Bump version to 0.1.3

Batching multiple updates to the leaky bucket at a fixed interval
improves the accuracy of the rate limiter. Previously the rate would
drop substantially over the course of the sync operation.

.gitignore

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

example_config.yml

@@ -13,3 +13,7 @@ spotify:
 # uncomment this block if you want to sync all playlists in the account with some exceptions
 #excluded_playlists:
 #  - spotify:playlist:1ABCDEqsABCD6EaABCDa0a
+
+# increasing these parameters should increase the search speed, while decreasing reduces likelihood of 429 errors
+max_concurrency: 10 # max concurrent connections at any given time
+rate_limit:      10 # max sustained connections per second

pyproject.toml

@@ -0,0 +1,23 @@
+[build-system]
+requires = ["setuptools >= 61.0", "wheel"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "spotify_to_tidal"
+version = "0.1.3"
+requires-python = ">= 3.10"
+
+dependencies = [
+  "spotipy~=2.21.0",
+  "tidalapi==0.7.6",
+  "pyyaml~=6.0",
+  "tqdm~=4.64",
+  "sqlalchemy~=2.0"
+]
+
+[tools.setuptools.packages."spotify_to_tidal"]
+where = "src"
+include = "spotify_to_tidal*"
+
+[project.scripts]
+spotify_to_tidal = "spotify_to_tidal.__main__:main"

readme.md

@@ -1,11 +1,11 @@
-A command line tool for importing your Spotify playlists into Tidal
+A command line tool for importing your Spotify playlists into Tidal. Due to various performance optimisations, it is particularly suited for periodic synchronisation of very large collections.
 
 Installation
 -----------
 Clone this git repository and then run:
 
 ```bash
-python3 -m pip install -r requirements.txt
+python3 -m pip install -e .
 ```
 
 Setup
@@ -18,18 +18,24 @@ Setup
 
 Usage
 ----
-To synchronize all of your Spotify playlists with your Tidal account run the following
+To synchronize all of your Spotify playlists with your Tidal account run the following from the project root directory
 
 ```bash
-python3 sync.py
+spotify_to_tidal
 ```
 
-This will take a long time because the Tidal API is really slow.
-
 You can also just synchronize a specific playlist by doing the following:
 
 ```bash
-python3 sync.py --uri 1ABCDEqsABCD6EaABCDa0a
+spotify_to_tidal --uri 1ABCDEqsABCD6EaABCDa0a # accepts playlist id or full playlist uri
 ```
 
-See example_config.yml for more configuration options, and `sync.py --help` for more options.
+See example_config.yml for more configuration options, and `spotify_to_tidal --help` for more options.
+
+---
+
+#### Join our amazing community as a code contributor
+<br><br>
+<a href="https://github.com/spotify2tidal/spotify_to_tidal/graphs/contributors">
+  <img class="dark-light" src="https://contrib.rocks/image?repo=spotify2tidal/spotify_to_tidal&anon=0&columns=25&max=100&r=true" />
+</a>

requirements.txt

@@ -1,5 +0,0 @@
-spotipy==2.21.0
-requests>=2.28.1 # for tidalapi
-tidalapi==0.7.2
-pyyaml==6.0
-tqdm==4.64.1

src/spotify_to_tidal/__main__.py

@@ -0,0 +1,37 @@
+import yaml
+import argparse
+import sys
+
+from . import sync as _sync
+from . import auth as _auth
+
+def main():
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--config', default='config.yml', help='location of the config file')
+    parser.add_argument('--uri', help='synchronize a specific URI instead of the one in the config')
+    args = parser.parse_args()
+
+    with open(args.config, 'r') as f:
+        config = yaml.safe_load(f)
+    print("Opening Spotify session")
+    spotify_session = _auth.open_spotify_session(config['spotify'])
+    print("Opening Tidal session")
+    tidal_session = _auth.open_tidal_session()
+    if not tidal_session.check_login():
+        sys.exit("Could not connect to Tidal")
+    if args.uri:
+        # if a playlist ID is explicitly provided as a command line argument then use that
+        spotify_playlist = spotify_session.playlist(args.uri)
+        tidal_playlists = _sync.get_tidal_playlists_dict(tidal_session)
+        tidal_playlist = _sync.pick_tidal_playlist_for_spotify_playlist(spotify_playlist, tidal_playlists)
+        _sync.sync_list(spotify_session, tidal_session, [tidal_playlist], config)
+    elif config.get('sync_playlists', None):
+        # if the config contains a sync_playlists list of mappings then use that
+        _sync.sync_list(spotify_session, tidal_session, _sync.get_playlists_from_config(spotify_session, tidal_session, config), config)
+    else:
+        # otherwise just use the user playlists in the Spotify account
+        _sync.sync_list(spotify_session, tidal_session, _sync.get_user_playlist_mappings(spotify_session, tidal_session, config), config)
+
+if __name__ == '__main__':
+    main()
+    sys.exit(0)

src/spotify_to_tidal/auth.py

@@ -6,12 +6,18 @@ import tidalapi
 import webbrowser
 import yaml
 
-def open_spotify_session(config):
+__all__ = [
+    'open_spotify_session',
+    'open_tidal_session'
+]
+
+def open_spotify_session(config) -> spotipy.Spotify:
     credentials_manager = spotipy.SpotifyOAuth(username=config['username'],
 				       scope='playlist-read-private',
 				       client_id=config['client_id'],
 				       client_secret=config['client_secret'],
-				       redirect_uri=config['redirect_uri'])
+				       redirect_uri=config['redirect_uri'],
+				       requests_timeout=2)
     try:
         credentials_manager.get_access_token(as_dict=False)
     except spotipy.SpotifyOauthError:
@@ -19,7 +25,7 @@ def open_spotify_session(config):
 
     return spotipy.Spotify(oauth_manager=credentials_manager)
 
