add spotify export script + README

README.org

@@ -0,0 +1,73 @@
+#+begin_src python :dir src :results drawer :exports results
+import spotifyexport.export as E; return E.make_parser().prog
+#+end_src
+
+#+RESULTS:
+:results:
+Export your personal Spotify data: playlists, saved tracks/albums/shows, etc. as JSON.
+:end:
+
+* Setting up
+1. The easiest way is =pip3 install --user git+https://github.com/karlicoss/spotifyexport=.
+
+ Alternatively, use =git clone --recursive=, or =git pull && git submodules update --init=. After that, you can use =pip3 install --editable=.
+2. To use the API, you need to create a new app on [[https://developer.spotify.com/dashboard/applications][Spotify for Developers]]
+
+ For =redirect_uri=: you can pick pretty much anything, e.g. =https://github.com=. After that you'll get =client_id= and =client_secret=.
+
+3. On the first script run, you'll be prompted to approve the script access.
+
+ Once approved, the token is saved to =cache_path= (will be created if doesn't exist). After that you won't need to enter the password again as long as you pass the same =cache_path=.
+
+
+* Exporting
+
+#+begin_src python :dir src :results drawer :exports results
+import spotifyexport.export as E; return E.make_parser().epilog
+#+end_src
+
+#+RESULTS:
+:results:
+
+Usage:
+
+*Recommended*: create =secrets.py= keeping your api parameters, e.g.:
+
+
+: client_id = "CLIENT_ID"
+: client_secret = "CLIENT_SECRET"
+: redirect_uri = "REDIRECT_URI"
+: cache_path = "CACHE_PATH"
+
+
+After that, use:
+
+: python3 -m spotifyexport.export --secrets /path/to/secrets.py
+
+That way you type less and have control over where you keep your plaintext secrets.
+
+*Alternatively*, you can pass parameters directly, e.g.
+
+: python3 -m spotifyexport.export --client_id <client_id> --client_secret <client_secret> --redirect_uri <redirect_uri> --cache_path <cache_path>
+
+However, this is verbose and prone to leaking your keys/tokens/passwords in shell history.
+
+
+
+I *highly* recommend checking exported files at least once just to make sure they contain everything you expect from your export. If not, please feel free to ask or raise an issue!
+
+:end:
+
+* API limitations
+
+- you might want to do a [[https://www.spotify.com/uk/privacy/#privacy-center-control-section][GDPR export]] in addition, just in case
+
+- [[https://developer.spotify.com/documentation/web-api/reference/player/get-recently-played]["Recently played"]] API endpoint **only returns the 50 most recent tracks**, which makes it kind of useless unless you export the data every hour or so.
+
+ If you care about them, might be a good idea to connect Spotify to Last.FM.
+
+ GPDR export has more tracks, but also seems incomplete (e.g. my data is missing first few years).
+
+* Using the data
+
+** TODO need to implement the data access bit

setup.py

@@ -12,6 +12,7 @@ def main():
  package_data={pkg: ['py.typed']},
 
  install_requires=[
+ 'spotipy', # API
  ],
  extras_require={
  'testing': ['pytest'],

src/spotifyexport/export.py

@@ -0,0 +1,111 @@
+#!/usr/bin/env python3
+import argparse
+import json
+from functools import lru_cache
+import sys
+from typing import List
+
+import spotipy # type: ignore[import]
+
+from .exporthelpers import logging_helper
+from .exporthelpers.export_helper import Json
+
+
+logger = logging_helper.logger('spotifyexport')
+
+
+def _cleanup(j: Json) -> Json:
+ '''
+ Clean up irrelevant (hopefully?) stuff from the data.
+ '''
+ # NOTE: for now not used.. maybe make it an optional cmdline flag?
+ artists = j['track']['album']['artists'] + j['track']['artists']
+ for k in ('external_urls', ):
+ for a in artists:
+ del a[k]
+ for k in ('available_markets', 'images', 'external_urls', 'href', 'uri', 'release_date_precision'):
+ del j['track']['album'][k]
+ for k in ('available_markets', 'preview_url', 'external_ids', 'external_urls', 'href', 'uri'):
+ del j['track'][k]
+ return j
+
+
+def as_list(api_method) -> List[Json]:
+ results: List[Json] = []
+ while True:
+ offset = len(results)
+ cres = api_method(limit=50, offset=offset)
+ chunk = cres['items']
+ total = cres['total']
+ logger.debug('%s: collected: %d/%d', api_method, len(results), total)
+ if len(results) >= total:
+ break
+ results.extend(chunk)
+ # todo log?
+ return results
+
+
+class Exporter:
+ SCOPE = 'playlist-read-private,user-library-read,user-read-recently-played'
+
+ def __init__(self, **kwargs) -> None:
+ kw = {
+ 'scope' : self.SCOPE,
+ 'open_browser': False,
+ }
+ kw.update(kwargs)
+ auth = spotipy.oauth2.SpotifyOAuth(**kw)
+ self.api = spotipy.Spotify(auth_manager=auth)
+
+ def export_json(self) -> Json:
+ playlists = as_list(self.api.current_user_playlists)
+ for p in playlists:
+ pid = p['id']
+ p['tracks'] = as_list(lambda *args, **kwargs: self.api.playlist_items(*args, playlist_id=pid, **kwargs))
+ # todo cleanup stuff??
+ return dict(
+ saved_tracks=as_list(self.api.current_user_saved_tracks),
+ saved_albums=as_list(self.api.current_user_saved_albums),
+ saved_shows =as_list(self.api.current_user_saved_shows),
+ # NOTE: seems that only supports the most recent 50
+ # https://developer.spotify.com/documentation/web-api/reference/player/get-recently-played
+ recently_played=self.api.current_user_recently_played(limit=50)['items'],
+ playlists =playlists,
+ )
+
+
+def get_json(**params):
+ return Exporter(**params).export_json()
+
+
+def main() -> None:
+ p = make_parser()
+ args = p.parse_args()
+
+ params = args.params
+ dumper = args.dumper
+ j = get_json(**params)
+ js = json.dumps(j, ensure_ascii=False, indent=1)
+ dumper(js)
+
+
+def make_parser():
+ from .exporthelpers.export_helper import setup_parser, Parser
+ p = Parser('Export your personal Spotify data: playlists, saved tracks/albums/shows, etc. as JSON.')
+ setup_parser(
+ parser=p,
+ params=[
+ 'client_id' ,
+ 'client_secret',
+ 'redirect_uri' ,
+ 'cache_path' ,
+ ]
+ )
+ return p
+
+
+if __name__ == "__main__":
+ main()
+
+
+# todo https://stackoverflow.com/a/30557896 in case of too may requests