Merge pull request #73 from alexander-nitsche/feature-handle-hyphenat…

…ions-case-insensitive FEATURE: Use `\hyphenations` case-insensitive (like `\patterns`)
vanderlee · May 25, 2023 · 8ee452b · 8ee452b
2 parents d47fe93 + 774ad9a
commit 8ee452b
Show file tree

Hide file tree

Showing 2 changed files with 42 additions and 5 deletions.
diff --git a/src/Syllable.php b/src/Syllable.php
@@ -748,19 +748,51 @@ private function parseWord($word)
     {
         $wordLength = mb_strlen($word);
 
-        // Is this word smaller than the minimal length requirement?
         if ($wordLength < $this->minHyphenLeft + $this->minHyphenRight
             || $wordLength < $this->minWordLength) {
             return [$word];
         }
 
-        // Is it a pre-hyphenated word?
-        if (isset($this->hyphenation[$word])) {
-            return mb_split('-', $this->hyphenation[$word]);
+        $wordLowerCased = mb_strtolower($word);
+
+        if (isset($this->hyphenation[$wordLowerCased])) {
+            return $this->parseWordByHyphenation($word, $wordLowerCased);
+        } else {
+            return $this->parseWordByPatterns($word, $wordLength, $wordLowerCased);
         }
+    }
+
+    private function parseWordByHyphenation($word, $wordLowerCased = null)
+    {
+        $wordLowerCased = $wordLowerCased ?: mb_strtolower($word);
+
+        $hyphenation = $this->hyphenation[$wordLowerCased];
+        $hyphenationLength = mb_strlen($hyphenation);
+
+        $parts = [];
+        $part = '';
+        for ($i = 0, $j = 0; $i < $hyphenationLength; $i++) {
+            if (mb_substr($hyphenation, $i, 1) !== '-') {
+                $part .= mb_substr($word, $j++, 1);
+            } else {
+                $parts[] = $part;
+                $part = '';
+            }
+        }
+        if (!empty($part)) {
+            $parts[] = $part;
+        }
+
+        return $parts;
+    }
+
+    private function parseWordByPatterns($word, $wordLength = 0, $wordLowerCased = null)
+    {
+        $wordLength = $wordLength > 0 ? $wordLength : mb_strlen($word);
+        $wordLowerCased = $wordLowerCased ?: mb_strtolower($word);
 
         // Convenience array
-        $text = '.'.mb_strtolower($word).'.';
+        $text = '.'.$wordLowerCased.'.';
         $textLength = $wordLength + 2;
         $patternLength = $this->maxPattern < $textLength
             ? $this->maxPattern

diff --git a/tests/src/SyllableTest.php b/tests/src/SyllableTest.php
@@ -638,8 +638,13 @@ public function testCaseInsensitivity()
     {
         $this->object->setHyphen('-');
 
+        // Patterns
         $this->assertEquals(['IN', 'EX', 'PLIC', 'A', 'BLE'], $this->object->splitText('INEXPLICABLE'));
         $this->assertEquals(['in', 'ex', 'plic', 'a', 'ble'], $this->object->splitText('inexplicable'));
+
+        // Hyphenations
+        $this->assertEquals(['as', 'so', 'ciate'], $this->object->splitText('associate'));
+        $this->assertEquals(['AS', 'SO', 'CIATE'], $this->object->splitText('ASSOCIATE'));
     }
 
     /**