Fix model name mismatch and update Ollama connection logic

.env

@@ -1,6 +1,7 @@
 # Hugging Face Settings
 HF_TOKEN=your_huggingface_token_here
 HF_API_ENDPOINT_URL=https://api-inference.huggingface.co/v1/
+USE_FALLBACK=false
 
 # API Keys
 TAVILY_API_KEY=your_tavily_api_key_here
@@ -12,7 +13,9 @@ REDIS_HOST=localhost
 REDIS_PORT=6379
 REDIS_USERNAME=
 REDIS_PASSWORD=
+REDIS_RETRIES=3
+REDIS_RETRY_DELAY=1
 
-# Model Configuration
-LOCAL_MODEL_NAME=mistral-7b
+# Model Configuration - Use the exact model name from Ollama
+LOCAL_MODEL_NAME=mistral:latest
 OLLAMA_HOST=https://ace32bd59aef.ngrok-free.app

app.py

@@ -1,4 +1,4 @@
-# Force redeploy trigger - version 1.7
+# Force redeploy trigger - version 1.8
 import streamlit as st
 from utils.config import config
 import requests
@@ -9,7 +9,7 @@ from core.memory import load_user_state, check_redis_health
 # Set page config
 st.set_page_config(page_title="AI Life Coach", page_icon="🧘", layout="centered")
 
-# Initialize session state for ngrok URL
+# Initialize session state
 if 'ngrok_url' not in st.session_state:
     st.session_state.ngrok_url = config.ollama_host
 
@@ -37,36 +37,39 @@ if st.sidebar.button("Update Ngrok URL"):
     st.sidebar.success("Ngrok URL updated!")
     st.experimental_rerun()
 
-# Model selection
-st.sidebar.markdown("---")
-st.sidebar.subheader("Model Selection")
-
 # Headers to skip ngrok browser warning
 NGROK_HEADERS = {
     "ngrok-skip-browser-warning": "true",
     "User-Agent": "AI-Life-Coach-App"
 }
 
-# Fetch available models when we have a valid connection
-if st.session_state.ngrok_url and st.session_state.model_status != "unreachable":
+# Fetch available models
+def fetch_available_models(ngrok_url):
     try:
         response = requests.get(
-            f"{st.session_state.ngrok_url}/api/tags", 
+            f"{ngrok_url}/api/tags", 
             headers=NGROK_HEADERS,
             timeout=5
         )
         if response.status_code == 200:
             models_data = response.json().get("models", [])
-            model_names = [m.get("name") for m in models_data]
-            if model_names:
-                st.session_state.available_models = model_names
-                # If current selected model not in list, select the first one
-                if st.session_state.selected_model not in model_names:
-                    st.session_state.selected_model = model_names[0]
-    except Exception as e:
-        pass  # Silently fail, we'll handle this in the main logic
+            return [m.get("name") for m in models_data]
+    except Exception:
+        pass
+    return []
+
+# Update available models
+if st.session_state.ngrok_url and st.session_state.model_status != "unreachable":
+    model_names = fetch_available_models(st.session_state.ngrok_url)
+    if model_names:
+        st.session_state.available_models = model_names
+        # If current selected model not in list, select the first one
+        if st.session_state.selected_model not in model_names:
+            st.session_state.selected_model = model_names[0]
 
 # Model selector dropdown
+st.sidebar.markdown("---")
+st.sidebar.subheader("Model Selection")
 if st.session_state.available_models:
     selected_model = st.sidebar.selectbox(
         "Select Model", 
@@ -84,13 +87,12 @@ else:
 st.sidebar.markdown("---")
 
 # Get environment info
-BASE_URL = os.environ.get("SPACE_ID", "")  # Will be set in HF Spaces
+BASE_URL = os.environ.get("SPACE_ID", "")
 IS_HF_SPACE = bool(BASE_URL)
 
 # Fetch Ollama status
 def get_ollama_status(ngrok_url):
     try:
-        # Try to connect to the remote Ollama service directly
         response = requests.get(
             f"{ngrok_url}/api/tags", 
             headers=NGROK_HEADERS,
@@ -102,7 +104,6 @@ def get_ollama_status(ngrok_url):
             st.session_state.available_models = model_names
             
             if models:
-                # Check if our selected model is available
                 selected_model_available = st.session_state.selected_model in model_names
                 return {
                     "running": True, 
@@ -121,7 +122,6 @@ def get_ollama_status(ngrok_url):
                 }
     except Exception as e:
         st.session_state.model_status = "unreachable"
-        # If direct connection fails, return error info
         return {
             "running": False, 
             "model_loaded": None,
@@ -129,7 +129,7 @@ def get_ollama_status(ngrok_url):
             "remote_host": ngrok_url
         }
 
-# After user selects name, load conversation history
+# Load conversation history
 def get_conversation_history(user_id):
     try:
         user_state = load_user_state(user_id)
@@ -139,7 +139,7 @@ def get_conversation_history(user_id):
         st.warning(f"Could not load conversation history: {e}")
     return []
 
-# Check Ollama status with the current ngrok URL
+# Check Ollama status
 ollama_status = get_ollama_status(st.session_state.ngrok_url)
 