-def open_tidal_session(config = None):
+def open_tidal_session(config = None) -> tidalapi.Session:
     try:
         with open('.session.yml', 'r') as session_file:
             previous_session = yaml.safe_load(session_file)

src/spotify_to_tidal/cache.py

@@ -0,0 +1,84 @@
+import datetime
+import sqlalchemy
+from sqlalchemy import Table, Column, String, DateTime, MetaData, insert, select, update, delete
+from typing import Dict, List, Sequence, Set, Mapping
+
+
+class MatchFailureDatabase:
+    """ 
+    sqlite database of match failures which persists between runs
+    this can be used concurrently between multiple processes
+    """
+
+    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: datetime.datetime | None = None) -> datetime.datetime:
+        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: str):
+        """ 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: str) -> bool:
+        """ 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: str):
+        """ 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)
+
+
+class TrackMatchCache:
+    """
+    Non-persistent mapping of spotify ids -> tidal_ids
+    This should NOT be accessed concurrently from multiple processes
+    """
+    data: Dict[str, int] = {}
+
+    def get(self, track_id: str) -> int | None:
+        return self.data.get(track_id, None)
+
+    def insert(self, mapping: tuple[str, int]):
+        self.data[mapping[0]] = mapping[1]
+
+
+# Main singleton instance
+failure_cache = MatchFailureDatabase()
+track_match_cache = TrackMatchCache()

