December 12, 2020

319 words 2 mins read

typesense/showcase-recipe-search

typesense/showcase-recipe-search

Instantly search 2M cooking recipes using Typesense Search (an open source alternative to Algolia / ElasticSearch)

repo name typesense/showcase-recipe-search
repo link https://github.com/typesense/showcase-recipe-search
homepage https://recipe-search.typesense.org
language JavaScript
size (curr.) 642 kB
stars (curr.) 360
created 2020-12-09
license Apache License 2.0

🥘 Instant Recipe Search, powered by Typesense

This is a demo that showcases some of Typesense’s features using a 2 Million database of recipes.

View it live here: recipe-search.typesense.org

Tech Stack

This search experience is powered by Typesense which is a blazing-fast, open source typo-tolerant search-engine. It is an open source alternative to Algolia and an easier-to-use alternative to ElasticSearch.

The recipe dataset is from Glorf/recipenlg 🙏!

The dataset is 2.2 GB on disk, with ~2.2 million rows. It took 8 minutes to index this dataset on a 3-node Typesense cluster with 4vCPUs per node and the index was 2.7GB in RAM.

The app was built using the Typesense Adapter for InstantSearch.js and is hosted on S3, with CloudFront for a CDN.

The search backend is powered by a geo-distributed 3-node Typesense cluster running on Typesense Cloud, with nodes in Oregon, Frankfurt and Mumbai.

Repo structure

  • src/ and index.html - contain the frontend UI components, built with Typesense Adapter for InstantSearch.js
  • scripts/indexer - contains the script to index the recipe data into Typesense.
  • scripts/data - contains a 1K sample subset of the recipes database. But you can download the full dataset from the link above.

Development

To run this project locally, install the dependencies and run the local server:

yarn
bundle # JSON parsing takes a while to run using JS when indexing, so we're using Ruby just for indexing

yarn run typesenseServer

ln -s .env.development .env

BATCH_SIZE=1000 yarn run indexer:transformDataset # This will output a JSONL file
yarn run indexer:importToTypesense # This will import the JSONL file into Typesense

yarn start

Open http://localhost:3000 to see the app.

Deployment

The app is hosted on S3, with Cloudfront for a CDN.

yarn build
yarn deploy
comments powered by Disqus