Skip to content

Commit

Permalink
[5.5][WIP] Sync Str with Stringy 3.0.1 (#18974)
Browse files Browse the repository at this point in the history
* Synced Stringy to 3.0.1

* Add language support for ascii

* Fix Style

* Beautify lang check

* Fix styleci

* Strict check

* Readability changes
  • Loading branch information
fernandobandeira authored and taylorotwell committed Apr 28, 2017
1 parent 75026aa commit 6a2b893
Show file tree
Hide file tree
Showing 2 changed files with 135 additions and 91 deletions.
218 changes: 127 additions & 91 deletions src/Illuminate/Support/Str.php
Original file line number Diff line number Diff line change
Expand Up @@ -33,10 +33,17 @@ class Str
* Transliterate a UTF-8 value to ASCII.
*
* @param string $value
* @param string $language
* @return string
*/
public static function ascii($value)
public static function ascii($value, $language = 'en')
{
$languageSpecific = static::languageSpecificCharsArray($language);

if (! is_null($languageSpecific)) {
$value = str_replace($languageSpecific[0], $languageSpecific[1], $value);
}

foreach (static::charsArray() as $key => $val) {
$value = str_replace($val, $key, $value);
}
Expand Down Expand Up @@ -339,17 +346,21 @@ public static function singular($value)
*
* @param string $title
* @param string $separator
* @param string $language
* @return string
*/
public static function slug($title, $separator = '-')
public static function slug($title, $separator = '-', $language = 'en')
{
$title = static::ascii($title);
$title = static::ascii($title, $language);

// Convert all dashes/underscores into separator
$flip = $separator == '-' ? '_' : '-';

$title = preg_replace('!['.preg_quote($flip).']+!u', $separator, $title);

// Replace @ with the word 'at'
$title = str_replace('@', $separator.'at'.$separator, $title);

// Remove all characters that are not the separator, letters, numbers, or whitespace.
$title = preg_replace('![^'.preg_quote($separator).'\pL\pN\s]+!u', '', mb_strtolower($title));

Expand Down Expand Up @@ -449,7 +460,7 @@ public static function ucfirst($string)
*
* Note: Adapted from Stringy\Stringy.
*
* @see https://github.com/danielstjules/Stringy/blob/2.3.1/LICENSE.txt
* @see https://github.com/danielstjules/Stringy/blob/3.0.1/LICENSE.txt
*
* @return array
*/
Expand All @@ -462,46 +473,45 @@ protected static function charsArray()
}

