Files
yellowjacket/backend/explore/lyrics.go
T
yonluandClaude Opus 5 e7748f1fd5
CI / check (push) Successful in 3m7s
CI / e2e (push) Canceled after 1m45s
feat(database): shape the library like files, and shrink the catalog
Plans 013 and 014, the album page that prompted them, and the smaller
fixes they turned up. Changelog, largest first.

## The local library is shaped like files, not like MusicBrainz

`audio_files` carries its own tags and points at `albums` and
`artists`; `file_genres` is the one real many-to-many. `recordings`,
`release_group_recordings`, `artist_credit`, `artist_credit_artist`,
`recording_genres`, `release_groups` and `release_to_rg` are gone from
the local side, and with them a six-way join in every read, a
`MIN(release_group_id)` subquery in eleven queries and a
first-credited-artist subquery in nine. Measured on a real 25,966-file
library, every many-to-many that model expressed was 1:1 in the data.

- Ownership is a file. `GetFilePathsByRecordingMBIDs`,
  `LibraryMBIDIndex.CheckMBIDs`, `collectLibraryEntities` and
  `pruneStaleLocalCrossReferences` all join `audio_files`, so the 812
  orphaned recordings, 216 release groups and 260 artists that library
  carried are now structurally impossible.
- One projection: every track query selects from the `track_metadata`
  view, one row type, one mapper. Nine hand-rolled copies had drifted
  far enough to report different years on different screens.
- `library_id = 0` means every library, so each list query exists once
  instead of scoped and unscoped with a branch at every call site.
- No migration chain. `sql/schemas/` is the one description of the
  shape; `sql/migrations/`, `applyMigrations` and `schema_migrations`
  are squashed away, along with the drift between them that had sqlc
  generating against a stale schema.
- `database.InsertTestTrack` is the one test seeder; twenty test files
  had been assembling the old FK chain each in its own order.

## The catalog stores its ids as bytes

`explore_index`'s three 36-char MBID columns and its entity-type text
are 16 raw bytes and a small integer. The table and its six indexes go
780 MB to 405 MB on a real 2,052,200-row catalog, which is why a fresh
install is ~0.6 GB rather than ~1.0 GB.

- `backend/explore/mbid.go` is the only place the encoding is known;
  everything above it speaks dashed strings.
- `CHECK(length(mbid) = 16)` makes a stringly write fail at the insert
  rather than silently returning no rows, since SQLite does not coerce
  between TEXT and BLOB.
- The importer asks the artifact what encoding it carries and converts
  on the way in, so the artifact already published keeps working and no
  format bump is needed.
- `indexRowColumns`/`scanIndexRow` replace four copies of a 22-column
  list, and `TestStoredEncodingRoundTrips` sweeps every read path.

## An album page that says how much of the album is yours

- One question, asked once: is there a file. `filePaths` is filled by a
  single batched lookup when the tracklist settles, and the badge, the
  Play count, the dimmed rows and every menu item read it — replacing
  four claims of decreasing confidence that could show a green tick on
  an album whose every action did nothing.
- Play, Play 7 of 12, or no play button at all.
- `total_tracks` on `explore_index` (~2 bytes over 400,677 release
  groups) and on `audio_files` from tags that have always carried it:
  a complete MBID-matched album now makes no catalog call at all, where
  it used to spend the most expensive request the app makes.
- A merged cluster shows the running order the most releases agree on,
  and the version list marks the release you own rather than standing a
  synthetic entry in for it.
- `AlbumReleasesFailed`: a slow fetch is no longer reported as a failed
  one by a 12-second timer.
- Rows not in the library are dimmed in place (with `aria-disabled`)
  instead of the owned ones wearing a green tick and a legend.

## Caches and cover art get ceilings

- Only the three tiers of a cover are stored; the full-resolution copy
  nothing rendered was 1,134 MB of a 1.4 GB covers directory.
- One artist portrait is downloaded and the rest are remembered as
  URLs — 4.1 GB of a 5.3 GB cache was candidates no code path reads.
- `browsedArtBudget` and `httpCacheBudget` bound what an age cannot:
  the same install held art for 5,770 artists in a 1,301-artist
  library.
- `OrphanedArtistImagesJob` joined a bare MBID onto a sharded
  directory, so it deleted the rows that were the only record of the
  files it left behind. `explore.ArtistImageDir` is that layout's one
  definition now.

