app.py

import gradio as gr
from huggingface_hub import InferenceClient

# Using Qwen 2.5 Coder
model_id = "Qwen/Qwen2.5-Coder-32B-Instruct"

def respond(message, history, system_message, temperature):
    # Initialize the client
    client = InferenceClient(model_id)
    
    # 1. Start with the System Message
    messages = [{"role": "system", "content": system_message}]
    
    # 2. Add the history
    for user_msg, bot_msg in history:
        messages.append({"role": "user", "content": user_msg})
        messages.append({"role": "assistant", "content": bot_msg})
    
    # 3. Add the current user message
    messages.append({"role": "user", "content": message})

    # Generate response
    response_text = ""
    try:
        stream = client.chat_completion(
            messages, 
            max_tokens=2048, 
            stream=True, 
            temperature=temperature,
            top_p=0.9
        )
        for chunk in stream:
            content = chunk.choices[0].delta.content
            if content:
                response_text += content
                yield response_text
    except Exception as e:
        yield f"Error: {str(e)}. The model might be busy."

# Build the UI
with gr.Blocks(fill_height=True) as demo:
    with gr.Sidebar():
        gr.Markdown("# AI Coding Assistant")
        gr.Markdown(f"Running **{model_id}**")
        gr.LoginButton("Sign in")
    
    gr.ChatInterface(
        respond,
        additional_inputs=[
            # System Message Input
            gr.Textbox(
                value="You are a helpful assistant.", 
                label="System Instruction",
                lines=2
            ),
            # Temperature Slider
            gr.Slider(
                minimum=0.1, 
                maximum=2.0, 
                value=0.7, 
                step=0.1, 
                label="Temperature"
            )
        ]
    )

demo.launch()