Merge pull request #767 from afoucret/fix-synonyms-parsing

Fix synonyms handling when number of words is > 2 or contains an hyphen.
Smile-SA · Feb 20, 2018 · 7fc64b7 · 7fc64b7
2 parents 30de2ce + e470298
commit 7fc64b7
Show file tree

Hide file tree

Showing 2 changed files with 19 additions and 5 deletions.
diff --git a/src/module-elasticsuite-thesaurus/Model/Index.php b/src/module-elasticsuite-thesaurus/Model/Index.php
@@ -39,7 +39,12 @@ class Index
     /**
      * @var string
      */
-    const WORD_DELIMITER = '-';
+    const WORD_DELIMITER = '_';
+
+    /**
+     * @var integer
+     */
+    const MAX_SIZE = 10;
 
     /**
      * @var \Elasticsearch\Client
@@ -218,6 +223,7 @@ private function getSynonymRewrites($storeId, $queryText, $type)
         foreach ($analysis['tokens'] as $token) {
             if ($token['type'] == 'SYNONYM') {
                 $positionKey = sprintf('%s_%s', $token['start_offset'], $token['end_offset']);
+                $token['token'] = str_replace('_', ' ', $token['token']);
                 $synonymByPositions[$positionKey][] = $token;
             }
         }

diff --git a/src/module-elasticsuite-thesaurus/Model/Indexer/IndexHandler.php b/src/module-elasticsuite-thesaurus/Model/Indexer/IndexHandler.php
@@ -107,9 +107,15 @@ private function getIndexSettings($synonyms, $expansions)
         ];
 
         $settings['analysis']['filter']['shingle'] = [
-            'type' => 'shingle',
-            'output_false' => true,
-            'token_separator' => ThesaurusIndex::WORD_DELIMITER,
+            'type'             => 'shingle',
+            'output_false'     => true,
+            'token_separator'  => ThesaurusIndex::WORD_DELIMITER,
+            'max_shingle_size' => ThesaurusIndex::MAX_SIZE,
+        ];
+
+        $settings['analysis']['filter']['type_filter'] = [
+            'type' => 'keep_types',
+            'types' => [ "SYNONYM" ],
         ];
 
         $settings = $this->addAnalyzerSettings($settings, 'synonym', $synonyms);
@@ -140,6 +146,8 @@ private function addAnalyzerSettings($settings, $type, $values)
             $settings['analysis']['analyzer'][$type]['filter'][] = $type;
         }
 
+        $settings['analysis']['analyzer'][$type]['filter'][] = 'type_filter';
+
         return $settings;
     }
 
@@ -154,7 +162,7 @@ private function addAnalyzerSettings($settings, $type, $values)
     private function prepareSynonymFilterData($rows)
     {
         $rowMaper = function ($row) {
-            return preg_replace('/([\w])\s(?=[\w])/', '\1-', $row);
+            return preg_replace('/([\w])[\s-](?=[\w])/', '\1_', $row);
         };
 
         return array_map($rowMaper, $rows);