Files
yellowjacket/backend/tagwriter/dbsync.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

388 lines
11 KiB
Go

package tagwriter
import (
"bytes"
"context"
"crypto/sha256"
"database/sql"
"encoding/hex"
"fmt"
"image"
"image/jpeg"
_ "image/png" // Register PNG decoder for cover art thumbnails.
"log/slog"
"os"
"path/filepath"
"golang.org/x/image/draw"
"yellowjacket/backend/coverart"
"yellowjacket/backend/database"
"yellowjacket/backend/database/sql/sqlcgen"
"yellowjacket/backend/metadata"
)
// dbSyncParams holds the context needed by syncDatabase to update the
// database after a successful file tag write.
type dbSyncParams struct {
audioFileID int64
filePath string
changes TagChanges
oldFile sqlcgen.AudioFile
}
// syncDatabase runs the database updates for a tag write inside a
// single transaction: the file's own tag columns, its album and artist
// links, its genres, its cover art, and the FTS row. If anything fails
// the whole transaction rolls back.
//
// This used to be four times longer, and three of the four parts were
// bookkeeping for tables that no longer exist: relinking a recording to
// a new artist_credit, unlinking and relinking release_group_recordings,
// and then three orphan sweeps to delete whichever of those rows the
// relink had stranded. Tags live on the file now, so changing a tag is
// an UPDATE, and nothing can be stranded by one.
func syncDatabase(
ctx context.Context,
logger *slog.Logger,
db *database.DB,
params dbSyncParams,
) error {
tx, err := db.BeginTx()
if err != nil {
return fmt.Errorf("begin sync tx: %w", err)
}
defer func() { _ = tx.Rollback() }() // no-op after commit
txq := db.Queries.WithTx(tx)
old := params.oldFile
// ------------------------------------------------------------------
// 1. Resolve the artist, if it changed.
// ------------------------------------------------------------------
artistCredit := old.ArtistCredit
artistID := old.ArtistID
if v, ok := params.changes[FieldArtist].(string); ok {
artistCredit = v
artist, artErr := txq.UpsertArtist(ctx, sqlcgen.UpsertArtistParams{Name: v})
if artErr != nil {
return fmt.Errorf("upsert artist: %w", artErr)
}
artistID = sql.NullInt64{Int64: artist.ID, Valid: true}
}
// ------------------------------------------------------------------
// 2. Resolve the album, if it changed.
// ------------------------------------------------------------------
albumID := old.AlbumID
if newAlbumName, ok := params.changes[FieldAlbum].(string); ok {
albumCredit := artistCredit
albumArtistID := artistID
if aav, aaOK := params.changes[FieldAlbumArtist].(string); aaOK && aav != "" {
albumCredit = aav
aaArtist, aaErr := txq.UpsertArtist(ctx, sqlcgen.UpsertArtistParams{Name: aav})
if aaErr != nil {
return fmt.Errorf("upsert album artist: %w", aaErr)
}
albumArtistID = sql.NullInt64{Int64: aaArtist.ID, Valid: true}
}
year := old.Year
if yv, yOK := asInt(params.changes[FieldYear]); yOK {
year = toNullInt64(yv)
}
album, albErr := txq.UpsertAlbum(ctx, sqlcgen.UpsertAlbumParams{
Name: newAlbumName,
ArtistCredit: albumCredit,
ArtistID: albumArtistID,
Year: year,
})
if albErr != nil {
return fmt.Errorf("upsert album: %w", albErr)
}
albumID = sql.NullInt64{Int64: album.ID, Valid: true}
}
// ------------------------------------------------------------------
// 3. Genres, if they changed.
// ------------------------------------------------------------------
if newGenre, ok := params.changes[FieldGenre].(string); ok {
if delErr := txq.DeleteFileGenres(ctx, params.audioFileID); delErr != nil {
return fmt.Errorf("delete file genres: %w", delErr)
}
for _, gName := range metadata.ParseGenres(newGenre) {
g, gErr := txq.UpsertGenre(ctx, gName)
if gErr != nil {
return fmt.Errorf("upsert genre %q: %w", gName, gErr)
}
if linkErr := txq.LinkFileGenre(ctx, sqlcgen.LinkFileGenreParams{
AudioFileID: params.audioFileID,
GenreID: g.ID,
}); linkErr != nil {
return fmt.Errorf("link file genre: %w", linkErr)
}
}
}
// ------------------------------------------------------------------
// 4. Cover art, if it changed. It belongs to the album, so a file
// with no album has nowhere to put it.
