- Solr is installed
- A Solr core is created and configured for your project, especially
schema.xml - The schema must contain a field 'key' with type string. This key is handled internally.
The URL is likely to look like http://localhost:8989/solr/REPLACEWITHCORENAME
- Use the plugin settings page to set the URL to your Solr Core. This can (and should) be set to an environment variable or alias.
- You can also use a config file named
solrconfig.php. This setting can be environment specific.
<?php
return [
'solrBaseUrl' => 'your solr base url here'
];
This model defines the search parameters used to query Solr. It should specify the default parameters, the dynamic parameters can be set when creating an instance.
The fl parameter (fields to be retrieved) must contain the key field, if formatting is enabled (see below).
Example:
<?php
namespace modules\main\models;
class MySearchParamsModel extends \wsydney76\solrsearch\models\SearchParamsModel
{
public $facet = 'on';
public $facetField = ['prodyear', 'seriestitle_exact', 'actress_exact'];
public $fl = 'score,key,id,type,slug,url,title,name,film,actress,seriestitle,prodyear,imagefile,station,remark';
public $hl = 'on';
public $hlFl = 'title,name,film,actress,seriestitle,prodyear,content,station,remark';
public $hlSimplePost = '</i></b>';
public $hlSimplePre = '<b><i>';
public $mm = '1<-2 6<70%';
public $pf = 'title^10.0 name^10.0 content^2.0 remark^1.0 film^1.0 actress 1^0 seriestitle^1.0 prodyear^1.0';
public $q = '';
public $qf = 'title^10.0 name^10.0 content^2.0 remark^1.0 film^1.0 actress 1^0 seriestitle^1.0 prodyear^1.0';
public $sort = 'score desc,seriestitle asc,title_sort asc';
public $spellcheck = 'true';
public $spellcheckCollate = 'true';
public $spellcheckExtendedResults = 'true';
}
In case you are using additional parameters to those defined in the base class, you will
need to overwrite the getParams() method to pass those parameters.
This can also be necessary if defaults are set in the solrconfig.xml file and should not be
set on every single request. However this is not recommended, this file should only contain system
settings and avoid project specifics.
Example:
{% set q = craft.app.request.getParam('q') %}
{% set searchParamsModel = create({class:'modules\\main\\models\\MySearchParamsModel', q:q}) %}
{% set results = craft.solrsearch.search(searchParamsModel) %}
Example:
$result = SolrSearch::$services->search->search(new MySearchParamsModel(['q' => 'searchterm']));
By default, Solr responses are formatted for easier use, highlighting results are merged into the single docs, and factes results are cleaned up.
Example:
[
'rcode' => 'OK'
'recordcount' => 2
'time' => 15
'docs' => [
0 => [
'id' => 1504
'title' => 'Erna Klawuppke'
'type' => 'person'
'url' => 'http://filmdb.local/erna-klawuppke'
'imagefile' => 'erna-klawuppke.jpg'
'score' => 1.5163089
'highlighting' => [
'title' => [
0 => 'Erna <b><i>Klawuppke</i></b>'
]
]
]
...
]
'facets' => [
'prodyear' => [
0 => [
'name' => '2018'
'count' => 1
]
]
...
]
'spellcheck' => [
'suggestions' => []
'correctlySpelled' => true
'collations' => []
]
]
You can get the raw Solr response by setting a second parameter to the search method: search(model, false)
Example:
[
'responseHeader' => [
'status' => 0
'QTime' => 13
'params' => [
'mm' => '1<-2 6<70%'
... all params echoed here
]
]
'response' => [
'numFound' => 2
'start' => 0
'maxScore' => 1.5163089
'docs' => [
0 => [
id' => 1504
'title' => 'Erna Klawuppke'
'type' => 'person'
'url' => 'http://filmdb.local/erna-klawuppke'
'imagefile' => 'erna-klawuppke.jpg'
'score' => 1.5163089
]
...
]
]
'facet_counts' => [
'facet_queries' => []
'facet_fields' => [
'prodyear' => [
0 => '1993'
1 => 0
...
50 => '2018'
51 => 1
52 => '2019'
53 => 0
]
...
]
'facet_dates' => []
'facet_ranges' => []
'facet_intervals' => []
'facet_heatmaps' => []
]
'highlighting' => [
1504 => [
'title' => [
0 => 'Erna <b><i>Klawuppke</i></b>'
]
]
]
'spellcheck' => [
'suggestions' => []
'correctlySpelled' => true
'collations' => []
]
]
The plugin uses an event to retrieve the data to be indexed. The event returns
an array that can be passed to the Solr add command. It has to match the fields specified
in Solr's schema.xml and may look like
$doc = [
'id' => $entry->id,
'title' => $entry->title,
'content' => $this->_getContent($entry),
'type' => $entry->section->handle,
'url' => $entry->url,
];
The doc must not contain a field called key. This is used internally as a unique key.
Example:
use wsydney76\solrsearch\events\GetSolrDocForElementEvent;
use wsydney76\solrsearch\services\SearchService;
Event::on(
SearchService::class,
SearchService::EVENT_GET_SOLR_DOC_FOR_ENTRY, function(GetSolrDocForElementEvent $event) {
$element = $event->element;
if (... do not index this entry)) {
$event->cancel = true;
return;
}
// Return the Solr doc, using a method that fits best for your project
$element->attachBehavior('solrBehavior', SolrSearchEntryBehavior::class);
$event->doc = $element->getSolrDoc();
if (!$event->doc) {
$event->cancel = true;
}
}
);
The entry passed by the event may or may not have eager loaded related elements (see below);
The Plugins listens to the following events for Entry/Product element types:
EVENT_AFTER_SAVE(adds / deletes the solr record depending on an entries status)EVENT_AFTER_DELETEEVENT_AFTER_RESTORE(Entry only)
The plugin uses an event to let the project specify the elements to be indexed when 'Update all' is called either via the Solr Search Utility in the Control Panel or via the CLI command.
use wsydney76\solrsearch\events\GetAllElementsForSolrSearchEvent;
use wsydney76\solrsearch\services\SearchService;
...
if (Craft::$app->plugins->isPluginEnabled('solrsearch')) {
Event::on(
SearchService::class,
SearchService::EVENT_GET_ALL_ENTRIES_FOR_SOLR_SEARCH, function(GetAllElementsForSolrSearchEvent $event) {
// Get entries, eg. via a service
$event->elements = $this->content->getEntriesForSolrSearch();
}
);
}
Eager loading related elements and matrix field should be used for better performance, the getDoc event has to be aware that related elements are eager loaded or not (when called if a single entry is saved.)
The plugin has no specific functionality for handling multi site content besides setting a unique key value.
The events will always be fired for all sites.
So if you save an entry that belongs to three sites, three events will be fired, referencing the entry for that specific site.
What you can do in your project:
- Define a Solr field for the site handle
- Set the field value when returning the Solr Doc for entry
- Add a filter query parameter to your search, e.g.
fq:'site:en'
./craft solrsearch/search/delete-all
./craft solrsearch/search/update-all
Raw update commands can be executed via the SolrService::command method.
The signature is
command($cmd, $async = false, $commit = true, $description = 'Execution Solr Command')
where
- $cmd is the command to execute, either as string or array that can be converted to a valid Solr json command
- $async = true pushes a background job
- $commit = false does not perform a commit. You can call
SolrService::commitwhen finished. - $description = Description for the background job, that will be displayd in queue manager.
Returns null if $async = true, else returns http response as instance of GuzzleHttp\Psr7\Response.
You can extract the Solr response with \craft\helpers\Json::decodeIfJson($returnValue->getBody()->getContents())