directus-extension-searchsync-2
v1.0.0
Published
Simple Directus extension that sync content with remote search engine (eg. Meilisearch, ElasticSearch, Algolia).
Downloads
10
Maintainers
Readme
Simple search engine indexer
Inspired by the dimitrov-adrian/directus-extension-searchsync, rewritten in TypeScript, supporting Directus 10.
Supported engines
- ✅ MeiliSearch
- 🚧 ElasticSearch (coming soon)
- 🚧 Algolia (coming soon)
CLI Commands
Usage: npx directus extension:searchsync <subdommand>
Subcommands:
index
- Reindex all documents from configuration
Configuration
The extension uses cosmiconfig for configuration loader with
searchsync
block or if EXTENSION_SEARCHSYNC_CONFIG_PATH
is set will try to use the file.
So, configuration should comes from one of next files:
- package.json
"searchsync":{...}
- .searchsyncrc
- .searchsyncrc.json
- .searchsyncrc.yaml
- .searchsyncrc.yml
- .searchsyncrc.js
- .searchsyncrc.cjs
- searchsync.config.js
- searchsync.config.cjs
Environment variables
References
server: object
Holds configuration for the search enginebatchLimit: number
Batch limit when performing index/reindex (defaults to 100)reindexOnStart: boolean
Performs full reindex of all documents upon Directus startscollections: object
Indexing data definitioncollections.<collection>.filter: object
The filter query in format like Directus on which item must match to be indexed (check Filter Rules )collections.<collection>.fields: array<string>
Fields that will be indexed in Directus formatcollections.<collection>.transform: function
(Could be defined only if config file is .js) A callback to return transformed/formatted data for indexing.collections.<collection>.indexName: string
Force collection name when storing in search indexcollections.<collection>.collectionField: string
If set, such field with value of the collection name will be added to the indexed document. Useful with conjuction with the indexName option
Examples
.searchsyncrc.json
{
"server": {
"type": "meilisearch",
"host": "http://search:7700/myindex",
"key": "the-private-key"
},
"batchLimit": 100,
"reindexOnStart": false,
"collections": {
"products": {
"filter": {
"status": "published",
"stock": "inStock"
},
"fields": [
"title",
"image.id",
"category.title",
"brand.title",
"tags",
"description",
"price",
"rating"
]
},
"posts": {
"indexName": "blog_posts",
"collectionField": "_collection",
"filter": {
"status": "published"
},
"fields": ["title", "teaser", "body", "thumbnail.id"]
}
}
}
.searchsyncrc.js
const config = {
server: {
type: "meilisearch",
host: "http://search:7700",
key: "the-private-key",
},
reindexOnStart: false,
batchLimit: 100,
collections: {
pages: {
filter: {
status: "published",
},
fields: ["title", "teaser", "body", "thumbnail.id"],
transform: (item, { flattenObject, striptags }) => {
return {
...flattenObject(item),
body: striptags(item.body),
someCustomValue: "Hello World!",
};
},
},
},
};
// Use as object.
module.exports = config;
Collection transformation callback description
/**
* @param {Object} item
* @param {{striptags, flattenObject, objectMap}} utils
* @param {String} collectionName
* @returns {Object}
*/
function (item, { striptags, flattenObject, objectMap }, collectionName) {
return item
}
Search engines config references
Meilisearch
{
"type": "meilisearch",
"host": "http://search:7700",
"key": "the-private-key"
}
Algolia (coming soon)
{
"type": "algolia",
"appId": "Application-Id",
"key": "secret-api-key"
}
ElasticSearch (coming soon)
New typeless behaviour, use collection names as index name.
{
"type": "elasticsearch",
"host": "http://search:9200/"
}
Use Authentification.
{
"type": "elasticsearch",
"host": "http://search:9200/",
"username": "elastic",
"password": "somepassword"
}
Ignore ssl-certificate-error.
{
"type": "elasticsearch",
"host": "http://search:9200/",
"ignore_cert": true
}
ElasticSearch for 5.x and 6.x (coming soon)
Old type behaviour, use collection names as types.
{
"type": "elasticsearch_legacy",
"host": "http://search:9200/projectindex"
}