Update app.py

app.py

@@ -3,7 +3,7 @@ import time
 import requests
 from streamlit.components.v1 import html
 import os
-import ollama
+
 
 # Import transformers and cache the help agent for performance
 @st.cache_resource
@@ -233,41 +233,22 @@ def ask_llama(conversation_history, category, is_final_guess=False):
         return "Could not generate question"
 
 # New function for the help AI assistant using the Hugging Face InferenceClient
+
 def ask_help_agent(query):
     try:
-        import ollama
-        
-        # System message setup
-        system_message = "You are a friendly Chatbot."
-        
-        # Build conversation history in Ollama format
-        messages = [{"role": "system", "content": system_message}]
+        API_URL = "https://api-inference.huggingface.co/models/mistralai/Mistral-7B-Instruct-v0.1"
+        headers = {"Authorization": "Bearer HF_TOKEN"}  # Get from huggingface.co/settings/tokens
         
-        if "help_conversation" in st.session_state:
-            for msg in st.session_state.help_conversation:
-                if msg.get("query"):
-                    messages.append({"role": "user", "content": msg["query"]})
-                if msg.get("response"):
-                    messages.append({"role": "assistant", "content": msg["response"]})
-        
-        # Add current query
-        messages.append({"role": "user", "content": query})
-        
-        # Get response from Ollama
-        response = ollama.chat(
-            model="llama3",  # Can also use "mistral", "phi3", etc.
-            messages=messages,
-            options={
-                "temperature": 0.7,
-                "top_p": 0.95,
-                "num_predict": 150  # Similar to max_tokens
-            }
-        )
-        
-        return response['message']['content']
+        prompt = f"""<s>[INST] <<SYS>>
+        You are a helpful assistant. Answer concisely.
+        <</SYS>>
+        {query} [/INST]"""
         
+        response = requests.post(API_URL, headers=headers, json={"inputs": prompt})
+        return response.json()[0]['generated_text'].split("[/INST]")[-1].strip()
     except Exception as e:
-        return f"Error in help agent: {str(e)}"
+        return f"Assistant unavailable: {str(e)}"
+
 # Main game logic with enhanced UI
 def main():
     inject_custom_css()