Page MenuHomeWickedGov Phorge
Files F584724

NumericUppercaseCollation.php
No OneTemporary
Actions

Size
3 KB
Referenced Files
None
Subscribers
None

NumericUppercaseCollation.php
View Options

<?php
/**
* @license GPL-2.0-or-later
* @file
*/
use MediaWiki\Language\Language;
use MediaWiki\Languages\LanguageFactory;
/**
* Collation that orders text with numbers "naturally", so that 'Foo 1' < 'Foo 2' < 'Foo 12'.
*
* Note that this only works in terms of sequences of digits, and the behavior for decimal fractions
* or pretty-formatted numbers may be unexpected.
*
* Digits will be based on the wiki's content language settings. If
* you change the content language of a wiki you will need to run
* updateCollation.php --force. Only English (ASCII 0-9) and the
* localized version will be counted. Localized digits from other languages
* or weird unicode digit equivalents (e.g. ４, 𝟜, ⓸ , ⁴, etc) will not count.
*
* @since 1.28
*/
class NumericUppercaseCollation extends UppercaseCollation {
/**
* @var Language How to convert digits (usually the content language)
*/
private $digitTransformLang;
/**
* @param LanguageFactory $languageFactory
* @param string|Language $digitTransformLang How to convert digits.
* For example, if given language "my" than ၇ is treated like 7.
* It is expected that usually this is given the content language.
*/
public function __construct(
LanguageFactory $languageFactory,
$digitTransformLang
) {
$this->digitTransformLang = $digitTransformLang instanceof Language
? $digitTransformLang
: $languageFactory->getLanguage( $digitTransformLang );
parent::__construct( $languageFactory );
}
/** @inheritDoc */
public function getSortKey( $string ) {
$sortkey = parent::getSortKey( $string );
$sortkey = $this->convertDigits( $sortkey );
// For each sequence of digits, insert the digit '0' and then the length of the sequence
// (encoded in two bytes) before it. That's all folks, it sorts correctly now! The '0' ensures
// correct position (where digits would normally sort), then the length will be compared putting
// shorter numbers before longer ones; if identical, then the characters will be compared, which
// generates the correct results for numbers of equal length.
$sortkey = preg_replace_callback( '/\d+/', static function ( $matches ) {
// Strip any leading zeros
$number = ltrim( $matches[0], '0' );
$len = strlen( $number );
// This allows sequences of up to 65536 numeric characters to be handled correctly. One byte
// would allow only for 256, which doesn't feel future-proof.
$prefix = chr( (int)floor( $len / 256 ) ) . chr( $len % 256 );
return '0' . $prefix . $number;
}, $sortkey );
return $sortkey;
}
/**
* Convert localized digits to english digits.
*
* based on Language::parseFormattedNumber but without commas.
*
* @param string $string sortkey to unlocalize digits of
* @return string Sortkey with all localized digits replaced with ASCII digits.
*/
private function convertDigits( $string ) {
$table = $this->digitTransformLang->digitTransformTable();
if ( $table ) {
$table = array_filter( $table );
$flipped = array_flip( $table );
// Some languages seem to also have commas in this table.
$flipped = array_filter( $flipped, 'is_numeric' );
$string = strtr( $string, $flipped );
}
return $string;
}
/** @inheritDoc */
public function getFirstLetter( $string ) {
$convertedString = $this->convertDigits( $string );
if ( preg_match( '/^\d/', $convertedString ) ) {
return wfMessage( 'category-header-numerals' )
->numParams( 0, 9 )
->text();
} else {
return parent::getFirstLetter( $string );
}
}
}

File Metadata

Mime Type
text/x-php
Expires
Thu, Apr 2, 12:58 (16 h, 39 m)
Storage Engine
local-disk
Storage Format
Raw Data
Storage Handle
dc/74/391bd8839fab9c499379df620747
Default Alt Text
NumericUppercaseCollation.php (3 KB)

Event Timeline