Algolia-search plugin for medusa e-commerce app

This plugin was copied and fixed from medusa.js official repository

Algolia

Provide powerful indexing and searching features in your commerce application with Algolia.

Medusa Website | Medusa Repository

Features

• Flexible configurations for specifying searchable and retrievable attributes.
• Ready-integration with Medusa's Next.js starter storefront.
• Utilize Algolia's powerful search functionalities including typo-tolerance, query suggestions, results ranking, and more.

Prerequisites

• Medusa backend
• Algolia account

How to Install

1. Run the following command in the directory of the Medusa backend:

npm install medusa-plugin-algolia-search

2. Set the following environment variables in .env:

ALGOLIA_APP_ID=<YOUR_APP_ID>
ALGOLIA_ADMIN_API_KEY=<YOUR_ADMIN_API_KEY>

3. In medusa-config.js add the following at the end of the plugins array:

const plugins = [
// ...
{
  resolve: `medusa-plugin-algolia-search`,
  options: {
    applicationId: process.env.ALGOLIA_APP_ID,
    adminApiKey: process.env.ALGOLIA_ADMIN_API_KEY,
    settings: {
      products: [
        indexSettings: {
          indexName: 'products'
          searchableAttributes: ["title", "description"],
          attributesToRetrieve: [
            "id",
            "title",
            "description",
            "handle",
            "thumbnail",
            "variants",
            "variant_sku",
            "options",
            "collection_title",
            "collection_handle",
            "images",
          ],
        },
        filter: (product) => product.status == "published",
        transformer: (product) => ({ 
          id: product.id, 
          // other attributes...
        }),
  	],
    },
  },
},
]

For further use learn our features

Filtering

Settings object may contain few types of entities medusa loads Now, medusa presents only products, which triggers update of all indexes declared under products array in settings object We decided to make settings type like this

settings: {
	[key: string]:{ // 'products' | string
		indexSettings: {
			indexName: string // algolia index name. Must be unique value
		}
		filter?: (document: any) => boolean
		transformer?: (document: any) => boolean
	}[]
}

This allow you to upload products in different indexes using different transformers and filters

For example

{
	resolve: 'medusa-plugin-algolia-search',
	options: {
		batch_size: 500,
		applicationId: process.env.ALGOLIA_APP_ID,
		adminApiKey: process.env.ALGOLIA_ADMIN_API_KEY,
		settings: {
			products: [
				{
					indexSettings: {
						indexName: 'products',
						searchableAttributes: ['title', 'description'],
						attributesToRetrieve: ['objectID', 'title'],
					},
					filter: (product) => {
						if (
							product?.status != null &&
							product.status != 'published'
						)
							return false

						return true
					},
					transformer: (product) => {
						return {
							objectID: product.id,
							title: product.title,
						}
					},
				},
				{
					indexSettings: {
						indexName: 'wowducts',
						searchableAttributes: ['title', 'description'],
						attributesToRetrieve: ['objectID', 'title'],
					},
					filter: (product) => {
						if (
							product?.status != null &&
							product.status != 'draft'
						)
							return false

						return true
					},
					transformer: (product) => {
						return {
							objectID: product.id,
							title: product.title + '2',
						}
					},
				},
			],
		},
	},
},

Scheduled refresh

type AlgoliaPluginOptions = {
	applicationId: string
	adminApiKey: string
	scheduledRefresh?: string // cron string or nothing to get refresh repeatly in settled period
	settings: {
		[key in IndexTypes]: IndexSettingsExtended[]
	}
}

You may schedule refresh job via cron string, which will emit SEARCH_INDEX_EVENT like on startup every time it will be executed.

Test the Plugin

1. Run the following command in the directory of the Medusa backend to run the backend:

npm run start

2. Try searching products either using your storefront or using the Store APIs.