import gradio as gr
import moondream as md
from PIL import Image
import torch

# Inisialisasi model secara lokal
model = md.vl(model="vikhyatk/moondream2")

def answer_question(image, question):
    if image is None:
        return "No image provided"
    
    image = image.convert("RGB")
    
    try:
        # Gunakan method .query() untuk versi library native
        result = model.query(image, question)
        
        if isinstance(result, dict) and 'answer' in result:
            return result['answer']
        return str(result)
        
    except Exception as e:
        return f"Error: {str(e)}"

# Interface Gradio
interface = gr.Interface(
    fn=answer_question,
    inputs=[
        gr.Image(type="pil", label="Upload Captcha"), 
        gr.Textbox(label="Question", value="What is the text or number in this image?")
    ],
    outputs=gr.Text(label="Result"),
    title="Moondream Solver (Local Stable)",
    description="Menggunakan library moondream native untuk stabilitas tinggi."
)

if __name__ == "__main__":
    # Menambahkan konfigurasi launch untuk stabilitas di Hugging Face
    interface.queue().launch(show_error=True)