Files
yellowjacket/backend/library/scan_queue.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

330 lines
8.2 KiB
Go

package library
import (
"fmt"
"yellowjacket/backend/events"
)
// scanQueueEntry holds the metadata needed to scan a single library.
type scanQueueEntry struct {
libraryID int64
libraryName string
libraryPath string
}
// ScanLibrary queues a scan for the library with the given database ID.
// If no scan is active the library is scanned immediately; otherwise it
// is appended to the queue. Duplicate requests (same library already
// scanning or already queued) are silently ignored.
func (l *Library) ScanLibrary(id int64) error {
lib, err := l.db.Queries.GetLibrary(l.ctx, id)
if err != nil {
return fmt.Errorf("could not get library %d: %w", id, err)
}
l.mu.Lock()
// Silent dedup: already scanning this library.
if l.currentScanLibraryID == id {
l.mu.Unlock()
return nil
}
// Silent dedup: already queued.
for _, entry := range l.scanQueue {
if entry.libraryID == id {
l.mu.Unlock()
return nil
}
}
entry := scanQueueEntry{
libraryID: lib.ID,
libraryName: lib.Name,
libraryPath: lib.Path,
}
if !l.scanActive {
l.scanActive = true
l.currentScanLibraryID = entry.libraryID
l.currentScanLibraryName = entry.libraryName
l.mu.Unlock()
go l.startScan(entry)
return nil
}
// A scan is already running — queue this library.
l.scanQueue = append(l.scanQueue, entry)
queueLength := len(l.scanQueue)
l.mu.Unlock()
l.emit(events.LibraryScanQueued, map[string]any{
"libraryId": lib.ID,
"libraryName": lib.Name,
"queueLength": queueLength,
})
// Registering the queued job takes l.mu again, so it must happen
// after the unlock above.
l.registerQueuedScanJob(entry)
return nil
}
// ScanAllLibraries queries all libraries from the database and queues
// each one for scanning. Existing dedup logic ensures no duplicates.
func (l *Library) ScanAllLibraries() error {
libs, err := l.db.Queries.GetAllLibraries(l.ctx)
if err != nil {
return fmt.Errorf("could not get all libraries: %w", err)
}
for _, lib := range libs {
if err := l.ScanLibrary(lib.ID); err != nil {
l.logger.Warn(
"could not queue library for scan",
"libraryID", lib.ID,
"libraryName", lib.Name,
"err", err,
)
}
}
return nil
}
// SoftScanAllLibraries performs a lightweight launch-time scan.
// First it claims any orphaned tracks (library_id=0) left over from
// the pre-multi-library schema. Then for each library it compares
// the number of audio files on disk against the track count in the
// database. Only libraries where the counts differ (files added or
// removed since last scan) are queued for a full scan. Libraries
// that are unchanged are silently skipped — no progress bar, no
// scan events.
func (l *Library) SoftScanAllLibraries() error {
libs, err := l.db.Queries.GetAllLibraries(l.ctx)
if err != nil {
return fmt.Errorf("could not get all libraries: %w", err)
}
// Claim orphaned tracks from the pre-multi-library schema.
// Tracks with library_id=0 exist from before the migration and
// need to be assigned to the library whose path matches.
// SAFETY: Hand-crafted UPDATE — path-prefix LIKE matching with
// dynamic library_id unsupported by sqlc. No user input.
for _, lib := range libs {
result, claimErr := l.db.ExecContext(
`UPDATE audio_files SET library_id = ?`+
` WHERE library_id = 0`+
` AND file_path LIKE ? || '%'`,
lib.ID, lib.Path+"/",
)
if claimErr != nil {
l.logger.Warn("soft scan: could not claim orphaned tracks",
"libraryID", lib.ID, "err", claimErr)
} else if claimed, _ := result.RowsAffected(); claimed > 0 {
l.logger.Info("soft scan: claimed orphaned tracks",
"libraryID", lib.ID, "libraryName", lib.Name,
"claimed", claimed)
}
}
for _, lib := range libs {
// A scan the user paused stays paused across restarts — the
// soft scan must not quietly start it again behind their back.