return $charsArray = [
'0' => ['°', '', '۰'],
'1' => ['¹', '', '۱'],
'2' => ['²', '', '۲'],
'3' => ['³', '', '۳'],
'4' => ['', '', '۴', '٤'],
'5' => ['', '', '۵', '٥'],
'6' => ['', '', '۶', '٦'],
'7' => ['', '', '۷'],
'8' => ['', '', '۸'],
'9' => ['', '', '۹'],
'a' => ['à', 'á', '', 'ã', '', 'ă', '', '', '', '', '', 'â', '', '', '', '', '', 'ā', 'ą', 'å', 'α', 'ά', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', 'ά', '', '', '', '', '', '', '', 'а', 'أ', '', '', '', 'ǻ', 'ǎ', 'ª', '', '', 'ا'],
'b' => ['б', 'β', 'Ъ', 'Ь', 'ب', '', ''],
'c' => ['ç', 'ć', 'č', 'ĉ', 'ċ'],
'd' => ['ď', 'ð', 'đ', 'ƌ', 'ȡ', 'ɖ', 'ɗ', '', '', '', 'д', 'δ', 'د', 'ض', '', '', ''],
'e' => ['é', 'è', '', '', '', 'ê', 'ế', '', '', '', '', 'ë', 'ē', 'ę', 'ě', 'ĕ', 'ė', 'ε', 'έ', '', '', '', '', '', '', '', 'έ', 'е', 'ё', 'э', 'є', 'ə', '', '', '', '', '', 'إ', 'ئ'],
'f' => ['ф', 'φ', 'ف', 'ƒ', ''],
'g' => ['ĝ', 'ğ', 'ġ', 'ģ', 'г', 'ґ', 'γ', '', '', 'گ'],
'h' => ['ĥ', 'ħ', 'η', 'ή', 'ح', 'ه', '', '', ''],
'i' => ['í', 'ì', '', 'ĩ', '', 'î', 'ï', 'ī', 'ĭ', 'į', 'ı', 'ι', 'ί', 'ϊ', 'ΐ', '', '', '', '', '', '', '', '', '', 'ί', '', '', '', 'ΐ', '', '', 'і', 'ї', 'и', '', '', '', 'ည်', 'ǐ', '', ''],
'j' => ['ĵ', 'ј', 'Ј', '', 'ج'],
'k' => ['ķ', 'ĸ', 'к', 'κ', 'Ķ', 'ق', 'ك', 'က', '', '', 'ک'],
'l' => ['ł', 'ľ', 'ĺ', 'ļ', 'ŀ', 'л', 'λ', 'ل', '', ''],
'm' => ['м', 'μ', 'م', '', ''],
'n' => ['ñ', 'ń', 'ň', 'ņ', 'ʼn', 'ŋ', 'ν', 'н', 'ن', '', ''],
'o' => ['ó', 'ò', '', 'õ', '', 'ô', '', '', '', '', '', 'ơ', '', '', '', '', '', 'ø', 'ō', 'ő', 'ŏ', 'ο', '', '', '', '', '', '', '', 'ό', 'о', 'و', 'θ', 'ို', 'ǒ', 'ǿ', 'º', '', ''],
'p' => ['п', 'π', '', '', 'پ'],
'q' => [''],
'r' => ['ŕ', 'ř', 'ŗ', 'р', 'ρ', 'ر', ''],
's' => ['ś', 'š', 'ş', 'с', 'σ', 'ș', 'ς', 'س', 'ص', '', 'ſ', ''],
't' => ['ť', 'ţ', 'т', 'τ', 'ț', 'ت', 'ط', '', '', 'ŧ', '', ''],
'u' => ['ú', 'ù', '', 'ũ', '', 'ư', '', '', '', '', '', 'û', 'ū', 'ů', 'ű', 'ŭ', 'ų', 'µ', 'у', '', '', '', 'ǔ', 'ǖ', 'ǘ', 'ǚ', 'ǜ', '', ''],
'v' => ['в', '', 'ϐ'],
'w' => ['ŵ', 'ω', 'ώ', '', ''],
'x' => ['χ', 'ξ'],
'y' => ['ý', '', '', '', '', 'ÿ', 'ŷ', 'й', 'ы', 'υ', 'ϋ', 'ύ', 'ΰ', 'ي', ''],
'z' => ['ź', 'ž', 'ż', 'з', 'ζ', 'ز', '', ''],
'0' => ['°', '', '۰', ''],
'1' => ['¹', '', '۱', ''],
'2' => ['²', '', '۲', ''],
'3' => ['³', '', '۳', ''],
'4' => ['', '', '۴', '٤', ''],
