MusicBrainz: Switch track lookup from browse to arid: search

The previous commit's `browse_artist_recordings` call passed
`inc=releases+artist-credits` — but MusicBrainz's recording browse
endpoint rejects `inc=releases` with HTTP 400. The adapter's error
handler returned an empty list, so the Tracks section stayed empty
even though the fix was supposed to populate it.

Browse without release info is useless for our search UI (tracks
would render with no album), so swap to the fielded Lucene search
`arid:<mbid>` on the `/recording` endpoint. That's the canonical MB
pattern for "find recordings by this artist WITH release context":
- arid: search accepts the artist MBID and returns recordings with
  `releases` (release-group, date, media) embedded in each result.
- One API call per lookup, same as browse would have been.

Renamed the method to `search_recordings_by_artist_mbid` so the name
matches its behaviour — it's a search, not a browse. Adapter updated
to call the new name; tests updated to match.

Verified against the live API: Metallica's MBID returns 5 recordings
in ~1.8 seconds (vs the previous 400 error).

core/musicbrainz_client.py

@@ -264,30 +264,35 @@ class MusicBrainzClient:
             return []
 
     @rate_limited
-    def browse_artist_recordings(self, artist_mbid: str,
-                                 limit: int = 100,
-                                 offset: int = 0,
-                                 includes: Optional[List[str]] = None) -> List[Dict[str, Any]]:
-        """Browse recordings (tracks) linked to an artist MBID.
+    def search_recordings_by_artist_mbid(self, artist_mbid: str,
+                                         limit: int = 100) -> List[Dict[str, Any]]:
+        """Search for recordings linked to an artist via Lucene `arid:` query.
 
-        Counterpart to `browse_artist_release_groups` — text search on
-        `/recording?query=...` matches recording TITLES, while browse follows
-        the artist→recording link directly.
+        This is the counterpart to `browse_artist_release_groups` for tracks.
+        The proper "browse" endpoint (`/recording?artist=<mbid>`) rejects
+        `inc=releases`, so we can't get album context per recording from
+        browse — only the track title/length/MBID. Without release info the
+        user would see tracks with no album, which is useless.
+
+        The search endpoint with a fielded `arid:<mbid>` query returns
+        recordings with the `releases` array already embedded (including
+        release-group, date, and media info), which is what the search-tab
+        UI needs.
 
         Args:
             artist_mbid: Artist's MusicBrainz ID
             limit: 1-100 (MB hard cap)
-            offset: Pagination offset
-            includes: e.g. ['releases', 'artist-credits'] to embed linked entities
 
         Returns:
-            List of recording dicts with `id`, `title`, `length`, `disambiguation`,
-            and optionally `releases` / `artist-credit` per includes.
+            List of recording dicts with `id`, `title`, `length`, `score`,
+            `artist-credit`, and `releases` (each with release-group + date).
         """
         try:
-            params = {'artist': artist_mbid, 'fmt': 'json', 'limit': min(limit, 100), 'offset': offset}
-            if includes:
-                params['inc'] = '+'.join(includes)
+            params = {
+                'query': f'arid:{artist_mbid}',
+                'fmt': 'json',
+                'limit': min(limit, 100),
+            }
 
             response = self.session.get(
                 f"{self.BASE_URL}/recording",
@@ -298,10 +303,10 @@ class MusicBrainzClient:
 
             data = response.json()
             recs = data.get('recordings', [])
-            logger.debug(f"Browsed {len(recs)} recordings for artist {artist_mbid}")
+            logger.debug(f"Found {len(recs)} recordings for artist {artist_mbid}")
             return recs
         except Exception as e:
-            logger.error(f"Error browsing recordings for artist {artist_mbid}: {e}")
+            logger.error(f"Error searching recordings for artist {artist_mbid}: {e}")
             return []
 
     @rate_limited

core/musicbrainz_search.py

@@ -429,16 +429,15 @@ class MusicBrainzSearchClient:
             if artist_name:
                 return self._search_tracks_text(title, artist_name, limit)
 
-            # Bare name → artist-first → browse.
+            # Bare name → artist-first → arid: search.
             top = self._resolve_top_artist(query)
             if top:
                 mbid = top.get('id', '')
                 tname = top.get('name', '') or query
-                recs = self._client.browse_artist_recordings(
-                    mbid,
-                    limit=100,
-                    includes=['releases', 'artist-credits'],
-                )
+                # /recording?artist=<mbid> (browse) rejects inc=releases,
+                # so we use the fielded Lucene search arid:<mbid> instead —
+                # that returns recordings with release context inline.
+                recs = self._client.search_recordings_by_artist_mbid(mbid, limit=100)
                 # Browse returns recordings unsorted. Dedupe by normalized
                 # title (MB has many live/compilation variants of the same
                 # song), then sort by release date desc so "newest" tracks

tests/test_musicbrainz_search.py

@@ -274,7 +274,7 @@ def test_search_tracks_bare_query_uses_browse_path():
     client = MusicBrainzSearchClient()
     client._client = MagicMock()
     client._client.search_artist.return_value = [_mk_artist('Metallica', 'mb-1', score=100)]
-    client._client.browse_artist_recordings.return_value = [
+    client._client.search_recordings_by_artist_mbid.return_value = [
         {'id': 'rec-1', 'title': 'One', 'length': 446000,
          'releases': [{'id': 'rel-1', 'title': '...And Justice for All', 'date': '1988',
                        'release-group': {'id': 'rg-1', 'primary-type': 'Album'}}],
@@ -287,7 +287,7 @@ def test_search_tracks_bare_query_uses_browse_path():
 
     tracks = client.search_tracks('metallica', limit=10)
 
-    client._client.browse_artist_recordings.assert_called_once()
+    client._client.search_recordings_by_artist_mbid.assert_called_once()
     client._client.search_recording.assert_not_called()
     assert len(tracks) == 2
     assert {t.name for t in tracks} == {'One', 'Battery'}
@@ -300,7 +300,7 @@ def test_search_tracks_dedupes_by_title():
     client = MusicBrainzSearchClient()
     client._client = MagicMock()
     client._client.search_artist.return_value = [_mk_artist('Metallica', 'mb-1', score=100)]
-    client._client.browse_artist_recordings.return_value = [
+    client._client.search_recordings_by_artist_mbid.return_value = [
         {'id': 'rec-1', 'title': 'One', 'length': 446000,
          'releases': [{'id': 'rel-1', 'title': '...And Justice for All', 'date': '1988'}],
          'artist-credit': [{'name': 'Metallica'}]},
@@ -328,7 +328,7 @@ def test_search_tracks_structured_query_uses_text_path():
 
     client._client.search_recording.assert_called_once()
     client._client.search_artist.assert_not_called()
-    client._client.browse_artist_recordings.assert_not_called()
+    client._client.search_recordings_by_artist_mbid.assert_not_called()
     assert len(tracks) == 1