LanguageCode.php

Go to the documentation of this file.

<?php
class LanguageCode {
    private const DEPRECATED_LANGUAGE_CODE_MAPPING = [
        // Note that als is actually a valid ISO 639 code (Tosk Albanian), but it
        // was previously used in MediaWiki for Alsatian, which comes under gsw
        'als' => 'gsw', // T25215
        'bat-smg' => 'sgs', // T27522
        'be-x-old' => 'be-tarask', // T11823
        'fiu-vro' => 'vro', // T31186
        'roa-rup' => 'rup', // T17988
        'zh-classical' => 'lzh', // T30443
        'zh-min-nan' => 'nan', // T30442
        'zh-yue' => 'yue', // T30441
    ];
 
    private const NON_STANDARD_LANGUAGE_CODE_MAPPING = [
        // All codes returned by LanguageNameUtils::getLanguageNames() validated
        // against IANA registry at
        //   https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry
        // with help of validator at
        //   http://schneegans.de/lv/
        'cbk-zam' => 'cbk', // T124657
        'de-formal' => 'de-x-formal',
        'eml' => 'egl', // T36217
        'en-rtl' => 'en-x-rtl',
        'es-formal' => 'es-x-formal',
        'hu-formal' => 'hu-x-formal',
        'map-bms' => 'jv-x-bms', // [[en:Banyumasan_dialect]] T125073
        'mo' => 'ro-Cyrl-MD', // T125073
        'nrm' => 'nrf', // [[en:Norman_language]] T25216
        'nl-informal' => 'nl-x-informal',
        'roa-tara' => 'nap-x-tara', // [[en:Tarantino_dialect]]
        'simple' => 'en-simple',
        'sr-ec' => 'sr-Cyrl', // T117845
        'sr-el' => 'sr-Latn', // T117845
 
        // Although these next codes aren't *wrong* per se, including
        // both the script and the country code helps compatibility with
        // other BCP 47 users. Note that MW also uses `zh-Hans`/`zh-Hant`,
        // without a country code, and those should be left alone.
        // (See $variantfallbacks in LanguageZh.php for Hans/Hant id.)
        'zh-cn' => 'zh-Hans-CN',
        'zh-sg' => 'zh-Hans-SG',
        'zh-my' => 'zh-Hans-MY',
        'zh-tw' => 'zh-Hant-TW',
        'zh-hk' => 'zh-Hant-HK',
        'zh-mo' => 'zh-Hant-MO',
    ];
 
    public static function getDeprecatedCodeMapping() {
        return self::DEPRECATED_LANGUAGE_CODE_MAPPING;
    }
 
    public static function getNonstandardLanguageCodeMapping() {
        $result = [];
        foreach ( self::DEPRECATED_LANGUAGE_CODE_MAPPING as $code => $ignore ) {
            $result[$code] = self::bcp47( $code );
        }
        foreach ( self::NON_STANDARD_LANGUAGE_CODE_MAPPING as $code => $ignore ) {
            $result[$code] = self::bcp47( $code );
        }
        return $result;
    }
 
    public static function replaceDeprecatedCodes( $code ) {
        return self::DEPRECATED_LANGUAGE_CODE_MAPPING[$code] ?? $code;
    }
 
    public static function bcp47( $code ) {
        $code = self::replaceDeprecatedCodes( strtolower( $code ) );
        if ( isset( self::NON_STANDARD_LANGUAGE_CODE_MAPPING[$code] ) ) {
            $code = self::NON_STANDARD_LANGUAGE_CODE_MAPPING[$code];
        }
        $codeSegment = explode( '-', $code );
        $codeBCP = [];
        foreach ( $codeSegment as $segNo => $seg ) {
            // when previous segment is x, it is a private segment and should be lc
            if ( $segNo > 0 && strtolower( $codeSegment[( $segNo - 1 )] ) == 'x' ) {
                $codeBCP[$segNo] = strtolower( $seg );
            // ISO 3166 country code
            } elseif ( ( strlen( $seg ) == 2 ) && ( $segNo > 0 ) ) {
                $codeBCP[$segNo] = strtoupper( $seg );
            // ISO 15924 script code
            } elseif ( ( strlen( $seg ) == 4 ) && ( $segNo > 0 ) ) {
                $codeBCP[$segNo] = ucfirst( strtolower( $seg ) );
            // Use lowercase for other cases
            } else {
                $codeBCP[$segNo] = strtolower( $seg );
            }
        }
        $langCode = implode( '-', $codeBCP );
        return $langCode;
    }
 