src/spotify_to_tidal/sync.py

@@ -0,0 +1,334 @@
+#!/usr/bin/env python3
+
+import asyncio
+from .cache import failure_cache, track_match_cache
+import datetime
+from functools import partial
+from typing import List, Sequence, Set, Mapping
+import math
+import requests
+import sys
+import spotipy
+import tidalapi
+from .tidalapi_patch import add_multiple_tracks_to_playlist, set_tidal_playlist
+import time
+from tqdm.asyncio import tqdm as atqdm
+import traceback
+import unicodedata
+import math
+
+from .type import spotify as t_spotify
+
+def normalize(s) -> str:
+    return unicodedata.normalize('NFD', s).encode('ascii', 'ignore').decode('ascii')
+
+def simple(input_string: str) -> str:
+    # only take the first part of a string before any hyphens or brackets to account for different versions
+    return input_string.split('-')[0].strip().split('(')[0].strip().split('[')[0].strip()
+
+def isrc_match(tidal_track: tidalapi.Track, spotify_track) -> bool:
+    if "isrc" in spotify_track["external_ids"]:
+        return tidal_track.isrc == spotify_track["external_ids"]["isrc"]
+    return False
+
+def duration_match(tidal_track: tidalapi.Track, spotify_track, tolerance=2) -> bool:
+    # the duration of the two tracks must be the same to within 2 seconds
+    return abs(tidal_track.duration - spotify_track['duration_ms']/1000) < tolerance
+
+def name_match(tidal_track, spotify_track) -> bool:
+    def exclusion_rule(pattern: str, tidal_track: tidalapi.Track, spotify_track: t_spotify.SpotifyTrack):
+        spotify_has_pattern = pattern in spotify_track['name'].lower()
+        tidal_has_pattern = pattern in tidal_track.name.lower() or (not tidal_track.version is None and (pattern in tidal_track.version.lower()))
+        return spotify_has_pattern != tidal_has_pattern
+
+    # handle some edge cases
+    if exclusion_rule("instrumental", tidal_track, spotify_track): return False
+    if exclusion_rule("acapella", tidal_track, spotify_track): return False
+    if exclusion_rule("remix", tidal_track, spotify_track): return False
+
+    # the simplified version of the Spotify track name must be a substring of the Tidal track name
+    # Try with both un-normalized and then normalized
+    simple_spotify_track = simple(spotify_track['name'].lower()).split('feat.')[0].strip()
+    return simple_spotify_track in tidal_track.name.lower() or normalize(simple_spotify_track) in normalize(tidal_track.name.lower())
+
+def artist_match(tidal_track: tidalapi.Track, spotify_track) -> bool:
+    def split_artist_name(artist: str) -> Sequence[str]:
+       if '&' in artist:
+           return artist.split('&')
+       elif ',' in artist:
+           return artist.split(',')
+       else:
+           return [artist]
+
+    def get_tidal_artists(tidal_track: tidalapi.Track, do_normalize=False) -> Set[str]:
+        result: list[str] = []
+        for artist in tidal_track.artists:
+            if do_normalize:
+                artist_name = normalize(artist.name)
+            else:
+                artist_name = artist.name
+            result.extend(split_artist_name(artist_name))
+        return set([simple(x.strip().lower()) for x in result])
+
+    def get_spotify_artists(spotify_track: t_spotify.SpotifyTrack, do_normalize=False) -> Set[str]:
+        result: list[str] = []
+        for artist in spotify_track['artists']:
+            if do_normalize:
+                artist_name = normalize(artist['name'])
+            else:
+                artist_name = artist['name']
+            result.extend(split_artist_name(artist_name))
+        return set([simple(x.strip().lower()) for x in result])
+    # There must be at least one overlapping artist between the Tidal and Spotify track
+    # Try with both un-normalized and then normalized
+    if get_tidal_artists(tidal_track).intersection(get_spotify_artists(spotify_track)) != set():
+        return True
+    return get_tidal_artists(tidal_track, True).intersection(get_spotify_artists(spotify_track, True)) != set()
+
+def match(tidal_track, spotify_track) -> bool:
+    if not spotify_track['id']: return False
+    return isrc_match(tidal_track, spotify_track) or (
+        duration_match(tidal_track, spotify_track)
+        and name_match(tidal_track, spotify_track)
+        and artist_match(tidal_track, spotify_track)
+    )
+
+async def tidal_search(spotify_track, rate_limiter, tidal_session: tidalapi.Session) -> tidalapi.Track | None:
+    def _search_for_track_in_album():
+        # 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])
+            for album in album_result['albums']:
+                album_tracks = album.tracks()
+                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
+    def _search_for_standalone_track():
+        # if album search 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
+    await rate_limiter.acquire()
+    album_search = await asyncio.to_thread( _search_for_track_in_album )
+    if album_search:
+        return album_search
+    await rate_limiter.acquire()
+    track_search = await asyncio.to_thread( _search_for_standalone_track )
+    if track_search:
+        return track_search
+    return None
+
+    # if none of the search modes succeeded then store the track id to the failure cache
+    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
+    print("Loading Tidal playlists... This may take some time.")
+    tidal_playlists = tidal_session.user.playlists()
+    output = {}
+    for playlist in tidal_playlists:
+        output[playlist.name] = playlist
+    return output 
+
+async def repeat_on_request_error(function, *args, remaining=5, **kwargs):
+    # utility to repeat calling the function up to 5 times if an exception is thrown
+    try:
+        return await function(*args, **kwargs)
+    except (tidalapi.exceptions.TooManyRequests, requests.exceptions.RequestException) as e:
+        if remaining:
+            print(f"{str(e)} occurred, retrying {remaining} times")
+        else:
+            print(f"{str(e)} could not be recovered")
+
+        if isinstance(e, requests.exceptions.RequestException) and not e.response is None:
+            print(f"Response message: {e.response.text}")
+            print(f"Response headers: {e.response.headers}")
+
+        if not remaining:
+            print("Aborting sync")
+            print(f"The following arguments were provided:\n\n {str(args)}")
+            print(traceback.format_exc())
+            sys.exit(1)
+        sleep_schedule = {5: 1, 4:10, 3:60, 2:5*60, 1:10*60} # sleep variable length of time depending on retry number
+        time.sleep(sleep_schedule.get(remaining, 1))
+        return await repeat_on_request_error(function, *args, remaining=remaining-1, **kwargs)
+
+async def get_tracks_from_spotify_playlist(spotify_session: spotipy.Spotify, spotify_playlist):
+    def _get_tracks_from_spotify_playlist(offset: int, spotify_session: spotipy.Spotify, playlist_id: str):
+        fields="next,total,limit,items(track(name,album(name,artists),artists,track_number,duration_ms,id,external_ids(isrc)))"
+        return spotify_session.playlist_tracks(playlist_id, fields, offset=offset)
+
+    output = []
+    print(f"Loading tracks from Spotify playlist '{spotify_playlist['name']}'")
+    results = _get_tracks_from_spotify_playlist( 0, spotify_session, spotify_playlist["id"] )
+    output.extend([r['track'] for r in results['items'] if r['track'] is not None])
+
+    # get all the remaining tracks in parallel
+    if results['next']:
+        offsets = [ results['limit'] * n for n in range(1, math.ceil(results['total']/results['limit'])) ]
+        extra_results = await atqdm.gather( *[asyncio.to_thread(_get_tracks_from_spotify_playlist, offset, spotify_session=spotify_session, playlist_id=spotify_playlist["id"]) for offset in offsets ] )
+        for extra_result in extra_results:
+            output.extend([r['track'] for r in extra_result['items'] if r['track'] is not None])
+    return output
+
+def populate_track_match_cache(spotify_tracks_: Sequence[t_spotify.SpotifyTrack], tidal_tracks_: Sequence[tidalapi.Track]):
+    """ Populate the track match cache with all the existing tracks in Tidal playlist corresponding to Spotify playlist """
+    def _populate_one_track_from_spotify(spotify_track: t_spotify.SpotifyTrack):
+        for idx, tidal_track in list(enumerate(tidal_tracks)):
+            if match(tidal_track, spotify_track):
+                track_match_cache.insert((spotify_track['id'], tidal_track.id))
+                tidal_tracks.pop(idx)
+                return
