Merge pull request #41 from spotify2tidal/feature/miss_cache

Add cache of match failures

.gitignore

@@ -1,7 +1,7 @@
 # Config and cache files
 config.yml
 config.yaml
-.cache-*
+.cache*
 .session.yml
 
 # Byte-compiled / optimized / DLL files

pyproject.toml

@@ -12,6 +12,7 @@ dependencies = [
   "tidalapi~=0.7",
   "pyyaml~=6.0",
   "tqdm~=4.64",
+  "sqlalchemy~=2.0"
 ]
 
 

src/spotify_to_tidal/database.py

@@ -0,0 +1,63 @@
+import datetime
+import sqlalchemy
+from sqlalchemy import Table, Column, String, DateTime, MetaData, insert, select, update, delete
+
+
+class Database:
+    def __init__(self, filename='.cache.db'):
+        self.engine = sqlalchemy.create_engine(f"sqlite:///{filename}")
+        meta = MetaData()
+        self.match_failures = Table('match_failures', meta,
+                                    Column('track_id', String,
+                                           primary_key=True),
+                                    Column('insert_time', DateTime),
+                                    Column('next_retry', DateTime),
+                                    sqlite_autoincrement=False)
+        meta.create_all(self.engine)
+
+    def _get_next_retry_time(self, insert_time=None):
+        if insert_time:
+            # double interval on each retry
+            interval = 2 * (datetime.datetime.now() - insert_time)
+        else:
+            interval = datetime.timedelta(days=7)
+        return datetime.datetime.now() + interval
+
+    def cache_match_failure(self, track_id):
+        """ notifies that matching failed for the given track_id """
+        fetch_statement = select(self.match_failures).where(
+            self.match_failures.c.track_id == track_id)
+        with self.engine.connect() as connection:
+            with connection.begin():
+                # Either update the next_retry time if track_id already exists, otherwise create a new entry
+                existing_failure = connection.execute(
+                    fetch_statement).fetchone()
+                if existing_failure:
+                    update_statement = update(self.match_failures).where(
+                        self.match_failures.c.track_id == track_id).values(next_retry=self._get_next_retry_time())
+                    connection.execute(update_statement)
+                else:
+                    connection.execute(insert(self.match_failures), {
+                                       "track_id": track_id, "insert_time": datetime.datetime.now(), "next_retry": self._get_next_retry_time()})
+
+    def has_match_failure(self, track_id):
+        """ checks if there was a recent search for which matching failed with the given track_id """
+        statement = select(self.match_failures.c.next_retry).where(
+            self.match_failures.c.track_id == track_id)
+        with self.engine.connect() as connection:
+            match_failure = connection.execute(statement).fetchone()
+            if match_failure:
+                return match_failure.next_retry > datetime.datetime.now()
+            return False
+
+    def remove_match_failure(self, track_id):
+        """ removes match failure from the database """
+        statement = delete(self.match_failures).where(
+            self.match_failures.c.track_id == track_id)
+        with self.engine.connect() as connection:
+            with connection.begin():
+                connection.execute(statement)
+
+
+# Main singleton instance
+failure_cache = Database()

src/spotify_to_tidal/sync.py

@@ -1,5 +1,6 @@
 #!/usr/bin/env python3
 
+from .database import failure_cache
 from functools import partial
 from typing import Sequence, Set, Mapping
 from multiprocessing import Pool
@@ -89,10 +90,12 @@ def match(tidal_track, spotify_track) -> bool:
         and artist_match(tidal_track, spotify_track)
     )
 
-
 def tidal_search(spotify_track_and_cache, tidal_session: tidalapi.Session) -> tidalapi.Track | None:
     spotify_track, cached_tidal_track = spotify_track_and_cache
     if cached_tidal_track: return cached_tidal_track
+    if spotify_track['id'] is None: return None
+    if failure_cache.has_match_failure(spotify_track['id']):
+        return None
     # search for album name and first album artist
     if 'album' in spotify_track and 'artists' in spotify_track['album'] and len(spotify_track['album']['artists']):
         album_result = tidal_session.search(simple(spotify_track['album']['name']) + " " + simple(spotify_track['album']['artists'][0]['name']), models=[tidalapi.album.Album])
@@ -101,11 +104,14 @@ def tidal_search(spotify_track_and_cache, tidal_session: tidalapi.Session) -> ti
             if len(album_tracks) >= spotify_track['track_number']:
                 track = album_tracks[spotify_track['track_number'] - 1]
                 if match(track, spotify_track):
+                    failure_cache.remove_match_failure(spotify_track['id'])
                     return track
     # if that fails then search for track name and first artist
     for track in tidal_session.search(simple(spotify_track['name']) + ' ' + simple(spotify_track['artists'][0]['name']), models=[tidalapi.media.Track])['tracks']:
         if match(track, spotify_track):
+            failure_cache.remove_match_failure(spotify_track['id'])
             return track
+    failure_cache.cache_match_failure(spotify_track['id'])
 
 def get_tidal_playlists_dict(tidal_session: tidalapi.Session) -> Mapping[str, tidalapi.Playlist]:
     # a dictionary of name --> playlist
@@ -284,4 +290,4 @@ def get_playlists_from_spotify(spotify_session: spotipy.Spotify, config):
 
 def get_playlists_from_config(config):
     # get the list of playlist sync mappings from the configuration file
-    return [(item['spotify_id'], item['tidal_id']) for item in config['sync_playlists']]
+    return [(item['spotify_id'], item['tidal_id']) for item in config['sync_playlists']]