From ae6cfde22c95e1188ad8b9aa503e0f25cdaabdc9 Mon Sep 17 00:00:00 2001
From: Romain Ruaud <romain.ruaud@smile.fr>
Date: Tue, 25 Apr 2023 17:36:20 +0200
Subject: [PATCH] Improve MLT query.

---
 .../Request/Query/Builder/MoreLikeThis.php    |  2 +
 .../Search/Request/Query/MoreLikeThis.php     | 56 +++++++++++++++++--
 2 files changed, 53 insertions(+), 5 deletions(-)

diff --git a/src/module-elasticsuite-core/Search/Adapter/Elasticsuite/Request/Query/Builder/MoreLikeThis.php b/src/module-elasticsuite-core/Search/Adapter/Elasticsuite/Request/Query/Builder/MoreLikeThis.php
index da88a5aea..0d2dbbb3f 100644
--- a/src/module-elasticsuite-core/Search/Adapter/Elasticsuite/Request/Query/Builder/MoreLikeThis.php
+++ b/src/module-elasticsuite-core/Search/Adapter/Elasticsuite/Request/Query/Builder/MoreLikeThis.php
@@ -45,6 +45,8 @@ public function buildQuery(QueryInterface $query)
             'min_doc_freq'         => $query->getMinDocFreq(),
             'max_doc_freq'         => $query->getMaxDocFreq(),
             'max_query_terms'      => $query->getMaxQueryTerms(),
+            'min_word_length'      => $query->getMinWordLength(),
+            'max_word_length'      => $query->getMaxWordLength(),
             'include'              => $query->includeOriginalDocs(),
         ];
 
diff --git a/src/module-elasticsuite-core/Search/Request/Query/MoreLikeThis.php b/src/module-elasticsuite-core/Search/Request/Query/MoreLikeThis.php
index 3c5449729..812880682 100644
--- a/src/module-elasticsuite-core/Search/Request/Query/MoreLikeThis.php
+++ b/src/module-elasticsuite-core/Search/Request/Query/MoreLikeThis.php
@@ -53,7 +53,17 @@ class MoreLikeThis implements QueryInterface
     /**
      * @var integer
      */
-    const DEFAULT_MAX_DOC_FREQ = 100;
+    const DEFAULT_MAX_DOC_FREQ = 2147483647;
+
+    /**
+     * @var integer
+     */
+    const DEFAULT_MIN_WORD_LENGTH = 0;
+
+    /**
+     * @var integer
+     */
+    const DEFAULT_MAX_WORD_LENGTH = 0;
 
     /**
      * @var string
@@ -100,6 +110,16 @@ class MoreLikeThis implements QueryInterface
      */
     private $maxDocFreq;
 
+    /**
+     * @var integer
+     */
+    private $minWordLength;
+
+    /**
+     * @var integer
+     */
+    private $maxWordLength;
+
     /**
      * @var integer
      */
@@ -124,6 +144,8 @@ class MoreLikeThis implements QueryInterface
      * @param integer      $minDocFreq          Minimum doc freq for a term to be considered.
      * @param integer      $maxDocFreq          Maximum doc freq for a term to be considered.
      * @param integer      $maxQueryTerms       Maximum number of term in generated queries.
+     * @param integer      $minWordLength       Minimum length of word to consider.
+     * @param integer      $maxWordLength       Maximum length of word to consider.
      * @param integer      $includeOriginalDocs Include original doc in the result set.
      * @param string       $name                Query name.
      * @param integer      $boost               Query boost.
@@ -137,6 +159,8 @@ public function __construct(
         $minDocFreq = self::DEFAULT_MIN_DOC_FREQ,
         $maxDocFreq = self::DEFAULT_MAX_DOC_FREQ,
         $maxQueryTerms = self::DEFAULT_MAX_QUERY_TERMS,
+        $minWordLength = self::DEFAULT_MIN_WORD_LENGTH,
+        $maxWordLength = self::DEFAULT_MAX_WORD_LENGTH,
         $includeOriginalDocs = false,
         $name = null,
         $boost = QueryInterface::DEFAULT_BOOST_VALUE
@@ -152,6 +176,8 @@ public function __construct(
         $this->name                = $name;
         $this->boost               = $boost;
         $this->includeOriginalDocs = $includeOriginalDocs;
+        $this->minWordLength       = $minWordLength;
+        $this->maxWordLength       = $maxWordLength;
     }
 
     /**
@@ -225,7 +251,7 @@ public function getBoostTerms()
      */
     public function getMinTermFreq()
     {
-        return $this->minTermFreq;
+        return (int) $this->minTermFreq;
     }
 
     /**
@@ -235,7 +261,7 @@ public function getMinTermFreq()
      */
     public function getMinDocFreq()
     {
-        return $this->minDocFreq;
+        return (int) $this->minDocFreq;
     }
 
     /**
@@ -245,7 +271,7 @@ public function getMinDocFreq()
      */
     public function getMaxDocFreq()
     {
-        return $this->maxDocFreq;
+        return (int) $this->maxDocFreq;
     }
 
     /**
@@ -255,7 +281,7 @@ public function getMaxDocFreq()
      */
     public function getMaxQueryTerms()
     {
-        return $this->maxQueryTerms;
+        return (int) $this->maxQueryTerms;
     }
 
     /**
@@ -267,4 +293,24 @@ public function includeOriginalDocs()
     {
         return $this->includeOriginalDocs;
     }
+
+    /**
+     * Minimum doc freq for a term to be considered.
+     *
+     * @return integer
+     */
+    public function getMinWordLength()
+    {
+        return (int) $this->minWordLength;
+    }
+
+    /**
+     * Maximum doc freq for a term to be considered.
+     *
+     * @return integer
+     */
+    public function getMaxWordLength()
+    {
+        return (int) $this->maxWordLength;
+    }
 }