[Core] No edge ngram decomposition in termVectors

To avoid false 'exact' matches on leading parts of a misspelled word. The logic is similar to the use of a distinct 'search_analyzer' on fields using 'standard_edge_ngram' as 'analyzer'.
Smile-SA · Oct 8, 2024 · e6a1acb · e6a1acb
1 parent b79e5dc
commit e6a1acb
Show file tree

Hide file tree

Showing 2 changed files with 15 additions and 0 deletions.
diff --git a/src/module-elasticsuite-core/Search/Adapter/Elasticsuite/Spellchecker.php b/src/module-elasticsuite-core/Search/Adapter/Elasticsuite/Spellchecker.php
@@ -164,6 +164,7 @@ private function getTermVectors(RequestInterface $request)
                 MappingInterface::DEFAULT_SPELLING_FIELD => $request->getQueryText(),
             ],
         ];
+        $perFieldAnalyzer = [];
 
         if ($request->isUsingReference()) {
             $doc['fields'][] = MappingInterface::DEFAULT_REFERENCE_FIELD . "." . FieldInterface::ANALYZER_REFERENCE;
@@ -172,9 +173,15 @@ private function getTermVectors(RequestInterface $request)
 
         if ($request->isUsingEdgeNgram()) {
             $doc['fields'][] = MappingInterface::DEFAULT_EDGE_NGRAM_FIELD . "." . FieldInterface::ANALYZER_EDGE_NGRAM;
+            $perFieldAnalyzer[MappingInterface::DEFAULT_EDGE_NGRAM_FIELD . "." . FieldInterface::ANALYZER_EDGE_NGRAM]
+                = FieldInterface::ANALYZER_STANDARD;
             $doc['doc'][MappingInterface::DEFAULT_EDGE_NGRAM_FIELD] = $request->getQueryText();
         }
 
+        if (!empty($perFieldAnalyzer)) {
+            $doc['per_field_analyzer'] = $perFieldAnalyzer;
+        }
+
         $docs = [];
 
         // Compute the mtermvector query on all indices.

diff --git a/src/module-elasticsuite-core/Test/Unit/Search/Adapter/Elasticsuite/SpellcheckerTest.php b/src/module-elasticsuite-core/Test/Unit/Search/Adapter/Elasticsuite/SpellcheckerTest.php
@@ -184,6 +184,10 @@ public function testEdgeNgramTermVectorsParams()
                             MappingInterface::DEFAULT_SEARCH_FIELD . "." . FieldInterface::ANALYZER_WHITESPACE,
                             MappingInterface::DEFAULT_EDGE_NGRAM_FIELD . "." . FieldInterface::ANALYZER_EDGE_NGRAM,
                         ],
+                        'per_field_analyzer' => [
+                            MappingInterface::DEFAULT_EDGE_NGRAM_FIELD . "." . FieldInterface::ANALYZER_EDGE_NGRAM
+                                => FieldInterface::ANALYZER_STANDARD,
+                        ],
                         'doc'             => [
                             MappingInterface::DEFAULT_SEARCH_FIELD   => $queryText,
                             MappingInterface::DEFAULT_SPELLING_FIELD => $queryText,
@@ -243,6 +247,10 @@ public function testReferenceAndEdgeNgramTermVectorsParams()
                             MappingInterface::DEFAULT_REFERENCE_FIELD . "." . FieldInterface::ANALYZER_REFERENCE,
                             MappingInterface::DEFAULT_EDGE_NGRAM_FIELD . "." . FieldInterface::ANALYZER_EDGE_NGRAM,
                         ],
+                        'per_field_analyzer' => [
+                            MappingInterface::DEFAULT_EDGE_NGRAM_FIELD . "." . FieldInterface::ANALYZER_EDGE_NGRAM
+                            => FieldInterface::ANALYZER_STANDARD,
+                        ],
                         'doc'             => [
                             MappingInterface::DEFAULT_SEARCH_FIELD   => $queryText,
                             MappingInterface::DEFAULT_SPELLING_FIELD => $queryText,