Update app.py

app.py

@@ -2,26 +2,33 @@ import gradio as gr
 import os
 from huggingface_hub import InferenceClient
 
+# Use Qwen 2.5 Coder
 model_id = "Qwen/Qwen2.5-Coder-32B-Instruct"
 
 def respond(message, history, system_message, temperature, request: gr.Request):
-    # 1. Try to get the user's token (if they logged in via the button)
     token = None
+    
+    # 1. Safely try to get the user's token (works on HF Spaces if logged in)
     if request:
-        token = request.token
+        # We use getattr() to avoid the "AttributeError" if running locally
+        token = getattr(request, "token", None)
     
     # 2. If no user token, fall back to the Space's secret token (HF_TOKEN)
     if token is None:
         token = os.getenv("HF_TOKEN")
 
+    # 3. If still no token, stop.
     if token is None:
-        yield "Error: No authentication token found. Please add 'HF_TOKEN' to Space Secrets."
+        yield "Error: No authentication token found. Please add 'HF_TOKEN' to Space Secrets or run locally with an env variable."
         return
 
-    # Initialize client with the found token
+    # Initialize client
     client = InferenceClient(model_id, token=token)
     
+    # Build the message format
     messages = [{"role": "system", "content": system_message}]
+    
+    # 'history' comes as tuples [(user, bot), (user, bot)...]
     for user_msg, bot_msg in history:
         messages.append({"role": "user", "content": user_msg})
         messages.append({"role": "assistant", "content": bot_msg})
@@ -49,7 +56,6 @@ with gr.Blocks(fill_height=True) as demo:
     with gr.Sidebar():
         gr.Markdown("# AI Coding Assistant")
         gr.Markdown(f"Running **{model_id}**")
-        # This button allows users to use their OWN token if they want
         gr.LoginButton("Sign in")
     
     gr.ChatInterface(