Healthy Home Meals

Healthy home meals is a web app which enables users to take a picture of their refrigerator, or items over the kitchen top, detects the ingredients present in the picture along with the ones they usually have at home, considers allergies and dietary styles like vegan, lactose-free etc, and recommends healthy recipes that they can make at home. They include detailed step-by-step instructions, youtube video, and other info like cuisine, and cooking time from a dataset of over 5000 Indian recipes.

Made with

Tech used	For
React.js	Frontend
Flask	Backend
MongoDB	Database
Tensorflow and CustomVision	ML model to detect ingredients in picture
SciKit Learn	ML model to find matching recipes
Azure	Hosting
Cloudflare	CDN for static data

Setup process

Frontend

• Open frontend/src/index.js and update window.APIROOT to the base URL for the backend. Default URL is given below.

window.APIROOT = "http://127.0.0.1:4950/"

• Run the below command in frontend folder.

npm install

Backend

• Run the below command in backend folder.

pip install -r requirements.txt

• Create an OAuth client ID in Google cloud console with the below info.

# Authorized JavaScript origins

http://127.0.0.1:4950
https://127.0.0.1:4950

# Authorized redirect URIs

http://127.0.0.1:4950/callback
https://127.0.0.1:4950/callback

• Download the client_secret.json file and save it in the backend folder.

• Create a YouTube Data API v3 key from Google cloud console.

• Include an attribute data in the client_secret.json file as below.

{
  "web": {
    // No changes here
  },
  "data": {
    "redirect_uri": "http://127.0.0.1:4950/callback",
    "home": "http://127.0.0.1:3000",
    "mongo": "MongoDB URL here",
    "youtube": "YouTube Data API v3 key here"
  }
}

Running process

Frontend

Run the below command in the frontend folder.

npm run start

Backend

Run the below command in the backend folder.

python main.py

ML Datasets

Google Drive