Counting two characters together as a word in CJK

[github/wallabag/wallabag.git] / src / Wallabag / CoreBundle / Tools / Utils.php

diff --git a/src/Wallabag/CoreBundle/Tools/Utils.php b/src/Wallabag/CoreBundle/Tools/Utils.php
index eba21c025d71cb86c79874f4c2ca26b978ffac66..e56e251e5ab31aeda4a2451912017e27f8b1c486 100644 (file)
--- a/src/Wallabag/CoreBundle/Tools/Utils.php
+++ b/src/Wallabag/CoreBundle/Tools/Utils.php
@@ -20,15 +20,14 @@ class Utils
     }
 
     /**
-     * For a given text, we calculate reading time for an article
-     * based on 200 words per minute.
+     * For a given text, we calculate reading time for an article based on 200 words per minute.
      *
-     * @param $text
+     * @param string $text
      *
      * @return float
      */
     public static function getReadingTime($text)
     {
-        return floor(count(preg_split('~[^\p{L}\p{N}\']+~u', strip_tags($text))) / 200);
+        return floor(\count(preg_split('~([^\p{L}\p{N}\']+|(\p{Han}|\p{Hiragana}|\p{Katakana}|\p{Hangul}){1,2})~u', strip_tags($text))) / 200);
     }
 }