app.py

import streamlit as st
import base64
from huggingface_hub import InferenceClient
import os
import requests

# Initialize Hugging Face Inference client using tokens from environment variables
client = InferenceClient(api_key=os.getenv("HF_API_TOKEN_DISH"))
client1 = InferenceClient(api_key=os.getenv("HF_API_TOKEN_DIET"))
groq_api_key = os.getenv("API_Groq")

# 1. Function to identify dish from image
def identify_dish(image_bytes):
    encoded_image = base64.b64encode(image_bytes).decode("utf-8")
    dish_name = ""

    for message in client.chat_completion(
        model="meta-llama/Llama-3.2-11B-Vision-Instruct",
        messages=[
            {
                "role": "You are a highly specialized food identification AI with extensive knowledge of global cuisines. Your sole task is to accurately identify dishes from images. Adhere strictly to these guidelines:\n1. Analyze the image thoroughly, focusing on ingredients, presentation, and cultural context.\n2. Provide ONLY the name of the main dish or dishes visible. Do not list individual ingredients or components.\n3. Use the most specific and widely recognized name for the dish.\n4. If multiple distinct dishes are present, list them separated by commas.\n5. If you cannot identify a dish with high confidence (>90%), respond with 'Unidentified dish'.\n6. Do not provide any explanations, descriptions, or additional commentary.\n7. Respond in a concise, list-like format.\nYour response should contain nothing but the dish name(s) or 'Unidentified dish'.",
                "content": [
                    {"type": "image_url", "image_url": {"url": f"data:image/jpeg;base64,{encoded_image}" }},
                    {"type": "text", "text": "Identify the dishes in the image and return only the names of the dishes."},
                ],
            }
        ],
        max_tokens=70,
        stream=True,
    ):
        if message.choices[0].delta.content:
            dish_name += message.choices[0].delta.content

    return dish_name.strip()

# 2. Function to calculate daily caloric needs
def calculate_metrics(age, gender, height_cm, weight_kg, weight_goal, activity_level, time_frame_months):
    bmi = weight_kg / ((height_cm / 100) ** 2)

    if gender == "male":
        bmr = 10 * weight_kg + 6.25 * height_cm - 5 * age + 5
    else:
        bmr = 10 * weight_kg + 6.25 * height_cm - 5 * age - 161

    activity_multipliers = {
        "sedentary": 1.2,
        "light": 1.375,
        "moderate": 1.55,
        "active": 1.725,
        "very active": 1.9
    }
    tdee = bmr * activity_multipliers[activity_level]

    if gender == "male":
        ibw = 50 + (0.91 * (height_cm - 152.4))
    else:
        ibw = 45.5 + (0.91 * (height_cm - 152.4))

    if weight_goal == "loss":
        daily_caloric_needs = tdee - 500
    elif weight_goal == "gain":
        daily_caloric_needs = tdee + 500
    else:
        daily_caloric_needs = tdee

    protein_calories = daily_caloric_needs * 0.2
    fat_calories = daily_caloric_needs * 0.25
    carbohydrate_calories = daily_caloric_needs * 0.55

    return {
        "BMI": bmi,
        "BMR": bmr,
        "TDEE": tdee,
        "IBW": ibw,
        "Daily Caloric Needs": daily_caloric_needs,
        "Protein Calories": protein_calories,
        "Fat Calories": fat_calories,
        "Carbohydrate Calories": carbohydrate_calories
    }

# 3. Function to generate diet plan
def generate_diet_plan(dish_name, calorie_intake_per_day, goal):
    user_input = f"""
    You are a certified Dietitian with 20 years of experience. Based on the following input, create an Indian diet plan that fits within the calculated calorie intake and assesses if the given dish is suitable for the user's goal.

