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( $langobj, $maincode,
  67                                                                 $variants = [],
  68                                                                 $variantfallbacks = [],
  69                                                                 $flags = [] ) {
  70                 parent::__construct( $langobj, $maincode,
  71                         $variants, $variantfallbacks, $flags );
  72
  73                 // No point delaying this since they're in code.
  74                 // Waiting until loadDefaultTables() means they never get loaded
  75                 // when the tables themselves are loaded from cache.
  76                 $this->loadExceptions();
  77         }
  78
  79         public $mCyrillicToLatin = [
  80
  81                 ## these are independent of location in the word, but have
  82                 ## to go first so other transforms don't bleed them
  83                 'гъ' => 'ğ', 'Гъ' => 'Ğ', 'ГЪ' => 'Ğ',
  84                 'къ' => 'q', 'Къ' => 'Q', 'КЪ' => 'Q',
  85                 'нъ' => 'ñ', 'Нъ' => 'Ñ', 'НЪ' => 'Ñ',
  86                 'дж' => 'c', 'Дж' => 'C', 'ДЖ' => 'C',
  87
  88                 'А' => 'A', 'а' => 'a', 'Б' => 'B', 'б' => 'b',
  89                 'В' => 'V', 'в' => 'v', 'Г' => 'G', 'г' => 'g',
  90                 'Д' => 'D', 'д' => 'd', 'Ж' => 'J', 'ж' => 'j',
  91                 'З' => 'Z', 'з' => 'z', 'И' => 'İ', 'и' => 'i',
  92                 'Й' => 'Y', 'й' => 'y', 'К' => 'K', 'к' => 'k',
  93                 'Л' => 'L', 'л' => 'l', 'М' => 'M', 'м' => 'm',
  94                 'Н' => 'N', 'н' => 'n', 'П' => 'P', 'п' => 'p',
  95                 'Р' => 'R', 'р' => 'r', 'С' => 'S', 'с' => 's',
  96                 'Т' => 'T', 'т' => 't', 'Ф' => 'F', 'ф' => 'f',
  97                 'Х' => 'H', 'х' => 'h', 'Ч' => 'Ç', 'ч' => 'ç',
  98                 'Ш' => 'Ş', 'ш' => 'ş', 'Ы' => 'I', 'ы' => 'ı',
  99                 'Э' => 'E', 'э' => 'e', 'Е' => 'E', 'е' => 'e',
 100                 'Я' => 'Â', 'я' => 'â', 'У' => 'U', 'у' => 'u',
 101                 'О' => 'O', 'о' => 'o',
 102
 103                 'Ё' => 'Yo', 'ё' => 'yo', 'Ю' => 'Yu', 'ю' => 'yu',
 104                 'Ц' => 'Ts', 'ц' => 'ts', 'Щ' => 'Şç', 'щ' => 'şç',
 105                 'Ь' => '', 'ь' => '', 'Ъ' => '', 'ъ' => '',
 106
 107         ];
 108
 109         public $mLatinToCyrillic = [
 110                 'Â' => 'Я', 'â' => 'я', 'B' => 'Б', 'b' => 'б',
 111                 'Ç' => 'Ч', 'ç' => 'ч', 'D' => 'Д', 'd' => 'д',
 112                 'F' => 'Ф', 'f' => 'ф', 'G' => 'Г', 'g' => 'г',
 113                 'H' => 'Х', 'h' => 'х', 'I' => 'Ы', 'ı' => 'ы',
 114                 'İ' => 'И', 'i' => 'и', 'J' => 'Ж', 'j' => 'ж',
 115                 'K' => 'К', 'k' => 'к', 'L' => 'Л', 'l' => 'л',
 116                 'M' => 'М', 'm' => 'м', 'N' => 'Н', 'n' => 'н',
 117                 'O' => 'О', 'o' => 'о', 'P' => 'П', 'p' => 'п',
 118                 'R' => 'Р', 'r' => 'р', 'S' => 'С', 's' => 'с',
 119                 'Ş' => 'Ш', 'ş' => 'ш', 'T' => 'Т', 't' => 'т',
 120                 'V' => 'В', 'v' => 'в', 'Z' => 'З', 'z' => 'з',
 121
 122                 'ya' => 'я', 'Ya' => 'Я', 'YA' => 'Я',
 123                 'ye' => 'е', 'YE' => 'Е', 'Ye' => 'Е',
 124
 125                 // hack, hack, hack
 126                 'A' => 'А', 'a' => 'а', 'E' => 'Е', 'e' => 'е',
 127                 'Ö' => 'Ё', 'ö' => 'ё', 'U' => 'У', 'u' => 'у',
 128                 'Ü' => 'Ю', 'ü' => 'ю', 'Y' => 'Й', 'y' => 'й',
 129                 'C' => 'Дж', 'c' => 'дж', 'Ğ' => 'Гъ', 'ğ' => 'гъ',
