[String] New locale aware casing methods

bram123 · fabpot · commit 2172dad98a21 · 2023-11-18T19:32:27.000+01:00
diff --git a/AbstractUnicodeString.php b/AbstractUnicodeString.php
@@ -220,6 +220,21 @@ public function lower(): static
         return $str;
     }
 
+    /**
+     * @param string $locale In the format language_region (e.g. tr_TR)
+     */
+    public function localeLower(string $locale): static
+    {
+        if (null !== $transliterator = $this->getLocaleTransliterator($locale, 'Lower')) {
+            $str = clone $this;
+            $str->string = $transliterator->transliterate($str->string);
+
+            return $str;
+        }
+
+        return $this->lower();
+    }
+
     public function match(string $regexp, int $flags = 0, int $offset = 0): array
     {
         $match = ((\PREG_PATTERN_ORDER | \PREG_SET_ORDER) & $flags) ? 'preg_match_all' : 'preg_match';
@@ -363,6 +378,21 @@ public function title(bool $allWords = false): static
         return $str;
     }
 
+    /**
+     * @param string $locale In the format language_region (e.g. tr_TR)
+     */
+    public function localeTitle(string $locale): static
+    {
+        if (null !== $transliterator = $this->getLocaleTransliterator($locale, 'Title')) {
+            $str = clone $this;
+            $str->string = $transliterator->transliterate($str->string);
+
+            return $str;
+        }
+
+        return $this->title();
+    }
+
     public function trim(string $chars = " \t\n\r\0\x0B\x0C\u{A0}\u{FEFF}"): static
     {
         if (" \t\n\r\0\x0B\x0C\u{A0}\u{FEFF}" !== $chars && !preg_match('//u', $chars)) {
@@ -450,6 +480,21 @@ public function upper(): static
         return $str;
     }
 
+    /**
+     * @param string $locale In the format language_region (e.g. tr_TR)
+     */
+    public function localeUpper(string $locale): static
+    {
+        if (null !== $transliterator = $this->getLocaleTransliterator($locale, 'Upper')) {
+            $str = clone $this;
+            $str->string = $transliterator->transliterate($str->string);
+
+            return $str;
+        }
+
+        return $this->upper();
+    }
+
     public function width(bool $ignoreAnsiDecoration = true): int
     {
         $width = 0;
@@ -587,4 +632,33 @@ private function wcswidth(string $string): int
 
         return $width;
     }
+
+    private function getLocaleTransliterator(string $locale, string $id): ?\Transliterator
+    {
+        $rule = $locale.'-'.$id;
+        if (\array_key_exists($rule, self::$transliterators)) {
+            return self::$transliterators[$rule];
+        }
+
+        if (null !== $transliterator = self::$transliterators[$rule] = \Transliterator::create($rule)) {
+            return $transliterator;
+        }
+
+        // Try to find a parent locale (nl_BE -> nl)
+        if (false === $i = strpos($locale, '_')) {
+            return null;
+        }
+
+        $parentRule = substr_replace($locale, '-'.$id, $i);
+
+        // Parent locale was already cached, return and store as current locale
+        if (\array_key_exists($parentRule, self::$transliterators)) {
+            return self::$transliterators[$rule] = self::$transliterators[$parentRule];
+        }
+
+        // Create transliterator based on parent locale and cache the result on both initial and parent locale values
+        $transliterator = \Transliterator::create($parentRule);
+
+        return self::$transliterators[$rule] = self::$transliterators[$parentRule] = $transliterator;
+    }
 }
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,6 +1,11 @@
 CHANGELOG
 =========
 
+7.1
+---
+
+ * Add `localeLower()`, `localeUpper()`, `localeTitle()` methods to `AbstractUnicodeString`
+
 6.2
 ---
 
diff --git a/Tests/AbstractUnicodeTestCase.php b/Tests/AbstractUnicodeTestCase.php
@@ -50,6 +50,48 @@ public function testAsciiClosureRule()
         $this->assertSame('Dieser Wert sollte grOEsser oder gleich', (string) $s->ascii([$rule]));
     }
 
