Update app.py

app.py

@@ -154,23 +154,44 @@ def ask_llama(conversation_history, category, is_final_guess=False):
         st.error(f"Error calling Llama API: {str(e)}")
         return "Could not generate question"
 
-# New function for the help AI assistant using a free Hugging Face conversational model
-# (replacing the original BlenderBot approach with DialoGPT for simplicity)
+# New function for the help AI assistant using the Hugging Face InferenceClient
 def ask_help_agent(query):
     try:
-        from transformers import pipeline, Conversation
-        # Use the conversational pipeline with DialoGPT-medium for dialogue context
-        help_chat = pipeline("conversational", model="microsoft/DialoGPT-medium")
-        # Create a Conversation object to hold the context
-        conv = Conversation(query)
-        # Generate a response by passing the conversation object
-        response = help_chat(conv)
-        # Extract and return the generated response from the conversation
-        return conv.generated_responses[-1]
+        from huggingface_hub import InferenceClient
+        # Initialize the client with the provided model
+        client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+        system_message = "You are a friendly Chatbot."
+        
+        # Build history from session state (if any)
+        history = []
+        if "help_conversation" in st.session_state:
+            for msg in st.session_state.help_conversation:
+                # Each history entry is a tuple: (user query, assistant response)
+                history.append((msg.get("query", ""), msg.get("response", "")))
+        
+        messages = [{"role": "system", "content": system_message}]
+        for user_msg, bot_msg in history:
+            if user_msg:
+                messages.append({"role": "user", "content": user_msg})
+            if bot_msg:
+                messages.append({"role": "assistant", "content": bot_msg})
+        messages.append({"role": "user", "content": query})
+        
+        response_text = ""
+        # Using streaming to collect the entire response from the model
+        for message in client.chat_completion(
+            messages,
+            max_tokens=150,
+            stream=True,
+            temperature=0.7,
+            top_p=0.95,
+        ):
+            token = message.choices[0].delta.content
+            response_text += token
+        return response_text
     except Exception as e:
         return f"Error in help agent: {str(e)}"
 
-
 # Main game logic
 def main():
     inject_custom_css()
@@ -272,7 +293,7 @@ def main():
                     
                     st.experimental_rerun()
 
-        # Side Help Option: independent chat with an AI help assistant (free Hugging Face model)
+        # Side Help Option: independent chat with an AI help assistant using Hugging Face model
         with st.expander("Need Help? Chat with AI Assistant"):
             help_query = st.text_input("Enter your help query:", key="help_query")
             if st.button("Send", key="send_help"):