Files
yellowjacket/backend/tagwriter/pipeline.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

379 lines
10 KiB
Go

package tagwriter
import (
"context"
"errors"
"fmt"
"log/slog"
"time"
"github.com/wailsapp/wails/v3/pkg/application"
"yellowjacket/backend/database"
"yellowjacket/backend/events"
)
// errNoChanges is returned when WriteTrackTags is called with an
// empty TagChanges map.
var errNoChanges = errors.New("tagwriter: no changes provided")
// BatchFailure records a single track that failed during a batch write.
type BatchFailure struct {
FilePath string `json:"filePath"`
Error string `json:"error"`
}
// BatchResult summarises the outcome of a batch tag write.
type BatchResult struct {
Total int `json:"total"`
Succeeded int `json:"succeeded"`
Failed int `json:"failed"`
Cancelled bool `json:"cancelled"`
Failures []BatchFailure `json:"failures"`
}
// PlayerStopper checks whether a file is currently playing and
// stops playback if needed. Defined as an interface to break the
// import cycle between tagwriter and player.
type PlayerStopper interface {
// CurrentFilePath returns the file path of the currently-
// loaded track, or empty string if nothing is loaded.
CurrentFilePath() string
// StopAndRelease stops playback and releases the file handle.
StopAndRelease()
}
// PipelineLocker abstracts the library's pipeline mutex for
// scan/write mutual exclusion.
type PipelineLocker interface {
AcquirePipelineLock()
ReleasePipelineLock()
}
// TagWriter orchestrates the complete tag writing pipeline:
// file write → DB sync → event emission.
type TagWriter struct {
logger *slog.Logger
db *database.DB
ctx context.Context // Wails context for event emission
player PlayerStopper
library PipelineLocker
cancelBatch chan struct{} // Signals batch cancellation.
suppressEvents bool // Suppresses per-track events during batch.
}
// NewTagWriter creates a TagWriter with the given dependencies.
// Call SetContext after the Wails runtime is available.
func NewTagWriter(
logger *slog.Logger,
db *database.DB,
player PlayerStopper,
library PipelineLocker,
) *TagWriter {
return &TagWriter{
logger: logger.WithGroup("tagwriter"),
db: db,
player: player,
library: library,
}
}
// ServiceStartup is v3's service lifecycle hook: it runs once the
// runtime exists, and ctx is cancelled when the app shuts down. It
// replaces v2's SetContext, which had to be called by hand from
// OnStartup and was exported, so it was also bound to the frontend.
func (tw *TagWriter) ServiceStartup(
ctx context.Context,
_ application.ServiceOptions,
) error {
tw.ctx = ctx
return nil
}
// WriteTrackTags is the single entry point for writing metadata to
// a track's audio file and synchronising all changes to the
// database. It accepts a track ID (audio_file.id) and a diff map
// of changed fields.
//
// The pipeline:
// 1. Look up track from DB.
// 2. Detect audio format.
// 3. Acquire pipeline lock (mutual exclusion with scan).
// 4. Stop player if this file is currently playing.
// 5. Write tags to file (format-specific writer).
// 6. Sync database (entity relink, FTS5, orphan cleanup).
// 7. Emit TrackMetadataChanged event.
func (tw *TagWriter) WriteTrackTags(trackID int64, changes TagChanges) error {
start := time.Now()
ctx := context.Background()
if len(changes) == 0 {
return errNoChanges
}
// 1. Look up track.
audioFile, err := tw.db.Queries.GetAudioFile(ctx, trackID)
if err != nil {
return fmt.Errorf("get audio file %d: %w", trackID, err)
}
// 2. Detect format.
format, err := DetectFormat(audioFile.FilePath)
if err != nil {
return fmt.Errorf("detect format: %w", err)
}
// 3. Acquire pipeline lock.
tw.library.AcquirePipelineLock()
defer tw.library.ReleasePipelineLock()
// 4. Player safety check.
if tw.player != nil && tw.player.CurrentFilePath() == audioFile.FilePath {
tw.logger.Info("stopping playback for tag write",
"path", audioFile.FilePath)
tw.player.StopAndRelease()
}
// 5. Write file tags.
switch format {
case FormatMP3:
err = writeMp3Tags(tw.logger, audioFile.FilePath, changes)
case FormatFLAC:
err = writeFlacTags(tw.logger, audioFile.FilePath, changes)
case FormatWAV:
err = writeWavTags(tw.logger, audioFile.FilePath, changes)
case FormatOGG:
err = writeOggTags(tw.logger, audioFile.FilePath, changes)
default:
err = fmt.Errorf("%w: %s", errUnsupportedFormat, format)
}
if err != nil {
return fmt.Errorf("write file tags: %w", err)
}
// 6. Sync database.
if syncErr := syncDatabase(ctx, tw.logger, tw.db, dbSyncParams{
audioFileID: audioFile.ID,
filePath: audioFile.FilePath,
changes: changes,
oldFile: audioFile,
}); syncErr != nil {
tw.logger.Error("db sync failed after successful file write",
"err", syncErr,
"trackID", trackID,
"path", audioFile.FilePath,
)
return fmt.Errorf("sync database: %w", syncErr)
}
// 7. Emit event (suppressed during batch writes).
if !tw.suppressEvents {
events.Emit(tw.ctx, events.TrackMetadataChanged,
map[string]any{
"trackId": trackID,
"filePath": audioFile.FilePath,
},
)
}
tw.logger.Info("tag write complete",
"trackID", trackID,
"path", audioFile.FilePath,
"duration", time.Since(start),
"changedFields", len(changes),
)
return nil
}
// WriteUntrackedFileTags writes tags to a file that is not in the
// library, skipping every step of the full pipeline that assumes it is:
// no audio_file lookup, no database sync, no event.