+    /**
+     * @dataProvider provideLocaleLower
+     *
+     * @requires extension intl
+     */
+    public function testLocaleLower(string $locale, string $expected, string $origin)
+    {
+        $instance = static::createFromString($origin)->localeLower($locale);
+
+        $this->assertNotSame(static::createFromString($origin), $instance);
+        $this->assertEquals(static::createFromString($expected), $instance);
+        $this->assertSame($expected, (string) $instance);
+    }
+
+    /**
+     * @dataProvider provideLocaleUpper
+     *
+     * @requires extension intl
+     */
+    public function testLocaleUpper(string $locale, string $expected, string $origin)
+    {
+        $instance = static::createFromString($origin)->localeUpper($locale);
+
+        $this->assertNotSame(static::createFromString($origin), $instance);
+        $this->assertEquals(static::createFromString($expected), $instance);
+        $this->assertSame($expected, (string) $instance);
+    }
+
+    /**
+     * @dataProvider provideLocaleTitle
+     *
+     * @requires extension intl
+     */
+    public function testLocaleTitle(string $locale, string $expected, string $origin)
+    {
+        $instance = static::createFromString($origin)->localeTitle($locale);
+
+        $this->assertNotSame(static::createFromString($origin), $instance);
+        $this->assertEquals(static::createFromString($expected), $instance);
+        $this->assertSame($expected, (string) $instance);
+    }
+
     public function provideCreateFromCodePoint(): array
     {
         return [
@@ -291,6 +333,78 @@ public static function provideLower(): array
         );
     }
 
+    public static function provideLocaleLower(): array
+    {
+        return [
+            // Lithuanian
+            // Introduce an explicit dot above when lowercasing capital I's and J's
+            // whenever there are more accents above.
+            // LATIN CAPITAL LETTER I WITH OGONEK -> LATIN SMALL LETTER I WITH OGONEK
+            ['lt', 'į', 'Į'],
+            // LATIN CAPITAL LETTER I WITH GRAVE -> LATIN SMALL LETTER I COMBINING DOT ABOVE
+            ['lt', 'i̇̀', 'Ì'],
+            // LATIN CAPITAL LETTER I WITH ACUTE -> LATIN SMALL LETTER I COMBINING DOT ABOVE COMBINING ACUTE ACCENT
+            ['lt', 'i̇́', 'Í'],
+            // LATIN CAPITAL LETTER I WITH TILDE -> LATIN SMALL LETTER I COMBINING DOT ABOVE COMBINING TILDE
+            ['lt', 'i̇̃', 'Ĩ'],
+
+            // Turkish and Azeri
+            // When lowercasing, remove dot_above in the sequence I + dot_above, which will turn into 'i'.
+            // LATIN CAPITAL LETTER I WITH DOT ABOVE -> LATIN SMALL LETTER I
+            ['tr', 'i', 'İ'],
+            ['tr_TR', 'i', 'İ'],
+            ['az', 'i', 'İ'],
+
+            // Default casing rules
+            // LATIN CAPITAL LETTER I WITH DOT ABOVE -> LATIN SMALL LETTER I COMBINING DOT ABOVE
+            ['en_US', 'i̇', 'İ'],
+            ['en', 'i̇', 'İ'],
+        ];
+    }
+
+    public static function provideLocaleUpper(): array
+    {
+        return [
+            // Turkish and Azeri
+            // When uppercasing, i turns into a dotted capital I
+            // LATIN SMALL LETTER I -> LATIN CAPITAL LETTER I WITH DOT ABOVE
+            ['tr', 'İ', 'i'],
+            ['tr_TR', 'İ', 'i'],
+            ['az', 'İ', 'i'],
+
+            // Greek
+            // Remove accents when uppercasing
+            // GREEK SMALL LETTER ALPHA WITH TONOS -> GREEK CAPITAL LETTER ALPHA
+            ['el', 'Α', 'ά'],
+            ['el_GR', 'Α', 'ά'],
+
+            // Default casing rules
+            // GREEK SMALL LETTER ALPHA WITH TONOS -> GREEK CAPITAL LETTER ALPHA WITH TONOS
+            ['en_US', 'Ά', 'ά'],
+            ['en', 'Ά', 'ά'],
+        ];
+    }
+
+    public static function provideLocaleTitle(): array
+    {
+        return [
+            // Greek
+            // Titlecasing words, should keep the accents on the first letter
+            ['el', 'Άδικος', 'άδικος'],
+            ['el_GR', 'Άδικος', 'άδικος'],
+            ['en', 'Άδικος', 'άδικος'],
+
+            // Dutch
+            // Title casing should treat 'ij' as one character
+            ['nl_NL', 'IJssel', 'ijssel'],
+            ['nl_BE', 'IJssel', 'ijssel'],
+            ['nl', 'IJssel', 'ijssel'],
+
+            // Default casing rules
+            ['en', 'Ijssel', 'ijssel'],
+        ];
+    }
+
     public static function provideUpper(): array
     {
         return array_merge(
