app.py

import gradio as gr
from huggingface_hub import InferenceClient
from dotenv import load_dotenv
import os

# carregar variáveis de ambiente
load_dotenv()
hf_token = os.getenv("TOKEN")

def respond(message, history, system_message, max_tokens, temperature, top_p):
    # criar cliente HF
    client = InferenceClient(
        model="lxcorp/WNL468M",
        token=hf_token
    )

    # mensagens no formato esperado
    messages = [{"role": "system", "content": system_message}]
    messages.extend([{"role": "user", "content": turn[0]} for turn in history])
    messages.append({"role": "user", "content": message})

    response = ""

    # streaming da resposta
    for msg in client.chat_completion(
        messages=messages,
        max_tokens=max_tokens,
        stream=True,
        temperature=temperature,
        top_p=top_p,
    ):
        if msg.choices and msg.choices[0].delta.get("content"):
            token = msg.choices[0].delta["content"]
            response += token
            yield response


# interface
chatbot = gr.ChatInterface(
    respond,
    type="messages",
    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"),
    ],
)

# com sidebar (opcional)
with gr.Blocks() as demo:
    with gr.Sidebar():
        gr.Markdown("### Hugging Face Chatbot")
    chatbot.render()

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