Update app.py

app.py

@@ -2,15 +2,18 @@ import gradio as gr
 from huggingface_hub import InferenceClient
 import traceback
 
-# Helper to extract text from Gradio 6's complex message format
+# --- Helper: Gradio 6.5 Text Extractor ---
 def get_text(content):
+    """Extracts text safely from Gradio 6's list-of-dict message format."""
     if isinstance(content, str):
         return content
     if isinstance(content, list):
-        # Extract text from multimodal blocks
         return "".join([block.get("text", "") for block in content if block.get("type") == "text"])
+    if isinstance(content, dict):
+        return content.get("text", str(content))
     return str(content)
 
+# --- Core Function: The Chat Logic ---
 def respond(
     message,
     history: list[dict],
@@ -20,52 +23,56 @@ def respond(
     top_p,
     hf_token: gr.OAuthToken, 
 ):
-    # 1. Check Token
+    # 1. Check Authentication
     if not hf_token or not hf_token.token:
         yield "⚠️ Please **Login** using the button in the sidebar to access the @frusto360 AI."
         return
 
     try:
-        client = InferenceClient(model="Frusto/llama-3.2-1b-frusto360-final", token=hf_token.token)
+        # 2. DIRECT ROUTING: Avoids the StopIteration error by bypassing the provider search
+        MODEL_ID = "Frusto/llama-3.2-1b-frusto360-final"
+        # We use the direct inference URL as the base_url
+        client = InferenceClient(
+            base_url=f"https://api-inference.huggingface.co/models/{MODEL_ID}",
+            token=hf_token.token
+        )
 
-        # 2. Build Llama 3.2 Prompt safely
+        # 3. Build Llama 3.2 Chat Template
         prompt = f"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\n{system_message}<|eot_id|>"
-        
         for msg in history:
             role = msg.get("role", "user")
             content = get_text(msg.get("content", ""))
             prompt += f"<|start_header_id|>{role}<|end_header_id|>\n\n{content}<|eot_id|>"
         
-        # Add current message (extract text if it's a dict)
-        curr_msg = get_text(message)
-        prompt += f"<|start_header_id|>user<|end_header_id|>\n\n{curr_msg}<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n"
+        prompt += f"<|start_header_id|>user<|end_header_id|>\n\n{get_text(message)}<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n"
 
-        # 3. Stream response
+        # 4. Stream the Response
         response = ""
-        # We wrap the generator to handle both string and object yields
-        stream = client.text_generation(
+        # Note: 'model' parameter is OMITTED because it's already in the base_url
+        for token in client.text_generation(
             prompt,
             max_new_tokens=max_tokens,
             stream=True,
             temperature=temperature,
             top_p=top_p,
             stop=["<|eot_id|>", "<|start_header_id|>"]
-        )
-
-        for token in stream:
-            # In some versions, token is an object, in others it's a string
+        ):
+            # Compatibility check for token format
             token_text = token if isinstance(token, str) else getattr(token, 'token', getattr(token, 'text', str(token)))
             response += token_text
             yield response
             
     except Exception as e:
-        # This will show you the ACTUAL error message now
-        error_details = traceback.format_exc()
-        print(error_details) # Check your "Logs" tab in HF Spaces
-        yield f"❌ **Error Details:**\n```\n{str(e)}\n```\n\n*Check the Space Logs for more info.*"
+        error_msg = str(e)
+        if "503" in error_msg:
+            yield "⏳ **Model is starting up.** Hugging Face is loading the weights. Please try again in 30 seconds!"
+        elif "404" in error_msg:
+            yield f"❌ **Error 404:** Model not found or Inference API disabled on the model page."
+        else:
+            yield f"❌ **Error:** {error_msg}\n\n*Check the Space Logs for details.*"
 
-# --- UI Setup ---
-chatbot = gr.ChatInterface(
+# --- Gradio UI Layout ---
+chatbot_interface = gr.ChatInterface(
     respond,
     additional_inputs=[
         gr.Textbox(value="You are a helpful assistant developed by @frusto360.", label="System message"),
@@ -75,14 +82,15 @@ chatbot = gr.ChatInterface(
     ],
 )
 
-with gr.Blocks() as demo:
+with gr.Blocks(fill_height=True) as demo:
     with gr.Sidebar():
-        gr.Markdown("### 🔐 Authentication")
+        gr.Markdown("## 🔐 @frusto360 Control")
         gr.LoginButton()
         gr.Markdown("---")
-        gr.Markdown("Developed by **@frusto360**")
+        gr.Markdown("Created by [@frusto360](https://youtube.com/@frusto360)")
     
-    chatbot.render()
+    chatbot_interface.render()
 
 if __name__ == "__main__":
-    demo.launch()
+    # Gradio 6.0+ prefers theme in launch()
+    demo.launch(theme="glass")