import gradio as gr
from huggingface_hub import InferenceClient

client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")

def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    messages = [{"role": "system", "content": system_message}] + history
    messages.append({"role": "user", "content": message})

    response = ""
    for part in client.chat_completion(
        messages, max_tokens=max_tokens, stream=True, temperature=temperature,
        top_p=top_p
    ):
        token = part.choices[0].delta.content
        if token:
            response += token

    history.append({"role": "user", "content": message})
    history.append({"role": "assistant", "content": response})

    return history, ""

with gr.Blocks() as demo:
    gr.Markdown("## Zephyr Chatbot Controls")

    role_dropdown = gr.Dropdown(choices=["SDE", "BA"], label="Select Role", value="SDE")

    system = gr.Textbox(value="You are a friendly chatbot.", label="System message")
    max_tokens = gr.Slider(1, 2048, value=512, label="Max tokens")
    temperature = gr.Slider(0.1, 4.0, value=0.7, label="Temperature", step=0.1)
    top_p = gr.Slider(0.1, 1.0, value=0.95, label="Top-p", step=0.05)

    with gr.Row():
        clear_btn = gr.Button("Clear Chat")
        dummy_btn = gr.Button("Dummy Action")

    clear_btn.click(lambda: gr.Info("Chat cleared!"))
    dummy_btn.click(lambda: gr.Info("Dummy action clicked!"))

if __name__ == "__main__":
    demo.launch()