languages/classes/LanguageCrh.php

   1 <?php
   2 /**
   3  * Crimean Tatar (Qırımtatarca) specific code.
   4  *
   5  * Adapted from https://crh.wikipedia.org/wiki/Qullan%C4%B1c%C4%B1:Don_Alessandro/Translit
   6  *
   7  * This program is free software; you can redistribute it and/or modify
   8  * it under the terms of the GNU General Public License as published by
   9  * the Free Software Foundation; either version 2 of the License, or
  10  * (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  15  * GNU General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU General Public License along
  18  * with this program; if not, write to the Free Software Foundation, Inc.,
  19  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  20  * http://www.gnu.org/copyleft/gpl.html
  21  *
  22  * @file
  23  * @ingroup Language
  24  */
  25
  26 /**
  27  * Crimean Tatar (Qırımtatarca) converter routines
  28  *
  29  * @ingroup Language
  30  */
  31 class CrhConverter extends LanguageConverter {
  32         // Defines working character ranges
  33
  34         // Cyrillic
  35         const C_UC = 'АБВГДЕЁЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЫЬЭЮЯ'; # Crimean Tatar Cyrillic uppercase
  36         const C_LC = 'абвгдеёжзийклмнопрстуфхцчшщъыьэюя'; # Crimean Tatar Cyrillic lowercase
  37         const C_CONS_UC = 'БВГДЖЗЙКЛМНПРСТФХЦЧШЩCÑ'; # Crimean Tatar Cyrillic + CÑ uppercase consonants
  38         const C_CONS_LC = 'бвгджзйклмнпрстфхцчшщcñ'; # Crimean Tatar Cyrillic + CÑ lowercase consonants
  39         const C_M_CONS = 'бгкмшcБГКМШC'; # Crimean Tatar Cyrillic M-type consonants
  40
  41         // Crimean Tatar Cyrillic + CÑ consonants
  42         const C_CONS = 'бвгджзйклмнпрстфхцчшщcñБВГДЖЗЙКЛМНПРСТФХЦЧШЩCÑ';
  43
  44         // Latin
  45         const L_UC = 'AÂBCÇDEFGĞHIİJKLMNÑOÖPQRSŞTUÜVYZ'; # Crimean Tatar Latin uppercase
  46         const L_LC = 'aâbcçdefgğhıijklmnñoöpqrsştuüvyz'; # Crimean Tatar Latin lowercase
  47         const L_N_CONS_UC = 'ÇNRSTZ'; # Crimean Tatar Latin N-type upper case consonants
  48         const L_N_CONS_LC = 'çnrstz'; # Crimean Tatar Latin N-type lower case consonants
  49         const L_N_CONS = 'çnrstzÇNRSTZ'; # Crimean Tatar Latin N-type consonants
  50         const L_M_CONS = 'bcgkmpşBCGKMPŞ'; # Crimean Tatar Latin M-type consonants
  51         const L_CONS_UC = 'BCÇDFGĞHJKLMNÑPQRSŞTVZ'; # Crimean Tatar Latin uppercase consonants
  52         const L_CONS_LC = 'bcçdfgğhjklmnñpqrsştvz'; # Crimean Tatar Latin lowercase consonants
  53         const L_CONS = 'bcçdfgğhjklmnñpqrsştvzBCÇDFGĞHJKLMNÑPQRSŞTVZ'; # Crimean Tatar Latin consonants
  54         const L_VOW_UC = 'AÂEIİOÖUÜ'; # Crimean Tatar Latin uppercase vowels
  55         const L_VOW = 'aâeıioöuüAÂEIİOÖUÜ'; # Crimean Tatar Latin vowels
  56         const L_F_UC = 'EİÖÜ'; # Crimean Tatar Latin uppercase front vowels
  57         const L_F = 'eiöüEİÖÜ'; # Crimean Tatar Latin front vowels
  58
  59         /**
  60          * @param Language $langobj
  61          * @param string $maincode
  62          * @param array $variants
  63          * @param array $variantfallbacks
  64          * @param array $flags
  65          */
  66         function __construct( Language $langobj, $maincode,
  67                 $variants = [],
  68                 $variantfallbacks = [],
  69                 $flags = []
  70         ) {
  71                 parent::__construct( $langobj, $maincode,
  72                         $variants, $variantfallbacks, $flags );
  73
  74                 // No point delaying this since they're in code.
