[String] Introduce a locale-aware Slugger in the String component with FrameworkBundle wiring

Author: tgalopin

AbstractUnicodeString.php

@@ -331,18 +331,6 @@ public function replaceMatches(string $fromPattern, $to): parent
         return $str;
     }
 
-    /**
-     * @return static
-     */
-    public function slug(string $separator = '-'): self
-    {
-        return $this
-            ->ascii()
-            ->replace('@', $separator.'at'.$separator)
-            ->replaceMatches('/[^A-Za-z0-9]++/', $separator)
-            ->trim($separator);
-    }
-
     public function snake(): parent
     {
         $str = $this->camel()->title();

Slugger/AsciiSlugger.php

@@ -0,0 +1,136 @@
+<?php
+
+/*
+ * This file is part of the Symfony package.
+ *
+ * (c) Fabien Potencier <fabien@symfony.com>
+ *
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ */
+
+namespace Symfony\Component\String\Slugger;
+
+use Symfony\Component\String\AbstractUnicodeString;
+use Symfony\Component\String\GraphemeString;
+use Symfony\Contracts\Translation\LocaleAwareInterface;
+
+/**
+ * @author Titouan Galopin <galopintitouan@gmail.com>
+ *
+ * @experimental in 5.0
+ */
+class AsciiSlugger implements SluggerInterface, LocaleAwareInterface
+{
+    private const LOCALE_TO_TRANSLITERATOR_ID = [
+        'am' => 'Amharic-Latin',
+        'ar' => 'Arabic-Latin',
+        'az' => 'Azerbaijani-Latin',
+        'be' => 'Belarusian-Latin',
+        'bg' => 'Bulgarian-Latin',
+        'bn' => 'Bengali-Latin',
+        'de' => 'de-ASCII',
+        'el' => 'Greek-Latin',
+        'fa' => 'Persian-Latin',
+        'he' => 'Hebrew-Latin',
+        'hy' => 'Armenian-Latin',
+        'ka' => 'Georgian-Latin',
+        'kk' => 'Kazakh-Latin',
+        'ky' => 'Kirghiz-Latin',
+        'ko' => 'Korean-Latin',
+        'mk' => 'Macedonian-Latin',
+        'mn' => 'Mongolian-Latin',
+        'or' => 'Oriya-Latin',
+        'ps' => 'Pashto-Latin',
+        'ru' => 'Russian-Latin',
+        'sr' => 'Serbian-Latin',
+        'sr_Cyrl' => 'Serbian-Latin',
+        'th' => 'Thai-Latin',
+        'tk' => 'Turkmen-Latin',
+        'uk' => 'Ukrainian-Latin',
+        'uz' => 'Uzbek-Latin',
+        'zh' => 'Han-Latin',
+    ];
+
+    private $defaultLocale;
+
+    /**
+     * Cache of transliterators per locale.
+     *
+     * @var \Transliterator[]
+     */
+    private $transliterators = [];
+
+    public function __construct(string $defaultLocale = null)
+    {
+        $this->defaultLocale = $defaultLocale;
+    }
+
+    /**
+     * {@inheritdoc}
+     */
+    public function setLocale($locale)
+    {
+        $this->defaultLocale = $locale;
+    }
+
+    /**
+     * {@inheritdoc}
+     */
+    public function getLocale()
+    {
+        return $this->defaultLocale;
+    }
+
+    /**
+     * {@inheritdoc}
+     */
+    public function slug(string $string, string $separator = '-', string $locale = null): AbstractUnicodeString
+    {
+        $locale = $locale ?? $this->defaultLocale;
+
+        $transliterator = [];
+        if ('de' === $locale || 0 === strpos($locale, 'de_')) {
+            // Use the shortcut for German in GraphemeString::ascii() if possible (faster and no requirement on intl)
+            $transliterator = ['de-ASCII'];
+        } elseif (\function_exists('transliterator_transliterate') && $locale) {
+            $transliterator = (array) $this->createTransliterator($locale);
+        }
+
+        return (new GraphemeString($string))
+            ->ascii($transliterator)
+            ->replace('@', $separator.'at'.$separator)
+            ->replaceMatches('/[^A-Za-z0-9]++/', $separator)
+            ->trim($separator)
+        ;
+    }
+
+    private function createTransliterator(string $locale): ?\Transliterator
+    {
+        if (isset($this->transliterators[$locale])) {
+            return $this->transliterators[$locale];
+        }
+
+        // Exact locale supported, cache and return
+        if ($id = self::LOCALE_TO_TRANSLITERATOR_ID[$locale] ?? null) {
+            return $this->transliterators[$locale] = \Transliterator::create($id.'/BGN') ?? \Transliterator::create($id);
+        }
+
+        // Locale not supported and no parent, fallback to any-latin
+        if (false === $str = strrchr($locale, '_')) {
+            return null;
+        }
+
+        // Try to use the parent locale (ie. try "de" for "de_AT") and cache both locales
+        $parent = substr($locale, 0, -\strlen($str));
+
+        if ($id = self::LOCALE_TO_TRANSLITERATOR_ID[$parent] ?? null) {
+            $transliterator = \Transliterator::create($id.'/BGN') ?? \Transliterator::create($id);
+            $this->transliterators[$locale] = $this->transliterators[$parent] = $transliterator;
+
+            return $transliterator;
+        }
+
+        return null;
+    }
+}