+
+    def _populate_one_track_from_tidal(tidal_track: tidalapi.Track):
+        for idx, spotify_track in list(enumerate(spotify_tracks)):
+            if match(tidal_track, spotify_track):
+                track_match_cache.insert((spotify_track['id'], tidal_track.id))
+                spotify_tracks.pop(idx)
+                return
+
+    # make a copy of the tracks to avoid modifying original arrays
+    spotify_tracks = [t for t in spotify_tracks_]
+    tidal_tracks = [t for t in tidal_tracks_]
+
+    # first populate from the tidal tracks
+    for track in tidal_tracks:
+        _populate_one_track_from_tidal(track)
+    # then populate from the subset of Spotify tracks that didn't match (to account for many-to-one style mappings)
+    for track in spotify_tracks:
+        _populate_one_track_from_spotify(track)
+
+def get_new_tracks_from_spotify_playlist(spotify_tracks: Sequence[t_spotify.SpotifyTrack], old_tidal_tracks: Sequence[tidalapi.Track]) -> list[t_spotify.SpotifyTrack]:
+    ''' Extracts only the new tracks in the Spotify playlist that are not already on Tidal or known match failures '''
+    populate_track_match_cache(spotify_tracks, old_tidal_tracks)
+    results = []
+    for spotify_track in spotify_tracks:
+        if not spotify_track['id']: continue
+        if not track_match_cache.get(spotify_track['id']) and not failure_cache.has_match_failure(spotify_track['id']):
+            results.append(spotify_track)
+    return results
+
+def get_tracks_for_new_tidal_playlist(spotify_tracks: Sequence[t_spotify.SpotifyTrack]) -> Sequence[int]:
+    ''' gets list of corresponding tidal track ids for each spotify track, ignoring duplicates '''
+    output = []
+    seen_tracks = set()
+    for spotify_track in spotify_tracks:
+        if not spotify_track['id']: continue
+        tidal_id = track_match_cache.get(spotify_track['id'])
+        if tidal_id and not tidal_id in seen_tracks:
+            output.append(tidal_id)
+            seen_tracks.add(tidal_id)
+    return output
+
+async def sync_playlist(spotify_session: spotipy.Spotify, tidal_session: tidalapi.Session, spotify_playlist, tidal_playlist: tidalapi.Playlist | None, config):
+    async def _run_rate_limiter(semaphore):
+        ''' Leaky bucket algorithm for rate limiting. Periodically releases items from semaphore at rate_limit'''
+        _sleep_time = config.get('max_concurrency', 10)/config.get('rate_limit', 10)/4 # aim to sleep approx time to drain 1/4 of 'bucket'
+        t0 = datetime.datetime.now()
+        while True:
+            await asyncio.sleep(_sleep_time)
+            t = datetime.datetime.now()
+            dt = (t - t0).total_seconds()
+            new_items = round(config.get('rate_limit', 10)*dt)
+            t0 = t
+            [semaphore.release() for i in range(new_items)] # leak new_items from the 'bucket'
+
+    # Create a new Tidal playlist if required
+    if not tidal_playlist:
+        print(f"No playlist found on Tidal corresponding to Spotify playlist: '{spotify_playlist['name']}', creating new playlist")
+        tidal_playlist =  tidal_session.user.create_playlist(spotify_playlist['name'], spotify_playlist['description'])
+
+    # Extract the new tracks from the playlist that we haven't already seen before
+    spotify_tracks = await get_tracks_from_spotify_playlist(spotify_session, spotify_playlist)
+    old_tidal_tracks = tidal_playlist.tracks()
+    tracks_to_search = get_new_tracks_from_spotify_playlist(spotify_tracks, old_tidal_tracks)
+    if not tracks_to_search:
+        print("No new tracks to search in Spotify playlist '{}'".format(spotify_playlist['name']))
+        return
+
+    # Search for each of the tracks on Tidal concurrently
+    task_description = "Searching Tidal for {}/{} tracks in Spotify playlist '{}'".format(len(tracks_to_search), len(spotify_tracks), spotify_playlist['name'])
+    semaphore = asyncio.Semaphore(config.get('max_concurrency', 10))
+    rate_limiter_task = asyncio.create_task(_run_rate_limiter(semaphore))
+    search_results = await atqdm.gather( *[ repeat_on_request_error(tidal_search, t, semaphore, tidal_session) for t in tracks_to_search ], desc=task_description )
+    rate_limiter_task.cancel()
+
+    # Add the search results to the cache
+    for idx, spotify_track in enumerate(tracks_to_search):
+        if search_results[idx]:
+            track_match_cache.insert( (spotify_track['id'], search_results[idx].id) )
+        else:
+            color = ('\033[91m', '\033[0m')
+            print(color[0] + "Could not find track {}: {} - {}".format(spotify_track['id'], ",".join([a['name'] for a in spotify_track['artists']]), spotify_track['name']) + color[1])
+
+    # Update the Tidal playlist if there are changes
+    old_tidal_track_ids = [t.id for t in old_tidal_tracks]
+    new_tidal_track_ids = get_tracks_for_new_tidal_playlist(spotify_tracks)
+    if new_tidal_track_ids == old_tidal_track_ids:
+        print("No changes to write to Tidal playlist")
+    elif new_tidal_track_ids[:len(old_tidal_track_ids)] == old_tidal_track_ids:
+        # Append new tracks to the existing playlist if possible
+        add_multiple_tracks_to_playlist(tidal_playlist, new_tidal_track_ids[len(old_tidal_track_ids):])
+    else:
+        # Erase old playlist and add new tracks from scratch if any reordering occured
+        set_tidal_playlist(tidal_playlist, new_tidal_track_ids)
+
+def sync_list(spotify_session: spotipy.Spotify, tidal_session: tidalapi.Session, playlists, config):
+  for spotify_playlist, tidal_playlist in playlists:
+    # sync the spotify playlist to tidal
+    asyncio.run(sync_playlist(spotify_session, tidal_session, spotify_playlist, tidal_playlist, config) )
+
+def pick_tidal_playlist_for_spotify_playlist(spotify_playlist, tidal_playlists: Mapping[str, tidalapi.Playlist]):
+    if spotify_playlist['name'] in tidal_playlists:
+      # if there's an existing tidal playlist with the name of the current playlist then use that
+      tidal_playlist = tidal_playlists[spotify_playlist['name']]
+      return (spotify_playlist, tidal_playlist)
+    else:
+      return (spotify_playlist, None)
+
+def get_user_playlist_mappings(spotify_session: spotipy.Spotify, tidal_session: tidalapi.Session, config):
+    results = []
+    spotify_playlists = asyncio.run(get_playlists_from_spotify(spotify_session, config))
+    tidal_playlists = get_tidal_playlists_dict(tidal_session)
+    for spotify_playlist in spotify_playlists:
+        results.append( pick_tidal_playlist_for_spotify_playlist(spotify_playlist, tidal_playlists) )
+    return results
+
+async def get_playlists_from_spotify(spotify_session: spotipy.Spotify, config):
+    # get all the user playlists from the Spotify account
+    playlists = []
+    print("Loading Spotify playlists")
+    results = spotify_session.user_playlists(config['spotify']['username'])
+    exclude_list = set([x.split(':')[-1] for x in config.get('excluded_playlists', [])])
+    playlists.extend([p for p in results['items'] if p['owner']['id'] == config['spotify']['username'] and not p['id'] in exclude_list])
+      
+    # get all the remaining playlists in parallel
+    if results['next']:
+        offsets = [ results['limit'] * n for n in range(1, math.ceil(results['total']/results['limit'])) ]
+        extra_results = await atqdm.gather( *[asyncio.to_thread(spotify_session.user_playlists, config['spotify']['username'], offset=offset) for offset in offsets ] )
+        for extra_result in extra_results:
+            playlists.extend([p for p in extra_result['items'] if p['owner']['id'] == config['spotify']['username'] and not p['id'] in exclude_list])
+    return playlists
+
+def get_playlists_from_config(spotify_session: spotipy.Spotify, tidal_session: tidalapi.Session, config):
+    # get the list of playlist sync mappings from the configuration file
+    def get_playlist_ids(config):
+        return [(item['spotify_id'], item['tidal_id']) for item in config['sync_playlists']]
+    output = []
+    for spotify_id, tidal_id in get_playlist_ids(config):
+        try:
+            spotify_playlist = spotify_session.playlist(spotify_id)
+        except spotipy.SpotifyException as e:
+            print(f"Error getting Spotify playlist {spotify_id}")
+            raise e
+        try:
+            tidal_playlist = tidal_session.playlist(tidal_id)
+        except Exception as e:
+            print(f"Error getting Tidal playlist {tidal_id}")
+            raise e
+        output.append((spotify_playlist, tidal_playlist))
+    return output
+