//
// This exists for the download import path, which tags files while they
// are still in the staging directory. Tagging before the move is what
// makes the import atomic from the library's point of view — the
// scanner only ever sees a finished, correctly tagged file, instead of
// ingesting a mislabelled one and being corrected afterwards.
//
// Callers are responsible for ensuring the file is not in a library
// path; using this on a tracked file would leave the database stale.
func (tw *TagWriter) WriteUntrackedFileTags(
filePath string,
changes TagChanges,
) error {
if len(changes) == 0 {
return errNoChanges
}
return WriteFileTags(tw.logger, filePath, changes)
}
// WriteFileTags writes tags straight to an audio file, with no
// database, player or lock involvement. It is the format-dispatch
// half of WriteUntrackedFileTags, exported so tooling that has no
// app to construct — the fixture generator in cmd/gentestdata — can
// tag files with the same writers the app uses, rather than growing a
// second tagger that is free to drift from this one.
//
// Callers owning a *TagWriter should use WriteUntrackedFileTags.
func WriteFileTags(
logger *slog.Logger,
filePath string,
changes TagChanges,
) error {
format, err := DetectFormat(filePath)
if err != nil {
return fmt.Errorf("detect format: %w", err)
}
switch format {
case FormatMP3:
err = writeMp3Tags(logger, filePath, changes)
case FormatFLAC:
err = writeFlacTags(logger, filePath, changes)
case FormatWAV:
err = writeWavTags(logger, filePath, changes)
case FormatOGG:
err = writeOggTags(logger, filePath, changes)
default:
err = fmt.Errorf("%w: %s", errUnsupportedFormat, format)
}
if err != nil {
return fmt.Errorf("write file tags: %w", err)
}
return nil
}
// WriteTrackTagsByPath resolves a file path to its audio_file.id and
// delegates to WriteTrackTags. This is the frontend-facing entry
// point since the frontend identifies tracks by FilePath.
func (tw *TagWriter) WriteTrackTagsByPath(filePath string, changes TagChanges) error {
ctx := context.Background()
audioFile, err := tw.db.Queries.GetAudioFileByPath(ctx, filePath)
if err != nil {
return fmt.Errorf("resolve track by path %q: %w", filePath, err)
}
return tw.WriteTrackTags(audioFile.ID, changes)
}
// CancelBatchWrite signals the in-progress batch write to stop after
// the current track completes.
func (tw *TagWriter) CancelBatchWrite() {
ch := tw.cancelBatch
if ch != nil {
select {
case <-ch:
// Already closed.
default:
close(ch)
}
}
}
// BatchWriteTrackTags applies the same TagChanges to every file in
// filePaths. It processes tracks sequentially, emits a
// BatchWriteProgress event after each track, and continues past
// individual failures. Returns a BatchResult summarising outcomes.
func (tw *TagWriter) BatchWriteTrackTags(
filePaths []string,
changes TagChanges,
) BatchResult {
start := time.Now()
total := len(filePaths)
result := BatchResult{
Total: total,
Failures: []BatchFailure{},
}
if total == 0 || len(changes) == 0 {
return result
}
// Set up cancellation channel.
tw.cancelBatch = make(chan struct{})
defer func() { tw.cancelBatch = nil }()
// Suppress per-track TrackMetadataChanged events — we emit one
// at the end instead.
tw.suppressEvents = true
defer func() { tw.suppressEvents = false }()
for i, filePath := range filePaths {
// Check for cancellation before each track.
select {
case <-tw.cancelBatch:
result.Cancelled = true
tw.logger.Info("batch write cancelled",
"at", i,
"total", total,
)
default:
}
if result.Cancelled {
break
}
err := tw.WriteTrackTagsByPath(filePath, changes)
if err != nil {
result.Failed++
result.Failures = append(result.Failures, BatchFailure{
FilePath: filePath,
Error: err.Error(),
})
tw.logger.Warn("batch track failed",
"path", filePath,
"err", err,
"index", i+1,
"total", total,
)
} else {
result.Succeeded++
}
// Emit progress after each track (success or failure).
events.Emit(tw.ctx,
events.BatchWriteProgress,
map[string]any{
"current": i + 1,
"total": total,
"filePath": filePath,
"succeeded": result.Succeeded,
"failed": result.Failed,
},
)
}
// Emit a single TrackMetadataChanged after the batch completes
// so the library store invalidates once rather than per-track.
events.Emit(tw.ctx, events.TrackMetadataChanged,
map[string]any{
"batch": true,
"total": result.Succeeded,
},
)
tw.logger.Info("batch write complete",
"total", total,
"succeeded", result.Succeeded,
"failed", result.Failed,
"cancelled", result.Cancelled,
"duration", time.Since(start),
)
return result
}