app.py

import tensorflow as tf
import gradio as gr
import numpy as np
import cv2
from PIL import Image
import requests
import torch
from transformers import AutoModelForImageClassification, AutoProcessor

# Step 1: Load Pre-trained Models
mobilenet_model = tf.keras.applications.MobileNetV2(weights='imagenet', include_top=True)
processor = AutoProcessor.from_pretrained("rajistics/finetuned-indian-food")
indian_food_model = AutoModelForImageClassification.from_pretrained("rajistics/finetuned-indian-food")

# USDA FoodData Central API Key
api_key = "k6zCqkjnVABP8WTcCJCQb6l7G7JXturqsudhra5h"

# Step 2: Image Preprocessing
def preprocess_image(image):
    img = Image.open(image).convert("RGB")
    img = img.resize((224, 224))
    img_array = tf.keras.preprocessing.image.img_to_array(img)
    img_array = np.expand_dims(img_array, axis=0)
    img_array = tf.keras.applications.mobilenet_v2.preprocess_input(img_array)
    return img_array

def detect_food(image):
    # Preprocess the image for MobileNetV2
    img_array = preprocess_image(image)
    preds = mobilenet_model.predict(img_array)
    decoded_preds = tf.keras.applications.mobilenet_v2.decode_predictions(preds, top=1)[0]
    food_item = decoded_preds[0][1]
    
    # Preprocess the image for the Indian food model
    inputs = processor(images=Image.open(image).convert("RGB"), return_tensors="pt")
    outputs = indian_food_model(**inputs)
    
    # Get the predicted class index (highest logit value)
    predicted_class_idx = outputs.logits.argmax(-1).item()
    
    # List of class labels for Indian food (modify this according to the model's classes)
    indian_food_labels = [
        "Samosa", "Chole Bhature", "Biryani", "Dosa", "Paneer Butter Masala", # etc.
        # Add all the food names corresponding to the model's output
    ]
    
    # Get the Indian food label based on the predicted index
    indian_food = indian_food_labels[predicted_class_idx]
    
    return food_item, indian_food

# Step 3: Retrieve Food Data
def get_food_data(food_item):
    url = f"https://api.nal.usda.gov/fdc/v1/foods/search?query={food_item}&pageSize=1&api_key={api_key}"
    response = requests.get(url)
    data = response.json()
    if 'foods' in data and data['foods']:
        return data['foods'][0].get('foodNutrients', [])
    return []

# Step 4: Calculate Nutrients Based on User Inputs
def calculate_nutrients(image, weight, volume):
    try:
        food_item, indian_food = detect_food(image)
        nutrients = get_food_data(food_item)
        result = f"Detected Food: {food_item} (Indian: {indian_food})\n"

        for nutrient in nutrients:
            name = nutrient['nutrientName']
            value = nutrient['value'] * (weight / 100)  # Scale by weight (default is per 100g)
            unit = nutrient['unitName']
            result += f"{name}: {value:.2f} {unit}\n"

        # Add additional volume-based calculations if necessary
        if volume:
            result += f"\nVolume: {volume} ml considered for additional calculations.\n"
        return result
    except Exception as e:
        return f"Error: {e}"

# Gradio Interface
interface = gr.Interface(
    fn=calculate_nutrients,
    inputs=[gr.Image(type="filepath"),          # Food Image
            gr.Slider(50, 500, value=100, step=10, label="Weight (g)"),  # Weight Slider
            gr.Slider(50, 500, value=100, step=10, label="Volume (ml)")  # Volume Slider
    ],
    outputs="text",
    title="Food Nutrient Calculator",
    description="Upload a food image and input its weight and volume to analyze the nutrient content."
)

# Launch the Gradio App
if __name__ == "__main__":
    interface.launch(share=True)