RediSearch client

This is a light weight, CMS/Framework agnostic client for RediSearch

For RediSearch version 1.0, please use version 1.0.2

How to use

First, you will need to install and configure Redis and RediSearch.

Then add this package to your project via composer require front/redisearch.

For all kind of operations, first step is to connect to redis server:

$client = \FKRediSearch\RediSearch\Setup::connect( $server, $port, $password, 0 );

Default values are: $server = '127.0.0.1', $port = 6379, $password = null, $database = 0

Field types

Redisearch supports some field types:

TEXT [NOSTEM] [WEIGHT {weight}] [PHONETIC {matcher}] ex: TEXT NOSTEM WEIGHT 5.6 PHONETIC {maycher}*
NUMERIC [SORTABLE] [NOINDEX]
GEO [SORTABLE] [NOINDEX]
TAG [SEPARATOR {sep}]

*matcher: Declaring a text field as PHONETIC will perform phonetic matching on it in searches by default. The obligatory {matcher} argument specifies the phonetic algorithm and language used. The following matchers are supported:

dm:en - Double Metaphone for English dm:fr - Double Metaphone for French dm:pt - Double Metaphone for Portuguese dm:es - Double Metaphone for Spanish

Create an index

To create index, first we need to specify indexName, then we can pass some optional flags and at the end, schema (fields).

Available commands are:

$index = new Index( $client ); // Initiate Index
$index->setIndexName('test'); // Set indexName

In the RediSearch version 2.0, the whole structure of the index have been changed. Unlike version 1.0 which used to store indexed documents into Redis HASHed, in version 2.0 the Redis HASHes are indexed automatically. This means:

While creating an index, you need to tell RediSearch which HASH prefixes should be indexed
The HASH key in which you want to index, you needs to have the same prefix specified for the index

Also you need to tell RediSearch which Redis type should be indexed. For now, only HASH is supported, but other data types might be added in the future.

$index->on('HASH');
$index->setPrefix('doc:');

In the new version of RediSearch, indexes can be created temporarily for a specified period of inactivity. The internal idle timer is reset whenever the index is searched or added to. This is specially useful for example: order history of individual users in a commerce website.

$index->setTemporary(3600); // The index will automatically be deleted after one hour. The underlying HASH values remain untouched.

RediSearch by default supports basic english stop-words. There is two option, you can totally disable stop-word exclusion or specify your own list:

$index->setStopWords( 0 ); // This will disables excluding of stop-words
// Or you can add your list of stop-words
$stop_words = array('this', 'that', 'it', 'what', 'is', 'are');
$index->setStopWords( $stop_words );

Some other flags are:

$index->setNoOffsetsEnabled( true ); // If set, term offsets won't be stored for documents (saves memory, does not allow exact searches or highlighting).
$index->setNoFieldsEnabled( true ); // If set, field bits for each term won't be stored. This saves memory, does not allow filtering by specific fields.
$index->setDefaultLang( 'english' ); // Set default language of the index. 
$index->setLangField( 'language' ); // Set document field used to specify individual documents language.
$index->setScore(1); // Set a default score for that specific index. 
$index->setScoreField('scoreField'); // Set a field which specifies each individual documents score.
$index->setPayloadField('payloadField'); // Document field that should be used as a binary safe payload string
$index->setMaxFields(34); // Maximum allowed number of fields. This is to preserve memory use
$index->setNoHighlight(); // This will deactivate highlighting feature. 
$index->setNoFreqs(); // Prevents storing term frequencies. 
$index->skipInitialScan(); // Cancels initial HASH scanning when index created.

And finally, we need to specify the schema (the fields, their types and flags):

$index->addTextField( $name, $weight = 1.0, $sortable = false, $noindex = false)
      ->addTagField( $name, $sortable = false, $noindex = false, $separator = ',')
      ->addNumericField( $name, $sortable = false, $noindex = false )
      ->addGeoField( $name, $noindex = false );
      
// Example 
$index->addTextField('title', 0.5, true, true) // `title TEXT WEIGHT 0.5, SORTABLE NOINDEX`
    ->addTextField('content') // `content TEXT WEIGHT 1.0`
    ->addTagField('category', true, true, ';') // `category TAG SEPARATOR ',' SORTABLE NOINDEX`
    ->addGeoField('location') // `location GEO`
    ->create(); // Finally, create the index.