 130                 'Ñ' => 'Нъ', 'ñ' => 'нъ', 'Q' => 'Къ', 'q' => 'къ',
 131
 132                 ];
 133
 134         public $mCyrl2LatnExceptions = [];
 135         public $mLatn2CyrlExceptions = [];
 136
 137         public $mCyrl2LatnPatterns = [];
 138         public $mLatn2CyrlPatterns = [];
 139
 140         public $mCyrlCleanUpRegexes = [];
 141
 142         public $mExceptionsLoaded = false;
 143
 144         function loadDefaultTables() {
 145                 $this->mTables = [
 146                         'crh-latn' => new ReplacementArray( $this->mCyrillicToLatin ),
 147                         'crh-cyrl' => new ReplacementArray( $this->mLatinToCyrillic ),
 148                         'crh' => new ReplacementArray()
 149                 ];
 150         }
 151
 152         function loadExceptions() {
 153                 if ( $this->mExceptionsLoaded ) {
 154                         return;
 155                 }
 156
 157                 $this->mExceptionsLoaded = true;
 158                 $crhExceptions = new MediaWiki\Languages\Data\CrhExceptions();
 159                 list( $this->mCyrl2LatnExceptions, $this->mLatn2CyrlExceptions,
 160                         $this->mCyrl2LatnPatterns, $this->mLatn2CyrlPatterns, $this->mCyrlCleanUpRegexes ) =
 161                         $crhExceptions->loadExceptions( self::L_LC . self::C_LC, self::L_UC . self::C_UC );
 162         }
 163
 164         /**
 165          * A function wrapper:
 166          *   - if there is no selected variant, leave the link
 167          *     names as they were
 168          *   - do not try to find variants for usernames
 169          *
 170          * @param string &$link
 171          * @param Title &$nt
 172          * @param bool $ignoreOtherCond
 173          */
 174         function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
 175                 // check for user namespace
 176                 if ( is_object( $nt ) ) {
 177                         $ns = $nt->getNamespace();
 178                         if ( $ns == NS_USER || $ns == NS_USER_TALK ) {
 179                                 return;
 180                         }
 181                 }
 182
 183                 $oldlink = $link;
 184                 parent::findVariantLink( $link, $nt, $ignoreOtherCond );
 185                 if ( $this->getPreferredVariant() == $this->mMainLanguageCode ) {
 186                         $link = $oldlink;
 187                 }
 188         }
 189
 190         /**
 191          *  It translates text into variant, specials:
 192          *    - omitting roman numbers
 193          *
 194          * @param string $text
 195          * @param bool $toVariant
 196          *
 197          * @throws MWException
 198          * @return string
 199          */
 200         function translate( $text, $toVariant ) {
 201                 switch ( $toVariant ) {
 202                         case 'crh-cyrl':
 203                         case 'crh-latn':
 204                                 break;
 205                         default:
 206                                 return $text;
 207                 }
 208
 209                 if ( !$this->mTablesLoaded ) {
 210                         $this->loadTables();
 211                 }
 212
 213                 if ( !isset( $this->mTables[$toVariant] ) ) {
 214                         throw new MWException( "Broken variant table: " . implode( ',', array_keys( $this->mTables ) ) );
 215                 }
 216
 217                 switch ( $toVariant ) {
 218                         case 'crh-cyrl':
 219                                 /* Check for roman numbers like VII, XIX...
