🍝 Eat or Pass 🙅

Your Tinder-like AI-powered food bot 🤤

Eat or Pass is a Telegram Bot application designed to fulfill your culinary curiosity. It will recommend some dishes from all around the world that may be of your liking, using AI to predict your tastes. Let it know you better!

Commands

/suggestion: The bot will reply with a dish trying to meet your tastes.
/likes: It will show a list of the latest dishes you liked.

To-do

I have developed this bot while learning NestJS so it's very basic. There are some more features that could be implemented in the bot, as the following:

User settings: Custom diet (vegetarian, vegan, gluten-free...)
User settings: Language selection to translate the content of the messages to the users language
Possibility to view a recipe for liked dishes

Contributions

You are welcome if you want to expand the bot features! I will be very pleased if you want to join the development; PRs are welcome.

To setup the dev environment follow this steps:

Clone the project

$ git clone git@github.com:Javiink/eat-or-pass.git

Fill the environment variables

You will need API keys for the third-party services that the bot will use, so copy the .env.example file and rename it to .env.development.local or .env and fill the secrets

Telegram Bot API: Talk to @BotFather in Telegram and create a bot. It will return an API key, paste it in the TG_BOT_API environment variable.
GROQ Cloud: To put it in a nutshell, this is the AI service that will generate the dishes. At this moment, their API is free to use. Go to GROQ Cloud, create an account and generate a key. Paste that key in the GROQ_API_KEY environment variable.
Gooogle Custom Search Engine: The bot will use this API to search for images for the dishes. We need 2 secrets here:
- Custom Search JSON API: Go to Google Custom Search Docs and follow the steps, the paste the API key in the GOOGLE_CUSTOMSEARCH_API_KEY environment variable.
- Custom Programmable Search Engine CX: Go to Google Programmable Search engine Control panel and add a new Custom Search. In the "What to search?" section, select "Search the entire web". In the "Search settings" section, make sure to toggle on those two switches. Create the search engine and put the Search engine ID in the GOOGLE_CUSTOMSEARCH_CX environment variable.
MongoDB URL: You will need a working MongoDB database to store the users preferences. You can install it in your computer/server, use a docker image or the free plan of MongoDB Atlas. If you use this project's docker-compose file, please see details in the Docker section below. Put the MongoDB URL in the DB_MONGO_URL environment variable, should be in the mongodb://USER:PASSWORD@HOST:PORT/DATABASE format.

Run the project

$ npm run start:dev

Docker

Setup

If you want to run the project using Docker, there is a docker-compose.yml file for you to run. But first you need to setup some environment variables.

You may want to change the MongoDB port that is exposed by Docker. You can change it in the eat-or-pass-mongodb ports section. The number before the colon (:) is the one you may change to your liking. Or you can remove the ports section entirely if you don't want to connect to the database externally.
You must fill the environment variables marked in the .env.example file. This variables must go in a .env file in the root of the project. The MONGO_INITDB_ROOT_USERNAME and MONGO_INITDB_ROOT_PASSWORD are the credentials you want to use in the DB_MONGO_URL variable. You can use this URL, just replace USER and PASSWORD with the ones you types in the MONGO_INITDB_ROOT_ fields: mongodb://USER:PASSWORD@eat-or-pass-mongodb/eat-or-pass

Run the project

$ docker compose up