app.py

import gradio as gr
from inference import evo_chat_predict, get_gpt_response
from logger import log_feedback
import subprocess

# Global chat history
chat_history = []

# 🧠 Handle main chat logic
def chat_fn(user_input, option1, option2, user_vote=None):
    global chat_history

    # Validate
    if not user_input or not option1 or not option2:
        return "❗ Please enter your question and both options.", chat_history

    options = [option1.strip(), option2.strip()]

    # Evo prediction
    evo_result = evo_chat_predict(chat_history, user_input, options)

    # GPT fallback (background comparison)
    gpt_response = get_gpt_response(user_input)

    # Format response
    evo_msg = f"**Answer:** {evo_result['answer']}  \n**Reasoning:** {evo_result['reasoning']}"
    chat_history.append(f"👤 User: {user_input}")
    chat_history.append(f"🤖 Evo: {evo_msg}")
    chat_history.append(f"🧠 GPT: {gpt_response}")

    # Logging for Evo retraining
    log_feedback(
        question=user_input,
        option1=option1,
        option2=option2,
        context=evo_result['context'],
        evo_output=evo_result['answer'],
        gpt_output=gpt_response,
        evo_reasoning=evo_result['reasoning'],
        user_preference=user_vote
    )

    return evo_msg, chat_history

# 🔁 Clear chat state
def clear_fn():
    global chat_history
    chat_history = []
    return "", "", "", None, []

# 📈 Live retrain
def retrain_model():
    try:
        subprocess.run(["python", "retrain_from_feedback.py"], check=True)
        return "✅ Evo retrained successfully."
    except Exception as e:
        return f"❌ Retraining failed: {str(e)}"

# 🌐 Gradio UI
with gr.Blocks(title="EvoRAG – Real-Time Adaptive Reasoning AI") as demo:
    gr.Markdown("## 🧬 EvoRAG – Real-Time Adaptive Reasoning AI")
    gr.Markdown("Ask Evo a question, give two options. Evo chooses with reasoning. Compare with GPT. Feedback fuels evolution.")

    with gr.Row():
        with gr.Column(scale=4):
            user_input = gr.Textbox(label="Your Question", lines=2)
            option1 = gr.Textbox(label="Option 1")
            option2 = gr.Textbox(label="Option 2")
            user_vote = gr.Radio(["Evo", "GPT"], label="🗳️ Who gave the better answer?", info="Optional – improves Evo.")
            submit = gr.Button("🧠 Ask Evo")
            clear = gr.Button("🔁 Clear")
            retrain = gr.Button("📈 Retrain Evo from Feedback")

        with gr.Column(scale=6):
            evo_reply = gr.Markdown()
            chat_display = gr.HighlightedText(label="Conversation History")

    submit.click(fn=chat_fn, inputs=[user_input, option1, option2, user_vote],
                 outputs=[evo_reply, chat_display])
    clear.click(fn=clear_fn, outputs=[user_input, option1, option2, user_vote, chat_display])
    retrain.click(fn=retrain_model, outputs=evo_reply)

demo.launch()