D3Public/daux.io
8
0
Fork 0
Code Pull-Requests Releases Aktivität

Use transliteration for slug #183

Quellcode durchsuchen
Dieser Commit ist enthalten in:
Stéphane Goetz 2019-12-29 20:41:13 +01:00
Ursprung 21eb4992d3
Commit 695d2db311
4 geänderte Dateien mit 13 neuen und 167 gelöschten Zeilen
Statistiken anzeigen Patch-Datei herunterladen Vergleichs-Datei herunterladen Alle Dateien ausklappen Alle Dateien einklappen

145
libs/DauxHelper.php
Datei anzeigen

@ -224,9 +224,8 @@ class DauxHelper
public static function slug($title)
{
// Convert to ASCII
foreach (static::charsArray() as $key => $value) {
$title = str_replace($value, $key, $title);
}
$title = transliterator_transliterate("Any-Latin; NFD; [:Nonspacing Mark:] Remove; NFC;", $title);
$title = iconv("utf-8", "ASCII//TRANSLIT//IGNORE", $title);
// Remove unsupported characters
$title = preg_replace('/[^\x20-\x7E]/u', '', $title);
@ -245,146 +244,6 @@ class DauxHelper
return trim($title, $separator);
}
/**
* Returns the replacements for the slug() method.
*
* Taken from Stringy
*
* @return array An array of replacements.
*/
public static function charsArray()
{
static $charsArray;
if (isset($charsArray)) {
return $charsArray;
}
return $charsArray = [
'a' => [
'à', 'á', 'ả', 'ã', 'ạ', 'ă', 'ắ', 'ằ', 'ẳ', 'ẵ',
'ặ', 'â', 'ấ', 'ầ', 'ẩ', 'ẫ', 'ậ', 'ä', 'ā', 'ą',
'å', 'α', 'ά', 'ἀ', 'ἁ', 'ἂ', 'ἃ', 'ἄ', 'ἅ', 'ἆ',
'ἇ', 'ᾀ', 'ᾁ', 'ᾂ', 'ᾃ', 'ᾄ', 'ᾅ', 'ᾆ', 'ᾇ', 'ὰ',
'ά', 'ᾰ', 'ᾱ', 'ᾲ', 'ᾳ', 'ᾴ', 'ᾶ', 'ᾷ', 'а', 'أ', ],
'b' => ['б', 'β', 'Ъ', 'Ь', 'ب'],
'c' => ['ç', 'ć', 'č', 'ĉ', 'ċ'],
'd' => ['ď', 'ð', 'đ', 'ƌ', 'ȡ', 'ɖ', 'ɗ', 'ᵭ', 'ᶁ', 'ᶑ',
'д', 'δ', 'د', 'ض', ],
'e' => ['é', 'è', 'ẻ', 'ẽ', 'ẹ', 'ê', 'ế', 'ề', 'ể', 'ễ',
'ệ', 'ë', 'ē', 'ę', 'ě', 'ĕ', 'ė', 'ε', 'έ', 'ἐ',
'ἑ', 'ἒ', 'ἓ', 'ἔ', 'ἕ', 'ὲ', 'έ', 'е', 'ё', 'э',
'є', 'ə', ],
'f' => ['ф', 'φ', 'ف'],
'g' => ['ĝ', 'ğ', 'ġ', 'ģ', 'г', 'ґ', 'γ', 'ج'],
'h' => ['ĥ', 'ħ', 'η', 'ή', 'ح', 'ه'],
'i' => ['í', 'ì', 'ỉ', 'ĩ', 'ị', 'î', 'ï', 'ī', 'ĭ', 'į',
'ı', 'ι', 'ί', 'ϊ', 'ΐ', 'ἰ', 'ἱ', 'ἲ', 'ἳ', 'ἴ',
'ἵ', 'ἶ', 'ἷ', 'ὶ', 'ί', 'ῐ', 'ῑ', 'ῒ', 'ΐ', 'ῖ',
'ῗ', 'і', 'ї', 'и', ],
'j' => ['ĵ', 'ј', 'Ј'],
'k' => ['ķ', 'ĸ', 'к', 'κ', 'Ķ', 'ق', 'ك'],
'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', 'ỏ', 'õ', 'ọ', 'ô', 'ố', 'ồ', 'ổ', 'ỗ',
'ộ', 'ơ', 'ớ', 'ờ', 'ở', 'ỡ', 'ợ', 'ø', 'ō', 'ő',
'ŏ', 'ο', 'ὀ', 'ὁ', 'ὂ', 'ὃ', 'ὄ', 'ὅ', 'ὸ', 'ό',
'ö', 'о', 'و', 'θ', ],
'p' => ['п', 'π'],
'r' => ['ŕ', 'ř', 'ŗ', 'р', 'ρ', 'ر'],
