explicit searching for VA releases when it seems likely

beets/autotag/__init__.py

@@ -74,6 +74,9 @@ SD_PATTERNS = [
     (r'(, )?(pt\.|part) .+', 0.2),
 ]
 
+# Artist signals that indicate "various artists".
+VA_ARTISTS = (u'', u'various artists', u'va', u'unknown')
+
 # Autotagging exceptions.
 class AutotagError(Exception):
     pass
@@ -230,7 +233,7 @@ def _plurality(objs):
             max_freq = freq
             res = obj
 
-    return res
+    return res, len(freqs) <= 1
 
 def current_metadata(items):
     """Returns the most likely artist and album for a set of Items.
@@ -238,10 +241,11 @@ def current_metadata(items):
     """
     keys = 'artist', 'album'
     likelies = {}
+    consensus = {}
     for key in keys:
         values = [getattr(item, key) for item in items]
-        likelies[key] = _plurality(values)
+        likelies[key], consensus[key] = _plurality(values)
-    return likelies['artist'], likelies['album']
+    return likelies['artist'], likelies['album'], consensus['artist']
 
 def order_items(items, trackinfo):
     """Orders the items based on how they match some canonical track
@@ -325,7 +329,7 @@ def distance(items, info):
     """Determines how "significant" an album metadata change would be.
     Returns a float in [0.0,1.0]. The list of items must be ordered.
     """
-    cur_artist, cur_album = current_metadata(items)
+    cur_artist, cur_album, _ = current_metadata(items)
     cur_artist = cur_artist or ''
     cur_album = cur_album or ''
     
@@ -492,7 +496,7 @@ def tag_album(items, search_artist=None, search_album=None):
     May raise an AutotagError if existing metadata is insufficient.
     """
     # Get current metadata.
-    cur_artist, cur_album = current_metadata(items)
+    cur_artist, cur_album, artist_consensus = current_metadata(items)
     log.debug('Tagging %s - %s' % (cur_artist, cur_album))
     
     # The output result tuples (keyed by MB album ID).
@@ -525,6 +529,14 @@ def tag_album(items, search_artist=None, search_album=None):
     else:
         candidates = []
 
+    # Possibly add "various artists" search.
+    if search_album and ((not artist_consensus) or \
+                         (search_artist.lower() in VA_ARTISTS) or \
+                         any(item.comp for item in items)):
+        log.debug('Possibly Various Artists; adding matches.')
+        candidates.extend(mb.match_album(None, search_album, len(items),
+                                         MAX_CANDIDATES))
+
     # Get candidates from plugins.
     candidates.extend(plugins.candidates(items))
     

beets/autotag/mb.py

@@ -226,7 +226,12 @@ def match_album(artist, album, tracks=None, limit=SEARCH_LIMIT):
     optionally, a number of tracks on the album.
     """
     # Build search criteria.
-    criteria = {'artist':  artist, 'release': album}
+    criteria = {'release': album}
+    if artist is not None:
+        criteria['artist'] = artist
+    else:
+        # Various Artists search.
+        criteria['arid'] = VARIOUS_ARTISTS_ID
     if tracks is not None:
         criteria['tracks'] = str(tracks)
 

test/test_autotag.py

@@ -28,26 +28,39 @@ from beets.library import Item
 class PluralityTest(unittest.TestCase):
     def test_plurality_consensus(self):
         objs = [1, 1, 1, 1]
-        obj = autotag._plurality(objs)
+        obj, consensus = autotag._plurality(objs)
         self.assertEqual(obj, 1)
+        self.assertTrue(consensus)
 
     def test_plurality_near_consensus(self):
         objs = [1, 1, 2, 1]
-        obj = autotag._plurality(objs)
+        obj, consensus = autotag._plurality(objs)
         self.assertEqual(obj, 1)
+        self.assertFalse(consensus)
 
     def test_plurality_conflict(self):
         objs = [1, 1, 2, 2, 3]
-        obj = autotag._plurality(objs)
+        obj, consensus = autotag._plurality(objs)
         self.assert_(obj in (1, 2))
+        self.assertFalse(consensus)
 
     def test_current_metadata_finds_pluralities(self):
         items = [Item({'artist': 'The Beetles', 'album': 'The White Album'}),
                  Item({'artist': 'The Beatles', 'album': 'The White Album'}),
                  Item({'artist': 'The Beatles', 'album': 'Teh White Album'})]
-        l_artist, l_album = autotag.current_metadata(items)
+        l_artist, l_album, artist_consensus = autotag.current_metadata(items)
         self.assertEqual(l_artist, 'The Beatles')
         self.assertEqual(l_album, 'The White Album')
+        self.assertFalse(artist_consensus)
+
+    def test_current_metadata_artist_consensus(self):
+        items = [Item({'artist': 'The Beatles', 'album': 'The White Album'}),
+                 Item({'artist': 'The Beatles', 'album': 'The White Album'}),
+                 Item({'artist': 'The Beatles', 'album': 'Teh White Album'})]
+        l_artist, l_album, artist_consensus = autotag.current_metadata(items)
+        self.assertEqual(l_artist, 'The Beatles')
+        self.assertEqual(l_album, 'The White Album')
+        self.assertTrue(artist_consensus)
 
 class AlbumDistanceTest(unittest.TestCase):
     def item(self, title, track, artist='some artist'):