app.py

import gradio as gr
from huggingface_hub import InferenceClient
from transformers import AutoTokenizer, AutoModelForQuestionAnswering, pipeline
from sentence_transformers import SentenceTransformer, util

# Modelos
model_name = "deepset/roberta-base-squad2"
qa_pipeline = pipeline("question-answering", model=model_name, tokenizer=model_name)

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

# Modelo para comparação semântica (cosine similarity)
similarity_model = SentenceTransformer("all-MiniLM-L6-v2")


def get_qa_pipeline_answer(question, context):
    return qa_pipeline({"question": question, "context": context})["answer"]


def get_zephyr_answer(question, context):
    messages = [
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": f"Context: {context}\n\nQuestion: {question}"},
    ]
    response = client.chat_completion(
        messages,
        max_tokens=512,
        temperature=0.7,
        top_p=0.95,
    )
    return response.choices[0].message.content.strip()


def compare_answers(answer1, answer2):
    emb1 = similarity_model.encode(answer1, convert_to_tensor=True)
    emb2 = similarity_model.encode(answer2, convert_to_tensor=True)
    similarity = util.cos_sim(emb1, emb2).item()
    return round(similarity, 3)


def respond(question, context):
    answer1 = get_qa_pipeline_answer(question, context)
    answer2 = get_zephyr_answer(question, context)
    similarity_score = compare_answers(answer1, answer2)

    return (
        f"📘 Roberta-base-squad2:\n{answer1}\n\n"
        f"🧠 Zephyr-7b:\n{answer2}\n\n"
        f"🔍 Similaridade Semântica: **{similarity_score}**"
    )


# Interface Gradio
with gr.Blocks() as demo:
    gr.Markdown("# 🔎 Perguntas com dois modelos\nCompare duas respostas e veja a similaridade.")
    with gr.Row():
        question = gr.Textbox(label="Pergunta")
        context = gr.Textbox(label="Contexto")
    submit_btn = gr.Button("Obter Respostas")
    output = gr.Textbox(label="Respostas e Similaridade")

    submit_btn.click(respond, inputs=[question, context], outputs=output)

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