Some notes:

field weight is float value and only applies to the field. Default value is 1.0
Don't forget to call create() method at the end

Synonym

Synonym matching is supported as well:

$synonym = array(
  array( 'boy', 'child', 'baby' ),
  array( 'girl', 'child', 'baby' ),
  array( 'man', 'person', 'adult' )
);

$index->synonymAdd( $synonym );

Add document to the index (indexing)

$document = new Document();
$document->setScore(0.3); // The document's rank based on the user's ranking. This must be between 0.0 and 1.0. Default value is 1.0
$document->setLanguage('english'); // This is usefull for stemming
$document->setId('doc:123'); // This is the HASH key. RediSearch uses the prefix of document ID to index the document
// And the fields 
$document->setFields(
  array(
    'title'       => 'Document title like post title',
    'category'    => 'search, fuzzy, synonym, phonetic',
    'date'        => strtotime( '2019-01-14 01:12:00' ),
    'location'    => new GeoLocation(-77.0366, 38.8977),
  )
);

$index->add( $document ); // Finally, add document to the index (in other term, index the document)

Persistence

After indexing documents, index can be written to the disk and in case of network issues, you won't miss your index.

$index->writeToDisk();

Query builder

Class Query\QueryBuilder is designed to help with construction of redisearch queries used in search and aggregation. It uses addCondition(), addGenericCondition() and addSubcondition() methods to add conditions to a query and also enables using partial search, fuzzy search, escaping, tokenization and stop words. Example:

$query = new QueryBuilder();
$query->setTokenize()
  ->setFuzzyMatching()
  ->addCondition('field', ['value1', 'value2'], 'OR');
$condition = $query->buildRedisearchQuery();

Search

And here is how to search:

$search = new Query( $client, 'indexName' );
$results = $search
        ->sortBy( $fieldName, $order = 'ASC' )
        ->geoFilter( $fieldName, $longitude, $latitude, $radius, $distanceUnit = 'km' )
        ->numericFilter( $fieldName, $min, $max = null )
        ->withScores() // If set, we also return the relative internal score of each document. this can be used to merge results from multiple instances
        ->withSortKey() // Returns the value of the sorting key
        ->verbatim() // if set, we do not try to use stemming for query expansion but search the query terms verbatim.
        ->withPayloads() // If set, we retrieve optional document payloads (see FT.ADD). the payloads follow the document id, and if WITHSCORES was set, follow the scores
        ->noStopWords() //  If set, we do not filter stopwords from the query
        ->slop() // If set, we allow a maximum of N intervening number of unmatched offsets between phrase terms. (i.e the slop for exact phrases is 0)
        ->inKeys( $number, $keys ) // If set, we limit the result to a given set of keys specified in the list. the first argument must be the length of the list, and greater than zero. Non-existent keys are ignored - unless all the keys are non-existent.
        ->inFields( $number, $fields ) // If set, filter the results to ones appearing only in specific fields of the document, like title or URL. num is the number of specified field arguments
        ->limit( $offset, $pageSize = 10 ) // If set, we limit the results to the offset and number of results given. The default is 0 10
        ->highlight( $fields, $openTag = '<strong>', $closeTag = '</strong>') 
        ->summarize( $fields = array(), $fragmentCount = 3, $fragmentLength = 50, $separator = '...') // Use this option to return only the sections of the field which contain the matched text
        ->return( $fields ) // Use this keyword to limit which fields from the document are returned. num is the number of fields following the keyword. If num is 0, it acts like NOCONTENT.
        ->noContent() // If it appears after the query, we only return the document ids and not the content. This is useful if RediSearch is only an index on an external document collection
        ->search( $query, $documentsAsArray = false ); // By default, return values will be object, but if TRUE is passed as `$documentsAsArray` results will return as array

There is two methods available to apply on search results which are:

$results->getCount(); // Returns search results numbers
$results->getDocuments(); // Returns search results object or array

Delete an index

$index->drop($deleteHashes); // This method accepts one param, if it set, the underlying HASHes will be deleted as well.

Notes:

NUMBERIC, TAG and GEO fields only can be used as filter and matching does not work on them.

Todos:

Add support for suggestion (auto-complete).
Implement individual document deletion
Implement filter on index
Implement document updating