 220                                  * Only need to split on Roman numerals when converting to Cyrillic
 221                                  * Lookahead assertion ensures $roman doesn't match the empty string, and
 222                                  * non-period after first "Roman" character allows initials to be converted
 223                                  */
 224                                 $roman = '(?=[MDCLXVI]([^.]|$))M{0,4}(C[DM]|D?C{0,3})(X[LC]|L?X{0,3})(I[VX]|V?I{0,3})';
 225
 226                                 $breaks = '([^\w\x80-\xff])';
 227
 228                                 // allow for multiple Roman numerals in a row; rare but it happens
 229                                 $romanRegex = '/^' . $roman . '$|^(' . $roman . $breaks . ')+|(' . $breaks . $roman . ')+$|' .
 230                                         $breaks . '(' . $roman . $breaks . ')+/';
 231
 232                                 $matches = preg_split( $romanRegex, $text, -1, PREG_SPLIT_OFFSET_CAPTURE );
 233                                 $mstart = 0;
 234                                 $ret = '';
 235                                 foreach ( $matches as $m ) {
 236                                         // copy over Roman numerals
 237                                         $ret .= substr( $text, $mstart, $m[1] - $mstart );
 238
 239                                         // process everything else
 240                                         if ( $m[0] !== '' ) {
 241                                                 $ret .= $this->regsConverter( $m[0], $toVariant );
 242                                         }
 243
 244                                         $mstart = $m[1] + strlen( $m[0] );
 245                                 }
 246
 247                                 return $ret;
 248                         default:
 249                                 // Just process the whole string in one go
 250                                 return $this->regsConverter( $text, $toVariant );
 251                 }
 252         }
 253
 254         private function regsConverter( $text, $toVariant ) {
 255                 if ( $text == '' ) return $text;
 256
 257                 $pat = [];
 258                 $rep = [];
 259                 switch ( $toVariant ) {
 260                         case 'crh-latn':
 261                                 $text = strtr( $text, $this->mCyrl2LatnExceptions );
 262                                 foreach ( $this->mCyrl2LatnPatterns as $pat => $rep ) {
 263                                         $text = preg_replace( $pat, $rep, $text );
 264                                 }
 265                                 $text = parent::translate( $text, $toVariant );
 266                                 $text = strtr( $text, [ '«' => '"', '»' => '"', ] );
 267                                 return $text;
 268                         case 'crh-cyrl':
 269                                 $text = strtr( $text, $this->mLatn2CyrlExceptions );
 270                                 foreach ( $this->mLatn2CyrlPatterns as $pat => $rep ) {
 271                                         $text = preg_replace( $pat, $rep, $text );
 272                                 }
 273                                 $text = parent::translate( $text, $toVariant );
 274                                 $text = strtr( $text, [ '“' => '«', '”' => '»', ] );
 275                                 foreach ( $this->mCyrlCleanUpRegexes as $pat => $rep ) {
 276                                         $text = preg_replace( $pat, $rep, $text );
 277                                 }
 278                                 return $text;
 279                         default:
 280                                 return $text;
 281                 }
 282         }
 283
 284 }
 285
 286 /**
 287  * Crimean Tatar (Qırımtatarca)
 288  *
 289  * @ingroup Language
 290  */
 291 class LanguageCrh extends Language {
 292
 293         function __construct() {
 294                 parent::__construct();
 295
 296                 $variants = [ 'crh', 'crh-cyrl', 'crh-latn' ];
 297                 $variantfallbacks = [
 298                         'crh' => 'crh-latn',
 299                         'crh-cyrl' => 'crh-latn',
 300                         'crh-latn' => 'crh-cyrl',
 301                 ];
 302
 303                 $this->mConverter = new CrhConverter( $this, 'crh', $variants, $variantfallbacks );
 304         }
 305 }