app.py

"""
Gradio UI application for Batik Classification
Optimized for Hugging Face Spaces deployment
"""
import gradio as gr
import torch
import torch.nn as nn
from torchvision import transforms, models
from PIL import Image
import json
import numpy as np
from typing import Tuple, Dict
from huggingface_hub import hf_hub_download
import os

# Global variables
model = None
class_names = []
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
transform = None


def load_model():
    global model, class_names, transform

    try:
        # Load model configuration
        with open('model_config.json', 'r') as f:
            config = json.load(f)

        num_classes = config['num_classes']
        class_names = config['class_names']
        image_size = config.get('image_size', 224)

        # Initialize VGG16 model
        model = models.vgg16(weights=None)
        # Modify classifier to match saved model architecture
        model.classifier[3] = nn.Linear(4096, num_classes)
        model.classifier = nn.Sequential(*list(model.classifier.children())[:4])

        # Download model from Hugging Face Hub
        print("📥 Downloading model from Hugging Face Hub...")
        model_path = hf_hub_download(
            repo_id="RimsJ/Batik-Classifier",
            filename="vgg16_batik_best.pth"
        )

        # Load trained weights
        checkpoint = torch.load(model_path, map_location=device)

        # Extract state_dict
        if isinstance(checkpoint, dict) and 'model_state_dict' in checkpoint:
            state_dict = checkpoint['model_state_dict']
        else:
            state_dict = checkpoint

        # Remove '_orig_mod.' prefix if present
        new_state_dict = {}
        for key, value in state_dict.items():
            if key.startswith('_orig_mod.'):
                new_key = key.replace('_orig_mod.', '')
                new_state_dict[new_key] = value
            else:
                new_state_dict[key] = value

        model.load_state_dict(new_state_dict)
        model = model.to(device)
        model.eval()

        # Define image preprocessing
        transform = transforms.Compose([
            transforms.Resize((image_size, image_size)),
            transforms.ToTensor(),
            transforms.Normalize(mean=[0.485, 0.456, 0.406],
                               std=[0.229, 0.224, 0.225])
        ])

        print(f"✅ Model loaded successfully on {device}")
        print(f"📊 Number of classes: {num_classes}")

    except Exception as e:
        print(f"❌ Error loading model: {str(e)}")
        raise


def predict_image(image):
    """
    Predict batik class from image

    Args:
        image: PIL Image

    Returns:
        Tuple of (top_k_dict, formatted_text)
    """
    global model, transform, class_names

    try:
        if image is None:
            return None, "❌ Silakan upload gambar batik terlebih dahulu"

        if model is None:
            return None, "❌ Model belum dimuat. Silakan refresh halaman."

        # Convert to RGB if needed
        if image.mode != 'RGB':
            image = image.convert('RGB')

        # Transform and predict
        input_tensor = transform(image).unsqueeze(0).to(device)

        with torch.no_grad():
            outputs = model(input_tensor)
            probabilities = torch.nn.functional.softmax(outputs, dim=1)
            top_probs, top_indices = torch.topk(probabilities, min(5, len(class_names)), dim=1)

        # Get top prediction
        predicted_class = class_names[top_indices[0][0].item()]
        confidence = top_probs[0][0].item() * 100

        # Format top-5 results
        results = {}
        for i in range(min(5, len(class_names))):
            class_name = class_names[top_indices[0][i].item()]
            conf = top_probs[0][i].item()
            results[class_name] = float(conf)

        # Format output text
        result_text = f"""
## 🎯 Hasil Prediksi

**Motif Batik:** `{predicted_class}`
**Confidence:** `{confidence:.2f}%`

---

### 📊 Top 5 Prediksi:
"""

        for idx, (class_name, conf) in enumerate(list(results.items())[:5], 1):
            bar = "█" * int(conf * 20)
            result_text += f"\n{idx}. **{class_name}** - {conf*100:.2f}%  \n   {bar}"

        return results, result_text

    except Exception as e:
        import traceback
        traceback.print_exc()
        return None, f"❌ Error: {str(e)}"


# Load model at startup
print("🔄 Loading model...")
load_model()
print("✅ Model ready!")

# Create Gradio interface
with gr.Blocks(
    title="Batik Classification",
    theme=gr.themes.Soft(),
    css=".gradio-container {max-width: 1200px; margin: auto;}"
) as demo:

    gr.Markdown("""
    # 🎨 Klasifikasi Motif Batik Indonesia

    Upload gambar batik untuk mengetahui motif dan asalnya!
    **Total 111 motif batik** dari berbagai daerah di Indonesia 🇮🇩
    """)

    with gr.Row():
        with gr.Column(scale=1):
            input_image = gr.Image(
                type="pil",
                label="📤 Upload Gambar Batik",
                height=400
            )
            predict_btn = gr.Button(
                "🔍 Prediksi Motif Batik",
                variant="primary",
                size="lg"
            )

            gr.Markdown("""
            ### 💡 Tips:
            - Gunakan gambar dengan kualitas baik
            - Pastikan motif batik terlihat jelas
            - Format: JPG, PNG, JPEG
            """)

        with gr.Column(scale=1):
            output_text = gr.Markdown(label="Hasil Prediksi")
            output_label = gr.Label(
                label="📊 Confidence Score",
                num_top_classes=5
            )

    # Event handler
    predict_btn.click(
        fn=predict_image,
        inputs=input_image,
        outputs=[output_label, output_text]
    )

    gr.Markdown("""
    ---
    ### 📋 Tentang Model
    - **Dataset:** 111 Motif Batik Indonesia
    - **Kategori:** Batik dari Jawa Tengah, Jawa Timur, Jawa Barat, Bali, Jakarta, Kalimantan, Lampung

    ### 🎨 Contoh Motif:
    Parang Kusumo, Megamendung, Kawung, Truntum, Semarangan, dan banyak lagi!

    ---
    **Made with ❤️ for Indonesian Batik Heritage**
    """)


# Launch
if __name__ == "__main__":
    demo.launch(server_name="0.0.0.0", server_port=7860)