Update app.py

app.py

@@ -1,39 +1,61 @@
-# Import necessary libraries
 import gradio as gr
-from smolagents import CodeAgent,HfApiModel  # Replace with your actual import if different
+from smolagents import Agent  # adjust the import to your actual smolagents module
 
-# 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
-    return agent
-
-# Create the agent instance once so that it persists across user interactions
-agent = create_agent()
+# Step 1: Initialize your smolagents agent.
+# Replace "zephyr-7b-beta" with your desired model name/configuration.
+agent = Agent(model="zephyr-7b-beta")  
 
-# Step 2: Define a function that uses the agent to process user input
-def process_input(user_input):
+def respond(message, history: list[tuple[str, str]], system_message, max_tokens, temperature, top_p):
     """
-    This function receives user input from the Gradio interface,
-    processes it with the smolagents agent, and returns the agent's response.
+    This function builds the conversation history, calls the smolagents agent,
+    and streams the response back to Gradio.
     """
-    # Use your agent's method to generate a response; here we assume a 'run' method exists.
-    response = agent.run(user_input)
-    return response
+    # Build the conversation messages list, starting with the system prompt.
+    messages = [{"role": "system", "content": system_message}]
+    for user_msg, assistant_msg in history:
+        if user_msg:
+            messages.append({"role": "user", "content": user_msg})
+        if assistant_msg:
+            messages.append({"role": "assistant", "content": assistant_msg})
+    # Add the latest user input.
+    messages.append({"role": "user", "content": message})
+    
+    # Initialize an empty response.
+    complete_response = ""
+    
+    # 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
+        yield complete_response
 
-# Step 3: Build the Gradio interface
-iface = gr.Interface(
-    fn=process_input,  # This is the function Gradio will call when a user submits input.
-    inputs=gr.Textbox(lines=2, placeholder="Enter your prompt here..."),
-    outputs="text",
-    title="Smolagents Agent via Gradio",
-    description="This interface integrates a smolagents-based agent with Gradio to process text inputs."
+# Step 3: Create the Gradio ChatInterface.
+demo = gr.ChatInterface(
+    fn=respond,
+    additional_inputs=[
+        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
+        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
+        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
+        gr.Slider(
+            minimum=0.1,
+            maximum=1.0,
+            value=0.95,
+            step=0.05,
+            label="Top-p (nucleus sampling)"
+        ),
+    ],
 )
 
-# Step 4: Launch the Gradio app
+# Step 4: Launch the Gradio app.
 if __name__ == "__main__":
-    iface.launch()
+    demo.launch()