Skip to main content

Self-Hosted typesense Service

· 2 min read
Xiaohai Huang

typesense is a open source Algolia alternative. This blog shows how to run a self-hosted typesense server on my own server.

Setup a typesense Server Using Docker​

Run the server with docker-compose up

docker-compose.yml
version: "3.9"
services:
  typesense:
    image: typesense/typesense:0.22.2
    environment:
      - TYPESENSE_DATA_DIR=/data
      - TYPESENSE_ENABLE_CORS=true
      - TYPESENSE_API_KEY=test-key
    ports:
      - "8108:8108"
    volumes:
      - /t-data:/data # /t-data is used to store search engine data

see "how to configure the server using environment variables".

Crawl the Site​

  1. Create a docsearch-config.json config file which contains the information about the document website.

    1. Download config file.
    2. Update the index_name and domain name.

  2. Create a .env file which contains the information about the typesense server created in the previous step.

.env
TYPESENSE_API_KEY=test-key
TYPESENSE_HOST=xiaohai-huang.net
TYPESENSE_PORT=8108
TYPESENSE_PROTOCOL=http
  1. Run the scraper using Docker.
CONFIG="$(cat docsearch-config.json)" # Retrieved from step 1
docker run -i --rm \
        --env-file=$(pwd)/.env \ # .env created at step 2
        -e CONFIG="${CONFIG}" \
        typesense/docsearch-scraper

Install the package.

yarn add docusaurus-theme-search-typesense@next

Add the following to docusaurus.config.js file:

{
  themes: ['docusaurus-theme-search-typesense'],
  themeConfig: {
    typesense: {
      typesenseCollectionName: 'xiaohai-mind-palace-index', // the index_name in `docsearch-config.json`

      typesenseServerConfig: {
        nodes: [ // typesense server info
          {
            host: 'xiaohai-huang.net',
            port: 8108,
            protocol: 'http',
          },

        ],
        apiKey: 'test-key', // TYPESENSE_API_KEY specified in docker-compose.yml
      },

      // Optional: Typesense search parameters: https://typesense.org/docs/0.21.0/api/documents.md#search-parameters
      typesenseSearchParameters: {},

      // Optional
      contextualSearch: true,
    },
  }
}

References​