Update app.py

app.py

@@ -21,6 +21,7 @@ from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain_community.vectorstores import Cassandra
 from langchain_openai import OpenAIEmbeddings
 from PIL import Image, ImageDraw, ImageFont
+from astrapy.db import AstraDB as DataAPIClient
 
 # Load environment variables
 load_dotenv()
@@ -79,54 +80,20 @@ def init_astra_db():
         astra_db_region = os.getenv("ASTRA_DB_REGION")
         astra_db_keyspace = os.getenv("ASTRA_DB_KEYSPACE")
         astra_db_application_token = os.getenv("ASTRA_DB_APPLICATION_TOKEN")
-        astra_db_endpoint = os.getenv("ASTRA_DB_ENDPOINT","https://8e3fd85c-5f28-4e1f-8538-9dd28a3ea2b0-us-east-2.apps.astra.datastax.com")
+        astra_db_endpoint = os.getenv("ASTRA_DB_ENDPOINT", "https://8e3fd85c-5f28-4e1f-8538-9dd28a3ea2b0-us-east-2.apps.astra.datastax.com")
         
         # Initialize the client
-        client = DataAPIClient(astra_db_token)
+        client = DataAPIClient(astra_db_application_token)
         db = client.get_database_by_api_endpoint(
             astra_db_endpoint,
             keyspace=astra_db_keyspace
         )
         
-        # Create keyspace if it doesn't exist
-        session.execute(f"""
-            CREATE KEYSPACE IF NOT EXISTS {astra_db_keyspace}
-            WITH replication = {{'class': 'SimpleStrategy', 'replication_factor': '3'}}
-        """)
-        
-        # Create table for vector embeddings if it doesn't exist
-        session.execute(f"""
-            CREATE TABLE IF NOT EXISTS {astra_db_keyspace}.product_embeddings (
-                id text PRIMARY KEY,
-                product_type text,
-                content text,
-                embedding_vector list<float>,
-                metadata text
-            )
-        """)
-        
         # Get or create collections
         product_embeddings = db.get_collection("product_embeddings")
         query_analytics = db.get_collection("query_analytics")
         product_images = db.get_collection("product_images")
         
-        # Create vector index for product_embeddings if it doesn't exist
-        try:
-            # Check if vector search is already set up
-            vector_indexes = product_embeddings.get_vector_indexes()
-            has_vector_index = any(index.get("name") == "embedding_vector_index" for index in vector_indexes)
-            
-            if not has_vector_index:
-                # Create vector index
-                product_embeddings.create_vector_index(
-                    vector_field="$vector", 
-                    dimension=1536,  # Adjust dimension based on your embedding model
-                    index_name="embedding_vector_index"
-                )
-                print("Vector index created for product_embeddings collection")
-        except Exception as e:
-            print(f"Warning: Could not create vector index: {e}")
-        
         print(f"Connected to Astra DB: {db.list_collection_names()}")
         
         # Return DB client and collections for use in application
@@ -681,7 +648,14 @@ def setup_and_update():
     
     # Initialize database and other services
     global astra_session, astra_keyspace, s3_client, embeddings_model
-    astra_session, astra_keyspace = init_astra_db()
+    astra_db_result = init_astra_db()
+    if astra_db_result:
+        astra_session = astra_db_result.get("db")
+        astra_keyspace = astra_db_result.get("keyspace")
+    else:
+        astra_session = None
+        astra_keyspace = None
+    
     s3_client = init_s3_client()
     embeddings_model = get_embeddings_model()