's' => ['ś', 'š', 'ş', 'с', 'σ', 'ș', 'ς', 'س', 'ص'],
't' => ['ť', 'ţ', 'т', 'τ', 'ț', 'ت', 'ط'],
'u' => ['ú', 'ù', 'ủ', 'ũ', 'ụ', 'ư', 'ứ', 'ừ', 'ử', 'ữ',
'ự', 'ü', 'û', 'ū', 'ů', 'ű', 'ŭ', 'ų', 'µ', 'у', ],
'v' => ['в'],
'w' => ['ŵ', 'ω', 'ώ'],
'x' => ['χ'],
'y' => ['ý', 'ỳ', 'ỷ', 'ỹ', 'ỵ', 'ÿ', 'ŷ', 'й', 'ы', 'υ',
'ϋ', 'ύ', 'ΰ', 'ي', ],
'z' => ['ź', 'ž', 'ż', 'з', 'ζ', 'ز'],
'aa' => ['ع'],
'ae' => ['æ'],
'ch' => ['ч'],
'dj' => ['ђ', 'đ'],
'dz' => ['џ'],
'gh' => ['غ'],
'kh' => ['х', 'خ'],
'lj' => ['љ'],
'nj' => ['њ'],
'oe' => ['œ'],
'ps' => ['ψ'],
'sh' => ['ш'],
'shch' => ['щ'],
'ss' => ['ß'],
'th' => ['þ', 'ث', 'ذ', 'ظ'],
'ts' => ['ц'],
'ya' => ['я'],
'yu' => ['ю'],
'zh' => ['ж'],
'(c)' => ['©'],
'A' => ['Á', 'À', 'Ả', 'Ã', 'Ạ', 'Ă', 'Ắ', 'Ằ', 'Ẳ', 'Ẵ',
'Ặ', 'Â', 'Ấ', 'Ầ', 'Ẩ', 'Ẫ', 'Ậ', 'Ä', 'Å', 'Ā',
'Ą', 'Α', 'Ά', 'Ἀ', 'Ἁ', 'Ἂ', 'Ἃ', 'Ἄ', 'Ἅ', 'Ἆ',
'Ἇ', 'ᾈ', 'ᾉ', 'ᾊ', 'ᾋ', 'ᾌ', 'ᾍ', 'ᾎ', 'ᾏ', 'Ᾰ',
'Ᾱ', 'Ὰ', 'Ά', 'ᾼ', 'А', ],
'B' => ['Б', 'Β'],
'C' => ['Ç', 'Ć', 'Č', 'Ĉ', 'Ċ'],
'D' => ['Ď', 'Ð', 'Đ', 'Ɖ', 'Ɗ', 'Ƌ', 'ᴅ', 'ᴆ', 'Д', 'Δ'],
'E' => ['É', 'È', 'Ẻ', 'Ẽ', 'Ẹ', 'Ê', 'Ế', 'Ề', 'Ể', 'Ễ',
'Ệ', 'Ë', 'Ē', 'Ę', 'Ě', 'Ĕ', 'Ė', 'Ε', 'Έ', 'Ἐ',
'Ἑ', 'Ἒ', 'Ἓ', 'Ἔ', 'Ἕ', 'Έ', 'Ὲ', 'Е', 'Ё', 'Э',
'Є', 'Ə', ],
'F' => ['Ф', 'Φ'],
'G' => ['Ğ', 'Ġ', 'Ģ', 'Г', 'Ґ', 'Γ'],
'H' => ['Η', 'Ή'],
'I' => ['Í', 'Ì', 'Ỉ', 'Ĩ', 'Ị', 'Î', 'Ï', 'Ī', 'Ĭ', 'Į',
'İ', 'Ι', 'Ί', 'Ϊ', 'Ἰ', 'Ἱ', 'Ἳ', 'Ἴ', 'Ἵ', 'Ἶ',
'Ἷ', 'Ῐ', 'Ῑ', 'Ὶ', 'Ί', 'И', 'І', 'Ї', ],
'K' => ['К', 'Κ'],
'L' => ['Ĺ', 'Ł', 'Л', 'Λ', 'Ļ'],
'M' => ['М', 'Μ'],
'N' => ['Ń', 'Ñ', 'Ň', 'Ņ', 'Ŋ', 'Н', 'Ν'],
'O' => ['Ó', 'Ò', 'Ỏ', 'Õ', 'Ọ', 'Ô', 'Ố', 'Ồ', 'Ổ', 'Ỗ',
'Ộ', 'Ơ', 'Ớ', 'Ờ', 'Ở', 'Ỡ', 'Ợ', 'Ö', 'Ø', 'Ō',
'Ő', 'Ŏ', 'Ο', 'Ό', 'Ὀ', 'Ὁ', 'Ὂ', 'Ὃ', 'Ὄ', 'Ὅ',
'Ὸ', 'Ό', 'О', 'Θ', 'Ө', ],
'P' => ['П', 'Π'],
'R' => ['Ř', 'Ŕ', 'Р', 'Ρ'],
'S' => ['Ş', 'Ŝ', 'Ș', 'Š', 'Ś', 'С', 'Σ'],
'T' => ['Ť', 'Ţ', 'Ŧ', 'Ț', 'Т', 'Τ'],
'U' => ['Ú', 'Ù', 'Ủ', 'Ũ', 'Ụ', 'Ư', 'Ứ', 'Ừ', 'Ử', 'Ữ',
'Ự', 'Û', 'Ü', 'Ū', 'Ů', 'Ű', 'Ŭ', 'Ų', 'У', ],
'V' => ['В'],
'W' => ['Ω', 'Ώ'],
'X' => ['Χ'],
'Y' => ['Ý', 'Ỳ', 'Ỷ', 'Ỹ', 'Ỵ', 'Ÿ', 'Ῠ', 'Ῡ', 'Ὺ', 'Ύ',
'Ы', 'Й', 'Υ', 'Ϋ', ],
'Z' => ['Ź', 'Ž', 'Ż', 'З', 'Ζ'],
'AE' => ['Æ'],
'CH' => ['Ч'],
'DJ' => ['Ђ'],
'DZ' => ['Џ'],
'KH' => ['Х'],
'LJ' => ['Љ'],
'NJ' => ['Њ'],
'PS' => ['Ψ'],
'SH' => ['Ш'],
'SHCH' => ['Щ'],
'SS' => ['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", ],
];
}
/**
* @param string $from
* @param string $to