'5' => ['', '', '۵', '٥', ''],
'6' => ['', '', '۶', '٦', ''],
'7' => ['', '', '۷', ''],
'8' => ['', '', '۸', ''],
'9' => ['', '', '۹', ''],
'a' => ['à', 'á', '', 'ã', '', 'ă', '', '', '', '', '', 'â', '', '', '', '', '', 'ā', 'ą', 'å', 'α', 'ά', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', 'ά', '', '', '', '', '', '', '', 'а', 'أ', '', '', '', 'ǻ', 'ǎ', 'ª', '', '', 'ا', '', 'ä'],
'b' => ['б', 'β', 'Ъ', 'Ь', 'ب', '', '', ''],
'c' => ['ç', 'ć', 'č', 'ĉ', 'ċ', ''],
'd' => ['ď', 'ð', 'đ', 'ƌ', 'ȡ', 'ɖ', 'ɗ', '', '', '', 'д', 'δ', 'د', 'ض', '', '', '', ''],
'e' => ['é', 'è', '', '', '', 'ê', 'ế', '', '', '', '', 'ë', 'ē', 'ę', 'ě', 'ĕ', 'ė', 'ε', 'έ', '', '', '', '', '', '', '', 'έ', 'е', 'ё', 'э', 'є', 'ə', '', '', '', '', '', 'إ', 'ئ', ''],
'f' => ['ф', 'φ', 'ف', 'ƒ', '', ''],
'g' => ['ĝ', 'ğ', 'ġ', 'ģ', 'г', 'ґ', 'γ', '', '', 'گ', ''],
'h' => ['ĥ', 'ħ', 'η', 'ή', 'ح', 'ه', '', '', '', ''],
'i' => ['í', 'ì', '', 'ĩ', '', 'î', 'ï', 'ī', 'ĭ', 'į', 'ı', 'ι', 'ί', 'ϊ', 'ΐ', '', '', '', '', '', '', '', '', '', 'ί', '', '', '', 'ΐ', '', '', 'і', 'ї', 'и', '', '', '', 'ည်', 'ǐ', '', '', 'ی', ''],
'j' => ['ĵ', 'ј', 'Ј', '', 'ج', ''],
'k' => ['ķ', 'ĸ', 'к', 'κ', 'Ķ', 'ق', 'ك', 'က', '', '', 'ک', ''],
'l' => ['ł', 'ľ', 'ĺ', 'ļ', 'ŀ', 'л', 'λ', 'ل', '', '', ''],
'm' => ['м', 'μ', 'م', '', '', ''],
'n' => ['ñ', 'ń', 'ň', 'ņ', 'ʼn', 'ŋ', 'ν', 'н', 'ن', '', '', ''],
'o' => ['ó', 'ò', '', 'õ', '', 'ô', '', '', '', '', '', 'ơ', '', '', '', '', '', 'ø', 'ō', 'ő', 'ŏ', 'ο', '', '', '', '', '', '', '', 'ό', 'о', 'و', 'θ', 'ို', 'ǒ', 'ǿ', 'º', '', '', '', 'ö'],
'p' => ['п', 'π', '', '', 'پ', ''],
'q' => ['', ''],
'r' => ['ŕ', 'ř', 'ŗ', 'р', 'ρ', 'ر', '', ''],
's' => ['ś', 'š', 'ş', 'с', 'σ', 'ș', 'ς', 'س', 'ص', '', 'ſ', '', ''],
't' => ['ť', 'ţ', 'т', 'τ', 'ț', 'ت', 'ط', '', '', 'ŧ', '', '', ''],
'u' => ['ú', 'ù', '', 'ũ', '', 'ư', '', '', '', '', '', 'û', 'ū', 'ů', 'ű', 'ŭ', 'ų', 'µ', 'у', '', '', '', 'ǔ', 'ǖ', 'ǘ', 'ǚ', 'ǜ', '', '', '', 'ў', 'ü'],
'v' => ['в', '', 'ϐ', ''],
'w' => ['ŵ', 'ω', 'ώ', '', '', ''],
'x' => ['χ', 'ξ', ''],
'y' => ['ý', '', '', '', '', 'ÿ', 'ŷ', 'й', 'ы', 'υ', 'ϋ', 'ύ', 'ΰ', 'ي', '', ''],
'z' => ['ź', 'ž', 'ż', 'з', 'ζ', 'ز', '', '', ''],
'aa' => ['ع', '', 'آ'],
'ae' => ['ä', 'æ', 'ǽ'],
'ae' => ['æ', 'ǽ'],
'ai' => [''],
'at' => ['@'],
'ch' => ['ч', '', '', 'چ'],
