fixing paths

config/settings.py

@@ -4,38 +4,30 @@ from dotenv import load_dotenv
 
 load_dotenv()
 
-# Get base project directory
-BASE_DIR = Path(__file__).parent.parent
-
 class Settings:
+    # Base Paths
+    BASE_DIR = Path(__file__).parent.parent
+    
     # Deployment Mode
-    DEPLOYMENT_MODE = os.getenv('DEPLOYMENT_MODE', 'local')  # 'local' or 'huggingface'
+    DEPLOYMENT_MODE = os.getenv('DEPLOYMENT_MODE', 'local')
     
     # API Keys
     OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
     HF_TOKEN = os.getenv("HF_TOKEN")
     
-    # Directory Paths
+    # HuggingFace Settings
+    HF_SPACE = "SongLift/LyrGen2"
+    HF_DATASET = "SongLift/LyrGen2_DB"
+    
+    # Local Settings
     LYRICS_DIR = BASE_DIR / "data" / "raw" / "lyrics"
     EMBEDDINGS_DIR = BASE_DIR / "data" / "processed" / "embeddings"
+    CHROMA_DIR = EMBEDDINGS_DIR / "chroma"
     
     # Model Settings
     EMBEDDING_MODEL = "text-embedding-ada-002"
     LLM_MODEL = "gpt-4"
     
-    # ChromaDB Settings
-    @classmethod
-    def get_chroma_settings(cls) -> dict:
-        """Get ChromaDB settings based on deployment mode"""
-        return {
-            "anonymized_telemetry": False,
-            "persist_directory": str(cls.get_embeddings_path() / "chroma")
-        }
-    
-    # HuggingFace Settings
-    HF_SPACE = "SongLift/LyrGen2"
-    HF_DATASET = "SongLift/LyrGen2_DB"
-    
     @classmethod
     def is_huggingface(cls) -> bool:
         """Check if running in HuggingFace environment"""
@@ -43,10 +35,29 @@ class Settings:
     
     @classmethod
     def get_embeddings_path(cls) -> Path:
-        """Get appropriate embeddings path based on deployment mode"""
+        """Get the base embeddings path"""
         if cls.is_huggingface():
-            # Use absolute path for HuggingFace
+            # HuggingFace: Use absolute path in persistent storage
             return Path("/data/processed/embeddings")
-        
-        # For local development, use project-relative path
-        return cls.EMBEDDINGS_DIR
+        # Local: Use project-relative path
+        return cls.BASE_DIR / "data" / "processed" / "embeddings"
+    
+    @classmethod
+    def get_chroma_path(cls) -> Path:
+        """Get the Chroma DB path"""
+        return cls.get_embeddings_path() / "chroma"
+    
+    @classmethod
+    def ensure_embedding_paths(cls) -> None:
+        """Ensure all embedding-related directories exist"""
+        if not cls.is_huggingface():  # Only create directories locally
+            cls.get_embeddings_path().mkdir(parents=True, exist_ok=True)
+            cls.get_chroma_path().mkdir(parents=True, exist_ok=True)
+    
+    @classmethod
+    def get_chroma_settings(cls) -> dict:
+        """Get ChromaDB settings"""
+        return {
+            "anonymized_telemetry": False,
+            "persist_directory": str(cls.get_chroma_path())
+        }

src/generator/generator.py

@@ -13,8 +13,16 @@ class LyricGenerator:
     def __init__(self):
         """Initialize the generator with embeddings"""
         print("Initializing LyricGenerator...")
+        print(f"Deployment mode: {Settings.DEPLOYMENT_MODE}")
+        
+        # Ensure paths exist (if local)
+        Settings.ensure_embedding_paths()
+        
+        # Get and log paths
         self.embeddings_dir = Settings.get_embeddings_path()
+        self.chroma_dir = Settings.get_chroma_path()
         print(f"Embeddings directory: {self.embeddings_dir}")
+        print(f"Chroma directory: {self.chroma_dir}")
         
         # Initialize OpenAI embeddings
         print("Setting up OpenAI embeddings...")