  75                 // Waiting until loadDefaultTables() means they never get loaded
  76                 // when the tables themselves are loaded from cache.
  77                 $this->loadExceptions();
  78         }
  79
  80         public $mCyrillicToLatin = [
  81
  82                 ## these are independent of location in the word, but have
  83                 ## to go first so other transforms don't bleed them
  84                 'гъ' => 'ğ', 'Гъ' => 'Ğ', 'ГЪ' => 'Ğ',
  85                 'къ' => 'q', 'Къ' => 'Q', 'КЪ' => 'Q',
  86                 'нъ' => 'ñ', 'Нъ' => 'Ñ', 'НЪ' => 'Ñ',
  87                 'дж' => 'c', 'Дж' => 'C', 'ДЖ' => 'C',
  88
  89                 'А' => 'A', 'а' => 'a', 'Б' => 'B', 'б' => 'b',
  90                 'В' => 'V', 'в' => 'v', 'Г' => 'G', 'г' => 'g',
  91                 'Д' => 'D', 'д' => 'd', 'Ж' => 'J', 'ж' => 'j',
  92                 'З' => 'Z', 'з' => 'z', 'И' => 'İ', 'и' => 'i',
  93                 'Й' => 'Y', 'й' => 'y', 'К' => 'K', 'к' => 'k',
  94                 'Л' => 'L', 'л' => 'l', 'М' => 'M', 'м' => 'm',
  95                 'Н' => 'N', 'н' => 'n', 'П' => 'P', 'п' => 'p',
  96                 'Р' => 'R', 'р' => 'r', 'С' => 'S', 'с' => 's',
  97                 'Т' => 'T', 'т' => 't', 'Ф' => 'F', 'ф' => 'f',
  98                 'Х' => 'H', 'х' => 'h', 'Ч' => 'Ç', 'ч' => 'ç',
  99                 'Ш' => 'Ş', 'ш' => 'ş', 'Ы' => 'I', 'ы' => 'ı',
 100                 'Э' => 'E', 'э' => 'e', 'Е' => 'E', 'е' => 'e',
 101                 'Я' => 'Â', 'я' => 'â', 'У' => 'U', 'у' => 'u',
 102                 'О' => 'O', 'о' => 'o',
 103
 104                 'Ё' => 'Yo', 'ё' => 'yo', 'Ю' => 'Yu', 'ю' => 'yu',
 105                 'Ц' => 'Ts', 'ц' => 'ts', 'Щ' => 'Şç', 'щ' => 'şç',
 106                 'Ь' => '', 'ь' => '', 'Ъ' => '', 'ъ' => '',
 107
 108         ];
 109
 110         public $mLatinToCyrillic = [
 111                 'Â' => 'Я', 'â' => 'я', 'B' => 'Б', 'b' => 'б',
 112                 'Ç' => 'Ч', 'ç' => 'ч', 'D' => 'Д', 'd' => 'д',
 113                 'F' => 'Ф', 'f' => 'ф', 'G' => 'Г', 'g' => 'г',
 114                 'H' => 'Х', 'h' => 'х', 'I' => 'Ы', 'ı' => 'ы',
 115                 'İ' => 'И', 'i' => 'и', 'J' => 'Ж', 'j' => 'ж',
 116                 'K' => 'К', 'k' => 'к', 'L' => 'Л', 'l' => 'л',
 117                 'M' => 'М', 'm' => 'м', 'N' => 'Н', 'n' => 'н',
 118                 'O' => 'О', 'o' => 'о', 'P' => 'П', 'p' => 'п',
 119                 'R' => 'Р', 'r' => 'р', 'S' => 'С', 's' => 'с',
 120                 'Ş' => 'Ш', 'ş' => 'ш', 'T' => 'Т', 't' => 'т',
 121                 'V' => 'В', 'v' => 'в', 'Z' => 'З', 'z' => 'з',
 122
 123                 'ya' => 'я', 'Ya' => 'Я', 'YA' => 'Я',
 124                 'ye' => 'е', 'YE' => 'Е', 'Ye' => 'Е',
 125
 126                 // hack, hack, hack
 127                 'A' => 'А', 'a' => 'а', 'E' => 'Е', 'e' => 'е',
 128                 'Ö' => 'Ё', 'ö' => 'ё', 'U' => 'У', 'u' => 'у',
 129                 'Ü' => 'Ю', 'ü' => 'ю', 'Y' => 'Й', 'y' => 'й',
 130                 'C' => 'Дж', 'c' => 'дж', 'Ğ' => 'Гъ', 'ğ' => 'гъ',
 131                 'Ñ' => 'Нъ', 'ñ' => 'нъ', 'Q' => 'Къ', 'q' => 'къ',
 132
 133                 ];
 134
 135         public $mCyrl2LatnExceptions = [];
 136         public $mLatn2CyrlExceptions = [];
 137
 138         public $mCyrl2LatnPatterns = [];
 139         public $mLatn2CyrlPatterns = [];