    public static function bcp47ToInternal( $code ) {
        static $invertedLookup = [];
        if ( !$invertedLookup ) {
            // There should never be two different entries in
            // NON_STANDARD_LANGUAGE_CODE_MAPPING which map *different*
            // internal codes to the same external BCP-47 code.  That is,
            // BCP-47 should preserve all the information from the internal
            // code (discussed further above)[*].  But note the converse isn't
            // true: multiple BCP-47 codes can alias to the same internal code:
            //     BCP-47      internal
            //   zh-Hans-CN => zh-cn    (in NON_STANDARD_LANGUAGE_CODE_MAPPING)
            //   zh-Hans    => zh-hans  (not in " )
            //   zh-CN      => zh-cn    (not in " )
            //
            // [*] eml/egl are the "exception that proves the rule": `egl` *is*
            // (prematurely?) defined as an internal code, but only
            // eml.wikipedia.org exists, and it defines its language as `eml`;
            // for internal purposes `egl` should map back into `eml` until
            // `eml` is deprecated (aka an `eml => egl` entry is added to
            // DEPRECATED_LANGUAGE_CODE_MAPPING): T36217.
            foreach ( self::NON_STANDARD_LANGUAGE_CODE_MAPPING as $internal => $bcp47 ) {
                $invertedLookup[strtolower( $bcp47 )] = $internal;
            }
            // We deliberately do *not* use DEPRECATED_LANGUAGE_CODE_MAPPING
            // here: deprecated codes are no longer valid mediawiki internal
            // codes and we should never return them.
        }
        // Internal codes are all lowercase.  This also achieves
        // case-insensitivity in the lookup.
        $code = strtolower( $code );
        return $invertedLookup[$code] ?? $code;
    }
 
    public static function isWellFormedLanguageTag( string $code, bool $lenient = false ): bool {
        $alpha = '[a-z]';
        $digit = '[0-9]';
        $alphanum = '[a-z0-9]';
        $x = 'x'; # private use singleton
        $singleton = '[a-wy-z]'; # other singleton
        $s = $lenient ? '[-_]' : '-';
 
        $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
        $script = "$alpha{4}"; # ISO 15924
        $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
        $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
        $extension = "$singleton(?:$s$alphanum{2,8})+";
        $privateUse = "$x(?:$s$alphanum{1,8})+";
 
        # Define certain legacy language tags (marked as “Type: grandfathered” in BCP 47),
        # since otherwise the regex is pretty useless.
        # Since these are limited, this is safe even later changes to the registry --
        # the only oddity is that it might change the type of the tag, and thus
        # the results from the capturing groups.
        # https://www.iana.org/assignments/language-subtag-registry
 
        $legacy = "en{$s}GB{$s}oed"
            . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
            . "|no{$s}(?:bok|nyn)"
            . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
            . "|zh{$s}min{$s}nan";
 
        $variantList = "$variant(?:$s$variant)*";
        $extensionList = "$extension(?:$s$extension)*";
 
        $langtag = "(?:($language)"
            . "(?:$s$script)?"
            . "(?:$s$region)?"
            . "(?:$s$variantList)?"
            . "(?:$s$extensionList)?"
            . "(?:$s$privateUse)?)";
 
        # Here is the final breakdown, with capturing groups for each of these components
        # The variants, extensions, legacy, and private-use may have interior '-'
 
        $root = "^(?:$langtag|$privateUse|$legacy)$";
 
        return preg_match( "/$root/", strtolower( $code ) );
    }
}