feat(server/m7-380): one-shot MBID backfill worker for existing libraries

Adds a boot-time goroutine that walks albums with NULL mbid, re-reads tags from one track per album via dhowden/tag, and persists album + artist MBIDs. Healed albums also get their cover_art_source='none' cleared so the enricher's next batch retries the MBCAA fetch. Caps at 5000 albums per boot to avoid stalling startup on huge libraries. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-05-04 19:40:27 -04:00
parent 26f4c7a79f
commit 34615fffbd
6 changed files with 228 additions and 0 deletions
@@ -83,6 +83,16 @@ func run() error {
 	})
 	coverEnricher := coverart.NewEnricher(pool, logger.With("component", "coverart"), coverFetcher, cfg.Library.CoverArtFromMBCAA)

+	// One-shot MBID backfill: walks albums with NULL mbid, reads tags from
+	// one track per album, persists the MBID. Filling in MBIDs unblocks the
+	// cover enricher for libraries that were imported before scanner-side
+	// MBID extraction landed (M7 #380).
+	go func() {
+		if _, err := library.BackfillMBIDs(ctx, pool, logger.With("component", "mbid_backfill"), 5000); err != nil {
+			logger.Warn("mbid backfill failed", "err", err)
+		}
+	}()
+
 	if cfg.Library.ScanOnStartup && len(cfg.Library.ScanPaths) > 0 {
 		go func() {
 			if _, err := scanner.Scan(ctx); err != nil {
@@ -347,6 +347,59 @@ func (q *Queries) ListAlbumsByGenre(ctx context.Context, arg ListAlbumsByGenrePa
 	return items, nil
 }

+const listAlbumsMissingMbidWithTrack = `-- name: ListAlbumsMissingMbidWithTrack :many
+SELECT a.id          AS album_id,
+       a.artist_id   AS artist_id,
+       a.title       AS title,
+       t.file_path   AS track_file_path
+  FROM albums a
+  JOIN LATERAL (
+    SELECT file_path
+      FROM tracks
+     WHERE album_id = a.id
+     ORDER BY disc_number NULLS LAST, track_number NULLS LAST, id
+     LIMIT 1
+  ) t ON true
+ WHERE a.mbid IS NULL
+ ORDER BY a.created_at ASC
+ LIMIT $1
+`
+
+type ListAlbumsMissingMbidWithTrackRow struct {
+	AlbumID       pgtype.UUID
+	ArtistID      pgtype.UUID
+	Title         string
+	TrackFilePath string
+}
+
+// One-shot MBID backfill: returns each album where mbid IS NULL alongside
+// one of its tracks' file_path so the worker can re-read tags. LIMIT
+// supplied by caller for batching/progress purposes.
+func (q *Queries) ListAlbumsMissingMbidWithTrack(ctx context.Context, limit int32) ([]ListAlbumsMissingMbidWithTrackRow, error) {
+	rows, err := q.db.Query(ctx, listAlbumsMissingMbidWithTrack, limit)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	var items []ListAlbumsMissingMbidWithTrackRow
+	for rows.Next() {
+		var i ListAlbumsMissingMbidWithTrackRow
+		if err := rows.Scan(
+			&i.AlbumID,
+			&i.ArtistID,
+			&i.Title,
+			&i.TrackFilePath,
+		); err != nil {
+			return nil, err
+		}
+		items = append(items, i)
+	}
+	if err := rows.Err(); err != nil {
+		return nil, err
+	}
+	return items, nil
+}
+
 const listAlbumsNewest = `-- name: ListAlbumsNewest :many
 SELECT id, title, sort_title, artist_id, release_date, mbid, cover_art_path, created_at, updated_at, cover_art_source FROM albums ORDER BY created_at DESC LIMIT $1 OFFSET $2
 `
@@ -25,6 +25,20 @@ func (q *Queries) ClearAlbumCover(ctx context.Context, id pgtype.UUID) error {
 	return err
 }

+const clearAlbumCoverNone = `-- name: ClearAlbumCoverNone :exec
+UPDATE albums
+   SET cover_art_source = NULL
+ WHERE id = $1 AND cover_art_source = 'none'
+`
+
+// M7 #380: when MBID backfill heals an album, clear its 'none' cover_art_source
+// back to NULL so the enricher's next batch retries the MBCAA fetch.
+// Idempotent — only updates rows currently stamped 'none'.
+func (q *Queries) ClearAlbumCoverNone(ctx context.Context, id pgtype.UUID) error {
+	_, err := q.db.Exec(ctx, clearAlbumCoverNone, id)
+	return err
+}
+
 const countAlbumCoverSources = `-- name: CountAlbumCoverSources :one
 SELECT
    COUNT(*) FILTER (WHERE cover_art_source IS NULL)     ::bigint AS untried,
@@ -104,3 +104,23 @@ UPDATE albums
   SET mbid = $2,
       updated_at = now()
 WHERE id = $1 AND mbid IS NULL;
+
+-- name: ListAlbumsMissingMbidWithTrack :many
+-- One-shot MBID backfill: returns each album where mbid IS NULL alongside
+-- one of its tracks' file_path so the worker can re-read tags. LIMIT
+-- supplied by caller for batching/progress purposes.
+SELECT a.id          AS album_id,
+       a.artist_id   AS artist_id,
+       a.title       AS title,
+       t.file_path   AS track_file_path
+  FROM albums a
+  JOIN LATERAL (
+    SELECT file_path
+      FROM tracks
+     WHERE album_id = a.id
+     ORDER BY disc_number NULLS LAST, track_number NULLS LAST, id
+     LIMIT 1
+  ) t ON true
+ WHERE a.mbid IS NULL
+ ORDER BY a.created_at ASC
+ LIMIT $1;
@@ -44,6 +44,14 @@ UPDATE albums
       cover_art_source = NULL
 WHERE id = $1;