'dj' => ['ђ', 'đ'],
'dz' => ['џ', ''],
Expand All @@ -512,7 +522,7 @@ protected static function charsArray()
'kh' => ['х', 'خ', ''],
'lj' => ['љ'],
'nj' => ['њ'],
'oe' => ['ö', 'œ', 'ؤ'],
'oe' => ['œ', 'ؤ'],
'oi' => [''],
'oii' => [''],
'ps' => ['ψ'],
Expand All @@ -522,59 +532,85 @@ protected static function charsArray()
'sx' => ['ŝ'],
'th' => ['þ', 'ϑ', 'ث', 'ذ', 'ظ'],
'ts' => ['ц', '', ''],
'ue' => ['ü'],
'uu' => [''],
'ya' => ['я'],
'yu' => ['ю'],
'zh' => ['ж', '', 'ژ'],
'(c)' => ['©'],
'A' => ['Á', 'À', '', 'Ã', '', 'Ă', '', '', '', '', '', 'Â', '', '', '', '', '', 'Å', 'Ā', 'Ą', 'Α', 'Ά', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', 'Ά', '', 'А', 'Ǻ', 'Ǎ'],
'B' => ['Б', 'Β', ''],
'C' => ['Ç', 'Ć', 'Č', 'Ĉ', 'Ċ'],
'D' => ['Ď', 'Ð', 'Đ', 'Ɖ', 'Ɗ', 'Ƌ', '', '', 'Д', 'Δ'],
'E' => ['É', 'È', '', '', '', 'Ê', '', '', '', '', '', 'Ë', 'Ē', 'Ę', 'Ě', 'Ĕ', 'Ė', 'Ε', 'Έ', '', '', '', '', '', '', 'Έ', '', 'Е', 'Ё', 'Э', 'Є', 'Ə'],
'F' => ['Ф', 'Φ'],
'G' => ['Ğ', 'Ġ', 'Ģ', 'Г', 'Ґ', 'Γ'],
'H' => ['Η', 'Ή', 'Ħ'],
'I' => ['Í', 'Ì', '', 'Ĩ', '', 'Î', 'Ï', 'Ī', 'Ĭ', 'Į', 'İ', 'Ι', 'Ί', 'Ϊ', '', '', '', '', '', '', 'Ἷ', '', '', '', 'Ί', 'И', 'І', 'Ї', 'Ǐ', 'ϒ'],
'K' => ['К', 'Κ'],
'L' => ['Ĺ', 'Ł', 'Л', 'Λ', 'Ļ', 'Ľ', 'Ŀ', ''],
'M' => ['М', 'Μ'],
'N' => ['Ń', 'Ñ', 'Ň', 'Ņ', 'Ŋ', 'Н', 'Ν'],
'O' => ['Ó', 'Ò', '', 'Õ', '', 'Ô', '', '', '', '', '', 'Ơ', '', '', '', '', '', 'Ø', 'Ō', 'Ő', 'Ŏ', 'Ο', 'Ό', '', '', '', '', '', '', '', 'Ό', 'О', 'Θ', 'Ө', 'Ǒ', 'Ǿ'],
'P' => ['П', 'Π'],
'R' => ['Ř', 'Ŕ', 'Р', 'Ρ', 'Ŗ'],
'S' => ['Ş', 'Ŝ', 'Ș', 'Š', 'Ś', 'С', 'Σ'],
'T' => ['Ť', 'Ţ', 'Ŧ', 'Ț', 'Т', 'Τ'],
'U' => ['Ú', 'Ù', '', 'Ũ', '', 'Ư', '', '', '', '', '', 'Û', 'Ū', 'Ů', 'Ű', 'Ŭ', 'Ų', 'У', 'Ǔ', 'Ǖ', 'Ǘ', 'Ǚ', 'Ǜ'],
'V' => ['В'],
'W' => ['Ω', 'Ώ', 'Ŵ'],
'X' => ['Χ', 'Ξ'],
'Y' => ['Ý', '', '', '', '', 'Ÿ', '', '', '', 'Ύ', 'Ы', 'Й', 'Υ', 'Ϋ', 'Ŷ'],
'Z' => ['Ź', 'Ž', 'Ż', 'З', 'Ζ'],
'AE' => ['Ä', 'Æ', 'Ǽ'],
'CH' => ['Ч'],
'DJ' => ['Ђ'],
'DZ' => ['Џ'],
'GX' => ['Ĝ'],
'HX' => ['Ĥ'],
'IJ' => ['IJ'],
'JX' => ['Ĵ'],
'KH' => ['Х'],
'LJ' => ['Љ'],
'NJ' => ['Њ'],
'OE' => ['Ö', 'Œ'],
'PS' => ['Ψ'],
'SH' => ['Ш'],
'SHCH' => ['Щ'],
'SS' => [''],
'TH' => ['Þ'],
'TS' => ['Ц'],
'UE' => ['Ü'],
'YA' => ['Я'],
'YU' => ['Ю'],
'ZH' => ['Ж'],
