Update qdrant_client.py

backend/qdrant_client.py

@@ -10,96 +10,92 @@ load_dotenv()
 # === Environment Variables ===
 QDRANT_HOST = os.getenv(
     "QDRANT_HOST",
-    "http://localhost:6333" # Default to local if not set, or your cloud URL
+    "https://9485db48-8672-469a-a917-41a4ebbfd533.us-east4-0.gcp.cloud.qdrant.io"  # Your cloud URL
 )
-QDRANT_API_KEY = os.getenv("QDRANT_API_KEY") # Only needed for cloud Qdrant
+QDRANT_API_KEY = os.getenv("QDRANT_API_KEY")  # Only needed for cloud Qdrant
 
 # === Collection Names ===
-KB_COLLECTION = "rag_collection" # For document embeddings
-CHAT_HISTORY_COLLECTION = "chat_history_collection" # For chat messages
+KB_COLLECTION = "rag_collection"  # For document embeddings
+CHAT_HISTORY_COLLECTION = "chat_history_collection"  # For chat messages
 
 # === Qdrant Client Initialization ===
 qdrant_client = QdrantClient(
     url=QDRANT_HOST,
     api_key=QDRANT_API_KEY,
-    prefer_grpc=False, # Set to True for gRPC, False for REST
-    timeout=30, # Increased timeout for potentially slow operations
-    # check_compatibility=False, # Uncomment if you face compatibility issues with Qdrant versions
+    prefer_grpc=False,  # Set to True for gRPC, False for REST
+    timeout=30,  # Increased timeout for potentially slow operations
+    check_compatibility=False,  # Skip version check to avoid warnings
 )
 
-print("✅ Connected to Qdrant via REST")
+print("✅ Connected to Qdrant Cloud")
 
 # === Collection Creation and Management ===
 def ensure_collection_exists(name: str, vector_size: int = 384):
     """
-    Ensures a Qdrant collection exists. If not, it creates it.
+    Guarantees that a collection exists; creates it only if it is missing.
+    Uses create_collection (non-destructive) so no delete permission is needed.
     """
     try:
-        # Check if collection exists
-        info = qdrant_client.get_collection(collection_name=name)
-        if info.status == CollectionStatus.GREEN:
-            print(f"✅ Collection exists and is ready: {name}")
-            return
+        qdrant_client.get_collection(collection_name=name)
+        print(f"✅ Collection '{name}' already exists.")
     except Exception:
-        # Collection does not exist, so proceed to create
-        pass
-
-    print(f"🆕 Creating collection: {name}")
-    qdrant_client.recreate_collection(
-        collection_name=name,
-        vectors_config=VectorParams(size=vector_size, distance=Distance.COSINE) # Cosine similarity for embeddings
-    )
-    print(f"🎉 Collection '{name}' created successfully!")
-
+        print(f"🆕 Creating collection: {name}")
+        try:
+            qdrant_client.create_collection(
+                collection_name=name,
+                vectors_config=VectorParams(
+                    size=vector_size,
+                    distance=Distance.COSINE,
+                ),
+            )
+            print(f"🎉 Collection '{name}' created successfully!")
+        except Exception as e:
+            print(f"❌ Failed to create collection '{name}': {e}")
 
 # === Payload Indexing Helper ===
 def create_index_if_needed(collection: str, field_name: str, schema_type: str):
     """
-    Creates a payload index on a specified field within a collection if it doesn't already exist.
-    Payload indexes speed up filtering operations (e.g., by session_id).
+    Creates a payload index on a field if it is not present already.
     """
     try:
         schema_enum = getattr(PayloadSchemaType, schema_type.upper())
         qdrant_client.create_payload_index(
             collection_name=collection,
             field_name=field_name,
-            field_schema=schema_enum
+            field_schema=schema_enum,
         )
         print(f"🔧 Indexed '{field_name}' as {schema_type} in '{collection}'")
     except Exception as e:
-        # This error often means the index already exists, which is fine.
-        if "already exists" in str(e): # More specific check
-            # print(f"ℹ️ Index for '{field_name}' in '{collection}' already exists.")
-            pass # Suppress common "already exists" error
+        if "already exists" in str(e):
+            pass  # Index is already there – ignore
         else:
-            print(f"⚠️ Failed to create index for '{field_name}' in '{collection}': {e}")
-
+            print(f"⚠️ Could not create index '{field_name}' on '{collection}': {e}")
 
 # === Data Cleanup Utility (for development/testing) ===
 def clean_collections():
     """
-    Deletes all data from both the RAG knowledge base and chat history collections.
-    USE WITH CAUTION: This will wipe your data!
+    Deletes ALL points from both collections.
+    Call it manually; do NOT run automatically in production.
     """
     print("🧹 Cleaning old data from all collections...")
-
+    
     # Selector to delete all points in a collection (empty Filter() means no specific filter)
     all_points_selector = FilterSelector(filter=Filter())
-
+    
     try:
         qdrant_client.delete(
             collection_name=KB_COLLECTION,
             points_selector=all_points_selector
         )
         print(f"🗑️ All data cleaned from '{KB_COLLECTION}'.")
-
+        
         qdrant_client.delete(
             collection_name=CHAT_HISTORY_COLLECTION,
             points_selector=all_points_selector
         )
         print(f"🗑️ All data cleaned from '{CHAT_HISTORY_COLLECTION}'.")
         print("🗑️ All old data cleaned from collections successfully.")
-
+        
     except Exception as e:
         print(f"❌ Error during collection cleanup: {e}")
 
@@ -111,15 +107,16 @@ ensure_collection_exists(CHAT_HISTORY_COLLECTION)
 # Create indexes for filtering and ordering chat history
 create_index_if_needed(CHAT_HISTORY_COLLECTION, "session_id", "keyword")
 create_index_if_needed(CHAT_HISTORY_COLLECTION, "turn_number", "integer")
-create_index_if_needed(CHAT_HISTORY_COLLECTION, "timestamp", "keyword") # Useful for sorting/filtering by time
+create_index_if_needed(CHAT_HISTORY_COLLECTION, "timestamp", "keyword")  # Useful for sorting/filtering by time
 
 # Create indexes for filtering knowledge base documents
 create_index_if_needed(KB_COLLECTION, "session_id", "keyword")
 create_index_if_needed(KB_COLLECTION, "upload_timestamp", "keyword")
 create_index_if_needed(KB_COLLECTION, "file_type", "keyword")
-create_index_if_needed(KB_COLLECTION, "source", "keyword") # Index source if you use it for filtering
+create_index_if_needed(KB_COLLECTION, "source", "keyword")  # Index source if you use it for filtering
 
-# === IMPORTANT: Wipe previous data (OPTIONAL - COMMENT OUT AFTER FIRST RUN!) ===
+# === IMPORTANT: Data wipe is now commented out ===
 # This line will wipe all your data from Qdrant EVERY TIME the backend starts.
 # It's useful for initial setup and debugging, but comment it out for persistence.
-clean_collections()
+# Uncomment ONLY when you need to clear data manually:
+# clean_collections()