Merge pull request #395 from puchsteink/fulltext-search-with-simple-query-string

FEATURE: Support `.simpleQueryStringFulltext()` simple query string query for fulltext search

Author: skurfuerst

Classes/Driver/QueryInterface.php

@@ -81,6 +81,16 @@ public function from(int $size): void;
      */
     public function fulltext(string $searchWord, array $options = []): void;
 
+    /**
+     * Match the search word against the fulltext index using the elasticsearch
+     * [simple query string query](https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-simple-query-string-query.html)
+     *
+     * @param string $searchWord
+     * @param array $options Options to configure the query_string
+     * @return void
+     */
+    public function simpleQueryStringFulltext(string $searchWord, array $options = []): void;
+
     /**
      * Configure Result Highlighting. Only makes sense in combination with fulltext(). By default, highlighting is enabled.
      * It can be disabled by calling "highlight(FALSE)".

Classes/Driver/Version6/Query/FilteredQuery.php

@@ -67,6 +67,20 @@ public function fulltext(string $searchWord, array $options = []): void
         ]);
     }
 
+    /**
+     * {@inheritdoc}
+     */
+    public function simpleQueryStringFulltext(string $searchWord, array $options = []): void
+    {
+        $this->appendAtPath('query.bool.must', [
+            'simple_query_string' => array_merge(
+                $this->queryStringParameters,
+                $options,
+                [ 'query' => $searchWord ]
+            )
+        ]);
+    }
+
     /**
      * {@inheritdoc}
      */

Classes/Eel/ElasticSearchQueryBuilder.php

@@ -672,7 +672,9 @@ public function count(): int
     }
 
     /**
-     * Match the searchword against the fulltext index
+     * Match the searchword against the fulltext index.
+     *
+     * NOTE: Please use {@see simpleQueryStringFulltext} instead, as it is more robust.
      *
      * @param string $searchWord
      * @param array $options Options to configure the query_string, see https://www.elastic.co/guide/en/elasticsearch/reference/7.6/query-dsl-query-string-query.html
@@ -689,6 +691,32 @@ public function fulltext(string $searchWord, array $options = []): QueryBuilderI
         return $this;
     }
 
+    /**
+     * Match the searchword against the fulltext index using the elasticsearch
+     * [simple query string query](https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-simple-query-string-query.html).
+     *
+     * This method has two main benefits over {@see fulltext()}:
+     * - Supports phrase searches like `"Neos and Flow"`, where
+     *   quotes mean "I want this exact phrase" (similar to many search engines).
+     * - do not crash if the user does not enter a fully syntactically valid query
+     *   (invalid query parts are ignored).
+     *
+     * This is exactly what we want for a good search field behavior.
+     *
+     * @param string $searchWord
+     * @param array $options Options to configure the query_string, see https://www.elastic.co/guide/en/elasticsearch/reference/7.6/query-dsl-query-string-query.html
+     * @return QueryBuilderInterface
+     * @api
+     */
+    public function simpleQueryStringFulltext(string $searchWord, array $options = []): QueryBuilderInterface
+    {
+        // We automatically enable result highlighting when doing fulltext searches. It is up to the user to use this information or not use it.
+        $this->request->simpleQueryStringFulltext($searchWord, $options);
+        $this->request->highlight(150, 2);
+
+        return $this;
+    }
+
     /**
      * Adds a prefix filter to the query
      * See: https://www.elastic.co/guide/en/elasticsearch/reference/7.6/query-dsl-prefix-query.html

Configuration/Settings.yaml

@@ -53,7 +53,7 @@ Flowpack:
                   - 'aggregations'
                   - 'suggest'
 
-                # Parameters for the query string query used by the fullText() operation
+                # Parameters for the query string query used by the fullText() and simpleQueryStringFulltext() operation
                 # See https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-query-string-query.html#query-string-multi-field
                 # for all available parameters
                 queryStringParameters:

README.md

@@ -441,18 +441,19 @@ As **value**, the following methods accept a simple type, a node object or a Dat
 |`from(5)`                                                        |Return the results starting from the 6th one|
 |`prefix('propertyName', 'prefix', [clauseType])`                 |Adds a prefix filter on the given field with the given prefix|
 |`geoDistance(propertyName, geoPoint, distance, [clauseType])`.   |Filters documents that include only hits that exists within a specific distance from a geo point.|
-|`fulltext('searchWord', options)`                                |Does a query_string query on the Fulltext index using the searchword and additional [options](https://www.elastic.co/guide/en/elasticsearch/reference/7.6/query-dsl-query-string-query.html) to the query_string|
+|`fulltext('searchWord', options)`                                |Does a query_string query on the Fulltext index using the searchword and additional [options](https://www.elastic.co/guide/en/elasticsearch/reference/7.6/query-dsl-query-string-query.html) to the query_string. Recommendation: **use simpleQueryStringFulltext instead, as it yields better results and is more tolerant to user input**.|
+|`simpleQueryStringFulltext('searchWord', options)`               |Does a simple_query_string query on the Fulltext index using the searchword and additional [options](https://www.elastic.co/guide/en/elasticsearch/reference/8.3/query-dsl-simple-query-string-query.html) to the simple_query_string. Supports phrase matching like `"firstname lastname"` and tolerates broken input without exceptions (in contrast to `fulltext()`)|
 |`highlight(fragmentSize, fragmentCount, noMatchSize, field)`     |Configure result highlighting for every fulltext field individually|
 
 ## Search Result Highlighting
 
-When using the `.fulltext()` oprtator to do a fulltext, **highlight snippets** ([elasticsearch documentation](https://www.elastic.co/guide/en/elasticsearch/reference/7.17/highlighting.html#highlighting)) are automatically queried. By default snippets with 150 characters are queried for all available fulltext fields with a 150 character fallback text. 
+When using the `.fulltext()` or `.simpleQueryStringFulltext()` operator to do a fulltext, **highlight snippets** ([elasticsearch documentation](https://www.elastic.co/guide/en/elasticsearch/reference/7.17/highlighting.html#highlighting)) are automatically queried. By default snippets with 150 characters are queried for all available fulltext fields with a 150 character fallback text. 
 
 To adjust this behavior you can first deactivate the default and then configure highlighting for every field individually:
 
 	Search.highlight(false).highlight(150, 2, 150, 'neos_fulltext.text').highlight(100, 1, 0, 'neos_fulltext.h2')
 
-This deactivates the default highlighting and then queries 2 snipets of 150 characters each from hits in `neos_fulltext.text`, with a fallback to 150 charchters of the beginning of the text if no match was found and additional 100 characters from `neos_fulltext.h2` without a fallback.
+This deactivates the default highlighting and then queries 2 snipets of 150 characters each from hits in `neos_fulltext.text`, with a fallback to 150 characters of the beginning of the text if no match was found and additional 100 characters from `neos_fulltext.h2` without a fallback.
 
 The highlight snippets can be accessed by