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         const WORD_BEGINS = '\r\s\"\'\(\)\-<>\[\]\/.,:;!?';
  34         const WORD_ENDS = '\r\s\"\'\(\)\-<>\[\]\/.,:;!?';
  35
  36         // Cyrillic
  37         const C_UC = 'АБВГДЕЁЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЫЬЭЮЯ'; # Crimean Tatar Cyrillic uppercase
  38         const C_LC = 'абвгдеёжзийклмнопрстуфхцчшщъыьэюя'; # Crimean Tatar Cyrillic lowercase
  39         const C_CONS_UC = 'БВГДЖЗЙКЛМНПРСТФХЦЧШЩCÑ'; # Crimean Tatar Cyrillic + CÑ uppercase consonants
  40         const C_CONS_LC = 'бвгджзйклмнпрстфхцчшщcñ'; # Crimean Tatar Cyrillic + CÑ lowercase consonants
  41         const C_M_CONS = 'бгкмпшcБГКМПШC'; # Crimean Tatar Cyrillic M-type consonants
  42
  43         # Crimean Tatar Cyrillic + CÑ consonants
  44         const C_CONS = 'бвгджзйклмнпрстфхцчшщcñБВГДЖЗЙКЛМНПРСТФХЦЧШЩCÑ';
  45
  46         // Latin
  47         const L_UC = 'AÂBCÇDEFGĞHIİJKLMNÑOÖPQRSŞTUÜVYZ'; # Crimean Tatar Latin uppercase
  48         const L_LC = 'aâbcçdefgğhıijklmnñoöpqrsştuüvyz'; # Crimean Tatar Latin lowercase
  49         const L_N_CONS_UC = 'ÇNRSTZ'; # Crimean Tatar Latin N-type upper case consonants
  50         const L_N_CONS_LC = 'çnrstz'; # Crimean Tatar Latin N-type lower case consonants
  51         const L_N_CONS = 'çnrstzÇNRSTZ'; # Crimean Tatar Latin N-type consonants
  52         const L_M_CONS = 'bcgkmpşBCGKMPŞ'; # Crimean Tatar Latin M-type consonants
  53         const L_CONS_UC = 'BCÇDFGHJKLMNÑPRSŞTVZ'; # Crimean Tatar Latin uppercase consonants
  54         const L_CONS_LC = 'bcçdfghjklmnñprsştvz'; # Crimean Tatar Latin lowercase consonants
  55         const L_CONS = 'bcçdfghjklmnñprsştvzBCÇDFGHJKLMNÑPRSŞTVZ'; # Crimean Tatar Latin consonants
  56         const L_VOW_UC = 'AÂEIİOÖUÜ'; # Crimean Tatar Latin uppercase vowels
  57         const L_VOW = 'aâeıioöuüAÂEIİOÖUÜ'; # Crimean Tatar Latin vowels
  58         const L_F_UC = 'EİÖÜ'; # Crimean Tatar Latin uppercase front vowels
  59         const L_F = 'eiöüEİÖÜ'; # Crimean Tatar Latin front vowels
  60
  61         public $mCyrillicToLatin = [
  62
  63                 ## these are independent of location in the word, but have
  64                 ## to go first so other transforms don't bleed them
  65                 'гъ' => 'ğ', 'Гъ' => 'Ğ', 'ГЪ' => 'Ğ',
  66                 'къ' => 'q', 'Къ' => 'Q', 'КЪ' => 'Q',
  67                 'нъ' => 'ñ', 'Нъ' => 'Ñ', 'НЪ' => 'Ñ',
  68                 'дж' => 'c', 'Дж' => 'C', 'ДЖ' => 'C',
  69
  70                 'А' => 'A', 'а' => 'a', 'Б' => 'B', 'б' => 'b',
  71                 'В' => 'V', 'в' => 'v', 'Г' => 'G', 'г' => 'g',
  72                 'Д' => 'D', 'д' => 'd', 'Ж' => 'J', 'ж' => 'j',
  73                 'З' => 'Z', 'з' => 'z', 'И' => 'İ', 'и' => 'i',
  74                 'Й' => 'Y', 'й' => 'y', 'К' => 'K', 'к' => 'k',
  75                 'Л' => 'L', 'л' => 'l', 'М' => 'M', 'м' => 'm',
  76                 'Н' => 'N', 'н' => 'n', 'П' => 'P', 'п' => 'p',
  77                 'Р' => 'R', 'р' => 'r', 'С' => 'S', 'с' => 's',
  78                 'Т' => 'T', 'т' => 't', 'Ф' => 'F', 'ф' => 'f',
  79                 'Х' => 'H', 'х' => 'h', 'Ч' => 'Ç', 'ч' => 'ç',
  80                 'Ш' => 'Ş', 'ш' => 'ş', 'Ы' => 'I', 'ы' => 'ı',
  81                 'Э' => 'E', 'э' => 'e', 'Е' => 'E', 'е' => 'e',
  82                 'Я' => 'Â', 'я' => 'â', 'У' => 'U', 'у' => 'u',
  83                 'О' => 'O', 'о' => 'o',
  84
  85                 'Ё' => 'Yo', 'ё' => 'yo', 'Ю' => 'Yu', 'ю' => 'yu',
  86                 'Ц' => 'Ts', 'ц' => 'ts', 'Щ' => 'Şç', 'щ' => 'şç',
  87                 'Ь' => '', 'ь' => '', 'Ъ' => '', 'ъ' => '',
  88
  89         ];
  90
  91         public $mLatinToCyrillic = [
  92                 'Â' => 'Я', 'â' => 'я', 'B' => 'Б', 'b' => 'б',
  93                 'Ç' => 'Ч', 'ç' => 'ч', 'D' => 'Д', 'd' => 'д',
  94                 'F' => 'Ф', 'f' => 'ф', 'G' => 'Г', 'g' => 'г',
  95                 'H' => 'Х', 'h' => 'х', 'I' => 'Ы', 'ı' => 'ы',
  96                 'İ' => 'И', 'i' => 'и', 'J' => 'Ж', 'j' => 'ж',
  97                 'K' => 'К', 'k' => 'к', 'L' => 'Л', 'l' => 'л',
  98                 'M' => 'М', 'm' => 'м', 'N' => 'Н', 'n' => 'н',
  99                 'O' => 'О', 'o' => 'о', 'P' => 'П', 'p' => 'п',
 100                 'R' => 'Р', 'r' => 'р', 'S' => 'С', 's' => 'с',
 101                 'Ş' => 'Ш', 'ş' => 'ш', 'T' => 'Т', 't' => 'т',