' ' => ["\xC2\xA0", "\xE2\x80\x80", "\xE2\x80\x81", "\xE2\x80\x82", "\xE2\x80\x83", "\xE2\x80\x84", "\xE2\x80\x85", "\xE2\x80\x86", "\xE2\x80\x87", "\xE2\x80\x88", "\xE2\x80\x89", "\xE2\x80\x8A", "\xE2\x80\xAF", "\xE2\x81\x9F", "\xE3\x80\x80"],
'A' => ['Á', 'À', '', 'Ã', '', 'Ă', '', '', '', '', '', 'Â', '', '', '', '', '', 'Å', 'Ā', 'Ą', 'Α', 'Ά', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', 'Ά', '', 'А', 'Ǻ', 'Ǎ', '', 'Ä'],
'B' => ['Б', 'Β', '', ''],
'C' => ['Ç', 'Ć', 'Č', 'Ĉ', 'Ċ', ''],
'D' => ['Ď', 'Ð', 'Đ', 'Ɖ', 'Ɗ', 'Ƌ', '', '', 'Д', 'Δ', ''],
'E' => ['É', 'È', '', '', '', 'Ê', '', '', '', '', '', 'Ë', 'Ē', 'Ę', 'Ě', 'Ĕ', 'Ė', 'Ε', 'Έ', '', '', '', '', '', '', 'Έ', '', 'Е', 'Ё', 'Э', 'Є', 'Ə', ''],
'F' => ['Ф', 'Φ', ''],
'G' => ['Ğ', 'Ġ', 'Ģ', 'Г', 'Ґ', 'Γ', ''],
'H' => ['Η', 'Ή', 'Ħ', ''],
'I' => ['Í', 'Ì', '', 'Ĩ', '', 'Î', 'Ï', 'Ī', 'Ĭ', 'Į', 'İ', 'Ι', 'Ί', 'Ϊ', '', '', '', '', '', '', 'Ἷ', '', '', '', 'Ί', 'И', 'І', 'Ї', 'Ǐ', 'ϒ', ''],
'J' => [''],
'K' => ['К', 'Κ', ''],
'L' => ['Ĺ', 'Ł', 'Л', 'Λ', 'Ļ', 'Ľ', 'Ŀ', '', ''],
'M' => ['М', 'Μ', ''],
'N' => ['Ń', 'Ñ', 'Ň', 'Ņ', 'Ŋ', 'Н', 'Ν', ''],
'O' => ['Ó', 'Ò', '', 'Õ', '', 'Ô', '', '', '', '', '', 'Ơ', '', '', '', '', '', 'Ø', 'Ō', 'Ő', 'Ŏ', 'Ο', 'Ό', '', '', '', '', '', '', '', 'Ό', 'О', 'Θ', 'Ө', 'Ǒ', 'Ǿ', '', 'Ö'],
'P' => ['П', 'Π', ''],
'Q' => [''],
'R' => ['Ř', 'Ŕ', 'Р', 'Ρ', 'Ŗ', ''],
'S' => ['Ş', 'Ŝ', 'Ș', 'Š', 'Ś', 'С', 'Σ', ''],
'T' => ['Ť', 'Ţ', 'Ŧ', 'Ț', 'Т', 'Τ', ''],
'U' => ['Ú', 'Ù', '', 'Ũ', '', 'Ư', '', '', '', '', '', 'Û', 'Ū', 'Ů', 'Ű', 'Ŭ', 'Ų', 'У', 'Ǔ', 'Ǖ', 'Ǘ', 'Ǚ', 'Ǜ', '', 'Ў', 'Ü'],
'V' => ['В', ''],
'W' => ['Ω', 'Ώ', 'Ŵ', ''],
'X' => ['Χ', 'Ξ', ''],
'Y' => ['Ý', '', '', '', '', 'Ÿ', '', '', '', 'Ύ', 'Ы', 'Й', 'Υ', 'Ϋ', 'Ŷ', ''],
'Z' => ['Ź', 'Ž', 'Ż', 'З', 'Ζ', ''],
'AE' => ['Æ', 'Ǽ'],
'Ch' => ['Ч'],
'Dj' => ['Ђ'],
'Dz' => ['Џ'],
'Gx' => ['Ĝ'],
'Hx' => ['Ĥ'],
'Ij' => ['IJ'],
'Jx' => ['Ĵ'],
'Kh' => ['Х'],
'Lj' => ['Љ'],
'Nj' => ['Њ'],
'Oe' => ['Œ'],
'Ps' => ['Ψ'],
'Sh' => ['Ш'],
'Shch' => ['Щ'],
'Ss' => [''],
'Th' => ['Þ'],
'Ts' => ['Ц'],
'Ya' => ['Я'],
'Yu' => ['Ю'],
'Zh' => ['Ж'],
' ' => ["\xC2\xA0", "\xE2\x80\x80", "\xE2\x80\x81", "\xE2\x80\x82", "\xE2\x80\x83", "\xE2\x80\x84", "\xE2\x80\x85", "\xE2\x80\x86", "\xE2\x80\x87", "\xE2\x80\x88", "\xE2\x80\x89", "\xE2\x80\x8A", "\xE2\x80\xAF", "\xE2\x81\x9F", "\xE3\x80\x80", "\xEF\xBE\xA0"],
];
}