 140
 141         public $mCyrlCleanUpRegexes = [];
 142
 143         public $mExceptionsLoaded = false;
 144
 145         function loadDefaultTables() {
 146                 $this->mTables = [
 147                         'crh-latn' => new ReplacementArray( $this->mCyrillicToLatin ),
 148                         'crh-cyrl' => new ReplacementArray( $this->mLatinToCyrillic ),
 149                         'crh' => new ReplacementArray()
 150                 ];
 151         }
 152
 153         function loadExceptions() {
 154                 if ( $this->mExceptionsLoaded ) {
 155                         return;
 156                 }
 157
 158                 $this->mExceptionsLoaded = true;
 159                 $crhExceptions = new MediaWiki\Languages\Data\CrhExceptions();
 160                 list( $this->mCyrl2LatnExceptions, $this->mLatn2CyrlExceptions,
 161                         $this->mCyrl2LatnPatterns, $this->mLatn2CyrlPatterns, $this->mCyrlCleanUpRegexes ) =
 162                         $crhExceptions->loadExceptions( self::L_LC . self::C_LC, self::L_UC . self::C_UC );
 163         }
 164
 165         /**
 166          * A function wrapper:
 167          *   - if there is no selected variant, leave the link
 168          *     names as they were
 169          *   - do not try to find variants for usernames
 170          *
 171          * @param string &$link
 172          * @param Title &$nt
 173          * @param bool $ignoreOtherCond
 174          */
 175         function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
 176                 // check for user namespace
 177                 if ( is_object( $nt ) ) {
 178                         $ns = $nt->getNamespace();
 179                         if ( $ns == NS_USER || $ns == NS_USER_TALK ) {
 180                                 return;
 181                         }
 182                 }
 183
 184                 $oldlink = $link;
 185                 parent::findVariantLink( $link, $nt, $ignoreOtherCond );
 186                 if ( $this->getPreferredVariant() == $this->mMainLanguageCode ) {
 187                         $link = $oldlink;
 188                 }
 189         }
 190
 191         /**
 192          *  It translates text into variant, specials:
 193          *    - omitting roman numbers
 194          *
 195          * @param string $text
 196          * @param bool $toVariant
 197          *
 198          * @throws MWException
 199          * @return string
 200          */
 201         function translate( $text, $toVariant ) {
 202                 switch ( $toVariant ) {
 203                         case 'crh-cyrl':
 204                         case 'crh-latn':
 205                                 break;
 206                         default:
 207                                 return $text;
 208                 }
 209
 210                 if ( !$this->mTablesLoaded ) {
 211                         $this->loadTables();
 212                 }
 213
 214                 if ( !isset( $this->mTables[$toVariant] ) ) {
 215                         throw new MWException( "Broken variant table: " . implode( ',', array_keys( $this->mTables ) ) );
 216                 }
 217
 218                 switch ( $toVariant ) {
 219                         case 'crh-cyrl':
 220                                 /* Check for roman numbers like VII, XIX...
