Update app.py

app.py

@@ -1,23 +1,37 @@
 import gradio as gr
-from smolagents import CodeAgent,HfApiModel  # adjust the import to your actual smolagents module
+from smolagents import CodeAgent, HfApiModel  # adjust the import to your actual smolagents module
 
-# Step 1: Set up your smolagents agent
+# Step 1: Set up your smolagents agent.
 def create_agent():
     """
     Initialize and return the agent.
     Adjust parameters like model type or configuration as needed.
     """
-    # For example, we initialize an Agent with a sample model
-    agent = CodeAgent(tools=[], model=HfApiModel(model_id='https://pflgm2locj2t89co.us-east-1.aws.endpoints.huggingface.cloud/')) # Change arguments as per your agent configuration
+    # For example, we initialize a CodeAgent with a sample model.
+    agent = CodeAgent(
+        tools=[], 
+        model=HfApiModel(model_id='https://pflgm2locj2t89co.us-east-1.aws.endpoints.huggingface.cloud/')
+    )
     return agent
 
-# Create the agent instance once so that it persists across user interactions
+# Create the agent instance once so that it persists across user interactions.
 agent = create_agent()
 
+def combine_messages(messages: list[dict]) -> str:
+    """
+    Helper function to combine a list of message dictionaries into a single string.
+    Each message is prefixed with its role.
+    """
+    conversation = ""
+    for msg in messages:
+        # Capitalize the role (e.g., 'User' instead of 'user') for clarity.
+        conversation += f"{msg['role'].capitalize()}: {msg['content']}\n"
+    return conversation.strip()
+
 def respond(message, history: list[tuple[str, str]], system_message, max_tokens, temperature, top_p):
     """
-    This function builds the conversation history, calls the smolagents agent,
-    and streams the response back to Gradio.
+    Build the conversation history, combine messages into a single string prompt,
+    call the smolagents agent, and stream the response back to Gradio.
     """
     # Build the conversation messages list, starting with the system prompt.
     messages = [{"role": "system", "content": system_message}]
@@ -29,24 +43,11 @@ def respond(message, history: list[tuple[str, str]], system_message, max_tokens,
     # Add the latest user input.
     messages.append({"role": "user", "content": message})
     
-    # Initialize an empty response.
-    complete_response = ""
+    # Combine the list of messages into a single string prompt.
+    prompt = combine_messages(messages)
     
-    # Step 2: Call the agent's chat_completion method.
-    # If your smolagents agent supports streaming (i.e. yielding tokens as they are generated),
-    # this loop will yield partial responses to update the UI.
-    # If streaming is not supported, you can simply do:
-    #   complete_response = agent.chat_completion(messages, max_tokens, temperature, top_p)
-    #   yield complete_response
-    # for token in agent.chat_completion(
-    #     messages,
-    #     max_tokens=max_tokens,
-    #     temperature=temperature,
-    #     top_p=top_p,
-    #     stream=True  # set to False if your agent does not support streaming
-    # ):
-    #     complete_response += token
-    complete_response=agent.run(messages)
+    # Now call the agent with the prompt.
+    complete_response = agent.run(prompt)
     yield complete_response
 
 # Step 3: Create the Gradio ChatInterface.