 # Update model status
@@ -166,7 +166,7 @@ else:
         st.sidebar.warning(f"🧠 Ollama Model: {model_status_msg} (selected model not available)")
     st.sidebar.info(f"Connected to: {ollama_status['remote_host']}")
 
-# Model status indicator
+# Status indicators
 model_status_container = st.sidebar.empty()
 if st.session_state.model_status == "ready":
     model_status_container.success("✅ Model Ready")
@@ -174,10 +174,9 @@ elif st.session_state.model_status == "checking":
     model_status_container.info("🔍 Checking model...")
 elif st.session_state.model_status == "no_models":
     model_status_container.warning("⚠️ No models found")
-else:  # unreachable
+else:
     model_status_container.error("❌ Ollama unreachable")
 
-# Redis status indicator
 redis_status_container = st.sidebar.empty()
 if check_redis_health():
     redis_status_container.success("✅ Redis Connected")
@@ -204,10 +203,15 @@ with st.expander("🔍 Connection Status"):
 # Function to send message to Ollama
 def send_to_ollama(user_input, conversation_history, ngrok_url, model_name):
     try:
+        # Use the correct chat endpoint with proper payload
         payload = {
             "model": model_name,
             "messages": conversation_history,
-            "stream": False
+            "stream": False,
+            "options": {
+                "temperature": 0.7,
+                "top_p": 0.9
+            }
         }
         
         response = requests.post(
@@ -231,23 +235,16 @@ def send_to_ollama(user_input, conversation_history, ngrok_url, model_name):
 # Function to send message to Hugging Face (fallback)
 def send_to_hf(user_input, conversation_history):
     try:
-        # Import here to avoid issues if not needed
         from core.llm import LLMClient
-        
-        # Initialize LLM client for Hugging Face
         llm_client = LLMClient(provider="huggingface")
         
-        # Format prompt for HF
-        prompt = ""
+        # Format for HF
+        prompt = "You are a helpful life coach. "
         for msg in conversation_history:
-            role = msg["role"]
-            content = msg["content"]
-            if role == "system":
-                prompt += f"System: {content}\n"
-            elif role == "user":
-                prompt += f"Human: {content}\n"
-            elif role == "assistant":
-                prompt += f"Assistant: {content}\n"
+            if msg["role"] == "user":
+                prompt += f"Human: {msg['content']} "
+            elif msg["role"] == "assistant":
+                prompt += f"Assistant: {msg['content']} "
         prompt += "Assistant:"
         
         response = llm_client.generate(prompt, max_tokens=500, stream=False)
@@ -274,7 +271,7 @@ if st.button("Send"):
 
         # Prepare conversation history
         conversation_history = [{"role": msg["role"], "content": msg["content"]} 
-                              for msg in conversation[-5:]]  # Last 5 messages
+                              for msg in conversation[-5:]]
         conversation_history.append({"role": "user", "content": user_input})
         
         # Send to appropriate backend
@@ -293,6 +290,5 @@ if st.button("Send"):
             
             if ai_response:
                 st.markdown(f"**AI Coach ({backend_used}):** {ai_response}")
-                # Note: In a production app, we'd save the conversation to Redis here
             else:
                 st.error(f"Failed to get response from {backend_used}.")

test_ollama_connection.py

@@ -0,0 +1,66 @@
+import requests
+import os
+from dotenv import load_dotenv
+
+# Load environment variables
+load_dotenv()
+
+OLLAMA_HOST = os.getenv("OLLAMA_HOST", "https://ace32bd59aef.ngrok-free.app")
+MODEL_NAME = os.getenv("LOCAL_MODEL_NAME", "mistral:latest")
+
+print(f"Testing Ollama connection to: {OLLAMA_HOST}")
+print(f"Using model: {MODEL_NAME}")
+print()
+
+# Headers to skip ngrok browser warning
+headers = {
+    "ngrok-skip-browser-warning": "true",
+    "User-Agent": "AI-Life-Coach-Test"
+}
+
+# Test 1: List models
+print("Test 1: Listing available models...")
+try:
+    response = requests.get(f"{OLLAMA_HOST}/api/tags", headers=headers, timeout=10)
+    print(f"Status Code: {response.status_code}")
+    
+    if response.status_code == 200:
+        data = response.json()
+        models = data.get("models", [])
+        print(f"Found {len(models)} models:")
+        for model in models:
+            print(f"  - {model['name']} ({model.get('size', 'Unknown size')})")
+    else:
+        print(f"Error: {response.text}")
+except Exception as e:
+    print(f"Connection failed: {e}")
+
+print()
+
+# Test 2: Simple chat test
+print("Test 2: Simple chat test...")
+try:
+    payload = {
+        "model": MODEL_NAME,
+        "messages": [
+            {"role": "user", "content": "Hello! Respond with just 'Hi there!'"}
+        ],
+        "stream": False
+    }
+    
+    response = requests.post(f"{OLLAMA_HOST}/api/chat", headers=headers, json=payload, timeout=30)
+    print(f"Status Code: {response.status_code}")
+    
+    if response.status_code == 200:
+        data = response.json()
+        message = data.get("message", {})
+        content = message.get("content", "")
+        print(f"Response: {content}")
+        print("✅ Chat test successful!")
+    else:
+        print(f"Error: {response.text}")
+except Exception as e:
+    print(f"Chat test failed: {e}")
+
+print()
+print("Test completed.")