Merge pull request #1 from luyadev/word-highlight

transliterate
luyadev · Mar 17, 2021 · dfe04f0 · dfe04f0
2 parents 7967443 + 6870a6a
commit dfe04f0
Show file tree

Hide file tree

Showing 3 changed files with 64 additions and 18 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -2,6 +2,10 @@
 
 All notable changes to this project will be documented in this file. This project adheres to [Semantic Versioning](http://semver.org/).
 
+## 1.1.0 (17. March 2021)
+
++ [#1](https://github.com/luyadev/yii-helpers/pull/1) Highlight word function works with text transliteration.
+
 ## 1.0.0 (2. February 2021)
 
 + First stable release
diff --git a/src/helpers/StringHelper.php b/src/helpers/StringHelper.php
@@ -295,38 +295,69 @@ public static function truncateMiddle($content, $word, $length, $affix = '..', $
     /**
      * Highlight a word within a content.
      *
-     * Since version 1.0.14 an array of words to highlight is possible.
+     * Since version 1.0.14 its possible to provide an array with words to highlight
      *
      * > This function IS NOT case sensitive!
-     *
-     *
+     * 
+     * ```php
+     * StringHelper::highlightWord('Hello John!', 'john');
+     * ```
+     * 
+     * The above example would return `Hello <b>John</b>!`.
      *
      * @param string $content The content to find the word.
      * @param string $word The word to find within the content.
      * @param string $markup The markup used wrap the word to highlight.
      */
     public static function highlightWord($content, $word, $markup = '<b>%s</b>')
     {
-        $word = (array) $word;
-        $content = strip_tags($content);
-        $latest = null;
-        foreach ($word as $needle) {
-            preg_match_all("/".preg_quote($needle, '/')."+/i", $content, $matches);
-            if (is_array($matches[0]) && count($matches[0]) >= 1) {
-                foreach ($matches[0] as $match) {
-                    // ensure if a word is found twice we don't replace again.
-                    if ($latest === $match) {
-                        continue;
-                    }
-                    $content = str_replace($match, sprintf($markup, $match), $content);
-                    $latest = $match;
-                }
+        $transliterateContent = Inflector::transliterate($content);
+        $highlights = [];
+        foreach ((array) $word as $word) {
+            // search in content
+            preg_match_all("/".preg_quote($word, '/')."+/i", $content, $matches);
+            foreach ($matches[0] as $word) {
+                $highlights[] = $word;
+                // if the word is covered already, do not process further in foreach and break here
+                break;
             }
+
+            // search in transliated content if not yet breaked from previous results
+            preg_match_all("/".preg_quote($word, '/')."+/i", $transliterateContent, $matches);
+            foreach ($matches[0] as $word) {
+                $highlights[] = self::sliceTransliteratedWord($word, $transliterateContent, $content);
+            }
+        }
+
+        foreach (array_unique($highlights) as $highlight) {
+            $content = str_replace($highlight, sprintf($markup, $highlight), $content);
         }
 
         return $content;
     }
 
+    /**
+     * Search a word within a transliterated text and cut out the original word in the original text.
+     * 
+     * For example when you search for the transliaterad word in text and want to return the original:
+     * 
+     * ```php
+     * StringHelper::sliceTransliteratedWord('frederic', 'Hello frederic', 'Hello fréderic');
+     * ```
+     * 
+     * The above example would return `fréderic`
+     *
+     * @param string $word
+     * @param string $transliteratedText
+     * @param string $originalText
+     * @return string
+     * @since 1.1.0
+     */
+    public static function sliceTransliteratedWord($word, $transliteratedText, $originalText)
+    {
+        return mb_substr($originalText, mb_strpos($transliteratedText, $word), mb_strlen($word));
+    }
+
     /**
      * Multibyte-safe str_split funciton.
      *

diff --git a/tests/helpers/StringHelperTest.php b/tests/helpers/StringHelperTest.php
@@ -139,7 +139,6 @@ public function testCut()
         $this->assertSame('..oo ABC ba..', StringHelper::truncateMiddle('foo ABC bar', 'abc', 3));
         $this->assertSame('foo abc..', StringHelper::truncateMiddle('foo abc bar', 'notfound', 4));
     }
-
     public function testHighlightWord()
     {
         $this->assertSame('foo <b>bar</b> foo', StringHelper::highlightWord('foo bar foo', 'bar'));
@@ -155,6 +154,18 @@ public function testHighlightWord()
 
         $this->assertStringContainsString(' unsere <b>kompetent</b>en und <b>motiviert</b>en', StringHelper::highlightWord($str, ['Kompetent', 'Motiviert']));
         $this->assertStringContainsString('vor <b>Ort</b> angebot', StringHelper::highlightWord($str, 'ort'));
+        $this->assertSame('test <b>référents</b>', StringHelper::highlightWord('test référents', 'referents'));
+        $this->assertSame('test <b>référents</b>', StringHelper::highlightWord('test référents', 'REFERENTS'));
+        $this->assertSame('test <b>référents</b>', StringHelper::highlightWord('test référents', 'référents'));
+        $this->assertSame('test <b>référents</b>', StringHelper::highlightWord('test référents', 'Référents'));
+        $this->assertSame('<b>foo</b> bar <b>foo</b> bar', StringHelper::highlightWord('foo bar foo bar', 'foo'));
+        $this->assertSame('<b>féé</b> bar <b>fee</b> bar', StringHelper::highlightWord('féé bar fee bar', 'fee'));
+    }
+
+    public function testTransliteratedWordSlice()
+    {
+        $this->assertSame('foo', StringHelper::sliceTransliteratedWord('foo', 'hello foo bar', 'hello foo bar'));
+        $this->assertSame('féé', StringHelper::sliceTransliteratedWord('fee', 'hello fee bar', 'hello féé bar'));
     }
 
     public function testCutAndHighlightWord()