src/spotify_to_tidal/tidalapi_patch.py

@@ -1,19 +1,21 @@
+from typing import List
+import tidalapi
 from tqdm import tqdm
 
-def _remove_indices_from_playlist(playlist, indices):
+def _remove_indices_from_playlist(playlist: tidalapi.UserPlaylist, indices: List[int]):
     headers = {'If-None-Match': playlist._etag}
     index_string = ",".join(map(str, indices))
-    playlist.requests.request('DELETE', (playlist._base_url + '/items/%s') % (playlist.id, index_string), headers=headers)
+    playlist.request.request('DELETE', (playlist._base_url + '/items/%s') % (playlist.id, index_string), headers=headers)
     playlist._reparse()
 
-def clear_tidal_playlist(playlist, chunk_size=20):
+def clear_tidal_playlist(playlist: tidalapi.UserPlaylist, chunk_size: int=20):
     with tqdm(desc="Erasing existing tracks from Tidal playlist", total=playlist.num_tracks) as progress:
         while playlist.num_tracks:
             indices = range(min(playlist.num_tracks, chunk_size))
             _remove_indices_from_playlist(playlist, indices)
             progress.update(len(indices))
     
-def add_multiple_tracks_to_playlist(playlist, track_ids, chunk_size=20):
+def add_multiple_tracks_to_playlist(playlist: tidalapi.UserPlaylist, track_ids: List[int], chunk_size: int=20):
     offset = 0
     with tqdm(desc="Adding new tracks to Tidal playlist", total=len(track_ids)) as progress:
         while offset < len(track_ids):
@@ -22,6 +24,6 @@ def add_multiple_tracks_to_playlist(playlist, track_ids, chunk_size=20):
             offset += count
             progress.update(count)
 
-def set_tidal_playlist(playlist, track_ids):
+def set_tidal_playlist(playlist: tidalapi.Playlist, track_ids: List[int]):
     clear_tidal_playlist(playlist)
     add_multiple_tracks_to_playlist(playlist, track_ids)

src/spotify_to_tidal/type/__init__.py

@@ -0,0 +1,25 @@
+from .config import SpotifyConfig, TidalConfig, PlaylistConfig, SyncConfig
+from .spotify import SpotifyTrack
+
+from spotipy import Spotify
+from tidalapi import Session, Track
+
+TidalID = str
+SpotifyID = str
+TidalSession = Session
+TidalTrack = Track
+SpotifySession = Spotify
+
+__all__ = [
+    "SpotifyConfig",
+    "TidalConfig",
+    "PlaylistConfig",
+    "SyncConfig",
+    "TidalPlaylist",
+    "TidalID",
+    "SpotifyID",
+    "SpotifySession",
+    "TidalSession",
+    "TidalTrack",
+    "SpotifyTrack",
+]

