danvics/soulsync
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 2

Improve album search normalization and fallback logic

Browse source 
Enhanced normalization by using _matching_engine when available for both search input and database values. Improved album search by considering artist name variations and deduplicating results. Added a fallback mechanism to search all albums by artist to address SQL accent sensitivity issues, increasing robustness in album matching.
This commit is contained in:
Broque Thomas 2026-01-08 16:42:09 -08:00
parent c6a9db234a
commit 8620b52709
1 changed files with 61 additions and 8 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

69
database/music_database.py
View file

@ -1271,8 +1271,12 @@ class MusicDatabase:
from unidecode import unidecode from unidecode import unidecode
# Normalize search terms # Normalize search terms
title_norm = unidecode(title).lower() if title else "" if _matching_engine:
artist_norm = unidecode(artist).lower() if artist else "" title_norm = _matching_engine.normalize_string(title) if title else ""
artist_norm = _matching_engine.normalize_string(artist) if artist else ""
else:
title_norm = unidecode(title).lower() if title else ""
artist_norm = unidecode(artist).lower() if artist else ""
# Try searching with normalized versions # Try searching with normalized versions
where_conditions = [] where_conditions = []
@ -1312,8 +1316,13 @@ class MusicDatabase:
# Filter results with proper Unicode normalization # Filter results with proper Unicode normalization
filtered_tracks = [] filtered_tracks = []
for row in rows: for row in rows:
db_title_norm = unidecode(row['title'].lower()) if row['title'] else ""
db_artist_norm = unidecode(row['artist_name'].lower()) if row['artist_name'] else "" if _matching_engine:
db_title_norm = _matching_engine.normalize_string(row['title']) if row['title'] else ""
db_artist_norm = _matching_engine.normalize_string(row['artist_name']) if row['artist_name'] else ""
else:
db_title_norm = unidecode(row['title'].lower()) if row['title'] else ""
db_artist_norm = unidecode(row['artist_name'].lower()) if row['artist_name'] else ""
title_matches = not title or title_norm in db_title_norm title_matches = not title or title_norm in db_title_norm
artist_matches = not artist or artist_norm in db_artist_norm artist_matches = not artist or artist_norm in db_artist_norm
@ -1707,7 +1716,16 @@ class MusicDatabase:
for variation in title_variations: for variation in title_variations:
# Search for this variation # Search for this variation
albums = self.search_albums(title=variation, artist=artist, limit=10, server_source=server_source) albums = []
artist_variations = self._get_artist_variations(artist)
for artist_variation in artist_variations:
found = self.search_albums(title=variation, artist=artist_variation, limit=10, server_source=server_source)
# Deduplicate by ID
existing_ids = {a.id for a in albums}
for album in found:
if album.id not in existing_ids:
albums.append(album)
existing_ids.add(album.id)
if albums: if albums:
logger.debug(f"📀 Found {len(albums)} albums for variation '{variation}'") logger.debug(f"📀 Found {len(albums)} albums for variation '{variation}'")
@ -1728,9 +1746,44 @@ class MusicDatabase:
if best_match and best_confidence >= confidence_threshold: if best_match and best_confidence >= confidence_threshold:
logger.debug(f"✅ Edition match found: '{title}' -> '{best_match.title}' (confidence: {best_confidence:.3f})") logger.debug(f"✅ Edition match found: '{title}' -> '{best_match.title}' (confidence: {best_confidence:.3f})")
return best_match, best_confidence return best_match, best_confidence
else:
logger.debug(f"❌ No confident edition match for '{title}' (best: {best_confidence:.3f}, threshold: {confidence_threshold})") # Fallback: Check ALL albums by this artist (resolves SQL accent sensitivity issues #101)
return None, best_confidence # If we haven't found a match yet, fetch broader list from artist and double check
if best_confidence < confidence_threshold:
logger.debug(f"⚠️ specific title search failed, trying broad artist search fallback for '{artist}'")
try:
# Get ALL albums by this artist (limit 100 to be safe)
# This bypasses SQL 'LIKE' limitations for diacritics (e.g. 'ă' vs 'a')
# And relies on Python-side normalization in _calculate_album_confidence
artist_albums = []
artist_variations = self._get_artist_variations(artist)
for artist_var in artist_variations:
found_albums = self.search_albums(title="", artist=artist_var, limit=100, server_source=server_source)
# Deduplicate
existing_ids = {a.id for a in artist_albums}
for album in found_albums:
if album.id not in existing_ids:
artist_albums.append(album)
existing_ids.add(album.id)
if artist_albums:
logger.debug(f" Found {len(artist_albums)} total albums for artist fallback")
for album in artist_albums:
confidence = self._calculate_album_confidence(title, artist, album, expected_track_count)
if confidence > best_confidence:
best_confidence = confidence
best_match = album
logger.debug(f" 🎯 Fallback match: '{album.title}' confidence: {confidence:.3f}")
except Exception as fallback_error:
logger.warning(f"Fallback artist search failed: {fallback_error}")
if best_match and best_confidence >= confidence_threshold:
logger.debug(f"✅ Fallback match succeeded: '{title}' -> '{best_match.title}' (confidence: {best_confidence:.3f})")
return best_match, best_confidence
logger.debug(f"❌ No confident edition match for '{title}' (best: {best_confidence:.3f}, threshold: {confidence_threshold})")
return None, best_confidence
except Exception as e: except Exception as e:
logger.error(f"Error in edition-aware album matching for '{title}' by '{artist}': {e}") logger.error(f"Error in edition-aware album matching for '{title}' by '{artist}': {e}")