fix: switch to spotapi (api-partner.spotify.com) to avoid rate limits

api.spotify.com/v1/ aggressively rate-limits server IPs regardless of TLS fingerprint or token. SpotipyFree was already hitting the correct endpoint (api-partner.spotify.com) but crashed on KeyError: 'ownerV2' due to its own outdated field parsing — not a network block. spotapi uses the same internal GraphQL API (Spotify's web player API) which is on a separate rate-limit bucket from the public REST endpoint. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-05-12 21:02:28 +02:00
1 parent d4c41a9f92
commit 4ec441af57
@@ -1,136 +1,136 @@
 #!/usr/bin/env python3
 """
-Resolves Spotify URLs via curl_cffi Chrome TLS impersonation + TOTP anonymous token.
-curl_cffi presents a real Chrome TLS fingerprint, which gets past Spotify's bot filters
-on open.spotify.com/api/server-time (plain Node.js/Python TLS does not).
+Resolves Spotify URLs via the Spotify partner GraphQL API (api-partner.spotify.com).
+This is the same internal API the Spotify web player uses — separate rate-limit bucket
+from the public api.spotify.com/v1/ which gets throttled on server IPs.
 """
-import sys, json, re, hmac, hashlib, struct, time
+import sys, json, re, time

 def eprint(*a): print(*a, file=sys.stderr)

-def _totp(secret_str, server_ts):
-    secret_bytes = [ord(c) for c in secret_str]
-    transformed = [b ^ ((i % 33) + 9) for i, b in enumerate(secret_bytes)]
-    key = bytes.fromhex(''.join(str(b) for b in transformed).encode('utf-8').hex())
-    counter_b = struct.pack('>Q', server_ts // 30)
-    digest = hmac.new(key, counter_b, hashlib.sha1).digest()
-    offset = digest[-1] & 0xf
-    code = ((digest[offset] & 0x7f) << 24 | (digest[offset+1] & 0xff) << 16 |
-            (digest[offset+2] & 0xff) << 8 | (digest[offset+3] & 0xff)) % 1_000_000
-    return str(code).zfill(6)
+# ─── RESPONSE PARSING ────────────────────────────────────────────────────────
+# Partner API track items look like:
+# { "itemV2": { "data": { "trackUnion": { "name":…, "artists":…, "albumOfTrack":…,
+#                                         "duration":…, "trackNumber":…, "contentRating":… } } } }

-def _get_token(session):
-    html = session.get('https://open.spotify.com/', headers={
-        'Accept': 'text/html,*/*;q=0.8', 'Accept-Language': 'en-US,en;q=0.9',
-        'Accept-Encoding': 'identity',
-    }, timeout=20).text
-    js_urls, seen = [], set()
-    for m in re.finditer(r'src="((?:https?://[^"]*|/[^"]*?)\.js(?:\?[^"]*)?)"', html):
-        u = m.group(1)
-        if not u.startswith('http'): u = 'https://open.spotify.com' + u
-        if u not in seen: seen.add(u); js_urls.append(u)
-    secret = version = None
-    for u in js_urls[:20]:
-        if secret: break
+def _parse_track(item, default_album='', default_art='', default_artist=''):
    try:
-            js = session.get(u, headers={'Accept':'*/*','Accept-Encoding':'identity'}, timeout=8).text
-            m = re.search(r'\{secret:\s*["\']([^"\']+)["\']\s*,\s*version:\s*(\d+)\s*\}', js)
-            if m: secret, version = m.group(1), int(m.group(2))
-        except: pass
-    if not secret: raise Exception('TOTP secret not found')
-    eprint(f'[TOTP] secret v{version} found')
-    st = session.get('https://open.spotify.com/api/server-time', headers={
-        'Accept': 'application/json', 'Referer': 'https://open.spotify.com/'}, timeout=10).json()
-    totp = _totp(secret, st['serverTime'])
-    tok = session.get(
-        f'https://open.spotify.com/api/token?reason=init&productType=web-player&totp={totp}&totpVer={version}&totpServer={totp}',
-        headers={'Accept': 'application/json', 'Referer': 'https://open.spotify.com/'}, timeout=10).json()
-    if 'accessToken' not in tok: raise Exception(f'No accessToken: {json.dumps(tok)[:100]}')
-    eprint('[TOTP] token ok')
-    return tok['accessToken']
+        inner = (item.get('itemV2') or item.get('track') or {}).get('data') or {}
+        t = inner.get('trackUnion') or inner
+        if not t.get('name'): return None

-def _auth_h(token):
-    return {'Authorization': f'Bearer {token}', 'app-platform': 'WebPlayer',
-            'spotify-app-version': '1.2.46.25.g7f189073', 'Origin': 'https://open.spotify.com',
-            'Referer': 'https://open.spotify.com/', 'Accept': 'application/json',
-            'Accept-Language': 'en-US,en;q=0.9'}
+        artist_items = (t.get('artists') or {}).get('items') or []
+        artists = ', '.join(
+            a.get('profile', {}).get('name', '') for a in artist_items
+        ) or default_artist

