From c8f82ceba5bf1682022a95369647a0614ac96827 Mon Sep 17 00:00:00 2001
From: Honey <me_honeyy@proton.me>
Date: Wed, 13 May 2026 01:48:37 +0200
Subject: [PATCH] fix: scrape Spotify OG tags as fallback for single track
 resolution
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

song_info returns GenericError for unauthenticated track lookups.
Fall back to scraping og:title/og:description/og:image from the public
Spotify track page — description format is "Song · Artist · Album · Year".

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
---
 backend/spotify_resolve.py | 65 ++++++++++++++++++++++++++------------
 1 file changed, 44 insertions(+), 21 deletions(-)

diff --git a/backend/spotify_resolve.py b/backend/spotify_resolve.py
index 7b264fb..850a359 100644
--- a/backend/spotify_resolve.py
+++ b/backend/spotify_resolve.py
@@ -4,7 +4,7 @@ Resolves Spotify URLs via the Spotify partner GraphQL API (api-partner.spotify.c
 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, time
+import sys, json, re, time, html as _html, urllib.request
 
 def eprint(*a): print(*a, file=sys.stderr)
 
@@ -65,34 +65,57 @@ def resolve(sp_type, sp_id):
 
     if sp_type == 'track':
         pub = sp_mod.Public()
-        # Try URI first, then full URL
-        raw = None
+        tr = None
+        # Try partner API first (works for some tracks)
         for arg in [f'spotify:track:{sp_id}', url]:
             try:
                 r = pub.song_info(arg)
                 tu_type = ((r.get('data') or {}).get('trackUnion') or {}).get('__typename', '')
                 if tu_type == 'GenericError':
-                    eprint(f'[spotapi] song_info({arg!r}) → GenericError, trying next')
+                    eprint(f'[spotapi] song_info({arg!r}) → GenericError')
                     continue
-                raw = r
-                break
+                d  = (r.get('data') or {})
+                tu = d.get('trackUnion') or d
+                tr = _parse_track({'itemV2': {'data': {'trackUnion': tu}}})
+                if tr and tr.get('trackName'):
+                    eprint(f'[spotapi] track resolved via song_info')
+                    break
+                tr = None
             except Exception as e:
                 eprint(f'[spotapi] song_info({arg!r}) failed: {e}')
-        if raw is None:
-            raise Exception(f'song_info returned GenericError for all inputs — track may require auth or be region-locked')
-        eprint(f'[spotapi] track raw keys={list(raw.keys())}')
-        d  = (raw.get('data') or {}) if isinstance(raw, dict) else {}
-        tu = d.get('trackUnion') or d
-        tr = _parse_track({'itemV2': {'data': {'trackUnion': tu}}}) or {
-            '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'),
-        }
+
+        # Fallback: scrape Open Graph tags from the public Spotify track page
+        if not tr:
+            eprint(f'[spotapi] falling back to OG scrape for track {sp_id}')
+            try:
+                req = urllib.request.Request(url, headers={'User-Agent': 'Mozilla/5.0'})
+                with urllib.request.urlopen(req, timeout=10) as resp:
+                    page = resp.read().decode('utf-8', errors='replace')
+                og = {m.group(1): _html.unescape(m.group(2))
+                      for m in re.finditer(r'<meta\s+property="og:([^"]+)"\s+content="([^"]*)"', page)}
+                # og:description format: "Song · Artist · Album · Year"
+                desc_parts = [p.strip() for p in og.get('description', '').split('·')]
+                track_name = og.get('title', '')
+                artist_name = desc_parts[1] if len(desc_parts) > 1 else ''
+                album_name  = desc_parts[2] if len(desc_parts) > 2 else ''
+                art_url     = og.get('image', '')
+                eprint(f'[spotapi] OG scrape: name={track_name!r} artist={artist_name!r} album={album_name!r}')
+                if track_name:
+                    tr = {
+                        'trackName':         track_name,
+                        'artistName':        artist_name,
+                        'collectionName':    album_name,
+                        'artworkUrl100':     art_url,
+                        'releaseDate':       '',
+                        'trackNumber':       None,
+                        'trackExplicitness': 'notExplicit',
+                        'durationMs':        None,
+                    }
+            except Exception as e:
+                eprint(f'[spotapi] OG scrape failed: {e}')
+
+        if not tr:
+            raise Exception(f'Track resolution failed: song_info returned GenericError and OG scrape found no title')
         return {'type': 'track', 'name': tr.get('trackName', ''),
                 'artist': tr.get('artistName', ''), 'artworkUrl': tr.get('artworkUrl100', ''),
                 'tracks': [tr]}