src/spotify_to_tidal/type/config.py

@@ -0,0 +1,26 @@
+from typing import TypedDict, Literal, List, Optional
+
+
+class SpotifyConfig(TypedDict):
+    client_id: str
+    client_secret: str
+    username: str
+    redirect_url: str
+
+
+class TidalConfig(TypedDict):
+    access_token: str
+    refresh_token: str
+    session_id: str
+    token_type: Literal["Bearer"]
+
+
+class PlaylistConfig(TypedDict):
+    spotify_id: str
+    tidal_id: str
+
+
+class SyncConfig(TypedDict):
+    spotify: SpotifyConfig
+    sync_playlists: Optional[List[PlaylistConfig]]
+    excluded_playlists: Optional[List[str]]

src/spotify_to_tidal/type/spotify.py

@@ -0,0 +1,69 @@
+from spotipy import Spotify
+from typing import TypedDict, List, Dict, Mapping, Literal, Optional
+
+
+class SpotifyImage(TypedDict):
+    url: str
+    height: int
+    width: int
+
+
+class SpotifyFollower(TypedDict):
+    href: str
+    total: int
+
+
+SpotifyID = str
+SpotifySession = Spotify
+
+
+class SpotifyArtist(TypedDict):
+    external_urls: Mapping[str, str]
+    followers: SpotifyFollower
+    genres: List[str]
+    href: str
+    id: str
+    images: List[SpotifyImage]
+    name: str
+    popularity: int
+    type: str
+    uri: str
+
+
+class SpotifyAlbum(TypedDict):
+    album_type: Literal["album", "single", "compilation"]
+    total_tracks: int
+    available_markets: List[str]
+    external_urls: Dict[str, str]
+    href: str
+    id: str
+    images: List[SpotifyImage]
+    name: str
+    release_date: str
+    release_date_precision: Literal["year", "month", "day"]
+    restrictions: Optional[Dict[Literal["reason"], str]]
+    type: Literal["album"]
+    uri: str
+    artists: List[SpotifyArtist]
+
+
+class SpotifyTrack(TypedDict):
+    album: SpotifyAlbum
+    artists: List[SpotifyArtist]
+    available_markets: List[str]
+    disc_number: int
+    duration_ms: int
+    explicit: bool
+    external_ids: Dict[str, str]
+    external_urls: Dict[str, str]
+    href: str
+    id: str
+    is_playable: bool
+    linked_from: Dict
+    restrictions: Optional[Dict[Literal["reason"], str]]
+    name: str
+    popularity: int
+    preview_url: str
+    track_number: int
+    type: Literal["track"]
+    uri: str

