update

app.py

@@ -15,7 +15,7 @@ import gradio as gr
 sys.path.append(os.path.dirname(__file__))
 
 # Import the main application components
-from demo import create_interface
+from demo import create_interface, initialize_nvidia_client, test_nvidia_api_connection
 
 # Set up environment variables for Hugging Face Spaces
 def setup_environment():
@@ -25,12 +25,15 @@ def setup_environment():
         print("🚀 Running on Hugging Face Spaces")
         
         # Check for NVIDIA API key
-        if not os.getenv("NVIDIA_API_KEY"):
+        api_key = os.getenv("NVIDIA_API_KEY")
+        if not api_key:
             print("⚠️  NVIDIA_API_KEY not set in Space secrets.")
             print("   Please set NVIDIA_API_KEY in the Space Repository Secrets.")
             print("   Some features may be limited without API access.")
         else:
             print("✅ NVIDIA_API_KEY found in Space secrets")
+            print(f"   Key length: {len(api_key)} characters")
+            print(f"   Key preview: {api_key[:10]}...{api_key[-4:] if len(api_key) > 14 else '***'}")
         
         # Set CUDA device for Spaces (usually limited resources)
         os.environ["CUDA_VISIBLE_DEVICES"] = "0"
@@ -41,15 +44,32 @@ def setup_environment():
     else:
         print("🏠 Running locally")
         # Check for local .env file
-        if not os.getenv("NVIDIA_API_KEY"):
+        api_key = os.getenv("NVIDIA_API_KEY")
+        if not api_key:
             print("ℹ️  NVIDIA_API_KEY not found. For local development, create a .env file")
             print("   or set the environment variable manually.")
+        else:
+            print(f"✅ NVIDIA_API_KEY found locally (length: {len(api_key)})")
 
 def main():
     """Main function to launch the application"""
     # Set up environment
     setup_environment()
     
+    # Reinitialize NVIDIA client after environment setup
+    print("🔄 Reinitializing NVIDIA API client...")
+    from demo import client
+    if client is None and os.getenv("NVIDIA_API_KEY"):
+        print("🔄 Attempting to reinitialize client with environment variables...")
+        # Reinitialize the client
+        import demo
+        demo.client = initialize_nvidia_client()
+    
+    # Test API connection
+    if os.getenv("NVIDIA_API_KEY"):
+        print("🧪 Testing NVIDIA API connection...")
+        test_nvidia_api_connection()
+    
     # Create the Gradio interface
     print("🎯 Creating RoutePilot interface...")
     demo = create_interface()

demo.py

@@ -97,17 +97,51 @@ if torch.cuda.is_available():
     print(f"CUDA device name: {torch.cuda.get_device_name(0)}")
 
 # Initialize OpenAI client for NVIDIA API
-if os.getenv("NVIDIA_API_KEY") is None:
-    print("❌ NVIDIA API key not found. Please create a .env file with your API key")
-    client = None
-else:
-    client = OpenAI(
-        base_url=NVIDIA_BASE_URL,
-        api_key=os.getenv("NVIDIA_API_KEY"),
-        timeout=60,
-        max_retries=2
-    )
-    print("✅ NVIDIA API client initialized successfully")
+def initialize_nvidia_client():
+    """Initialize the NVIDIA API client with proper error handling"""
+    api_key = os.getenv("NVIDIA_API_KEY")
+    if api_key is None:
+        print("❌ NVIDIA API key not found. Please create a .env file with your API key")
+        print("   For Hugging Face Spaces: Set NVIDIA_API_KEY in Repository Secrets")
+        return None
+    else:
+        try:
+            client = OpenAI(
+                base_url=NVIDIA_BASE_URL,
+                api_key=api_key,
+                timeout=60,
+                max_retries=2
+            )
+            print("✅ NVIDIA API client initialized successfully")
+            return client
+        except Exception as e:
+            print(f"❌ Failed to initialize NVIDIA API client: {e}")
+            return None
+
+# Initialize the client
+client = initialize_nvidia_client()
+
+def test_nvidia_api_connection():
+    """Test the NVIDIA API connection to verify authentication"""
+    if client is None:
+        print("❌ Cannot test API connection - client not initialized")
+        return False
+    
+    try:
+        print("🧪 Testing NVIDIA API connection...")
+        # Make a simple test call
+        test_response = client.chat.completions.create(
+            model="meta/llama-3.1-8b-instruct",
+            messages=[{"role": "user", "content": "Hello"}],
+            max_tokens=10,
+            temperature=0.0,
+            stream=False
+        )
+        print("✅ NVIDIA API connection test successful")
+        return True
+    except Exception as e:
+        print(f"❌ NVIDIA API connection test failed: {e}")
+        return False
 
 def model_prompting(
     llm_model: str,
@@ -134,7 +168,15 @@ def model_prompting(
     if client is None:
         raise Exception("NVIDIA API client not initialized. Please check your .env file contains NVIDIA_API_KEY")
     
+    # Debug information
+    api_key = os.getenv("NVIDIA_API_KEY")
+    if api_key:
+        print(f"🔑 API Key available: {api_key[:10]}...{api_key[-4:] if len(api_key) > 14 else '***'}")
+    else:
+        print("❌ No API key found in environment")
+    
     try:
+        print(f"🚀 Making API call to model: {llm_model}")
         completion = client.chat.completions.create(
             model=llm_model,
             messages=[{"role": "user", "content": prompt}],
@@ -151,7 +193,19 @@ def model_prompting(
         
         return response_text
     except Exception as e:
-        raise Exception(f"API call failed: {str(e)}")
+        error_msg = str(e)
+        print(f"❌ API call failed: {error_msg}")
+        
+        # Provide more specific error information
+        if "401" in error_msg or "Unauthorized" in error_msg:
+            print("🔍 Authentication Error Details:")
+            print(f"   - API Key present: {'Yes' if api_key else 'No'}")
+            print(f"   - API Key length: {len(api_key) if api_key else 0}")
+            print(f"   - Base URL: {NVIDIA_BASE_URL}")
+            print("   - For Hugging Face Spaces: Check if NVIDIA_API_KEY is set in Repository Secrets")
+            print("   - For local development: Check if .env file contains NVIDIA_API_KEY")
+        
+        raise Exception(f"API call failed: {error_msg}")
 
 # Initialize the Longformer model for embeddings (same as enhance_query_with_templates.py)
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")