## The autotag queue asks whether there is work

`tagging_items` was a row per album folder, not a queue, and no query
read the `tag_status` column that held the answer. The four queue
queries ask the files, which matters most where it is least visible:
`startPrefetch` was scoring every album in a tagged library against
MusicBrainz.

## Phantom playlist tracks resolve in place

An M3U8 imported before its files leaves phantom rows; they now match
by path and fall back to position, keep their place in the playlist
when resolved, and pair best-first so two phantoms cannot claim the
same file.

## Playing a track plays the list it is in

Double-click, and Play on a single row's menu, queue the list as
displayed with `startIndex` on that row — the album page and the track
list used to queue one track and discard the album around it. A
multi-row selection still plays exactly itself.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AfVYUVExXsx1nSWrXN8mAh
2026-08-16 13:58:15 -04:00

238 lines
6.6 KiB
Go

package explore
import (
"context"
"errors"
"yellowjacket/backend/database"
)
// LyricsResult is a single lyric-search hit, mapped from the DB layer
// into the camelCase shape the frontend consumes.
type LyricsResult struct {
AudioFileID int64 `json:"audioFileId"`
FilePath string `json:"filePath"`
LengthMs int64 `json:"lengthMs"`
Title string `json:"title"`
Artist string `json:"artist"`
Album string `json:"album"`
}
// TrackLyrics is the stored or freshly-fetched lyrics for one track.
// Source is "embedded" (from the file's tags / library DB), "lrclib"
// (fetched on demand), or "" when none are available.
type TrackLyrics struct {
Plain string `json:"plain"`
Synced string `json:"synced"`
Instrumental bool `json:"instrumental"`
Source string `json:"source"`
}
const (
// lyricsSearchLimit caps lyric-search hits returned to the UI.
lyricsSearchLimit = 30
// lyricsBackfillBatch is how many missing-lyrics recordings the
// background backfill processes per pass.
lyricsBackfillBatch = 200
// lyricsBackfillMaxPasses bounds a single backfill run so it can't
// loop forever on a huge library; the next launch resumes where
// this one left off (candidates with lyrics now filled are skipped).
lyricsBackfillMaxPasses = 25
)
// SearchLyrics finds library tracks whose lyrics contain the given
// fragment, ranked by relevance. Pure local FTS — no network.
func (e *Service) SearchLyrics(query string) []LyricsResult {
hits, err := e.db.SearchLyrics(query, lyricsSearchLimit)
if err != nil {
e.logger.Warn("lyrics search failed", "query", query, "err", err)
return nil
}
out := make([]LyricsResult, 0, len(hits))
for _, h := range hits {
out = append(out, LyricsResult{
AudioFileID: h.AudioFileID,
FilePath: h.FilePath,
LengthMs: h.LengthMilliseconds,
Title: h.Title,
Artist: h.Artist,
Album: h.Album,
})
}
return out
}
// GetTrackLyrics returns lyrics for a file. If the library
// already has them (from embedded tags) they're returned as-is;
// otherwise it fetches from LRCLIB, persists them (updating the FTS
// index), and returns them. Never returns an error to the frontend —
// a miss just yields an empty result.
func (e *Service) GetTrackLyrics(audioFileID int64) TrackLyrics {
stored, err := e.db.GetLyrics(audioFileID)
if err == nil && stored != "" {
return TrackLyrics{Plain: stored, Source: "embedded"}
}
lookup, err := e.db.FileLyricLookup(audioFileID)
if err != nil || lookup == nil {
return TrackLyrics{}
}
fetched := e.fetchAndStoreLyrics(e.ctx, *lookup)
if fetched == nil {
return TrackLyrics{}
}
return TrackLyrics{
Plain: fetched.Plain,
Synced: fetched.Synced,
Instrumental: fetched.Instrumental,
Source: "lrclib",
}
}