+-- name: ClearAlbumCoverNone :exec
+-- M7 #380: when MBID backfill heals an album, clear its 'none' cover_art_source
+-- back to NULL so the enricher's next batch retries the MBCAA fetch.
+-- Idempotent — only updates rows currently stamped 'none'.
+UPDATE albums
+   SET cover_art_source = NULL
+ WHERE id = $1 AND cover_art_source = 'none';
+
 -- name: CountAlbumCoverSources :one
 -- Diagnostic for the admin overview: counts by source, including NULL.
 -- Returns one row with named tallies.
@@ -0,0 +1,123 @@
+package library
+
+import (
+	"context"
+	"fmt"
+	"log/slog"
+	"os"
+
+	"github.com/dhowden/tag"
+	"github.com/jackc/pgx/v5/pgxpool"
+
+	"git.fabledsword.com/bvandeusen/minstrel/internal/db/dbq"
+)
+
+// BackfillMBIDsResult tallies what the one-shot worker did.
+type BackfillMBIDsResult struct {
+	Processed int // Albums considered (mbid IS NULL).
+	Healed    int // Albums whose mbid we set.
+	Skipped   int // Albums where the file was missing or had no MBID tag.
+}
+
+// BackfillMBIDs walks albums with NULL mbid, opens one track per album,
+// extracts MusicBrainz IDs from the tags, and persists them. Also heals
+// the artist's mbid when the same tag-read yields one. When a row is
+// healed, any 'none' cover_art_source on the album is cleared back to
+// NULL so the enricher batch retries.
+//
+// Intended use: one-shot at server boot. Re-running it after the first
+// pass costs the same N file reads but is a no-op for healed rows.
+//
+// limit caps how many albums are processed in one call. Pass -1 for no
+// cap. The caller normally batches (e.g. limit=500 per boot) to avoid
+// stalling startup on huge libraries.
+func BackfillMBIDs(ctx context.Context, pool *pgxpool.Pool, logger *slog.Logger, limit int) (BackfillMBIDsResult, error) {
+	q := dbq.New(pool)
+	queryLimit := int32(limit)
+	if limit < 0 {
+		queryLimit = 1<<31 - 1 // effectively unbounded
+	}
+
+	rows, err := q.ListAlbumsMissingMbidWithTrack(ctx, queryLimit)
+	if err != nil {
+		return BackfillMBIDsResult{}, fmt.Errorf("list missing mbid: %w", err)
+	}
+
+	var res BackfillMBIDsResult
+	for i, r := range rows {
+		if ctx.Err() != nil {
+			return res, ctx.Err()
+		}
+		res.Processed++
+
+		albumMBID, artistMBID := readMBIDsForFile(r.TrackFilePath, logger)
+		if albumMBID == "" {
+			res.Skipped++
+			continue
+		}
+
+		// Heal album.
+		m := albumMBID
+		if err := q.SetAlbumMbidIfNull(ctx, dbq.SetAlbumMbidIfNullParams{
+			ID: r.AlbumID, Mbid: &m,
+		}); err != nil {
+			logger.Warn("mbid backfill: set album mbid failed",
+				"album_id", r.AlbumID, "err", err)
+			res.Skipped++
+			continue
+		}
+
+		// Heal artist (if we got one).
+		if artistMBID != "" {
+			am := artistMBID
+			if err := q.SetArtistMbidIfNull(ctx, dbq.SetArtistMbidIfNullParams{
+				ID: r.ArtistID, Mbid: &am,
+			}); err != nil {
+				logger.Warn("mbid backfill: set artist mbid failed",
+					"artist_id", r.ArtistID, "err", err)
+			}
+		}
+
+		// Clear cover_art_source = 'none' so the cover enricher's next
+		// batch retries the MBCAA fetch for this album.
+		if err := q.ClearAlbumCoverNone(ctx, r.AlbumID); err != nil {
+			logger.Warn("mbid backfill: clear cover none failed",
+				"album_id", r.AlbumID, "err", err)
+		}
+
+		res.Healed++
+
+		// Progress log every 100 healed.
+		if (i+1)%100 == 0 {
+			logger.Info("mbid backfill progress",
+				"processed", res.Processed, "healed", res.Healed, "skipped", res.Skipped)
+		}
+	}
+
+	logger.Info("mbid backfill complete",
+		"processed", res.Processed,
+		"healed", res.Healed,
+		"skipped", res.Skipped)
+	return res, nil
+}
+
+// readMBIDsForFile opens a single audio file, reads tags via dhowden/tag,
+// and returns extracted album + artist MBIDs. Returns ("", "") when the
+// file can't be opened or the tag read fails — those errors are logged
+// at Warn level but not propagated, so a single broken file doesn't
+// halt the whole backfill.
+func readMBIDsForFile(path string, logger *slog.Logger) (albumMBID, artistMBID string) {
+	f, err := os.Open(path)
+	if err != nil {
+		logger.Warn("mbid backfill: open failed", "path", path, "err", err)
+		return "", ""
+	}
+	defer func() { _ = f.Close() }()
+
+	meta, err := tag.ReadFrom(f)
+	if err != nil {
+		logger.Warn("mbid backfill: tag read failed", "path", path, "err", err)
+		return "", ""
+	}
+	return extractMBIDs(meta.Raw())
+}