Update app.py

app.py

@@ -19,24 +19,27 @@ model = ChatOpenAI(
 # ✅ Initialize HuggingFace Embeddings
 embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
 
-# ✅ Initialize Chroma Vector Store (No need for chromadb.PersistentClient)
+# ✅ Initialize Chroma Vector Store
 vector_store = Chroma(
     collection_name="chat_collection",  # Specify the collection name
     embedding_function=embeddings,
-    persist_directory="./chroma_langchain_db",  # Directory to store data locally
+    persist_directory="./chroma_db",  # Directory to store data locally
 )
 
 # ✅ Step 1: Helper Functions for Chat Memory
 def get_chat_history(user_id):
     """Fetches stored messages for a given user from the vector store."""
-    # Use Chroma's retriever to fetch documents filtered by user_id
     retriever = vector_store.as_retriever(
-        search_type="mmr",  # Modify search type as needed (e.g., "mmr" for Maximal Marginal Relevance)
-        search_kwargs={"k": 100, "fetch_k": 100}  # Adjust the number of results to fetch and how many to filter
+        search_type="mmr",
+        search_kwargs={"k": 100, "fetch_k": 100}
     )
     
-    # Use the retriever to fetch results, filtered by the user_id
+    # Ensure filter is applied to user_id correctly
     results = retriever.invoke("Chat history", filter={"user_id": user_id})
+    
+    if not results:  # If no results, return empty string
+        return ""
+    
     # Extract the page content (chat messages) from the results
     user_history = [doc.page_content for doc in results]
     return "\n".join(user_history) if user_history else ""
@@ -52,24 +55,27 @@ def generate_response(username, user_input):
     """Generates a chatbot response using GPT-4 and stores chat history."""
     user_id = username.lower().strip()
     history = get_chat_history(user_id)
-    # Append the latest user input to the conversation history
     messages = [{"role": "system", "content": "You are a helpful AI assistant. Please provide answer in 20 words only"}]
+    
     if history:
         messages.append({"role": "user", "content": f"Chat history:\n{history}"})
-    # Include the latest user input in the conversation
+    
     messages.append({"role": "user", "content": user_input})
-    print(messages)
-    response = model.invoke(messages)  # Using the model to generate response
+    
+    # Generate the response
+    response = model.invoke(messages)
     bot_response = response.content
+    
     # Store the conversation for future reference
     store_chat_message(user_id, user_input, bot_response)
+    
     # Return the entire conversation including the user's input and bot's response
     return f"{history}\nUser: {user_input}\nBot: {bot_response}"
 
 # ✅ Step 3: Gradio UI with User Dropdown
 with gr.Blocks() as demo:
     gr.Markdown("# 🔥 Multi-User Chatbot with GPT-4 and Memory (ChromaDB)")
-
+    
     # Dropdown for selecting user
     username_input = gr.Dropdown(
         label="Select User",