    Input:
    - Dish Name: {dish_name}
    - Caloric Intake per Day: {calorie_intake_per_day} calories
    - Goal: {goal}
    """
    response = client1.chat_completion(
        model="meta-llama/Meta-Llama-3-8B-Instruct",
        messages=[{"role": "You are a certified Dietitian with 20 years of Experience", "content": user_input}],
        max_tokens=500
    )

    return response.choices[0].message.content

# 4. Function to generate a recipe using Groq API
def generate_recipe(recipe_name):
    url = "https://api.groq.com/openai/v1/chat/completions"
    headers = {
        "Authorization": f"Bearer {groq_api_key}",
        "Content-Type": "application/json"
    }
    data = {
        "model": "llama-3.1-70b-versatile",
        "messages": [
            {"role": "system", "content": "You are a skilled chef with 10 years of experience."},
            {"role": "user", "content": f"Create a detailed recipe for {recipe_name}"}
        ]
    }

    response = requests.post(url, headers=headers, json=data)

    if response.status_code == 200:
        result = response.json()
        generated_recipe = result['choices'][0]['message']['content']
        return generated_recipe
    else:
        return f"Error: {response.status_code}, {response.text}"

# Streamlit App Title
st.title("AI Diet Planner")

# Sidebar for user input
st.sidebar.title("User Input")
image_file = st.sidebar.file_uploader("Upload an image of the dish", type=["jpeg", "png"])
age = st.sidebar.number_input("Enter your age", min_value=1)
gender = st.sidebar.selectbox("Select your gender", ["male", "female"])
height_cm = st.sidebar.number_input("Enter your height (cm)", min_value=1.0)
weight_kg = st.sidebar.number_input("Enter your weight (kg)", min_value=1.0)
weight_goal = st.sidebar.selectbox("Weight goal", ["loss", "gain", "maintain"])
activity_level = st.sidebar.selectbox("Activity level", ["sedentary", "light", "moderate", "active", "very active"])
time_frame = st.sidebar.number_input("Time frame to achieve goal (months)", min_value=1)

# Submit button
submit = st.sidebar.button("Submit")

# Process the image and calculate metrics upon submission
if submit:
    if image_file:
        st.write("### Results")
        image_bytes = image_file.read()

        # Step 1: Identify the dish
        dish_name = identify_dish(image_bytes)
        st.markdown("<hr>", unsafe_allow_html=True)
        st.write("#### Dish Name Identified:")
        st.markdown(f"<div style='background-color: #d4edda; color: #155724; padding: 10px; border-radius: 10px;'>{dish_name}</div>", unsafe_allow_html=True)

        # Step 2: Perform Calculations
        metrics = calculate_metrics(age, gender, height_cm, weight_kg, weight_goal, activity_level, time_frame)
        st.markdown("<hr>", unsafe_allow_html=True)
        st.write("#### Metrics Calculated:")
        st.markdown(f"""
            <div style='background-color: #f8d7da; color: #721c24; padding: 10px; border-radius: 10px;'>
                <p><b>Your BMI:</b> {metrics['BMI']:.2f}</p>
                <p><b>Your BMR:</b> {metrics['BMR']:.2f} calories</p>
                <p><b>Your TDEE:</b> {metrics['TDEE']:.2f} calories</p>
                <p><b>Ideal Body Weight (IBW):</b> {metrics['IBW']:.2f} kg</p>
                <p><b>Daily Caloric Needs:</b> {metrics['Daily Caloric Needs']:.2f} calories</p>
            </div>
        """, unsafe_allow_html=True)

        # Step 3: Generate diet plan
        diet_plan = generate_diet_plan(dish_name, metrics["Daily Caloric Needs"], weight_goal)
        st.markdown("<hr>", unsafe_allow_html=True)
        st.write("#### Diet Plan Based on Dish & Goal:")
        st.markdown(f"<div style='background-color: #d1ecf1; color: #0c5460; padding: 10px; border-radius: 10px;'>{diet_plan}</div>", unsafe_allow_html=True)

        # Step 4: Ask for recipe generation
        st.markdown("<hr>", unsafe_allow_html=True)
        recipe_option = st.radio("Would you like a recipe for a dish?", ["No", "Yes"])

        if recipe_option == "Yes":
            recipe_name = st.text_input("Enter the dish name for the recipe:")
            if st.button("Generate Recipe"):
                recipe = generate_recipe(recipe_name)
                st.write("### Generated Recipe:")
                st.markdown(f"<div style='background-color: #fff3cd; color: #856404; padding: 10px; border-radius: 10px;'>{recipe}</div>", unsafe_allow_html=True)
    else:
        st.error("Please upload a valid image in JPEG or PNG format.")

# CSS styling
st.markdown("""
    <style>
    .stButton button { background-color: #4CAF50; color: white; }
    .stContainer { border: 1px solid #ddd; padding: 20px; margin-bottom: 20px; }
    hr { border: 1px solid #e9ecef; margin: 20px 0; }
    </style>
""", unsafe_allow_html=True)