Skip to content

Latest commit

 

History

History
544 lines (416 loc) · 14.2 KB

available-methods.md

File metadata and controls

544 lines (416 loc) · 14.2 KB
Raw

Available Methods

aggregate

This method can be used to aggregate data based on a search query;

$searchResult = Book::searchQuery()
    ->aggregate('max_price', [
        'max' => [
            'field' => 'price',
        ],
    ])
    ->execute();

Alternatively you can use the aggregateRaw method:

$searchResult = Book::searchQuery()
    ->aggregateRaw([
        'max_price' => [
            'max' => [
                'field' => 'price',
            ],
        ],
    ])
    ->execute();

You can retrieve the aggregated data from the search result as follows:

$aggregations = $searchResult->aggregations();
$maxPrice = $aggregations->get('max_price');

collapse

This method allows to collapse search results based on field values:

$searchResult = Book::searchQuery($query)
    ->collapse('author_id')
    ->sort('published', 'desc')
    ->execute();

There is also the collapseRaw method at your disposal:

$searchResult = Book::searchQuery($query)
    ->collapseRaw(['field' => 'author_id'])
    ->sort('price', 'asc')
    ->execute();

explain

When set to true every hit includes detailed information about score computation:

$searchResult = Book::searchQuery($query)
    ->explain(true)
    ->execute();

$hits = $searchResult->hits();
$explanation = $hits->first()->explanation();

// every explanation includes a value, a description and details
// it is also possible to get its raw representation
$value = $explanation->value();
$description = $explanation->description();
$details = $explanation->details();
$raw = $explanation->raw();

from

from defines the starting document offset:

$searchResult = Book::searchQuery($query)
    ->from(5)
    ->execute();

highlight

This method allows you to get highlighted snippets from one or more fields in your search results:

$searchResult = Book::searchQuery($query)
    ->highlight('title')
    ->execute();

Use highlightRaw method if you need more control:

$searchResult = Book::searchQuery($query)
    ->highlightRaw(['fields' => ['title' => ['number_of_fragments' => 3]]])
    ->execute();

Use highlights method to retrieve all highlights from the search result:

$highlights = $searchResult->highlights();

You can also get a highlight for every hit:

$hits = $searchResult->hits();
$highlight = $hits->first()->highlight();

The highlighted snippets can be retrieved as follows:

$snippets = $highlight->snippets('title');

It is also possible to get a raw highlight:

$raw = $highlight->raw();

join

This method enables multi indices search:

$query = Query::bool()
    ->should(Query::match()->field('name')->query('John'))
    ->should(Query::match()->field('title')->query('The Book'))
    ->minimumShouldMatch(1);

$searchResult = Author::searchQuery($query)
    ->join(Book::class)
    ->execute();

In the example above, we search for an author with name John or a book with title The Book in two different indices. Note that the result collection of models includes both types:

// every model is either Author or Book
$models = $searchResult->models();

When searching in multiple indices, you can boost results from a specific index by providing the second argument in join method:

$searchResult = Author::searchQuery($query)
    ->join(Book::class, 2)
    ->execute();

load

This method allows you to eager load model relations:

$searchResult = Book::searchQuery($query)
    ->load(['author'])
    ->execute();

When searching in multiple indices, you need to explicitly define the model you want the relations for:

$searchResult = Book::searchQuery($query)
    ->join(Author::class)
    ->load(['author'], Book::class)
    ->load(['books'], Author::class)
    ->execute();

minScore

This method allows you to set minimum score for matching documents:

$searchResult = Book::searchQuery($query)
    ->minScore(0.5)
    ->execute();

pointInTime

pointInTime allows you to set a point in time that should be used for the search:

$pit = Book::openPointInTime('1m');

$searchResult = Book::searchQuery($query)
    ->pointInTime($pit, '5m')
    ->execute();

Note, that join, preference and routing parameters are ignored when pointInTime is used.

postFilter

postFilter is used to filter search results:

$postFilter = Query::term()
    ->field('published')
    ->value('2020-06-07');

$searchResult = Book::searchQuery($query)
    ->postFilter($postFilter)
    ->execute();

You can also provide a raw query in the postFilter method:

$searchResult = Book::searchQuery($query)
    ->postFilter(['term' => ['published' => '2020-06-07']])
    ->execute();

preference

preference defines nodes and shards used for the search:

$searchResult = Book::searchQuery($query)
    ->preference('_local')
    ->execute();

refineModels

This method allows you to set the callback where you can modify the database query.

use Illuminate\Database\Eloquent\Builder as EloquentBuilder;

$models = Book::searchQuery($query)
   ->refineModels(function (EloquentBuilder $query) {
       $query->select(['id', 'title', 'description']);
   })
   ->execute()
   ->models();

When searching in multiple indices, you need to explicitly define the model for which you want to set the callback:

use Illuminate\Database\Eloquent\Builder as EloquentBuilder;

$models = Book::searchQuery($query)
    ->join(Author::class)
    ->refineModels(function (EloquentBuilder $query) {
        $query->select(['id', 'title', 'description']);
    }, Book::class)
    ->refineModels(function (EloquentBuilder $query) {
        $query->select(['id', 'name', 'last_name']);
    }, Author::class)
    ->execute()
    ->models();

rescore

This method allows you to rescore the search results. In addition, you can also use rescoreWeights and rescoreWindowSize to set query_weight, rescore_query_weight and window_size:

