diagnose_ollama.py

#!/usr/bin/env python3
"""
Diagnostic script to test Ollama connectivity
"""

import requests
from utils.config import config

def test_ollama_connectivity(custom_ngrok_url=None):
    """Test if Ollama is reachable from the current environment"""
    
    test_url = custom_ngrok_url if custom_ngrok_url else config.ollama_host
    
    print("=== Ollama Connectivity Diagnostic ===")
    print(f"Testing Ollama Host: {test_url}")
    print(f"Configured Model: {config.local_model_name}")
    print()
    
    # Headers to skip ngrok browser warning
    headers = {
        "ngrok-skip-browser-warning": "true",
        "User-Agent": "AI-Life-Coach-Diagnostic"
    }
    
    # Test 1: Check if we can reach the Ollama host
    print("Test 1: Checking Ollama host connectivity...")
    try:
        response = requests.get(
            f"{test_url}/api/tags", 
            headers=headers, 
            timeout=10
        )
        print(f"  Status Code: {response.status_code}")
        
        if response.status_code == 200:
            print("  ✓ Successfully connected to Ollama host")
            try:
                data = response.json()
                models = data.get("models", [])
                print(f"  Available Models: {len(models)}")
                for model in models:
                    print(f"    - {model.get('name', 'Unknown model')}")
                
                # Check if our configured model is available
                model_names = [m.get('name') for m in models]
                if config.local_model_name in model_names:
                    print(f"  ✓ Configured model '{config.local_model_name}' is available")
                else:
                    print(f"  ⚠ Configured model '{config.local_model_name}' not found")
                    if models:
                        print(f"    Available models: {', '.join(model_names)}")
                    else:
                        print(f"    No models found. Try running: ollama pull {config.local_model_name}")
                        
            except Exception as e:
                print(f"  Error parsing response: {e}")
                print(f"  Response text: {response.text[:200]}...")
        else:
            print(f"  ✗ Unexpected status code: {response.status_code}")
            print(f"  Response: {response.text[:200]}...")
            
    except requests.exceptions.Timeout:
        print("  ✗ Request timed out (took more than 10 seconds)")
        print("    This may indicate network issues or an unresponsive Ollama server")
    except requests.exceptions.ConnectionError as e:
        print(f"  ✗ Connection error: {e}")
        print("    This may indicate that the Ollama server is not running or the URL is incorrect")
    except Exception as e:
        print(f"  ✗ Unexpected error: {e}")
    
    print()
    
    # Test 2: Test model generation if we can connect
    print("Test 2: Testing model generation...")
    try:
        # First verify we can connect
        response = requests.get(
            f"{test_url}/api/tags", 
            headers=headers, 
            timeout=10
        )
        
        if response.status_code == 200:
            data = response.json()
            models = data.get("models", [])
            model_names = [m.get('name') for m in models]
            
            if config.local_model_name in model_names:
                print(f"  Testing generation with model: {config.local_model_name}")
                
                # Test generation
                generate_payload = {
                    "model": config.local_model_name,
                    "prompt": "Hello, please respond with just the word 'Success' in uppercase.",
                    "stream": False
                }
                
                generate_response = requests.post(
                    f"{test_url}/api/generate",
                    json=generate_payload,
                    headers=headers,
                    timeout=30
                )
                
                if generate_response.status_code == 200:
                    generate_data = generate_response.json()
                    response_text = generate_data.get("response", "").strip()
                    print(f"  ✓ Model generation successful")
                    print(f"  Response: {response_text}")
                else:
                    print(f"  ✗ Model generation failed with status {generate_response.status_code}")
                    print(f"  Response: {generate_response.text[:200]}...")
            else:
                print(f"  ⚠ Skipping generation test - model '{config.local_model_name}' not available")
                if not models:
                    print(f"    No models found. Try running: ollama pull {config.local_model_name}")
        else:
            print(f"  ⚠ Skipping generation test - cannot connect to Ollama host")
            
    except requests.exceptions.Timeout:
        print("  ✗ Generation test timed out (took more than 30 seconds)")
        print("    This may indicate the model is still loading or the server is under heavy load")
    except requests.exceptions.ConnectionError as e:
        print(f"  ✗ Generation test connection error: {e}")
    except Exception as e:
        print(f"  ✗ Generation test error: {e}")
    
    print()
    
    # Test 3: Check localhost as fallback
    print("Test 3: Checking localhost fallback (if different from configured host)...")
    if test_url != "http://localhost:11434":
        try:
            local_response = requests.get(
                "http://localhost:11434/api/tags", 
                headers=headers, 
                timeout=5
            )
            print(f"  Local Status Code: {local_response.status_code}")
            if local_response.status_code == 200:
                print("  ✓ Successfully connected to localhost Ollama")
                try:
                    local_data = local_response.json()
                    local_models = local_data.get("models", [])
                    print(f"  Local Available Models: {len(local_models)}")
                except Exception as e:
                    print(f"  Error parsing local response: {e}")
            else:
                print(f"  ✗ Local connection failed with status {local_response.status_code}")
        except Exception as e:
            print(f"  ✗ Local connection error: {e}")
    else:
        print("  Skipping (configured host is already localhost)")
    
    print()
    print("=== Diagnostic Complete ===")
    print()
    print("Troubleshooting Tips:")
    print("1. If connection fails, verify your Ollama server is running: ollama serve")
    print("2. If no models found, pull a model: ollama pull mistral")
    print("3. If ngrok issues, verify your tunnel is active and URL is correct")
    print("4. If timeout issues, check firewall settings and network connectivity")

if __name__ == "__main__":
    # Check if user provided a custom ngrok URL
    import sys
    custom_url = None
    if len(sys.argv) > 1:
        custom_url = sys.argv[1]
        print(f"Using custom ngrok URL: {custom_url}")
    
    test_ollama_connectivity(custom_url)