-def _rest(session, path, token):
-    url = path if path.startswith('http') else f'https://api.spotify.com/v1{path}'
-    r = session.get(url, headers=_auth_h(token), timeout=15)
-    if r.status_code == 429:
-        retry = int(r.headers.get('Retry-After', '30'))
-        eprint(f'[REST] 429 Retry-After={retry}s')
-        print(json.dumps({'error': 'RATE_LIMITED', 'retryAfter': retry}))
-        sys.exit(1)
-    if not r.ok: raise Exception(f'REST {r.status_code}: {path}')
-    return r.json()
+        alb       = t.get('albumOfTrack') or {}
+        alb_name  = alb.get('name', '') or default_album
+        alb_imgs  = (alb.get('images') or {}).get('items') or []
+        art       = alb_imgs[0].get('url', '') if alb_imgs else default_art
+        release   = ((alb.get('date') or {}).get('isoString') or '')[:10]
+
+        dur_ms    = (t.get('duration') or {}).get('totalMilliseconds')
+        rating    = (t.get('contentRating') or {}).get('label', '')

-def _fmt_track(t, default_album='', default_art='', default_artist=''):
-    if not t or not t.get('name'): return None
-    artists = ', '.join(a.get('name', '') for a in (t.get('artists') or [])) or default_artist
-    alb = t.get('album') or {}
-    imgs = alb.get('images') or []
-    art = sorted(imgs, key=lambda x: -(x.get('width') or 0))[0].get('url', '') if imgs else default_art
        return {
            'trackName':         t.get('name', ''),
            'artistName':        artists,
-        'collectionName':    alb.get('name', '') or default_album,
+            'collectionName':    alb_name,
            'artworkUrl100':     art,
-        'releaseDate':       alb.get('release_date', ''),
-        'trackNumber':       t.get('track_number'),
-        'trackExplicitness': 'explicit' if t.get('explicit') else 'notExplicit',
-        'durationMs':        t.get('duration_ms'),
+            'releaseDate':       release,
+            'trackNumber':       t.get('trackNumber'),
+            'trackExplicitness': 'explicit' if rating == 'EXPLICIT' else 'notExplicit',
+            'durationMs':        dur_ms,
        }
+    except Exception as e:
+        eprint(f'[parse] error on item: {e}')
+        return None

-def _best_art(imgs):
-    return sorted(imgs, key=lambda x: -(x.get('width') or 0))[0].get('url', '') if imgs else ''
+def _parse_images(imgs_obj):
+    items = (imgs_obj or {}).get('items') or []
+    return items[0].get('url', '') if items else ''
+
+# ─── RESOLVER ────────────────────────────────────────────────────────────────

 def resolve(sp_type, sp_id):
-    from curl_cffi import requests as cf
-    session = cf.Session(impersonate='chrome124')
-    token = _get_token(session)
+    from spotapi import PublicPlaylist, PublicAlbum, PublicSong
+
+    url = f'https://open.spotify.com/{sp_type}/{sp_id}'

    if sp_type == 'track':
-        t = _rest(session, f'/tracks/{sp_id}', token)
-        artist = ', '.join(a['name'] for a in (t.get('artists') or []))
-        art = _best_art((t.get('album') or {}).get('images') or [])
-        return {'type': 'track', 'name': t.get('name', ''), 'artist': artist, 'artworkUrl': art,
-                'tracks': [_fmt_track(t)]}
+        song = PublicSong(url)
+        raw = song.get_song_info()
+        eprint(f'[spotapi] track response keys: {list(raw.keys()) if isinstance(raw, dict) else type(raw)}')
+        tr = _parse_track(raw)
+        if not tr:
+            # Fallback: raw might already be a flat dict
+            tr = {
+                'trackName':         raw.get('name', ''),
+                'artistName':        raw.get('artist', ''),
+                'collectionName':    raw.get('album', ''),
+                'artworkUrl100':     raw.get('image', ''),
+                'releaseDate':       raw.get('release_date', ''),
+                'trackNumber':       raw.get('track_number'),
+                'trackExplicitness': 'notExplicit',
+                'durationMs':        raw.get('duration_ms'),
+            }
+        name   = tr.get('trackName', '')
+        artist = tr.get('artistName', '')
+        art    = tr.get('artworkUrl100', '')
+        return {'type': 'track', 'name': name, 'artist': artist, 'artworkUrl': art, 'tracks': [tr]}

    if sp_type == 'album':
-        alb = _rest(session, f'/albums/{sp_id}', token)
-        artist = ', '.join(a['name'] for a in (alb.get('artists') or []))
-        art = _best_art(alb.get('images') or [])
-        tracks, path = [], f'/albums/{sp_id}/tracks?limit=50'
-        while path:
-            pg = _rest(session, path, token)
-            tracks += [t for t in [_fmt_track(i, alb.get('name', ''), art, artist)
-                       for i in (pg.get('items') or []) if i and i.get('id')] if t]
-            path = pg.get('next')
-            if path: time.sleep(0.3)
-        eprint(f'[curl_cffi] album ok — {len(tracks)} tracks')
-        return {'type': 'album', 'name': alb.get('name', ''), 'artist': artist,
-                'artworkUrl': art, 'tracks': tracks}
+        alb      = PublicAlbum(url)
+        alb_info = alb.get_album_info()
+        eprint(f'[spotapi] album response type: {type(alb_info)}, keys: {list(alb_info.keys()) if isinstance(alb_info, dict) else "?"}')

