import os
import gradio as gr
from huggingface_hub import InferenceClient

MODEL_ID = "unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF"
token = os.environ.get("HF_TOKEN")
client = InferenceClient(model=MODEL_ID, token=token)

def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    # Monta o prompt tipo chat manualmente
    prompt = f"{system_message}\n\n"
    for user_msg, bot_msg in history:
        if user_msg:
            prompt += f"User: {user_msg}\n"
        if bot_msg:
            prompt += f"Assistant: {bot_msg}\n"
    prompt += f"User: {message}\nAssistant:"

    # Chama o endpoint de geração de texto normal, sem streaming
    response = client.text_generation(
        prompt,
        max_new_tokens=max_tokens,
        temperature=temperature,
        top_p=top_p,
    )
    # A resposta vem como string simples
    yield response

demo = gr.ChatInterface(
    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)"),
    ],
)

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