/**
* Returns the language specific replacements for the ascii method.
*
* Note: Adapted from Stringy\Stringy.
*
* @see https://github.com/danielstjules/Stringy/blob/3.0.1/LICENSE.txt
*
* @param string $language
* @return array|null
*/
protected static function languageSpecificCharsArray($language)
{
static $languageSpecific;

if (! isset($languageSpecific)) {
$languageSpecific = [
'de' => [
['ä', 'ö', 'ü', 'Ä', 'Ö', 'Ü'],
['ae', 'oe', 'ue', 'AE', 'OE', 'UE'],
],
];
}

return $languageSpecific[$language] ?? null;
}
}
8 changes: 8 additions & 0 deletions tests/Support/SupportStrTest.php
Original file line number Diff line number Diff line change
Expand Up @@ -38,6 +38,13 @@ public function testStringWithoutWordsDoesntProduceError()
$this->assertEquals($nbsp, Str::words($nbsp));
}

public function testStringAscii()
{
$this->assertEquals('@', Str::ascii('@'));
$this->assertEquals('u', Str::ascii('ü'));
$this->assertEquals('ue', Str::ascii('ü', 'de'));
}

public function testStartsWith()
{
$this->assertTrue(Str::startsWith('jason', 'jas'));
Expand Down Expand Up @@ -99,6 +106,7 @@ public function testSlug()
$this->assertEquals('hello-world', Str::slug('hello-world'));
$this->assertEquals('hello-world', Str::slug('hello_world'));
$this->assertEquals('hello_world', Str::slug('hello_world', '_'));
$this->assertEquals('user-at-host', Str::slug('user@host'));
}

public function testFinish()
Expand Down

0 comments on commit 6a2b893

Please sign in to comment.