fix(deezer): detect compilations credited to a single main artist

beetsplug/deezer.py

@@ -12,6 +12,7 @@
 from beets.autotag import AlbumInfo, TrackInfo
 from beets.dbcore import types
 from beets.exceptions import UserError
+from beets.importer.tasks import SINGLE_ARTIST_THRESH
 from beets.metadata_plugins import IDResponse, SearchApiMetadataSourcePlugin
 
 VARIOUS_ARTISTS_ID = 5080
@@ -112,7 +113,15 @@ def album_for_id(self, album_id: str) -> AlbumInfo | None:
         for track in tracks:
             track.medium_total = medium_totals[track.medium]
 
-        is_va = str(album_data["artist"]["id"]) == str(VARIOUS_ARTISTS_ID)
+        album_artist_id = str(album_data["artist"]["id"])
+        # Deezer assigns some compilations a single "main" artist instead of
+        # the Various Artists entity, leaving them undetected. Treat a release
+        # as VA when its album artist performs on fewer tracks than the
+        # importer's own single-artist plurality threshold.
+        is_va = album_artist_id == str(VARIOUS_ARTISTS_ID) or (
+            sum(t.artist_id == album_artist_id for t in tracks)
+            < len(tracks) * SINGLE_ARTIST_THRESH
+        )
         if is_va:
             va_name = config["va_name"].as_str()
             artist = va_name

docs/changelog.rst

@@ -48,6 +48,9 @@ Bug fixes
 - :doc:`plugins/lyrics`: ``beet lyrics`` no longer crashes with an
   ``AttributeError`` on tracks that have no stored lyrics when ``force`` is
   enabled; a missing lyrics body is now treated as empty text. :bug:`6860`
+- :doc:`plugins/deezer`: Detect compilations that Deezer credits to a single
+  "main" artist instead of "Various Artists", so they are tagged as such rather
+  than getting every track artist in the album artist field. :bug:`4057`
 - Flexible attributes whose names contain uppercase characters (for example
   ``beet import --set Tag_With_Uppercase=true``) can now be found by queries.
   Field names are lowercased when a query is parsed, so such attributes could

test/plugins/test_deezer.py

@@ -1,8 +1,12 @@
+from typing import Any
+
 import pytest
 
 from beets.library import Item
 from beetsplug.deezer import DeezerPlugin
 
+JSONDict = dict[str, Any]
+
 
 @pytest.fixture
 def plugin():
@@ -41,6 +45,19 @@ def test_album_query_omits_artist_for_various_artists(self, plugin):
         assert query == 'album:"Album"'
 
 
+def make_track(artist_id: int, position: int) -> JSONDict:
+    """Build a minimal Deezer track payload credited to ``artist_id``."""
+    return {
+        "title": f"Track {position}",
+        "id": 1000 + position,
+        "link": f"https://www.deezer.com/track/{1000 + position}",
+        "duration": 200,
+        "track_position": position,
+        "disk_number": 1,
+        "artist": {"id": artist_id, "name": f"Artist {artist_id}"},
+    }
+
+
 class TestGetTrack:
     def track_data(self, **fields):
         return {
@@ -72,3 +89,100 @@ def test_tolerates_missing_artist_and_contributors(self, plugin):
 
         assert track.artist is None
         assert track.artist_id is None
+
+
+class TestVariousArtistsDetection:
+    def make_album(
+        self, album_artist_id: int, contributor_ids: list[int]
+    ) -> JSONDict:
+        """Build a minimal Deezer album payload."""
+        return {
+            "title": "Some Album",
+            "link": "https://www.deezer.com/album/1",
+            "record_type": "album",
+            "label": "Some Label",
+            "release_date": "2017-01-01",
+            "artist": {
+                "id": album_artist_id,
+                "name": f"Artist {album_artist_id}",
+            },
+            "contributors": [
+                {"id": aid, "name": f"Artist {aid}"} for aid in contributor_ids
+            ],
+            "cover_xl": None,
+        }
+
+    def mock_album(
+        self,
+        requests_mock,
+        album_artist_id,
+        track_artist_ids,
+        contributor_ids=None,
+    ):
+        """Mock the album and album-tracks endpoints for album id 1.
+
+        ``contributor_ids`` overrides the album-level contributors, which
+        otherwise cover every artist involved.
+        """
+        if contributor_ids is None:
+            contributor_ids = list(
+                dict.fromkeys([album_artist_id, *track_artist_ids])
+            )
+        requests_mock.get(
+            f"{DeezerPlugin.album_url}1",
+            json=self.make_album(album_artist_id, contributor_ids),
+        )
+        requests_mock.get(
+            f"{DeezerPlugin.album_url}1/tracks",
+            json={
+                "data": [
+                    make_track(aid, i)
+                    for i, aid in enumerate(track_artist_ids, start=1)
+                ]
+            },
+        )
+
+    def test_compilation_with_non_va_album_artist(self, plugin, requests_mock):
+        # Album credited to a single "main" artist that performs on only one
+        # of many tracks: this is a compilation and should be flagged as VA.
+        self.mock_album(
+            requests_mock,
+            album_artist_id=100,
+            track_artist_ids=[100, 200, 300, 400, 500, 600],
+        )
+
+        album_info = plugin.album_for_id("1")
+
+        assert album_info is not None
+        assert album_info.va is True
+        assert album_info.artist == "Various Artists"
+
+    def test_plurality_artist_album_not_va(self, plugin, requests_mock):
+        # Album whose main artist performs on 2 of 5 tracks (40%) is above
+        # the importer's single-artist threshold and stays single-artist.
+        self.mock_album(
+            requests_mock,
+            album_artist_id=100,
+            track_artist_ids=[100, 100, 200, 300, 400],
+            contributor_ids=[100],
+        )
+
+        album_info = plugin.album_for_id("1")
+
+        assert album_info is not None
+        assert album_info.va is False
+        assert album_info.artist == "Artist 100"
+
+    def test_single_artist_album_not_va(self, plugin, requests_mock):
+        # Album whose main artist performs on every track is not a compilation.
+        self.mock_album(
+            requests_mock,
+            album_artist_id=100,
+            track_artist_ids=[100, 100, 100, 100],
+        )
+
+        album_info = plugin.album_for_id("1")
+
+        assert album_info is not None
+        assert album_info.va is False
+        assert album_info.artist == "Artist 100"