// ------------------------------------------------------------------
if _, hasCoverArt := params.changes[FieldCoverArt]; hasCoverArt && albumID.Valid {
coverArtID := sql.NullInt64{}
if data, isBytes := asBytes(params.changes[FieldCoverArt]); isBytes && len(data) > 0 {
newID, caErr := saveCoverArtAndSync(ctx, logger, txq, data)
if caErr != nil {
logger.Warn("cover art sync failed", "err", caErr)
} else {
coverArtID = sql.NullInt64{Int64: newID, Valid: true}
}
}
if upErr := txq.SetAlbumCoverArt(ctx, sqlcgen.SetAlbumCoverArtParams{
CoverArtID: coverArtID,
ID: albumID.Int64,
}); upErr != nil {
logger.Warn("update album cover art failed", "err", upErr, "albumID", albumID.Int64)
}
}
// ------------------------------------------------------------------
// 5. The file's own tag columns.
// ------------------------------------------------------------------
title := old.Title
if v, ok := params.changes[FieldTitle].(string); ok {
title = v
}
year := old.Year
if v, ok := asInt(params.changes[FieldYear]); ok {
year = toNullInt64(v)
}
trackNum := old.TrackNumber
if v, ok := asInt(params.changes[FieldTrackNumber]); ok {
trackNum = toNullInt64(v)
}
discNum := old.DiscNumber
if v, ok := asInt(params.changes[FieldDiscNumber]); ok {
discNum = toNullInt64(v)
}
composer := old.Composer
if v, ok := params.changes[FieldComposer].(string); ok {
composer = v
}
// Writing tags rewrites the file, changing its mtime and possibly
// its size. Recording the new values keeps the scan from mistaking
// YellowJacket's own edit for an external one and re-importing.
modifiedAt, fileSize := old.ModifiedAt, old.FileSize
if info, statErr := os.Stat(params.filePath); statErr != nil {
logger.Warn("could not stat file after tag write",
"path", params.filePath, "err", statErr)
} else {
modifiedAt, fileSize = info.ModTime().Unix(), info.Size()
}
if updErr := txq.UpdateAudioFileTags(ctx, sqlcgen.UpdateAudioFileTagsParams{
Title: title,
ArtistCredit: artistCredit,
ArtistID: artistID,
AlbumID: albumID,
TrackNumber: trackNum,
DiscNumber: discNum,
TotalTracks: old.TotalTracks,
Year: year,
Composer: composer,
Comment: old.Comment,
RecordingMbid: old.RecordingMbid,
SampleRate: old.SampleRate,
BitDepth: old.BitDepth,
Channels: old.Channels,
Bitrate: old.Bitrate,
FileSize: fileSize,
LengthMilliseconds: old.LengthMilliseconds,
ModifiedAt: modifiedAt,
ID: params.audioFileID,
}); updErr != nil {
return fmt.Errorf("update audio file tags: %w", updErr)
}
// ------------------------------------------------------------------
// 6. The FTS row.
// ------------------------------------------------------------------
album := ""
if albumID.Valid {
if row, albErr := txq.GetAlbum(ctx, albumID.Int64); albErr == nil {
album = row.Name
}
}
// SAFETY: FTS5 DELETE on contentless_delete=1 table. Parameterized rowid.
if _, ftsDelErr := tx.ExecContext(ctx,
"DELETE FROM search_index WHERE rowid = ?",
params.audioFileID,
); ftsDelErr != nil {
logger.Warn("FTS5 delete failed", "err", ftsDelErr,
"audioFileID", params.audioFileID)
}
// SAFETY: FTS5 INSERT into contentless table. All values parameterized.
if _, ftsInsErr := tx.ExecContext(ctx,
`INSERT INTO search_index(rowid, file_path, title, artist, album)
VALUES (?, ?, ?, ?, ?)`,
params.audioFileID, params.filePath, title, artistCredit, album,
); ftsInsErr != nil {
logger.Warn("FTS5 insert failed", "err", ftsInsErr,
"audioFileID", params.audioFileID)
}
// Albums and artists left empty by this write are swept by the
// library's own cleanup, which is one query each now.