18
libs/Format/HTML/ContentTypes/Markdown/TOC/Processor.php
Datei anzeigen

@ -14,6 +14,7 @@ use League\CommonMark\Node\Node;
use ReflectionMethod;
use Todaymade\Daux\Config;
use Todaymade\Daux\ContentTypes\Markdown\TableOfContents;
use Todaymade\Daux\DauxHelper;
class Processor
{
@ -71,21 +72,6 @@ class Processor
}
}
/**
* Get an escaped version of the link
* @param string $url
* @return string
*/
protected function escaped($url) {
$url = trim($url);
$url = preg_replace('~[^\\pL0-9_]+~u', '-', $url);
$url = trim($url, "-");
$url = iconv("utf-8", "ASCII//TRANSLIT//IGNORE", $url);
$url = preg_replace('~[^-a-zA-Z0-9_]+~', '', $url);
return $url;
}
protected function getUniqueId(Document $document, $proposed) {
if ($proposed == "page_") {
$proposed = "page_section_" . (count($document->heading_ids) + 1);
@ -138,7 +124,7 @@ class Processor
}
}
$node->data['attributes']['id'] = $this->getUniqueId($document, 'page_' . $this->escaped($text));
$node->data['attributes']['id'] = $this->getUniqueId($document, 'page_' . DauxHelper::slug($text));
}
/**

12
tests/Format/HTML/TableOfContentsTest.php
Datei anzeigen

@ -57,14 +57,14 @@ EXPECTED;
$expected = <<<EXPECTED
<ul class="TableOfContents">
<li>
<p><a href="#page_section_1">基础操作</a></p>
<p><a href="#page_ji_chu_cao_zuo">基础操作</a></p>
</li>
<li>
<p><a href="#page_section_2">操作基础</a></p>
<p><a href="#page_cao_zuo_ji_chu">操作基础</a></p>
</li>
</ul>
<h1 id="page_section_1">基础操作</h1>
<h1 id="page_section_2">操作基础</h1>
<h1 id="page_ji_chu_cao_zuo">基础操作</h1>
<h1 id="page_cao_zuo_ji_chu">操作基础</h1>
EXPECTED;
@ -99,10 +99,10 @@ EXPECTED;
$expected = <<<EXPECTED
<ul class="TableOfContents">
<li>
<p><a href="#page_TEST-Test">TEST : Test</a></p>
<p><a href="#page_TEST_Test">TEST : Test</a></p>
</li>
</ul>
<h1 id="page_TEST-Test">TEST : Test</h1>
<h1 id="page_TEST_Test">TEST : Test</h1>
EXPECTED;

5
tests/Tree/BuilderTest.php
Datei anzeigen

@ -85,7 +85,7 @@ class BuilderTest extends TestCase
// File, Url, Uri, Title
['A Page.md', 'dir/A_Page.html', 'A_Page.html', 'A Page'],
['Page#1.md', 'dir/Page1.html', 'Page1.html', 'Page#1'],
['你好世界.md', 'dir/你好世界.html', '你好世界.html', '你好世界']
['你好世界.md', 'dir/ni_hao_shi_jie.html', 'ni_hao_shi_jie.html', '你好世界']
];
}
@ -163,6 +163,7 @@ class BuilderTest extends TestCase
$structure = [
'Page.md' => 'another page',
'Button.md' => 'another page',
'你好世界.md' => 'another page',
'22.png' => ''
];
$root = vfsStream::setup('root', null, $structure);
@ -176,7 +177,7 @@ class BuilderTest extends TestCase
Builder::build($tree, []);
$this->assertEquals(
['22.png', 'Button.html', 'Page.html'],
['22.png', 'Button.html', 'Page.html', 'ni_hao_shi_jie.html'],
array_keys($tree->getEntries())
);
}