 102                 'V' => 'В', 'v' => 'в', 'Z' => 'З', 'z' => 'з',
 103
 104                 'ya' => 'я', 'Ya' => 'Я', 'YA' => 'Я',
 105                 'ye' => 'е', 'YE' => 'Е', 'Ye' => 'Е',
 106
 107                 // hack, hack, hack
 108                 'A' => 'А', 'a' => 'а', 'E' => 'Е', 'e' => 'е',
 109                 'Ö' => 'О', 'ö' => 'о', 'U' => 'У', 'u' => 'у',
 110                 'Ü' => 'У', 'ü' => 'у', 'Y' => 'Й', 'y' => 'й',
 111
 112                 'C' => 'Дж', 'c' => 'дж', 'Ğ' => 'Гъ', 'ğ' => 'гъ',
 113                 'Ñ' => 'Нъ', 'ñ' => 'нъ', 'Q' => 'Къ', 'q' => 'къ',
 114
 115                 ];
 116
 117         public $mExceptions = [];
 118         public $mCyrl2LatnPatterns = [];
 119         public $mLatn2CyrlPatterns = [];
 120         public $mCyrlCleanUpRegexes = [];
 121
 122         public $mExceptionsLoaded = false;
 123
 124         function loadDefaultTables() {
 125                 $this->mTables = [
 126                         'crh-latn' => new ReplacementArray( $this->mCyrillicToLatin ),
 127                         'crh-cyrl' => new ReplacementArray( $this->mLatinToCyrillic ),
 128                         'crh' => new ReplacementArray()
 129                 ];
 130         }
 131
 132         function postLoadTables() {
 133                 $this->loadExceptions();
 134         }
 135
 136         function loadExceptions() {
 137                 if ( $this->mExceptionsLoaded ) {
 138                         return;
 139                 }
 140
 141                 $this->mExceptionsLoaded = true;
 142                 $crhExceptions = new MediaWiki\Languages\Data\CrhExceptions();
 143                 list( $this->mExceptions, $this->mCyrl2LatnPatterns, $this->mLatn2CyrlPatterns,
 144                         $this->mCyrlCleanUpRegexes ) = $crhExceptions->loadExceptions( self::L_LC . self::C_LC,
 145                         self::L_UC . self::C_UC );
 146         }
 147
 148         /**
 149          * A function wrapper:
 150          *   - if there is no selected variant, leave the link
 151          *     names as they were
 152          *   - do not try to find variants for usernames
 153          *
 154          * @param string &$link
 155          * @param Title &$nt
 156          * @param bool $ignoreOtherCond
 157          */
 158         function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
 159                 // check for user namespace
 160                 if ( is_object( $nt ) ) {
 161                         $ns = $nt->getNamespace();
 162                         if ( $ns == NS_USER || $ns == NS_USER_TALK ) {
 163                                 return;
 164                         }
 165                 }
 166
 167                 $oldlink = $link;
 168                 parent::findVariantLink( $link, $nt, $ignoreOtherCond );
 169                 if ( $this->getPreferredVariant() == $this->mMainLanguageCode ) {
 170                         $link = $oldlink;
 171                 }
 172         }
 173
 174         /**
 175          *  It translates text into variant, specials:
 176          *    - ommiting roman numbers
 177          *
 178          * @param string $text
 179          * @param bool $toVariant
 180          *
 181          * @throws MWException
 182          * @return string
 183          */
 184         function translate( $text, $toVariant ) {
 185                 $letters = '';
 186                 switch ( $toVariant ) {
 187                         case 'crh-cyrl':
 188                                 $letters = self::L_UC . self::L_LC . "\'";
 189                                 break;
 190                         case 'crh-latn':
 191                                 $letters = self::C_UC . self::C_LC . "";
 192                                 break;
 193                         default:
 194                                 return $text;
 195                                 break;
 196                 }
 197
 198                 if ( !$this->mTablesLoaded ) {
 199                         $this->loadTables();
 200                 }
 201
 202                 if ( !isset( $this->mTables[$toVariant] ) ) {
 203                         throw new MWException( "Broken variant table: " . implode( ',', array_keys( $this->mTables ) ) );
 204                 }
 205
 206                 // check for roman numbers like VII, XIX...