sync.py

@@ -1,312 +0,0 @@
-#!/usr/bin/env python3
-
-import argparse
-from auth import open_tidal_session, open_spotify_session
-from functools import partial
-from multiprocessing import Pool
-import requests
-import sys
-import spotipy
-import tidalapi
-from tidalapi_patch import set_tidal_playlist
-import time
-from tqdm import tqdm
-import traceback
-import unicodedata
-import yaml
-
-def normalize(s):
-    return unicodedata.normalize('NFD', s).encode('ascii', 'ignore').decode('ascii')
-
-def simple(input_string):
-    # only take the first part of a string before any hyphens or brackets to account for different versions
-    return input_string.split('-')[0].strip().split('(')[0].strip().split('[')[0].strip()
-
-def isrc_match(tidal_track, spotify_track):
-    if "isrc" in spotify_track["external_ids"]:
-        return tidal_track.isrc == spotify_track["external_ids"]["isrc"]
-    return False
-
-def duration_match(tidal_track, spotify_track, tolerance=2):
-    # the duration of the two tracks must be the same to within 2 seconds
-    return abs(tidal_track.duration - spotify_track['duration_ms']/1000) < tolerance
-
-def name_match(tidal_track, spotify_track):
-    def exclusion_rule(pattern, tidal_track, spotify_track):
-        spotify_has_pattern = pattern in spotify_track['name'].lower()
-        tidal_has_pattern = pattern in tidal_track.name.lower() or (not tidal_track.version is None and (pattern in tidal_track.version.lower()))
-        return spotify_has_pattern != tidal_has_pattern
-
-    # handle some edge cases
-    if exclusion_rule("instrumental", tidal_track, spotify_track): return False
-    if exclusion_rule("acapella", tidal_track, spotify_track): return False
-    if exclusion_rule("remix", tidal_track, spotify_track): return False
-
-    # the simplified version of the Spotify track name must be a substring of the Tidal track name
-    # Try with both un-normalized and then normalized
-    simple_spotify_track = simple(spotify_track['name'].lower()).split('feat.')[0].strip()
-    return simple_spotify_track in tidal_track.name.lower() or normalize(simple_spotify_track) in normalize(tidal_track.name.lower())
-
-def artist_match(tidal_track, spotify_track):
-    def split_artist_name(artist):
-       if '&' in artist:
-           return artist.split('&')
-       elif ',' in artist:
-           return artist.split(',')
-       else:
-           return [artist]
-
-    def get_tidal_artists(tidal_track, do_normalize=False):
-        result = []
-        for artist in tidal_track.artists:
-            if do_normalize:
-                artist_name = normalize(artist.name)
-            else:
-                artist_name = artist.name
-            result.extend(split_artist_name(artist_name))
-        return set([simple(x.strip().lower()) for x in result])
-
-    def get_spotify_artists(spotify_track, do_normalize=False):
-        result = []
-        for artist in spotify_track['artists']:
-            if do_normalize:
-                artist_name = normalize(artist['name'])
-            else:
-                artist_name = artist['name']
-            result.extend(split_artist_name(artist_name))
-        return set([simple(x.strip().lower()) for x in result])
-    # There must be at least one overlapping artist between the Tidal and Spotify track
-    # Try with both un-normalized and then normalized
-    if get_tidal_artists(tidal_track).intersection(get_spotify_artists(spotify_track)) != set():
-        return True
-    return get_tidal_artists(tidal_track, True).intersection(get_spotify_artists(spotify_track, True)) != set()
-
-def match(tidal_track, spotify_track):
-    return isrc_match(tidal_track, spotify_track) or (
-        duration_match(tidal_track, spotify_track)
-        and name_match(tidal_track, spotify_track)
-        and artist_match(tidal_track, spotify_track)
-    )
-
-
-def tidal_search(spotify_track_and_cache, tidal_session):
-    spotify_track, cached_tidal_track = spotify_track_and_cache
-    if cached_tidal_track: return cached_tidal_track
-    # 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])
-        for album in album_result['albums']:
-            album_tracks = album.tracks()
-            if len(album_tracks) >= spotify_track['track_number']:
-                track = album_tracks[spotify_track['track_number'] - 1]
-                if match(track, spotify_track):
-                    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):
-            return track
-
-def get_tidal_playlists_dict(tidal_session):
-    # a dictionary of name --> playlist
-    tidal_playlists = tidal_session.user.playlists()
-    output = {}
-    for playlist in tidal_playlists:
-        output[playlist.name] = playlist
-    return output 
-
-def repeat_on_request_error(function, *args, remaining=5, **kwargs):
-    # utility to repeat calling the function up to 5 times if an exception is thrown
-    try:
-        return function(*args, **kwargs)
-    except requests.exceptions.RequestException as e:
-        if remaining:
-            print(f"{str(e)} occurred, retrying {remaining} times")
-        else:
-            print(f"{str(e)} could not be recovered")
-
-        if not e.response is None:
-            print(f"Response message: {e.response.text}")
-            print(f"Response headers: {e.response.headers}")
-
-        if not remaining:
-            print("Aborting sync")
-            print(f"The following arguments were provided:\n\n {str(args)}")
-            print(traceback.format_exc())
-            sys.exit(1)
-        sleep_schedule = {5: 1, 4:10, 3:60, 2:5*60, 1:10*60} # sleep variable length of time depending on retry number
-        time.sleep(sleep_schedule.get(remaining, 1))
-        return repeat_on_request_error(function, *args, remaining=remaining-1, **kwargs)
-
-def _enumerate_wrapper(value_tuple, function, **kwargs):
-    # just a wrapper which accepts a tuple from enumerate and returns the index back as the first argument
-    index, value = value_tuple
-    return (index, repeat_on_request_error(function, value, **kwargs))
-
-def call_async_with_progress(function, values, description, num_processes, **kwargs):
-    results = len(values)*[None]
-    with Pool(processes=num_processes) as process_pool:
-        for index, result in tqdm(process_pool.imap_unordered(partial(_enumerate_wrapper, function=function, **kwargs),
-                                  enumerate(values)), total=len(values), desc=description):
-            results[index] = result
-    return results
-
-def get_tracks_from_spotify_playlist(spotify_session, spotify_playlist):
-    output = []
-    results = spotify_session.playlist_tracks(
-        spotify_playlist["id"],
-        fields="next,items(track(name,album(name,artists),artists,track_number,duration_ms,id,external_ids(isrc)))",
-    )
-    while True:
-        output.extend([r['track'] for r in results['items'] if r['track'] is not None])
-        # move to the next page of results if there are still tracks remaining in the playlist
