Add album_artist, year, rating, file_path column to data layer across all providers

.gitignore

@@ -14,6 +14,7 @@ env/
 # IMPORTANT: Never commit your .env file with secrets!
 .env*
 !.env.example
+!.env.test.example
 # You can add an exception for an example file if you create one
 # !.env.example
 
@@ -35,6 +36,7 @@ env/
 # Test artifacts
 .pytest_cache/
 htmlcov/
+nul
 
 # Large model files in query folder
 /query/*.pt
@@ -64,3 +66,6 @@ student_clap/models/*.onnx
 student_clap/config.local.yaml
 student_clap/models/FMA_SONGS_LICENSE.md
 student_clap/models/FMA_SONGS_2247_LICENSE.md
+
+# Testing suite
+testing_suite/

Dockerfile

@@ -403,4 +403,4 @@ ENV PYTHONPATH=/usr/local/lib/python3/dist-packages:/app
 EXPOSE 8000
 
 WORKDIR /workspace
-CMD ["bash", "-c", "if [ -n \"$TZ\" ] && [ -f \"/usr/share/zoneinfo/$TZ\" ]; then ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo $TZ > /etc/timezone; elif [ -n \"$TZ\" ]; then echo \"Warning: timezone '$TZ' not found in /usr/share/zoneinfo\" >&2; fi; if [ \"$SERVICE_TYPE\" = \"worker\" ]; then echo 'Starting worker processes via supervisord...' && /usr/bin/supervisord -c /etc/supervisor/conf.d/supervisord.conf; else echo 'Starting web service...' && gunicorn --bind 0.0.0.0:8000 --workers 1 --timeout 120 app:app; fi"]
+CMD ["bash", "-c", "if [ -n \"$TZ\" ] && [ -f \"/usr/share/zoneinfo/$TZ\" ]; then ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo $TZ > /etc/timezone; elif [ -n \"$TZ\" ]; then echo \"Warning: timezone '$TZ' not found in /usr/share/zoneinfo\" >&2; fi; if [ \"$SERVICE_TYPE\" = \"worker\" ]; then echo 'Starting worker processes via supervisord...' && /usr/bin/supervisord -c /etc/supervisor/conf.d/supervisord.conf; else echo 'Starting web service...' && gunicorn --bind 0.0.0.0:8000 --workers 1 --timeout 300 app:app; fi"]

Dockerfile-noavx2

@@ -397,4 +397,4 @@ ENV PYTHONPATH=/usr/local/lib/python3/dist-packages:/app
 EXPOSE 8000
 
 WORKDIR /workspace
-CMD ["bash", "-c", "if [ \"$SERVICE_TYPE\" = \"worker\" ]; then echo 'Starting worker processes via supervisord...' && /usr/bin/supervisord -c /etc/supervisor/conf.d/supervisord.conf; else echo 'Starting web service...' && gunicorn --bind 0.0.0.0:8000 --workers 1 --timeout 120 app:app; fi"]
+CMD ["bash", "-c", "if [ \"$SERVICE_TYPE\" = \"worker\" ]; then echo 'Starting worker processes via supervisord...' && /usr/bin/supervisord -c /etc/supervisor/conf.d/supervisord.conf; else echo 'Starting web service...' && gunicorn --bind 0.0.0.0:8000 --workers 1 --timeout 300 app:app; fi"]

app_helper.py

@@ -87,7 +87,7 @@ def init_db():
         cur.execute('CREATE EXTENSION IF NOT EXISTS unaccent')
         cur.execute('CREATE EXTENSION IF NOT EXISTS pg_trgm')
         # Create 'score' table
-        cur.execute("CREATE TABLE IF NOT EXISTS score (item_id TEXT PRIMARY KEY, title TEXT, author TEXT, album TEXT, tempo REAL, key TEXT, scale TEXT, mood_vector TEXT)")
+        cur.execute("CREATE TABLE IF NOT EXISTS score (item_id TEXT PRIMARY KEY, title TEXT, author TEXT, album TEXT, album_artist TEXT, tempo REAL, key TEXT, scale TEXT, mood_vector TEXT)")
         # Add 'energy' column if not exists
         cur.execute("SELECT EXISTS (SELECT 1 FROM information_schema.columns WHERE table_name = 'score' AND column_name = 'energy')")
         if not cur.fetchone()[0]:
@@ -103,6 +103,26 @@ def init_db():
         if not cur.fetchone()[0]:
             logger.info("Adding 'album' column to 'score' table.")
             cur.execute("ALTER TABLE score ADD COLUMN album TEXT")
+        # Add 'album_artist' column if not exists
+        cur.execute("SELECT EXISTS (SELECT 1 FROM information_schema.columns WHERE table_name = 'score' AND column_name = 'album_artist')")
+        if not cur.fetchone()[0]:
+            logger.info("Adding 'album_artist' column to 'score' table.")
+            cur.execute("ALTER TABLE score ADD COLUMN album_artist TEXT")
+        # Add 'year' column if not exists
+        cur.execute("SELECT EXISTS (SELECT 1 FROM information_schema.columns WHERE table_name = 'score' AND column_name = 'year')")
+        if not cur.fetchone()[0]:
+            logger.info("Adding 'year' column to 'score' table.")
+            cur.execute("ALTER TABLE score ADD COLUMN year INTEGER")
+        # Add 'rating' column if not exists
+        cur.execute("SELECT EXISTS (SELECT 1 FROM information_schema.columns WHERE table_name = 'score' AND column_name = 'rating')")
+        if not cur.fetchone()[0]:
+            logger.info("Adding 'rating' column to 'score' table.")
+            cur.execute("ALTER TABLE score ADD COLUMN rating INTEGER")
+        # Add 'file_path' column if not exists
+        cur.execute("SELECT EXISTS (SELECT 1 FROM information_schema.columns WHERE table_name = 'score' AND column_name = 'file_path')")
+        if not cur.fetchone()[0]:
+            logger.info("Adding 'file_path' column to 'score' table.")
+            cur.execute("ALTER TABLE score ADD COLUMN file_path TEXT")
 
         # Add 'search_u' column if not exists (helps search)
         cur.execute("SELECT EXISTS (SELECT 1 FROM information_schema.columns WHERE table_name = 'score' AND column_name = 'search_u')")
@@ -441,7 +461,7 @@ def track_exists(item_id):
     cur.close()
     return row is not None
 
-def save_track_analysis_and_embedding(item_id, title, author, tempo, key, scale, moods, embedding_vector, energy=None, other_features=None, album=None):
+def save_track_analysis_and_embedding(item_id, title, author, tempo, key, scale, moods, embedding_vector, energy=None, other_features=None, album=None, album_artist=None, year=None, rating=None, file_path=None):
     """Saves track analysis and embedding in a single transaction."""
 
     def _sanitize_string(s, max_length=1000, field_name="field"):
@@ -479,19 +499,83 @@ def _sanitize_string(s, max_length=1000, field_name="field"):
     title = _sanitize_string(title, max_length=500, field_name="title")
     author = _sanitize_string(author, max_length=200, field_name="author")
     album = _sanitize_string(album, max_length=200, field_name="album")
+    album_artist = _sanitize_string(album_artist, max_length=200, field_name="album_artist")
     key = _sanitize_string(key, max_length=10, field_name="key")
     scale = _sanitize_string(scale, max_length=10, field_name="scale")
     other_features = _sanitize_string(other_features, max_length=2000, field_name="other_features")
 
+    # year: parse from various date formats and validate
+    def _parse_year_from_date(year_value):
+        """
+        Parse year from various date formats.
+        Supports: YYYY, YYYY-MM-DD, MM-DD-YYYY, DD-MM-YYYY (with - or / separators)
+        """
+        if year_value is None:
+            return None
+
+        year_str = str(year_value).strip()
+        if not year_str:
+            return None
+
+        # Try parsing as pure integer first (YYYY)
+        try:
+            year = int(year_str)
+            if 1000 <= year <= 2100:
+                return year
+        except (ValueError, TypeError):
+            pass
+
+        # Normalize separators
+        normalized = year_str.replace('/', '-')
+        parts = normalized.split('-')
+
+        if len(parts) == 3:
+            try:
+                # YYYY-MM-DD format
+                if len(parts[0]) == 4:
+                    year = int(parts[0])
+                    if 1000 <= year <= 2100:
+                        return year
+
+                # MM-DD-YYYY or DD-MM-YYYY format
+                if len(parts[2]) == 4:
+                    year = int(parts[2])
+                    if 1000 <= year <= 2100:
+                        return year
+
+                # 2-digit year (MM-DD-YY)
+                if len(parts[2]) == 2:
+                    year = int(parts[2])
+                    year += 2000 if year < 30 else 1900
+                    if 1000 <= year <= 2100:
+                        return year
+            except (ValueError, TypeError, IndexError):
+                pass
+
+        return None
+
+    year = _parse_year_from_date(year)
+
+    # rating: validate as integer 0-5 (5-star rating system)
+    if rating is not None:
+        try:
+            rating = int(rating)
+            if rating < 0 or rating > 5:
+                rating = None
+        except (ValueError, TypeError):
+            rating = None
+
+    file_path = _sanitize_string(file_path, max_length=1000, field_name="file_path")
+
     mood_str = ','.join(f"{k}:{v:.3f}" for k, v in moods.items())
 
     conn = get_db() # This now calls the function within this file
     cur = conn.cursor()
     try:
         # Save analysis to score table
         cur.execute("""
-            INSERT INTO score (item_id, title, author, tempo, key, scale, mood_vector, energy, other_features, album)
-            VALUES (%s, %s, %s, %s, %s, %s, %s, %s, %s, %s)
+            INSERT INTO score (item_id, title, author, tempo, key, scale, mood_vector, energy, other_features, album, album_artist, year, rating, file_path)
+            VALUES (%s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s)
             ON CONFLICT (item_id) DO UPDATE SET
                 title = EXCLUDED.title,
                 author = EXCLUDED.author,
@@ -501,8 +585,12 @@ def _sanitize_string(s, max_length=1000, field_name="field"):
                 mood_vector = EXCLUDED.mood_vector,
                 energy = EXCLUDED.energy,
                 other_features = EXCLUDED.other_features,
-                album = EXCLUDED.album
-        """, (item_id, title, author, tempo, key, scale, mood_str, energy, other_features, album))
+                album = EXCLUDED.album,
+                album_artist = EXCLUDED.album_artist,
+                year = EXCLUDED.year,
+                rating = EXCLUDED.rating,
+                file_path = EXCLUDED.file_path
+        """, (item_id, title, author, tempo, key, scale, mood_str, energy, other_features, album, album_artist, year, rating, file_path))
 
         # Save embedding
         if isinstance(embedding_vector, np.ndarray) and embedding_vector.size > 0:
@@ -589,7 +677,7 @@ def get_all_tracks():
     conn = get_db() # This now calls the function within this file
     cur = conn.cursor(cursor_factory=DictCursor)
     cur.execute("""
-        SELECT s.item_id, s.title, s.author, s.tempo, s.key, s.scale, s.mood_vector, s.energy, s.other_features, e.embedding
+        SELECT s.item_id, s.title, s.author, s.tempo, s.key, s.scale, s.mood_vector, s.energy, s.other_features, s.year, s.rating, s.file_path, e.embedding
         FROM score s
         LEFT JOIN embedding e ON s.item_id = e.item_id
     """)
@@ -620,7 +708,7 @@ def get_tracks_by_ids(item_ids_list):
     item_ids_str = [str(item_id) for item_id in item_ids_list]
 
     query = """
-        SELECT s.item_id, s.title, s.author, s.album, s.tempo, s.key, s.scale, s.mood_vector, s.energy, s.other_features, e.embedding
+        SELECT s.item_id, s.title, s.author, s.album, s.album_artist, s.tempo, s.key, s.scale, s.mood_vector, s.energy, s.other_features, s.year, s.rating, s.file_path, e.embedding
         FROM score s
         LEFT JOIN embedding e ON s.item_id = e.item_id
         WHERE s.item_id IN %s
@@ -648,7 +736,7 @@ def get_score_data_by_ids(item_ids_list):
     conn = get_db() # This now calls the function within this file
     cur = conn.cursor(cursor_factory=DictCursor)
     query = """
-        SELECT s.item_id, s.title, s.author, s.album, s.tempo, s.key, s.scale, s.mood_vector, s.energy, s.other_features
+        SELECT s.item_id, s.title, s.author, s.album, s.album_artist, s.tempo, s.key, s.scale, s.mood_vector, s.energy, s.other_features, s.year, s.rating, s.file_path
         FROM score s
         WHERE s.item_id IN %s
     """

app_voyager.py

@@ -113,7 +113,8 @@ def search_tracks_endpoint():
                     'item_id': r.get('item_id'),
                     'title': r.get('title'),
                     'author': r.get('author'),
-                    'album': album
+                    'album': album,
+                    'album_artist': (r.get('album_artist') or '').strip() or 'unknown'
                 })
             else:
                 results.append({'item_id': None, 'title': None, 'author': None, 'album': 'unknown'})
@@ -256,6 +257,7 @@ def get_similar_tracks_endpoint():
                     "title": track_info['title'],
                     "author": track_info['author'],
                     "album": (track_info.get('album') or 'unknown'),
+                    "album_artist": (track_info.get('album_artist') or 'unknown'),
                     "distance": distance_map[neighbor_id]
                 })
 
@@ -314,7 +316,8 @@ def get_track_endpoint():
         "item_id": d.get('item_id'),
         "title": d.get('title'),
         "author": d.get('author'),
-        "album": (d.get('album') or 'unknown')
+        "album": (d.get('album') or 'unknown'),
+        "album_artist": (d.get('album_artist') or 'unknown')
     }), 200
   except Exception as e:
     logger.error(f"Unexpected error fetching track {item_id}: {e}", exc_info=True)

config.py

@@ -465,6 +465,11 @@
 # }
 ENABLE_PROXY_FIX = os.environ.get("ENABLE_PROXY_FIX", "False").lower() == "true"
 
+# --- Instant Playlist Optimization ---
+# Max songs from a single artist in the instant playlist (diversity enforcement)
+MAX_SONGS_PER_ARTIST_PLAYLIST = int(os.environ.get("MAX_SONGS_PER_ARTIST_PLAYLIST", "5"))
+# Enable energy-arc shaping for playlist ordering (gentle start -> peak -> cool down)
+PLAYLIST_ENERGY_ARC = os.environ.get("PLAYLIST_ENERGY_ARC", "False").lower() == "true"
 # --- Authentication ---
 # Set all three to enable authentication. Leave any blank to disable (legacy mode).
 AUDIOMUSE_USER = os.environ.get("AUDIOMUSE_USER", "")

deployment/docker-compose-nvidia-local.yaml

@@ -29,6 +29,8 @@ services:
     build:
       context: ..
       dockerfile: Dockerfile
+      args:
+        BASE_IMAGE: nvidia/cuda:12.8.1-cudnn-runtime-ubuntu24.04
     image: audiomuse-ai:local-nvidia
     container_name: audiomuse-ai-flask-app
     ports:
@@ -52,6 +54,8 @@ services:
       OPENAI_MODEL_NAME: "${OPENAI_MODEL_NAME}"
       GEMINI_API_KEY: "${GEMINI_API_KEY}"
       MISTRAL_API_KEY: "${MISTRAL_API_KEY}"
+      OLLAMA_SERVER_URL: "${OLLAMA_SERVER_URL:-http://192.168.1.71:11434/api/generate}"
+      OLLAMA_MODEL_NAME: "${OLLAMA_MODEL_NAME:-qwen3:1.7b}"
       CLAP_ENABLED: "${CLAP_ENABLED:-true}"
       TEMP_DIR: "/app/temp_audio"
       # Authentication (optional) – leave blank to disable
@@ -78,6 +82,8 @@ services:
     build:
       context: ..
       dockerfile: Dockerfile
+      args:
+        BASE_IMAGE: nvidia/cuda:12.8.1-cudnn-runtime-ubuntu24.04
     image: audiomuse-ai:local-nvidia
     container_name: audiomuse-ai-worker-instance
     environment:
@@ -126,5 +132,7 @@ services:
 volumes:
   redis-data:
   postgres-data:
+    external: true
+    name: deployment_postgres-data
   temp-audio-flask:
   temp-audio-worker:

tasks/analysis.py

@@ -881,7 +881,7 @@ def get_missing_mulan_track_ids(track_ids):
                         logger.info(f"  - Other Features: {other_features}")
 
                         # Save MusiCNN score+embedding first (creates the 'score' row)
-                        save_track_analysis_and_embedding(item['Id'], item['Name'], item.get('AlbumArtist', 'Unknown'), musicnn_analysis['tempo'], musicnn_analysis['key'], musicnn_analysis['scale'], top_moods, musicnn_embedding, energy=musicnn_analysis['energy'], other_features=other_features, album=item.get('Album', None))
+                        save_track_analysis_and_embedding(item['Id'], item['Name'], item.get('AlbumArtist', 'Unknown'), musicnn_analysis['tempo'], musicnn_analysis['key'], musicnn_analysis['scale'], top_moods, musicnn_embedding, energy=musicnn_analysis['energy'], other_features=other_features, album=item.get('Album', None), album_artist=item.get('OriginalAlbumArtist', None), year=item.get('Year'), rating=item.get('Rating'), file_path=item.get('FilePath'))
 
                     # Save CLAP embedding AFTER score row exists (FK: clap_embedding.item_id → score.item_id)
                     if clap_embedding_for_track is not None and needs_clap:
@@ -1210,9 +1210,9 @@ def monitor_and_clear_jobs():
                         track_id_str = str(item['Id'])
                         try:
                             with get_db() as conn, conn.cursor() as cur:
-                                cur.execute("UPDATE score SET album = %s WHERE item_id = %s", (album.get('Name'), track_id_str))
+                                cur.execute("UPDATE score SET album = %s, album_artist = %s, year = %s, rating = %s, file_path = %s WHERE item_id = %s", (album.get('Name'), item.get('OriginalAlbumArtist'), item.get('Year'), item.get('Rating'), item.get('FilePath'), track_id_str))
                                 conn.commit()
-                            logger.info(f"[MainAnalysisTask] Updated album name for track '{item['Name']}' to '{album.get('Name')}' (main task)")
+                            logger.info(f"[MainAnalysisTask] Updated album/album_artist/year/rating/file_path for track '{item['Name']}' to '{album.get('Name')}' (main task)")
                         except Exception as e:
                             logger.warning(f"[MainAnalysisTask] Failed to update album name for '{item['Name']}': {e}")
                     albums_skipped += 1

tasks/chat_manager.py

@@ -1247,11 +1247,14 @@ def generate_final_sql_query(intent, strategy_info, found_artists, found_keyword
 - item_id (text)
 - title (text)
 - author (text)
+- album (text)
+- album_artist (text)
 - tempo (numeric 40-200)
 - mood_vector (text, format: 'pop:0.8,rock:0.3')
 - other_features (text, format: 'danceable:0.7,party:0.6')
 - energy (numeric 0-0.15, higher = more energetic)
-- **NOTE: NO YEAR OR DATE COLUMN EXISTS**
+- year (integer, e.g. 2005, NULL if unknown)
+- rating (integer 0-5, NULL if unrated, represents 5-star rating)
 
 **PROGRESSIVE FILTERING STRATEGY - CRITICAL:**
 The goal is to return EXACTLY {target_count} songs. Start with minimal filters and add more ONLY if needed.