Slugger/SluggerInterface.php

@@ -0,0 +1,29 @@
+<?php
+
+/*
+ * This file is part of the Symfony package.
+ *
+ * (c) Fabien Potencier <fabien@symfony.com>
+ *
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ */
+
+namespace Symfony\Component\String\Slugger;
+
+use Symfony\Component\String\AbstractUnicodeString;
+
+/**
+ * Creates a URL-friendly slug from a given string.
+ *
+ * @author Titouan Galopin <galopintitouan@gmail.com>
+ *
+ * @experimental in 5.0
+ */
+interface SluggerInterface
+{
+    /**
+     * Creates a slug for the given string and locale, using appropriate transliteration when needed.
+     */
+    public function slug(string $string, string $separator = '-', string $locale = null): AbstractUnicodeString;
+}

Tests/SluggerTest.php

@@ -0,0 +1,49 @@
+<?php
+
+/*
+ * This file is part of the Symfony package.
+ *
+ * (c) Fabien Potencier <fabien@symfony.com>
+ *
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ */
+
+namespace Symfony\Component\String\Tests;
+
+use PHPUnit\Framework\TestCase;
+use Symfony\Component\String\Slugger\AsciiSlugger;
+
+class SluggerTest extends TestCase
+{
+    /**
+     * @requires extension intl
+     * @dataProvider provideSlug
+     */
+    public function testSlug(string $string, string $locale, string $expectedSlug)
+    {
+        $slugger = new AsciiSlugger($locale);
+
+        $this->assertSame($expectedSlug, (string) $slugger->slug($string));
+    }
+
+    public static function provideSlug(): array
+    {
+        return [
+            ['Стойността трябва да бъде лъжа', 'bg', 'Stoinostta-tryabva-da-bude-luzha'],
+            ['Dieser Wert sollte größer oder gleich', 'de', 'Dieser-Wert-sollte-groesser-oder-gleich'],
+            ['Dieser Wert sollte größer oder gleich', 'de_AT', 'Dieser-Wert-sollte-groesser-oder-gleich'],
+            ['Αυτή η τιμή πρέπει να είναι ψευδής', 'el', 'Avti-i-timi-prepi-na-inai-psevdhis'],
+            ['该变量的值应为', 'zh', 'gai-bian-liang-de-zhi-ying-wei'],
+            ['該變數的值應為', 'zh_TW', 'gai-bian-shu-de-zhi-ying-wei'],
+        ];
+    }
+
+    public function testSeparatorWithoutLocale()
+    {
+        $slugger = new AsciiSlugger();
+
+        $this->assertSame('hello-world', (string) $slugger->slug('hello world'));
+        $this->assertSame('hello_world', (string) $slugger->slug('hello world', '_'));
+    }
+}