app.py

import gradio as gr
import tensorflow as tf
import numpy as np
import json
from pathlib import Path
from PIL import Image

# =========================
# 1. Load Model
# =========================
MODEL_PATH = "model/best_food_model.keras"
model = tf.keras.models.load_model(MODEL_PATH)

# =========================
# 2. Labels
# =========================
LABELS = [
    'apple_pie', 'baby_back_ribs', 'baklava', 'beef_carpaccio', 'beef_tartare',
    'beet_salad', 'beignets', 'bibimbap', 'bread_pudding', 'breakfast_burrito',
    'bruschetta', 'caesar_salad', 'cannoli', 'caprese_salad', 'carrot_cake',
    'ceviche', 'cheese_plate', 'cheesecake', 'chicken_curry',
    'chicken_quesadilla', 'chicken_wings', 'chocolate_cake',
    'chocolate_mousse', 'churros', 'clam_chowder', 'club_sandwich',
    'crab_cakes', 'creme_brulee', 'croque_madame', 'cup_cakes',
    'deviled_eggs', 'donuts', 'dumplings', 'edamame', 'eggs_benedict',
    'escargots', 'falafel', 'filet_mignon', 'fish_and_chips', 'foie_gras',
    'french_fries', 'french_onion_soup', 'french_toast', 'fried_calamari',
    'fried_rice', 'frozen_yogurt', 'garlic_bread', 'gnocchi', 'greek_salad',
    'grilled_cheese_sandwich', 'grilled_salmon', 'guacamole', 'gyoza',
    'hamburger', 'hot_and_sour_soup', 'hot_dog', 'huevos_rancheros',
    'hummus', 'ice_cream', 'lasagna', 'lobster_bisque',
    'lobster_roll_sandwich', 'macaroni_and_cheese', 'macarons',
    'miso_soup', 'mussels', 'nachos', 'omelette', 'onion_rings',
    'oysters', 'pad_thai', 'paella', 'pancakes', 'panna_cotta',
    'peking_duck', 'pho', 'pizza', 'pork_chop', 'poutine', 'prime_rib',
    'pulled_pork_sandwich', 'ramen', 'ravioli', 'red_velvet_cake',
    'risotto', 'samosa', 'sashimi', 'scallops', 'seaweed_salad',
    'shrimp_and_grits', 'spaghetti_bolognese', 'spaghetti_carbonara',
    'spring_rolls', 'steak', 'strawberry_shortcake', 'sushi', 'tacos',
    'takoyaki', 'tiramisu', 'tuna_tartare', 'waffles'
]

# =========================
# 3. Load Nutrition JSON
# =========================
NUTRITION_PATH = Path("nutrition_db.json")

if NUTRITION_PATH.exists():
    with open(NUTRITION_PATH, "r", encoding="utf-8") as f:
        NUTRITION_DB = json.load(f)
else:
    NUTRITION_DB = {}

# =========================
# 4. Prediction Function
# =========================
def predict_nutrition(img):
    if img is None:
        return {}, "Upload an image."

    # Ensure PIL RGB
    if isinstance(img, np.ndarray):
        img = Image.fromarray(img)
    img = img.convert("RGB").resize((224, 224))

    img_array = tf.keras.preprocessing.image.img_to_array(img)
    img_array = np.expand_dims(img_array, axis=0) / 255.0

    preds = model.predict(img_array, verbose=0)[0]

    # Top 3 predictions
    top_indices = np.argsort(preds)[-3:][::-1]
    confidences = {LABELS[i]: float(preds[i]) for i in top_indices}

    # Top 1 nutrition
    top_idx = int(np.argmax(preds))
    food_name = LABELS[top_idx]

    nutri = NUTRITION_DB.get(
        food_name,
        {"cal": 0, "protein": 0, "carbs": 0, "fat": 0}
    )

    clean_name = food_name.replace("_", " ").title()

    nutrition_md = f"""
### 🥗 Nutrition Facts — {clean_name}
*(Estimated per 100g)*

| Nutrient | Amount |
|---|---|
| Calories | {nutri['cal']} kcal |
| Protein | {nutri['protein']} g |
| Carbs | {nutri['carbs']} g |
| Fat | {nutri['fat']} g |
"""

    return confidences, nutrition_md

# =========================
# 5. Gradio UI
# =========================
with gr.Blocks(theme=gr.themes.Soft()) as demo:
    gr.Markdown("# 🍎 Food-101 Classifier")
    gr.Markdown("Upload food → get prediction + macros.")

    with gr.Row():
        with gr.Column():
            input_img = gr.Image(type="numpy", label="Upload Food Image")
            submit_btn = gr.Button("Analyze Meal", variant="primary")

        with gr.Column():
            output_chart = gr.Label(num_top_classes=3)
            output_nutri = gr.Markdown()

    submit_btn.click(
        fn=predict_nutrition,
        inputs=input_img,
        outputs=[output_chart, output_nutri]
    )

    gr.Markdown("---")
    gr.Markdown("Educational demo. Not medical advice.")

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