Download the PHP package la-haute-societe/craft-elasticsearch without Composer

Elasticsearch plugin for Craft CMS

Bring the power of Elasticsearch to your Craft CMS projects.

Requirements

This plugin works with Craft CMS 3 or 4.

In order to index data, you will need an Elasticsearch 6.0 (or later) instance, with the Ingest attachment processor plugin activated.

Installation

The easy way

Just install the plugin from the Craft Plugin Store.

Using Composer

• Install with Composer from your project directory: composer require la-haute-societe/craft-elasticsearch
• In the Craft Control Panel, go to Settings → Plugins and click the Install button for Elasticsearch.

Elasticsearch plugin Overview

Elasticsearch plugin will automatically index each entry and Craft Commerce product (if installed) on your site(s).

It will figure out the best Elasticsearch mapping for you based on your site(s)' language.

Supported languages

Configuring the Elasticsearch plugin

You can configure the Elasticsearch plugin from the Craft Control Panel (some settings only), of from the config/elasticsearch.php file in your Craft installation (all settings). If a setting is defined both in the CP and in the configuration file, the latter takes precedence.

The src/config.php, file is a configuration template to be copied to config/elasticsearch.php.

💡 If you currently don't have an Elasticsearch server handy, here's how you can set one up.

In both the configuration file and the CP

elasticsearchEndpoint

Type: string

The Elasticsearch instance endpoint URL (with protocol, host and port). Ignored if elasticsearchComponentConfig is set.

isAuthEnabled

Type: bool

A boolean indicating whether authentication is required on the Elasticsearch instance. Ignored if elasticsearchComponentConfig is set.

username

Type: string

The username used to authenticate on the Elasticsearch instance if it's protected by X-Pack Security. Ignored if isAuthEnabled is set to false or elasticsearchComponentConfig is set.

password

Type: string

The password used to authenticate on the Elasticsearch instance if it's protected by X-Pack Security. Ignored if isAuthEnabled is set to false or elasticsearchComponentConfig is set.

indexNamePrefix

Type: string

Index name prefix used to avoid index name collision when using a single Elasticsearch instance with several Craft instances.
Up to 5 characters, all lowercase.

highlight

Type: array

The elasticsearch configuration used to highlight query results. Only pre_tags and post_tags are configurable in the CP, advanced config must be done in the file. For more options, refer to the elasticsearch documentation.

blacklistedEntryTypes

Type: string[]

An array of entry type handles. Entries of those types won't be indexed.

blacklistedAssetVolumes

Type: string[]

An array of asset volume handles. Assets in those volumes won't be indexed.

Only in the configuration file

contentExtractorCallback

Type: callable

A callback function (function(string $entryContent): string) used to extract the content to be indexed from the full HTML source of the entry's page.

The default is to extract the HTML code between those 2 comments: <!-- BEGIN elasticsearch indexed content --> and <!-- END elasticsearch indexed content -->.

elementContentCallback

Type: callable

A callback function (function (\craft\base\ElementInterface $element): string) used to get the HTML content for the given element to index.

Note:

• If this parameter is not set or null, the default Guzzle client implementation will be used to get the HTML content of the element. If set, you will have to handle that part yourself.
• Content should be returned as HTML content in order to be correctly indexed.

resultFormatterCallback

Type: callable

A callback function (function (array $formattedResult, $result): array) used to prepare and format the Elasticsearch result object in order to be used by the results twig view.

elasticsearchComponentConfig

Type: array

An associative array passed to the yii2-elasticsearch component Connection class constructor. All public properties of the yii2-elasticsearch component Connection class can be set. If this is set, the elasticsearchEndpoint, username, password and isAuthEnabled settings will be ignored.

extraFields

Type: array

An associative array allowing to declare additional fields to be indexed along with the defaults ones. See Index additional data for more details.

Indexable content

By default, the content indexed in each entry is between the <!-- BEGIN elasticsearch indexed content --> and <!-- END elasticsearch indexed content --> HTML comments in the source of the entry page.

If you're using semantic HTML in your templates, then putting your <main> or <article> element between those comments should be ideal.

If you need more control over what is indexed, you'll have to set up a custom contentExtractorCallback.

Running a search

The search feature can be used from a frontend template file by calling the craft.elasticsearch.search('Something to search') variable. For instance, in a template search/index.twig:

Each entry consists of the following attributes:

• id: unique ID of the result
• title: page title
• url: full url to the page
• postDate : date of element publication start date
• expiryDate : date of element expiration date
• elementHandle: the element handle name. Can be either entry or product
• score: entry result score
• highlights: array of highlighted content matching the query terms
• rawResult: the ElasticsearchRecord raw result object

Notes:

• To add additional attributes, see Index additional data for more details.
• To customize the Elasticsearch query, see More complex way to get even more control

Auto indexing

The plugin automatically indexes entries and Craft Commerce products (created, updated or removed), as long as they're not in a blacklisted entry types, or disabled.

