bugfix: artist matching improvements (spotDL#1417)

Author: xnetcat

spotdl/providers/provider_utils.py

@@ -21,7 +21,7 @@ def _match_percentage(str1: str, str2: str, score_cutoff: float = 0) -> float:
 
     # ! this will throw an error if either string contains a UTF-8 encoded emoji
     try:
-        return fuzz.partial_ratio(str1, str2, score_cutoff=score_cutoff)
+        return fuzz.partial_ratio(str1, str2, processor=None, score_cutoff=score_cutoff)
 
     # ! we build new strings that contain only alphanumerical characters and spaces
     # ! and return the partial_ratio of that
@@ -38,7 +38,7 @@ def _match_percentage(str1: str, str2: str, score_cutoff: float = 0) -> float:
             if each_letter.isalnum() or each_letter.isspace()
         )
 
-        return fuzz.partial_ratio(new_str1, new_str2, score_cutoff=score_cutoff)
+        return fuzz.partial_ratio(new_str1, new_str2, processor=None, score_cutoff=score_cutoff)
 
 
 def _parse_duration(duration: str) -> float:
@@ -135,7 +135,9 @@ def _parse_path_template(path_template, song_object, output_format, short=False)
         artist=_sanitize_filename(song_object.contributing_artists[0]),
         title=_sanitize_filename(song_object.song_name),
         album=_sanitize_filename(song_object.album_name),
-        playlist=_sanitize_filename(song_object.playlist_name) if song_object.playlist_name else "",
+        playlist=_sanitize_filename(song_object.playlist_name)
+        if song_object.playlist_name
+        else "",
         artists=_sanitize_filename(
             ", ".join(song_object.contributing_artists)
             if short is False

spotdl/providers/ytm_provider.py

@@ -154,45 +154,38 @@ def _order_ytm_results(
 
         # Find artist match
         # ! match  = (no of artist names in result) / (no. of artist names on spotify) * 100
-        artist_match_number = 0
+        artist_match_number = 0.0
 
         # ! we use fuzzy matching because YouTube spellings might be mucked up
         if result["type"] == "song":
             for artist in song_artists:
-                if _match_percentage(
-                    unidecode(artist.lower()), unidecode(result["artist"]).lower(), 85
-                ):
-                    artist_match_number += 1
+                artist_match_number += _match_percentage(
+                    unidecode(artist.lower()), unidecode(result["artist"]).lower()
+                )
         else:
             # ! i.e if video
             for artist in song_artists:
                 # ! something like _match_percentage('rionos', 'aiobahn, rionos Motivation
                 # ! (remix)' would return 100, so we're absolutely corrent in matching
                 # ! artists to song name.
-                if _match_percentage(
-                    unidecode(artist.lower()), unidecode(result["name"]).lower(), 85
-                ):
-                    artist_match_number += 1
+                artist_match_number += _match_percentage(
+                    unidecode(artist.lower()), unidecode(result["name"]).lower()
+                )
 
             # we didn't find artist in the video title, so we fallback to
             # detecting song artist in the channel name
             # I am not sure if this won't create false positives
             if artist_match_number == 0:
                 for artist in song_artists:
-                    if _match_percentage(
+                    artist_match_number += _match_percentage(
                         unidecode(artist.lower()),
                         unidecode(result["artist"].lower()),
-                        85,
-                    ):
-                        artist_match_number += 1
+                    )
 
-        # ! Skip if there are no artists in common, (else, results like 'Griffith Swank -
-        # ! Madness' will be the top match for 'Ruelle - Madness')
-        if artist_match_number == 0:
+        artist_match = artist_match_number / len(song_artists)
+        if artist_match < 70:
             continue
 
-        artist_match = (artist_match_number / len(song_artists)) * 100
-
         song_title = _create_song_title(song_name, song_artists).lower()
 
         # Find name match and drop results below 60%

tests/cassettes/test_download_a_playlist.yaml

@@ -7,7 +7,7 @@ interactions:
       Accept-Encoding:
       - gzip, deflate
       Authorization:
-      - Bearer BQBjXTkWDqZSM5LBXmnnjNmgd6G43iNiKTni8W1mriEGpuuEz6zHA2ovAPCO2NFT3mmhad8J_PuRDC6BmfE
+      - Bearer BQAzYDtBJuYcuT7R6wj4mCduSrbeVHFdATVu6hZIPxK4Spv02qOAoAqOZXSPy4yBtobnSGUv37CLzOmhqXk
       Connection:
       - keep-alive
       Content-Type:
@@ -48,7 +48,7 @@ interactions:
       content-type:
       - application/json; charset=utf-8
       date:
-      - Fri, 22 Oct 2021 19:20:01 GMT
+      - Sat, 20 Nov 2021 12:46:27 GMT
       server:
       - envoy
       strict-transport-security: