feat: use HuggingFace persistent storage (/data) for database persistence

- Use /data directory for persistent database storage in HF Spaces
- Unify download/upload repo to channelcorp/ko-tts-arena-db
- Database survives Space restarts and code pushes
- Fall back to instance/ if persistent storage unavailable

app.py

@@ -11,25 +11,44 @@ month = datetime.now().month
 
 # Check if running in a Hugging Face Space
 IS_SPACES = False
+PERSISTENT_DATA_DIR = None  # Will be set if persistent storage is available
+DATABASE_REPO_ID = "channelcorp/ko-tts-arena-db"  # Single source for DB
+
 if os.getenv("SPACE_REPO_NAME"):
     print("Running in a Hugging Face Space 🤗")
     IS_SPACES = True
 
-    # Setup database sync for HF Spaces
-    if not os.path.exists("instance/tts_arena.db"):
-        os.makedirs("instance", exist_ok=True)
+    # Check for persistent storage availability (/data directory)
+    # HuggingFace Spaces provides /data as persistent storage
+    if os.path.exists("/data") and os.access("/data", os.W_OK):
+        PERSISTENT_DATA_DIR = "/data"
+        print("Persistent storage available at /data ✅")
+    else:
+        # Fallback to instance directory (non-persistent)
+        PERSISTENT_DATA_DIR = "instance"
+        print("⚠️ Warning: Persistent storage (/data) not available. Using 'instance/' (data may be lost on restart)")
+
+    # Define database path
+    db_path = os.path.join(PERSISTENT_DATA_DIR, "tts_arena.db")
+    
+    # Setup database - download only if it doesn't exist in persistent storage
+    if not os.path.exists(db_path):
+        os.makedirs(PERSISTENT_DATA_DIR, exist_ok=True)
         try:
-            print("Database not found, downloading from HF dataset...")
+            print(f"Database not found at {db_path}, downloading from HF dataset ({DATABASE_REPO_ID})...")
             hf_hub_download(
-                repo_id="channelcorp/ko-tts-arena-db",
+                repo_id=DATABASE_REPO_ID,
                 filename="tts_arena.db",
                 repo_type="dataset",
-                local_dir="instance",
+                local_dir=PERSISTENT_DATA_DIR,
                 token=os.getenv("HF_TOKEN"),
             )
-            print("Database downloaded successfully ✅")
+            print(f"Database downloaded successfully to {db_path} ✅")
         except Exception as e:
             print(f"Error downloading database from HF dataset: {str(e)} ⚠️")
+            print("A new database will be created.")
+    else:
+        print(f"Database found at {db_path} (persistent storage) ✅")
 
 from flask import (
     Flask,
@@ -117,9 +136,17 @@ if not IS_SPACES:
 
 app = Flask(__name__)
 app.config["SECRET_KEY"] = os.getenv("SECRET_KEY", os.urandom(24))
-app.config["SQLALCHEMY_DATABASE_URI"] = os.getenv(
-    "DATABASE_URI", "sqlite:///tts_arena.db"
-)
+
+# Configure database path - use persistent storage in HF Spaces
+if IS_SPACES and PERSISTENT_DATA_DIR:
+    # Use persistent storage path for HF Spaces
+    app.config["SQLALCHEMY_DATABASE_URI"] = f"sqlite:///{PERSISTENT_DATA_DIR}/tts_arena.db"
+    print(f"Using database at: {PERSISTENT_DATA_DIR}/tts_arena.db")
+else:
+    app.config["SQLALCHEMY_DATABASE_URI"] = os.getenv(
+        "DATABASE_URI", "sqlite:///tts_arena.db"
+    )
+
 app.config["SQLALCHEMY_TRACK_MODIFICATIONS"] = False
 app.config["SESSION_COOKIE_SECURE"] = True
 app.config["SESSION_COOKIE_SAMESITE"] = (
@@ -959,13 +986,15 @@ def setup_periodic_tasks():
     if not IS_SPACES:
         return
 
-    db_path = app.config["SQLALCHEMY_DATABASE_URI"].replace("sqlite:///", "instance/") # Get relative path
+    # Get database path from config (handles both persistent storage and fallback)
+    db_path = app.config["SQLALCHEMY_DATABASE_URI"].replace("sqlite:///", "")
     preferences_repo_id = "TTS-AGI/arena-v2-preferences"
-    database_repo_id = "TTS-AGI/database-arena-v2"
+    # Use the same repo for download and upload (consistency)
+    database_repo_id = DATABASE_REPO_ID
     votes_dir = "./votes"
 
     def sync_database():
-        """Uploads the database to HF dataset"""
+        """Uploads the database to HF dataset (backup to cloud)"""
         with app.app_context(): # Ensure app context for logging
             try:
                 if not os.path.exists(db_path):
@@ -979,7 +1008,7 @@ def setup_periodic_tasks():
                     repo_id=database_repo_id,
                     repo_type="dataset",
                 )
-                app.logger.info(f"Database uploaded to {database_repo_id} at {datetime.utcnow()}")
+                app.logger.info(f"Database backed up to {database_repo_id} at {datetime.utcnow()}")
             except Exception as e:
                 app.logger.error(f"Error uploading database to {database_repo_id}: {str(e)}")
 
@@ -1282,8 +1311,11 @@ def check_for_coordinated_campaigns():
 
 if __name__ == "__main__":
     with app.app_context():
-        # Ensure ./instance and ./votes directories exist
-        os.makedirs("instance", exist_ok=True)
+        # Ensure directories exist
+        if IS_SPACES and PERSISTENT_DATA_DIR:
+            os.makedirs(PERSISTENT_DATA_DIR, exist_ok=True)
+        else:
+            os.makedirs("instance", exist_ok=True)
         os.makedirs("./votes", exist_ok=True) # Create votes directory if it doesn't exist
         os.makedirs(CACHE_AUDIO_DIR, exist_ok=True) # Ensure cache audio dir exists
 
@@ -1302,22 +1334,8 @@ if __name__ == "__main__":
         except Exception as e:
              app.logger.error(f"Error clearing cache directory {CACHE_AUDIO_DIR}: {e}")
 
-
-        # Download database if it doesn't exist (only on initial space start)
-        if IS_SPACES and not os.path.exists(app.config["SQLALCHEMY_DATABASE_URI"].replace("sqlite:///", "")):
-             try:
-                print("Database not found, downloading from HF dataset...")
-                hf_hub_download(
-                    repo_id="TTS-AGI/database-arena-v2",
-                    filename="tts_arena.db",
-                    repo_type="dataset",
-                    local_dir="instance", # download to instance/
-                    token=os.getenv("HF_TOKEN"),
-                )
-                print("Database downloaded successfully ✅")
-             except Exception as e:
-                 print(f"Error downloading database from HF dataset: {str(e)} ⚠️")
-
+        # Note: Database download is handled at module load time for HF Spaces
+        # This ensures DB is ready before app initialization
 
         db.create_all()  # Create tables if they don't exist
         insert_initial_models()