-    pl = _rest(session, f'/playlists/{sp_id}?fields=name,description,owner,images', token)
-    owner = (pl.get('owner') or {}).get('display_name', '')
-    art = _best_art(pl.get('images') or [])
-    tracks, skipped, path = [], 0, f'/playlists/{sp_id}/tracks?limit=100'
-    while path:
-        pg = _rest(session, path, token)
-        for item in (pg.get('items') or []):
-            t = item.get('track')
-            if not t or not t.get('id'): continue
-            if t.get('is_local'): skipped += 1; continue
-            tr = _fmt_track(t)
+        # alb_info may be a dict with metadata + items, or just items
+        meta   = alb_info if isinstance(alb_info, dict) else {}
+        name   = meta.get('name', '')
+        artist = meta.get('artist', '') or meta.get('artists', '')
+        art    = meta.get('image', '') or meta.get('cover', '')
+
+        raw_tracks = meta.get('tracks') or meta.get('items') or (alb_info if isinstance(alb_info, list) else [])
+        tracks = [t for t in (_parse_track(i) for i in raw_tracks) if t]
+        eprint(f'[spotapi] album ok — {len(tracks)} tracks')
+        return {'type': 'album', 'name': name, 'artist': artist, 'artworkUrl': art, 'tracks': tracks}
+
+    # playlist
+    pl = PublicPlaylist(url)
+    pl_meta = pl.get_playlist_info(limit=1)  # first call often returns metadata too
+    eprint(f'[spotapi] playlist meta type: {type(pl_meta)}')
+
+    tracks, skipped = [], 0
+    offset = 0
+    while True:
+        page = pl.get_playlist_info(limit=100, offset=offset)
+        items = page if isinstance(page, list) else (page.get('items') or page.get('tracks') or [])
+        if not items: break
+        for item in items:
+            if (item.get('itemV2') or item.get('track', {}).get('data', {})).get('__typename') == 'LocalTrackResponseWrapper':
+                skipped += 1; continue
+            tr = _parse_track(item)
            if tr: tracks.append(tr)
-        path = pg.get('next')
-        if path: time.sleep(0.3)
-    eprint(f'[curl_cffi] playlist ok — {len(tracks)} tracks')
-    return {'type': 'playlist', 'name': pl.get('name', ''), 'artist': owner,
+        if len(items) < 100: break
+        offset += len(items)
+        time.sleep(0.3)
+
+    # Extract playlist name/art from first page if available
+    name  = ''
+    owner = ''
+    art   = ''
+    if isinstance(pl_meta, dict):
+        name  = pl_meta.get('name', '')
+        owner = pl_meta.get('owner', '') or pl_meta.get('owner_name', '')
+        art   = pl_meta.get('image', '') or pl_meta.get('cover', '')
+
+    eprint(f'[spotapi] playlist ok — {len(tracks)} tracks')
+    return {'type': 'playlist', 'name': name, 'artist': owner,
            'artworkUrl': art, 'skippedLocal': skipped, 'tracks': tracks}

+# ─── MAIN ────────────────────────────────────────────────────────────────────
+
 def main():
    if len(sys.argv) < 2: eprint('No URL'); sys.exit(1)
    url = sys.argv[1]
@@ -143,7 +143,7 @@ def main():
        print(json.dumps(result, ensure_ascii=False))
    except SystemExit: raise
    except Exception as e:
-        eprint(f'[curl_cffi] failed: {type(e).__name__}: {e}')
+        eprint(f'[spotapi] failed: {type(e).__name__}: {e}')
        sys.exit(1)

 if __name__ == '__main__':
@@ -34,7 +34,7 @@ services:
      sh -c "
        apt-get update && apt-get install -y --no-install-recommends python3 python3-pip ffmpeg curl git && rm -rf /var/lib/apt/lists/* &&
        curl -sL https://github.com/yt-dlp/yt-dlp/releases/latest/download/yt-dlp -o /usr/local/bin/yt-dlp && chmod a+rx /usr/local/bin/yt-dlp &&
-        pip install --break-system-packages SpotipyFree 'git+https://github.com/TzurSoffer/spotify-downloader' &&
+        pip install --break-system-packages spotapi 'git+https://github.com/TzurSoffer/spotify-downloader' &&
        mkdir -p /deps && cd /deps && npm install express &&
        cd /app && NODE_PATH=/deps/node_modules node novasonic-server.js
      "