All entries are reindexed (in the background) when plugin settings are saved.

Elasticsearch plugin utilities

If your Elasticsearch index becomes out of sync with your sites contents, you can go to Utilities → Elasticsearch then click the Reindex all button.

Elasticsearch plugin console commands

The plugin provides an extension to the Craft console command that lets you reindex all entries or recreate empty indexes.

Recreate empty indexes

Remove index & create an empty one for all sites

`

Reindex all sites

`

Notes:

• The command will probably fail in case you don't affect a specific domain to a given site, for instance, avoid to use @web as a base URL.

Indexing of additional data

Simple way using the configuration file

To easily index additional data (elasticsearch fields), you can declare them using the extraFields parameter in the plugin configuration file.

Each field should be declared by using associative array with keys representing fields names and value as an associative array to configure the field behavior:

• mapping (optional): an associative array providing the elasticsearch mapping definition for the field. For more complex mapping, you can also use a callback (function (\lhs\elasticsearch\records\ElasticsearchRecord $esRecord)) to return the associative array. For example:

• highlighter (optional): an object defining the elasticsearch highlighter behavior for the field. To know more about that configuration, refer to the documentation here.
• value: either a string or a callable function taking one argument of craft\base\Element type and returning the value of the field. Second argument can be used to access the related lhs\elasticsearch\records\ElasticsearchRecord instance.

For example, to declare a color field in the configuration file, one could do:

More complex way to get even more control

You can get even more control over your additional data by listening to the following events in some project module:

• lhs\elasticsearch\record\ElasticsearchRecord::EVENT_BEFORE_CREATE_INDEX: That event is triggered before the Elasticsearch index is created. Once you get a reference to the Elasticsearch Record instance, the following methods can be used to customise the schema as needed:

  • getSchema() method can be used to get the current default Elasticsearch schema.
  • setSchema(array $schema) method can be used to set the customized schema

  For example, if you want to add a 'color' field, you could do something like:

  Note: Do not alter the following defaults properties: title, url, elementHandle and content. Also, do not alter the default attachment processor.

• lhs\elasticsearch\record\ElasticsearchRecord::EVENT_INIT: That event can be used to add additional attributes to the ElasticsearchRecord instance to handle your indexes. You can use the addAttributes(array $additionalAttributes) to add the list of additional attributes. This is mandatory in order to get or set any additional fields you declared in your schema in the previous step. For example, if you wish to declare the 'color' field, you could do:

• lhs\elasticsearch\record\ElasticsearchRecord::EVENT_BEFORE_SAVE: By listening to that event, you get a chance to set the value of your additional fields declared in the previous step. You can access the related Craft Element by using the getElement() method. For example, if you wish to set the value the 'color' attribute to be indexed, given that 'color' attribute corresponds to a Craft color field type, you could do:

• lhs\elasticsearch\record\ElasticsearchRecord::EVENT_BEFORE_SEARCH: That lhs\elasticsearch\events\SearchEvent event type is triggered just before a query is sent to Elasticsearch, typically when you call the craft.elasticsearch.search('Something to search') variable from your Twig views. During that event, you can use the following methods to forge your Elasticsearch query request to your needs:

  • getQueryParams($query) and setQueryParams($queryParams) can be used to alter the default Elasticsearch query parameters (see example below)
  • getHighlightParams() and setHighlightParams($highlightParams) can be used to alter the default Elasticsearch highlighter parameters (see example below) For example, if you wish to add the 'color' field to your query, given that 'color' attribute is a Craft color field type, you could do:

  Note: By using the resultFormatterCallback configuration callback property, you can also add the related results accessible to your Twig page search results.

  For example, to add the 'color' field result you could do:

Tips

Enable fuzziness

As of the other search query parameters, you can set fuzziness by altering default search query as follow:

Troubleshooting

• When saving the settings, if you get a Could not connect to the Elasticsearch instance at http://somedomain:9200. Please check the endpoint URL and authentication settings. error :

  • If you are using an external Elasticsearch instance, check:
    • that the instance is running
    • the domain name/IP address and ports are correct
    • your firewall settings allow access to the server and port

  • If you are using a docker-compose setup, check:

    • that your containers are running correctly once the docker-compose up has been executed. Execute a docker ps command to see if the containers appear in the list. If they are not, chances are there are some issues preventing the containers from starting and you should check those errors in the first place.

    • that you can access the instance by going to http://localhost:9200. You should see some default json payload like the following:

    • that all the containers are running on the same shared network (see the networks parameter in the docker-compose.yml file, in our example, it is named app_tier).

Elasticsearch is a trademark of Elasticsearch BV, registered in the U.S. and in other countries.