$searchResult = Book::searchQuery($query)
    ->rescore('match_phrase', [
        'message' => [
            'query' => 'the quick brown',
            'slop' => 2,
        ],
    ])
    ->rescoreWeights(0.7, 1.2)
    ->rescoreWindowSize(10)
    ->execute();

Alternatively you can use rescoreRaw:

$searchResult = Book::searchQuery($query)
   ->rescoreRaw([
       'window_size' => 50,
       'query' => [
           'rescore_query' => [
               'match_phrase' => [
                   'message' => [
                       'query' => 'the quick brown',
                       'slop' => 2,
                   ],
               ],
           ],
           'query_weight' => 0.7,
           'rescore_query_weight' => 1.2,
       ]
   ])
   ->execute();

routing

This method allows you to search with custom routing:

$searchResult = Book::searchQuery($query)
    ->routing(['author1', 'author2'])
    ->execute();

searchAfter

You can use searchAfter to retrieve the next page of hits using a set of sort values from the previous page:

$firstPage = Book::searchQuery($query)
    ->pointInTime($pit)
    ->execute();

$searchAfter = $firstPage->hits()->last()->sort();

$secondPage = Book::searchQuery($query)
    ->pointInTime($pit)
    ->searchAfter($searchAfter)
    ->execute();

searchType

searchType defines how distributed term frequencies are calculated for relevance scoring:

$searchResult = Book::searchQuery($query)
    ->searchType('query_then_fetch')
    ->execute();

size

size method limits the number of hits to return:

$searchResult = Book::searchQuery($query)
    ->size(2)
    ->execute();

sort

This method sorts the search results:

$searchResult = Book::searchQuery($query)
    ->sort('price', 'asc')
    ->execute();

In case you need more advanced sorting algorithm use sortRaw:

$searchResult = Book::searchQuery($query)
    ->sortRaw([['price' => 'asc'], ['published' => 'asc']])
    ->execute();

source

This method allows you to select what document fields of the source are returned:

$searchResult = Book::searchQuery($query)
    ->source(['title', 'description'])
    ->execute();

sourceRaw allows you to use a single wildcard pattern, an array of fields or a boolean value in case you want to exclude document source from the result:

$searchResult = Book::searchQuery($query)
    ->sourceRaw(false)
    ->execute();

suggest

This method can be used to get similar looking terms based on the provided text:

$searchResult = Book::searchQuery(Query::matchNone())
    ->suggest('title_suggest', ['text' => 'book', 'term' => ['field' => 'title']])
    ->execute();

The same query with suggestRaw method:

$searchResult = Book::searchQuery(Query::matchNone())
    ->suggestRaw(['title_suggest' => ['text' => 'book', 'term' => ['field' => 'title']]])
    ->execute();

You can use the suggestions method to retrieve suggestions from the search result:

$suggestions = $searchResult->suggestions();

Each key of this collection is a suggestion name, each element is a collection of suggested terms:

$titleSuggestions = $suggestions->get('title_suggest');

Each suggestion contains various information about the term:

$firstSuggestion = $titleSuggestions->first();

// the suggestion text
$text = $firstSuggestion->text();
// the start offset and the length of the suggested text
$offset = $firstSuggestion->offset();
$length = $firstSuggestion->length();
// an arbitrary number of options
$options = $firstSuggestion->options();
// related models (only some suggesters support this feature)
$models = $firstSuggestion->models();
// an array representation of the suggestion
$raw = $firstSuggestion->raw();

trackScores

This method forces scores to be computed and tracked:

$searchResult = Book::searchQuery($query)
    ->trackScores(true)
    ->execute();

trackTotalHits

This method allows you to control how the total number of hits should be tracked:

$searchResult = Book::searchQuery($query)
    ->trackTotalHits(true)
    ->execute();

unless

This method will execute the given callback unless the first argument given to the method evaluates to true:

$searchResult = Book::searchQuery($query)
    ->unless($orderBy, function ($builder, $orderBy) {
        return $builder->sort($orderBy, 'asc');
    })
    ->execute();

You may also pass another closure as a third argument to the unless method. This closure will be only executed if the first argument evaluates to true:

$searchResult = Book::searchQuery($query)
    ->unless($orderBy, function ($builder, $orderBy) {
        return $builder->sort($orderBy, 'asc');
    }, function ($builder) {
         return $builder->sort('price', 'asc');
     })
    ->execute();

when

This method will execute the given callback when the first argument given to the method evaluates to true:

$searchResult = Book::searchQuery($query)
    ->when($orderBy, function ($builder, $orderBy) {
        return $builder->sort($orderBy, 'asc');
    })
    ->execute();

You may also pass another closure as a third argument to the when method. This closure will be only executed if the first argument evaluates to false:

$searchResult = Book::searchQuery($query)
    ->when($orderBy, function ($builder, $orderBy) {
        return $builder->sort($orderBy, 'asc');
    }, function ($builder) {
         return $builder->sort('price', 'asc');
     })
    ->execute();