// RebuildLyricsIndex rebuilds the FTS lyrics index from the current
// library. Cheap; safe to call after every scan.
func (e *Service) RebuildLyricsIndex() {
if err := e.db.RebuildLyricsIndex(); err != nil {
e.logger.Warn("lyrics index rebuild failed", "err", err)
return
}
e.index.setMeta(lyricsIndexReadyKey, "1")
e.logger.Info("lyrics index rebuilt")
}
// RebuildLyricsIndexIfNeeded rebuilds the lyrics FTS only when it has not
// been built since the last library change. The backfill keeps the index
// in sync incrementally thereafter, so on an unchanged library the full
// rebuild is redundant; the scan-completion path calls the unconditional
// form.
func (e *Service) RebuildLyricsIndexIfNeeded() {
if e.index.hasMeta(lyricsIndexReadyKey) {
return
}
e.RebuildLyricsIndex()
}
// BackfillLibraryLyrics fetches lyrics from LRCLIB for library tracks
// that don't have them, in the background. Idempotent and bounded —
// each recording is tried once (a miss is cached), and a run stops
// after a fixed number of passes, resuming on the next launch.
func (e *Service) BackfillLibraryLyrics() {
go e.backfillLibraryLyrics(e.ctx)
}
func (e *Service) backfillLibraryLyrics(ctx context.Context) {
total := 0
// LRCLIB has its own limiter, so this starves nothing — but it is
// still work nobody asked for, and it was the last background pass
// with no way to see or stop it. The job is registered lazily,
// after the first batch proves there is something to do, because on
// a covered library every launch would otherwise put an empty job
// in the indicator.
ctx = WithBackgroundPriority(ctx)
var job *backfillJob
defer func() { job.finish(ctx) }()
for range lyricsBackfillMaxPasses {
if ctx.Err() != nil {
return
}
candidates, err := e.db.FilesMissingLyrics(lyricsBackfillBatch)
if err != nil {
e.logger.Warn("lyrics backfill: query failed", "err", err)
return
}
if len(candidates) == 0 {
break
}
if job == nil {
job, ctx = startBackfillJob(
ctx, e.index.jobRegistry(), lyricsBackfillJobID,
"Looking up lyrics",
"Tracks in your library with no lyrics yet",
len(candidates),
)
}
filled := 0
for i, c := range candidates {
if ctx.Err() != nil {
return
}
job.progress(i, len(candidates))
if e.fetchAndStoreLyrics(ctx, c) != nil {
filled++
total++
}
}
// If a whole batch produced no stored lyrics, every remaining
// candidate is a cached miss with no new data — stop early
// rather than spinning through identical misses.
if filled == 0 {
break
}
}
if total > 0 {
e.logger.Info("lyrics backfill complete", "filled", total)
}
}
// fetchAndStoreLyrics looks up a single candidate on LRCLIB and, on a
// non-instrumental hit with plain lyrics, persists it to the recording
// (which also updates the FTS index). Returns the fetched lyrics, or
// nil on any miss/error. Instrumental hits are recorded as an empty
// lyrics string so they still count as "resolved" and aren't retried.
func (e *Service) fetchAndStoreLyrics(
ctx context.Context, c database.LyricsCandidate,
) *Lyrics {
durationSec := int(c.LengthMilliseconds / 1000) //nolint:mnd
lyrics, err := e.lrclib.GetLyrics(ctx, c.Artist, c.Title, c.Album, durationSec)
if err != nil {
if !errors.Is(err, ErrLyricsNotFound) {
e.logger.Debug("lyrics fetch failed",
"artist", c.Artist, "title", c.Title, "err", err,
)
}
return nil
}
if lyrics.Instrumental || lyrics.Plain == "" {
return nil
}
// Marked `lrclib` rather than `tag`: these came off the network and
// a rebuild that discards them pays for them again.
if err := e.db.SetLyrics(
c.AudioFileID, lyrics.Plain, "lrclib", c.RecordingMBID,
); err != nil {
e.logger.Warn("lyrics store failed", "audioFileId", c.AudioFileID, "err", err)
return nil
}
return lyrics
}