 207                 $roman = '/^M{0,3}(C[DM]|D{0,1}C{0,3})(X[LC]|L{0,1}X{0,3})(I[VX]|V{0,1}I{0,3})$/u';
 208
 209                 # match any sub-string of the relevant letters and convert it
 210                 $matches = preg_split( '/(\b|^)[^' . $letters . ']+(\b|$)/u',
 211                         $text, -1, PREG_SPLIT_OFFSET_CAPTURE );
 212                 $mstart = 0;
 213                 $ret = '';
 214                 foreach ( $matches as $m ) {
 215                         # copy over the non-matching bit
 216                         $ret .= substr( $text, $mstart, $m[1] - $mstart );
 217                         # skip certain classes of strings
 218
 219                         if ( array_key_exists( $m[0], $this->mExceptions ) ) {
 220                                 # if it's an exception, just copy down the right answer
 221                                 $ret .= $this->mExceptions[$m[0]];
 222                         } elseif ( ! $m[0] || # empty strings
 223                                          preg_match( $roman, $m[0] ) || # roman numerals
 224                                          preg_match( '/[^' . $letters . ']/', $m[0] ) # mixed orthography
 225                                         ) {
 226                                 $ret .= $m[0];
 227                         } else {
 228                                 # convert according to the rules
 229                                 $token = $this->regsConverter( $m[0], $toVariant );
 230                                 $ret .= parent::translate( $token, $toVariant );
 231                         }
 232                         $mstart = $m[1] + strlen( $m[0] );
 233                 }
 234
 235                 # pick up stray quote marks
 236                 switch ( $toVariant ) {
 237                         case 'crh-cyrl':
 238                                 $ret = strtr( $ret, [ '“' => '«', '”' => '»', ] );
 239                                 $ret = $this->regsConverter( $ret, 'cyrl-cleanup' );
 240                                 break;
 241                         case 'crh-latn':
 242                                 $ret = strtr( $ret, [ '«' => '"', '»' => '"', ] );
 243                                 break;
 244                 }
 245
 246                 return $ret;
 247         }
 248
 249         private function regsConverter( $text, $toVariant ) {
 250                 if ( $text == '' ) return $text;
 251
 252                 $pat = [];
 253                 $rep = [];
 254                 switch ( $toVariant ) {
 255                         case 'crh-latn':
 256                                 foreach ( $this->mCyrl2LatnPatterns as $pat => $rep ) {
 257                                         $text = preg_replace( $pat, $rep, $text );
 258                                 }
 259                                 return $text;
 260                         case 'crh-cyrl':
 261                                 foreach ( $this->mLatn2CyrlPatterns as $pat => $rep ) {
 262                                         $text = preg_replace( $pat, $rep, $text );
 263                                 }
 264                                 return $text;
 265                         case 'cyrl-cleanup':
 266                                 foreach ( $this->mCyrlCleanUpRegexes as $pat => $rep ) {
 267                                         $text = preg_replace( $pat, $rep, $text );
 268                                 }
 269                                 return $text;
 270                         default:
 271                                 return $text;
 272                 }
 273         }
 274
 275 }
 276
 277 /**
 278  * Crimean Tatar (Qırımtatarca)
 279  *
 280  * @ingroup Language
 281  */
 282 class LanguageCrh extends Language {
 283
 284         function __construct() {
 285                 parent::__construct();
 286
 287                 $variants = [ 'crh', 'crh-cyrl', 'crh-latn' ];
 288                 $variantfallbacks = [
 289                         'crh' => 'crh-latn',
 290                         'crh-cyrl' => 'crh-latn',
 291                         'crh-latn' => 'crh-cyrl',
 292                 ];
 293
 294                 $this->mConverter = new CrhConverter( $this, 'crh', $variants, $variantfallbacks );
 295         }
 296 }