app.py

import os
import gradio as gr
from watchdog import manual_retrain
from inference import generate_response
from logger import log_feedback
from dashboard import display_dashboard

# ✅ Ensure model is initialized
if not os.path.exists("trained_model/config.json"):
    print("📦 No model found in 'trained_model/'. Initializing...")
    exec(open("init_save.py").read())

print("===== Application Startup at 2025-07-14 12:XX:XX =====")

# ✅ Interface logic
def evo_chat(goal, sol1, sol2):
    response = generate_response(goal, sol1, sol2)
    return response

def retrain_model():
    success = manual_retrain()
    return "✅ Retrained successfully!" if success else "⚠️ Retrain failed."

def log_user_feedback(goal, sol1, sol2, correct):
    log_feedback(goal, sol1, sol2, correct)
    return "✅ Feedback logged. Thank you!"

# ✅ UI
with gr.Blocks(title="EvoTransformer v2.1") as demo:
    gr.Markdown("### 🧠 EvoTransformer v2.1 – Compare Options and Learn")

    with gr.Row():
        goal = gr.Textbox(label="Goal")
    with gr.Row():
        sol1 = gr.Textbox(label="Option 1")
        sol2 = gr.Textbox(label="Option 2")

    output = gr.Textbox(label="Model Suggestion")
    run_btn = gr.Button("🔍 Compare")
    run_btn.click(fn=evo_chat, inputs=[goal, sol1, sol2], outputs=output)

    with gr.Row():
        correct = gr.Radio(["Solution 1", "Solution 2"], label="Which was better?")
        log_btn = gr.Button("✅ Log Feedback")
        log_btn.click(fn=log_user_feedback, inputs=[goal, sol1, sol2, correct], outputs=None)

    with gr.Accordion("📊 Dashboard", open=False):
        display_dashboard()

    retrain_btn = gr.Button("♻️ Retrain Evo")
    retrain_status = gr.Textbox(label="Retrain Status")
    retrain_btn.click(fn=retrain_model, outputs=retrain_status)

demo.launch()