Pennylane Tech Test - Larder Lookup

This is a recipe searching app built for the Pennylane tech test. The app has been built as a monorepo containing a Ruby on Rails 7 API with a React front-end.

A demo of the app loaded with a subset of the data (5065/10013) can be found at:

https://larder-lookup-2.fly.dev/

Approach

The main objective was to search for ingredients and return recipes that include the selected ingredients. The obvious and most straightforward approach would have been to save each recipe's ingredient as a string in an Ingredients table and search using ILIKE. Instead, I have decided to do some of the work upfront and parse the ingredients into FoodItems. The reason behind this upfront parsing is to make the results of the frontend's autocomplete cleaner, and by splitting up the measurements from the actual ingredient items, it becomes possible to introduce features like changing measurement types (i.e., from imperial to metric), or automatically adjust quantities in order to make more or less servings per recipe.

Backend

Callable Services used to contain and isolate logic
Queries have been decoupled from models
Rake task to parse and populate data
PostgreSQL Trigram's for better search capabilities

Potential Improvements

Presenter/View layer
Request specs
Validations for services
Add pagination

Frontend

Autocomplete
Control Panel to search Ingredients or preset search queries
Control panel designed to be extendable and allow additional features like sorting.

Potential Improvements

State management
Test coverage
Removal of Stimulas
Add sorting/pagination
More effective data fetching mechanism

Setup

Get code

git clone https://github.com/8bithero/larder-lookup.git
cd larder-lookup

Install gems/packages

bundle install
yarn install

Create database

rails db:setup
rails recipes:import

Populate data

For all data you can run:

rails recipes:import

If you would like to populate the database with a subset of the data, you can run the below, replacing the number with the number of records you want to populate the database with.

rails 'recipes:import[500]'

Start Server

bin/dev

Please note you will need to start the server with bin/dev rather than rails s to ensure the React frontend is also started.

Original Pennylane Test Brief

Problem statement

It's dinner time ! Create an application that helps users find the most relevant recipes that they can prepare with the ingredients that they have at home

Objective

Deliver an application prototype to answer the above problem statement.

By prototype, we mean:

something usable, yet as simple as possible
UI / design is not important
we do not expect features which are outside the basic scope of the problem

We expect to use this prototype as a starting point to discuss current implenentation details, as well as ideas for improvement.

Tech must-haves

MySQL / PostgreSQL or any other MySQL-compatible database.
A backend application which responds to queries
A web interface (can be VERY simple)
Ruby on Rails (if you're not familiar with Ruby on Rails, use something you're familiar with)

Bonus points

React
Application is hosted on fly.io

Data

We provide two datasets to choose from:

french-language recipes scraped from www.marmiton.org with python-marmiton
english-language recipes scraped from www.allrecipes.com with recipe-scrapers

Download it with this command if the above link doesn't work:

wget https://pennylane-interviewing-assets-20220328.s3.eu-west-1.amazonaws.com/recipes-en.json.gz && gzip -dc recipes-en.json.gz > recipes-en.json

Deliverable

the 2-3 user stories which the application implements
the codebase : in a git repo (share it with @quentindemetz @tdeo @soyoh @lucasbonin @sforsell @clemalfroy @dmilon @pointcom @evangelos-fotis @thecodehunter)
the database structure
the application, running on fly.io or on a personal server.
please submit links to the GitHub repo and the hosted application via this form and if you're on Mac, make sure your browser has permission to share the screen