return tx.Commit()
}
// toNullInt64 converts an int to sql.NullInt64, treating 0 as null.
func toNullInt64(v int) sql.NullInt64 {
if v == 0 {
return sql.NullInt64{}
}
return sql.NullInt64{Int64: int64(v), Valid: true}
}
// saveCoverArtAndSync saves cover art bytes to the covers cache
// directory (with content-hash deduplication), generates sized
// thumbnails, upserts a cover_art DB row, and returns the row ID.
func saveCoverArtAndSync(
ctx context.Context,
logger *slog.Logger,
txq *sqlcgen.Queries,
data []byte,
) (int64, error) {
coverDir, err := coverart.CoversDir()
if err != nil {
return 0, fmt.Errorf("resolve covers dir: %w", err)
}
if err := os.MkdirAll(coverDir, 0o755); err != nil {
return 0, fmt.Errorf("create covers dir: %w", err)
}
// Content-hash filename (same scheme as library/coverart.go): the
// tiers are the only thing stored, and the largest is the cover's
// canonical path. The full-resolution bytes stay in the file the
// user just wrote them to.
hash := sha256.Sum256(data)
hashStr := hex.EncodeToString(hash[:8])
mime := detectMIME(data)
filePath := filepath.Join(
coverDir, coverart.SizedFilename(hashStr, largestTierSuffix),
)
generateSizedVariants(logger, data, coverDir, hashStr)
// Upsert cover_art DB row.
ca, err := txq.UpsertCoverArt(ctx, sqlcgen.UpsertCoverArtParams{
IsEmbedded: true,
FilePath: filePath,
MimeType: mime,
})
if err != nil {
return 0, fmt.Errorf("upsert cover art: %w", err)
}
return ca.ID, nil
}
// thumbnailTier defines a single size tier for generated thumbnails.
type thumbnailTier struct {
suffix string
maxSize int
quality int
}
// largestTierSuffix is the tier stored as a cover's canonical path.
const largestTierSuffix = "_lg"
// thumbnailTiers matches the tiers in library/coverart.go.
var thumbnailTiers = []thumbnailTier{
{suffix: "_sm", maxSize: 100, quality: 75},
{suffix: "_md", maxSize: 200, quality: 80},
{suffix: "_lg", maxSize: 400, quality: 85},
}
// generateSizedVariants creates all thumbnail tiers for the given image.
func generateSizedVariants(logger *slog.Logger, imgData []byte, dir, hashStr string) {
src, _, err := image.Decode(bytes.NewReader(imgData))
if err != nil {
logger.Warn("could not decode image for thumbnails", "err", err)
return
}
bounds := src.Bounds()
srcW := bounds.Dx()
srcH := bounds.Dy()
for _, tier := range thumbnailTiers {
tierPath := filepath.Join(dir, fmt.Sprintf("%s%s.jpg", hashStr, tier.suffix))
// Skip if already exists.
if _, statErr := os.Stat(tierPath); statErr == nil {
continue
}
w, h := fitDimensions(srcW, srcH, tier.maxSize)
dst := image.NewRGBA(image.Rect(0, 0, w, h))
draw.ApproxBiLinear.Scale(dst, dst.Bounds(), src, src.Bounds(), draw.Over, nil)
var buf bytes.Buffer
if encErr := jpeg.Encode(&buf, dst, &jpeg.Options{Quality: tier.quality}); encErr != nil {
logger.Warn("could not encode thumbnail", "tier", tier.suffix, "err", encErr)
continue
}
if writeErr := os.WriteFile(tierPath, buf.Bytes(), 0o644); writeErr != nil {
logger.Warn("could not write thumbnail", "tier", tier.suffix, "err", writeErr)
}
}
}
// fitDimensions calculates output dimensions that fit within maxSize
// while preserving aspect ratio.
func fitDimensions(srcW, srcH, maxSize int) (int, int) {
if srcW <= maxSize && srcH <= maxSize {
return srcW, srcH
}
w, h := maxSize, maxSize
if srcW > srcH {
h = srcH * maxSize / srcW
} else {
w = srcW * maxSize / srcH
}
return w, h
}