fix(library): extract recording MBID → tracks.mbid; unblock LB similarity

tracks.mbid was 100% NULL: the scanner only extracted album + artist
MBIDs, never the recording MBID. The ListenBrainz similarity worker is
gated on tracks.mbid IS NOT NULL, so track_similarity could never
populate — starving For-You/radio's strongest candidate source
(lb_similar) on every library.

- mbids.go: add extractRecordingMBID (mbz.Recording / Picard
  musicbrainz_recordingid). Separate fn so extractMBIDs' signature +
  unit tests stay untouched.
- scanner.go: persist recording MBID via UpsertTrack (heals on the
  file_path conflict, so re-scans backfill for free).
- BackfillTrackMBIDs: one-shot pass mirroring BackfillMBIDs, wired as
  scan Stage 2b (idempotent via SetTrackMbidIfNull, gated by
  BackfillCap, log-only progress).
- migration 0029: tracks_mbid_unique (0002, written when the column
  was always NULL) wrongly assumes one MBID == one track. A recording
  appears on multiple releases, so rows legitimately share a recording
  MBID. Replace with a non-unique partial index. Zero-risk: column is
  100% NULL at migration time.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

internal/db/dbq/tracks.sql.go

@@ -376,6 +376,41 @@ func (q *Queries) ListTracksByAlbum(ctx context.Context, arg ListTracksByAlbumPa
 	return items, nil
 }
 
+const listTracksMissingMbidWithPath = `-- name: ListTracksMissingMbidWithPath :many
+SELECT id, file_path
+  FROM tracks
+ WHERE mbid IS NULL
+ ORDER BY id
+ LIMIT $1
+`
+
+type ListTracksMissingMbidWithPathRow struct {
+	ID       pgtype.UUID
+	FilePath string
+}
+
+// Track recording-MBID backfill: tracks with NULL mbid that still have
+// a file to re-read. $1 caps the batch (mirrors the album backfill).
+func (q *Queries) ListTracksMissingMbidWithPath(ctx context.Context, limit int32) ([]ListTracksMissingMbidWithPathRow, error) {
+	rows, err := q.db.Query(ctx, listTracksMissingMbidWithPath, limit)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	var items []ListTracksMissingMbidWithPathRow
+	for rows.Next() {
+		var i ListTracksMissingMbidWithPathRow
+		if err := rows.Scan(&i.ID, &i.FilePath); err != nil {
+			return nil, err
+		}
+		items = append(items, i)
+	}
+	if err := rows.Err(); err != nil {
+		return nil, err
+	}
+	return items, nil
+}
+
 const searchTracks = `-- name: SearchTracks :many
 SELECT id, title, album_id, artist_id, track_number, disc_number, duration_ms, file_path, file_size, file_format, bitrate, mbid, genre, added_at, updated_at FROM tracks
 WHERE title ILIKE '%' || $1::text || '%'
@@ -437,6 +472,24 @@ func (q *Queries) SearchTracks(ctx context.Context, arg SearchTracksParams) ([]T
 	return items, nil
 }
 
+const setTrackMbidIfNull = `-- name: SetTrackMbidIfNull :exec
+UPDATE tracks
+   SET mbid = $2, updated_at = now()
+ WHERE id = $1 AND mbid IS NULL
+`
+
+type SetTrackMbidIfNullParams struct {
+	ID   pgtype.UUID
+	Mbid *string
+}
+
+// Heal a track's recording MBID only while still NULL — idempotent, so
+// re-running the backfill is a no-op for already-healed rows.
+func (q *Queries) SetTrackMbidIfNull(ctx context.Context, arg SetTrackMbidIfNullParams) error {
+	_, err := q.db.Exec(ctx, setTrackMbidIfNull, arg.ID, arg.Mbid)
+	return err
+}
+
 const upsertTrack = `-- name: UpsertTrack :one
 INSERT INTO tracks (
     title, album_id, artist_id, track_number, disc_number,

internal/db/migrations/0029_tracks_mbid_nonunique.down.sql

@@ -0,0 +1,6 @@
+-- Reverts to the unique partial index. This FAILS if any recording
+-- MBID is shared across releases in the library (the exact case 0029
+-- exists to allow) — dedupe tracks.mbid before rolling back.
+
+DROP INDEX IF EXISTS tracks_mbid_idx;
+CREATE UNIQUE INDEX tracks_mbid_unique ON tracks (mbid) WHERE mbid IS NOT NULL;

internal/db/migrations/0029_tracks_mbid_nonunique.up.sql

@@ -0,0 +1,15 @@
+-- tracks.mbid holds the MusicBrainz *recording* MBID — the id the
+-- ListenBrainz similarity pipeline (ListPlayedTracksNeedingSimilarity,
+-- GetTracksByMBIDs) matches on. A single recording legitimately appears
+-- on multiple releases (album + compilation + single), so multiple
+-- tracks rows share one recording MBID.
+--
+-- tracks_mbid_unique (added in 0002, when this column was always NULL
+-- and never populated) wrongly assumes one MBID == one track. Now that
+-- the scanner extracts the recording MBID, that uniqueness throws
+-- 23505 for any recording present on >1 release. Replace it with a
+-- non-unique partial index serving the same lookups. Zero-risk: the
+-- column is 100% NULL at migration time.
+
+DROP INDEX IF EXISTS tracks_mbid_unique;
+CREATE INDEX IF NOT EXISTS tracks_mbid_idx ON tracks (mbid) WHERE mbid IS NOT NULL;

internal/db/queries/tracks.sql

@@ -19,6 +19,22 @@ ON CONFLICT (file_path) DO UPDATE SET
     updated_at = now()
 RETURNING *;
 
+-- name: ListTracksMissingMbidWithPath :many
+-- Track recording-MBID backfill: tracks with NULL mbid that still have
+-- a file to re-read. $1 caps the batch (mirrors the album backfill).
+SELECT id, file_path
+  FROM tracks
+ WHERE mbid IS NULL
+ ORDER BY id
+ LIMIT $1;
+
+-- name: SetTrackMbidIfNull :exec
+-- Heal a track's recording MBID only while still NULL — idempotent, so
+-- re-running the backfill is a no-op for already-healed rows.
+UPDATE tracks
+   SET mbid = $2, updated_at = now()
+ WHERE id = $1 AND mbid IS NULL;
+
 -- name: GetTrackByID :one
 SELECT * FROM tracks WHERE id = $1;