xml:"cueLine,omitempty" json:"cueLine,omitempty"` + Offset *int64 `xml:"offset,attr,omitempty" json:"offset,omitempty"` + Synced bool `xml:"synced,attr" json:"synced"` } type StructuredLyrics []StructuredLyric diff --git a/tests/fixtures/bom-test.ttml b/tests/fixtures/bom-test.ttml new file mode 100644 index 000000000..319ab1f07 --- /dev/null +++ b/tests/fixtures/bom-test.ttml @@ -0,0 +1,2 @@ + +

BOM test line

 diff --git a/tests/fixtures/bom-utf16-test.ttml b/tests/fixtures/bom-utf16-test.ttml new file mode 100644 index 000000000..a5621ef5d Binary files /dev/null and b/tests/fixtures/bom-utf16-test.ttml differ diff --git a/tests/fixtures/test-enhanced.lrc b/tests/fixtures/test-enhanced.lrc new file mode 100644 index 000000000..8f7b60f8c --- /dev/null +++ b/tests/fixtures/test-enhanced.lrc @@ -0,0 +1,6 @@ +[ar:Test Artist] +[ti:Enhanced Test] +[lang:eng] +[00:01.00]<00:01.00>Some <00:01.50>lyrics <00:02.00>here +[00:03.00]<00:03.00>More <00:03.50>words +[00:05.00]Plain line without inline markers diff --git a/tests/fixtures/test-metadata.ttml b/tests/fixtures/test-metadata.ttml new file mode 100644 index 000000000..c0243c18f --- /dev/null +++ b/tests/fixtures/test-metadata.ttml @@ -0,0 +1,25 @@ + + + + + + + + Hola + + + + + konni + + + + + + +
+

こんにちは

+

こんばんは

+
+ + diff --git a/tests/fixtures/test.elrc b/tests/fixtures/test.elrc new file mode 100644 index 000000000..01c3d2cdd --- /dev/null +++ b/tests/fixtures/test.elrc @@ -0,0 +1,5 @@ +[ar:ELRC Artist] +[ti:ELRC Song] +[lang:eng] +[00:01.00]<00:01.00>Lead <00:01.50>words +[00:03.00]Fallback line diff --git a/tests/fixtures/test.srt b/tests/fixtures/test.srt new file mode 100644 index 000000000..3c9c09a39 --- /dev/null +++ b/tests/fixtures/test.srt @@ -0,0 +1,7 @@ +1 +00:00:18,800 --> 00:00:22,800 +We're from subtitles + +2 +00:00:22,801 --> 00:00:26,000 +Another subtitle line diff --git a/tests/fixtures/test.ttml b/tests/fixtures/test.ttml new file mode 100644 index 000000000..a85673a1b --- /dev/null +++ b/tests/fixtures/test.ttml @@ -0,0 +1,12 @@ + + + +
+

We're no strangers to love

+

You know the rules and so do I

+
+
+

Nao somos estranhos ao amor

+
+ +