// RestorePausedScans has already surfaced it in the jobs panel
// with a resume button.
if l.isScanPausedPersistently(lib.ID) {
l.logger.Info(
"soft scan: library scan is paused, skipping",
"libraryID", lib.ID,
"libraryName", lib.Name,
)
continue
}
dbCount, countErr := l.db.Queries.CountAudioFiles(
l.ctx, lib.ID,
)
if countErr != nil {
l.logger.Warn(
"soft scan: could not count DB tracks, queueing full scan",
"libraryID", lib.ID,
"libraryName", lib.Name,
"err", countErr,
)
_ = l.ScanLibrary(lib.ID)
continue
}
dbModTime, modErr := l.db.Queries.GetLibraryMaxModifiedAt(
l.ctx, lib.ID,
)
if modErr != nil {
l.logger.Warn(
"soft scan: could not read newest mtime, queueing full scan",
"libraryID", lib.ID,
"libraryName", lib.Name,
"err", modErr,
)
_ = l.ScanLibrary(lib.ID)
continue
}
// Excluded paths are on disk and deliberately not in the
// database, so the survey must skip them or the two counts
// disagree forever and every launch queues a full scan.
diskCount, diskModTime := surveyAudioFiles(
lib.Path, l.excludedPathSet(lib.ID),
)
// A newer file on disk than anything on record means something
// was edited in place since the last scan. An older newest-mtime
// is not evidence of a change: deleting the newest file lowers it
// while the count check already covers that case.
staleTags := diskModTime > dbModTime
if diskCount == dbCount && !staleTags {
l.logger.Info(
"soft scan: library unchanged, skipping",
"libraryID", lib.ID,
"libraryName", lib.Name,
"tracks", dbCount,
)
continue
}
l.logger.Info(
"soft scan: library changed, queueing scan",
"libraryID", lib.ID,
"libraryName", lib.Name,
"diskFiles", diskCount,
"dbTracks", dbCount,
"diskModTime", diskModTime,
"dbModTime", dbModTime,
"reason", softScanReason(diskCount != dbCount, staleTags),
)
if err := l.ScanLibrary(lib.ID); err != nil {
l.logger.Warn(
"soft scan: could not queue library",
"libraryID", lib.ID,
"libraryName", lib.Name,
"err", err,
)
}
}
return nil
}
// softScanReason labels why the soft scan queued a library, for the log.
func softScanReason(countChanged, staleTags bool) string {
switch {
case countChanged && staleTags:
return "file count mismatch and modified files"
case countChanged:
return "file count mismatch"
default:
return "modified files"
}
}
// CancelCurrentScan cancels only the currently scanning library.
// The next queued library (if any) starts automatically when the
// current scan's goroutine completes.
func (l *Library) CancelCurrentScan() {
l.mu.Lock()
cancel := l.scanCancel
l.mu.Unlock()
if cancel != nil {
cancel()
}
}
// CancelAllScans cancels the current scan and clears the entire
// queue so no further libraries are scanned.
func (l *Library) CancelAllScans() {
l.mu.Lock()
l.scanQueue = nil
cancel := l.scanCancel
l.mu.Unlock()
if cancel != nil {
cancel()
}
}
// GetScanQueueLength returns the number of libraries waiting in the
// scan queue (excludes the currently scanning library).
func (l *Library) GetScanQueueLength() int {
l.mu.Lock()
defer l.mu.Unlock()
return len(l.scanQueue)
}
// QueuedLibraryNames returns the display names of libraries waiting
// in the scan queue, in FIFO order.
func (l *Library) QueuedLibraryNames() []string {
l.mu.Lock()
defer l.mu.Unlock()
names := make([]string, len(l.scanQueue))
for i, entry := range l.scanQueue {
names[i] = entry.libraryName
}
return names
}
// startScan runs the scan for a single library entry and then drains
// the queue. It is always called in a new goroutine.
func (l *Library) startScan(entry scanQueueEntry) {
l.scanInternal(entry.libraryID, entry.libraryName, entry.libraryPath)
l.drainQueue()
}
// drainQueue is called after each scan completes. If the queue is
// non-empty the next entry is popped and scanned; otherwise the
// scan pipeline is marked idle.
func (l *Library) drainQueue() {
l.mu.Lock()
if len(l.scanQueue) > 0 {
next := l.scanQueue[0]
l.scanQueue = l.scanQueue[1:]
l.currentScanLibraryID = next.libraryID
l.currentScanLibraryName = next.libraryName
l.mu.Unlock()
go l.startScan(next)
return
}
l.currentScanLibraryID = 0
l.currentScanLibraryName = ""
l.scanActive = false
hooks := l.scanHooks
l.mu.Unlock()
l.emit(events.LibraryScanQueueDrained)
if hooks.OnAllScansComplete != nil {
hooks.OnAllScansComplete()
}
}