-        if results['next']:
-            results = spotify_session.next(results)
-        else:
-            return output
-
-class TidalPlaylistCache:
-    def __init__(self, playlist):
-        self._data = playlist.tracks()
-
-    def _search(self, spotify_track):
-        ''' check if the given spotify track was already in the tidal playlist.'''
-        results = []
-        for tidal_track in self._data:
-            if match(tidal_track, spotify_track):
-                return tidal_track
-        return None
-
-    def search(self, spotify_session, spotify_playlist):
-        ''' Add the cached tidal track where applicable to a list of spotify tracks '''
-        results = []
-        cache_hits = 0
-        work_to_do = False
-        spotify_tracks = get_tracks_from_spotify_playlist(spotify_session, spotify_playlist)
-        for track in spotify_tracks:
-            cached_track = self._search(track)
-            if cached_track:
-                results.append( (track, cached_track) )
-                cache_hits += 1
-            else:
-                results.append( (track, None) )
-        return (results, cache_hits)
-
-def tidal_playlist_is_dirty(playlist, new_track_ids):
-    old_tracks = playlist.tracks()
-    if len(old_tracks) != len(new_track_ids):
-        return True
-    for i in range(len(old_tracks)):
-        if old_tracks[i].id != new_track_ids[i]:
-            return True
-    return False
-
-def sync_playlist(spotify_session, tidal_session, spotify_id, tidal_id, config):
-    try:
-        spotify_playlist = spotify_session.playlist(spotify_id)
-    except spotipy.SpotifyException as e:
-        print("Error getting Spotify playlist " + spotify_id)
-        print(e)
-        results.append(None)
-        return
-    if tidal_id:
-        # if a Tidal playlist was specified then look it up
-        try:
-            tidal_playlist = tidal_session.playlist(tidal_id)
-        except Exception as e:
-            print("Error getting Tidal playlist " + tidal_id)
-            print(e)
-            return
-    else:
-        # create a new Tidal playlist if required
-        print(f"No playlist found on Tidal corresponding to Spotify playlist: '{spotify_playlist['name']}', creating new playlist")
-        tidal_playlist =  tidal_session.user.create_playlist(spotify_playlist['name'], spotify_playlist['description'])
-    tidal_track_ids = []
-    spotify_tracks, cache_hits = TidalPlaylistCache(tidal_playlist).search(spotify_session, spotify_playlist)
-    if cache_hits == len(spotify_tracks):
-        print("No new tracks to search in Spotify playlist '{}'".format(spotify_playlist['name']))
-        return
-
-    task_description = "Searching Tidal for {}/{} tracks in Spotify playlist '{}'".format(len(spotify_tracks) - cache_hits, len(spotify_tracks), spotify_playlist['name'])
-    tidal_tracks = call_async_with_progress(tidal_search, spotify_tracks, task_description, config.get('subprocesses', 50), tidal_session=tidal_session)
-    for index, tidal_track in enumerate(tidal_tracks):
-        spotify_track = spotify_tracks[index][0]
-        if tidal_track:
-            tidal_track_ids.append(tidal_track.id)
-        else:
-            color = ('\033[91m', '\033[0m')
-            print(color[0] + "Could not find track {}: {} - {}".format(spotify_track['id'], ",".join([a['name'] for a in spotify_track['artists']]), spotify_track['name']) + color[1])
-
-    if tidal_playlist_is_dirty(tidal_playlist, tidal_track_ids):
-        set_tidal_playlist(tidal_playlist, tidal_track_ids)
-    else:
-        print("No changes to write to Tidal playlist")
-
-def sync_list(spotify_session, tidal_session, playlists, config):
-  results = []
-  for spotify_id, tidal_id in playlists:
-    # sync the spotify playlist to tidal
-    repeat_on_request_error(sync_playlist, spotify_session, tidal_session, spotify_id, tidal_id, config)
-    results.append(tidal_id)
-  return results
-
-def pick_tidal_playlist_for_spotify_playlist(spotify_playlist, tidal_playlists):
-    if spotify_playlist['name'] in tidal_playlists:
-      # if there's an existing tidal playlist with the name of the current playlist then use that
-      tidal_playlist = tidal_playlists[spotify_playlist['name']]
-      return (spotify_playlist['id'], tidal_playlist.id)
-    else:
-      return (spotify_playlist['id'], None)
-    
-
-def get_user_playlist_mappings(spotify_session, tidal_session, config):
-  results = []
-  spotify_playlists = get_playlists_from_spotify(spotify_session, config)
-  tidal_playlists = get_tidal_playlists_dict(tidal_session)
-  for spotify_playlist in spotify_playlists:
-      results.append( pick_tidal_playlist_for_spotify_playlist(spotify_playlist, tidal_playlists) )
-  return results
-
-def get_playlists_from_spotify(spotify_session, config):
-    # get all the user playlists from the Spotify account
-    playlists = []
-    spotify_results = spotify_session.user_playlists(config['spotify']['username'])
-    exclude_list = set([x.split(':')[-1] for x in config.get('excluded_playlists', [])])
-    while True:
-        for spotify_playlist in spotify_results['items']:
-            if spotify_playlist['owner']['id'] == config['spotify']['username'] and not spotify_playlist['id'] in exclude_list:
-                playlists.append(spotify_playlist)
-        # move to the next page of results if there are still playlists remaining
-        if spotify_results['next']:
-            spotify_results = spotify_session.next(spotify_results)
-        else:
-            break
-    return playlists
-
-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']]
-
-if __name__ == '__main__':
-    parser = argparse.ArgumentParser()
-    parser.add_argument('--config', default='config.yml', help='location of the config file')
-    parser.add_argument('--uri', help='synchronize a specific URI instead of the one in the config')
-    args = parser.parse_args()
-
-    with open(args.config, 'r') as f:
-        config = yaml.safe_load(f)
-    spotify_session = open_spotify_session(config['spotify'])
-    tidal_session = open_tidal_session()
-    if not tidal_session.check_login():
-        sys.exit("Could not connect to Tidal")
-    if args.uri:
-        # if a playlist ID is explicitly provided as a command line argument then use that
-        spotify_playlist = spotify_session.playlist(args.uri)
-        tidal_playlists = get_tidal_playlists_dict(tidal_session)
-        tidal_playlist = pick_tidal_playlist_for_spotify_playlist(spotify_playlist, tidal_playlists)
-        sync_list(spotify_session, tidal_session, [tidal_playlist], config)
-    elif config.get('sync_playlists', None):
-        # if the config contains a sync_playlists list of mappings then use that
-        sync_list(spotify_session, tidal_session, get_playlists_from_config(config), config)
-    else:
-        # otherwise just use the user playlists in the Spotify account
-        sync_list(spotify_session, tidal_session, get_user_playlist_mappings(spotify_session, tidal_session, config), config)