 221                                  * Only need to split on Roman numerals when converting to Cyrillic
 222                                  * Lookahead assertion ensures $roman doesn't match the empty string, and
 223                                  * non-period after first "Roman" character allows initials to be converted
 224                                  */
 225                                 $roman = '(?=[MDCLXVI]([^.]|$))M{0,4}(C[DM]|D?C{0,3})(X[LC]|L?X{0,3})(I[VX]|V?I{0,3})';
 226
 227                                 $breaks = '([^\w\x80-\xff])';
 228
 229                                 // allow for multiple Roman numerals in a row; rare but it happens
 230                                 $romanRegex = '/^' . $roman . '$|^(' . $roman . $breaks . ')+|(' . $breaks . $roman . ')+$|' .
 231                                         $breaks . '(' . $roman . $breaks . ')+/';
 232
 233                                 $matches = preg_split( $romanRegex, $text, -1, PREG_SPLIT_OFFSET_CAPTURE );
 234                                 $mstart = 0;
 235                                 $ret = '';
 236                                 foreach ( $matches as $m ) {
 237                                         // copy over Roman numerals
 238                                         $ret .= substr( $text, $mstart, $m[1] - $mstart );
 239
 240                                         // process everything else
 241                                         if ( $m[0] !== '' ) {
 242                                                 $ret .= $this->regsConverter( $m[0], $toVariant );
 243                                         }
 244
 245                                         $mstart = $m[1] + strlen( $m[0] );
 246                                 }
 247
 248                                 return $ret;
 249                         default:
 250                                 // Just process the whole string in one go
 251                                 return $this->regsConverter( $text, $toVariant );
 252                 }
 253         }
 254
 255         private function regsConverter( $text, $toVariant ) {
 256                 if ( $text == '' ) return $text;
 257
 258                 $pat = [];
 259                 $rep = [];
 260                 switch ( $toVariant ) {
 261                         case 'crh-latn':
 262                                 $text = strtr( $text, $this->mCyrl2LatnExceptions );
 263                                 foreach ( $this->mCyrl2LatnPatterns as $pat => $rep ) {
 264                                         $text = preg_replace( $pat, $rep, $text );
 265                                 }
 266                                 $text = parent::translate( $text, $toVariant );
 267                                 $text = strtr( $text, [ '«' => '"', '»' => '"', ] );
 268                                 return $text;
 269                         case 'crh-cyrl':
 270                                 $text = strtr( $text, $this->mLatn2CyrlExceptions );
 271                                 foreach ( $this->mLatn2CyrlPatterns as $pat => $rep ) {
 272                                         $text = preg_replace( $pat, $rep, $text );
 273                                 }
 274                                 $text = parent::translate( $text, $toVariant );
 275                                 $text = strtr( $text, [ '“' => '«', '”' => '»', ] );
 276                                 foreach ( $this->mCyrlCleanUpRegexes as $pat => $rep ) {
 277                                         $text = preg_replace( $pat, $rep, $text );
 278                                 }
 279                                 return $text;
 280                         default:
 281                                 return $text;
 282                 }
 283         }
 284
 285 }
 286
 287 /**
 288  * Crimean Tatar (Qırımtatarca)
 289  *
 290  * @ingroup Language
 291  */
 292 class LanguageCrh extends Language {
 293
 294         function __construct() {
 295                 parent::__construct();
 296
 297                 $variants = [ 'crh', 'crh-cyrl', 'crh-latn' ];
 298                 $variantfallbacks = [
 299                         'crh' => 'crh-latn',
 300                         'crh-cyrl' => 'crh-latn',
 301                         'crh-latn' => 'crh-cyrl',
 302                 ];
 303
 304                 $this->mConverter = new CrhConverter( $this, 'crh', $variants, $variantfallbacks );
 305         }
 306 }