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         /**
  62          * @param Language $langobj
  63          * @param string $maincode
  64          * @param array $variants
  65          * @param array $variantfallbacks
  66          * @param array $flags
  67          */
  68         function __construct( $langobj, $maincode,
  69                                                                 $variants = [],
  70                                                                 $variantfallbacks = [],
  71                                                                 $flags = [] ) {
  72                 parent::__construct( $langobj, $maincode,
  73                         $variants, $variantfallbacks, $flags );
  74
  75                 // No point delaying this since they're in code.
  76                 // Waiting until loadDefaultTables() means they never get loaded
  77                 // when the tables themselves are loaded from cache.
  78                 $this->loadExceptions();
  79         }
  80
  81         public $mCyrillicToLatin = [
  82
  83                 ## these are independent of location in the word, but have
  84                 ## to go first so other transforms don't bleed them
  85                 'гъ' => 'ğ', 'Гъ' => 'Ğ', 'ГЪ' => 'Ğ',
  86                 'къ' => 'q', 'Къ' => 'Q', 'КЪ' => 'Q',
  87                 'нъ' => 'ñ', 'Нъ' => 'Ñ', 'НЪ' => 'Ñ',
  88                 'дж' => 'c', 'Дж' => 'C', 'ДЖ' => 'C',
  89
  90                 'А' => 'A', 'а' => 'a', 'Б' => 'B', 'б' => 'b',
  91                 'В' => 'V', 'в' => 'v', 'Г' => 'G', 'г' => 'g',
  92                 'Д' => 'D', 'д' => 'd', 'Ж' => 'J', 'ж' => 'j',
  93                 'З' => 'Z', 'з' => 'z', 'И' => 'İ', 'и' => 'i',
  94                 'Й' => 'Y', 'й' => 'y', 'К' => 'K', 'к' => 'k',
  95                 'Л' => 'L', 'л' => 'l', 'М' => 'M', 'м' => 'm',
  96                 'Н' => 'N', 'н' => 'n', 'П' => 'P', 'п' => 'p',
  97                 'Р' => 'R', 'р' => 'r', 'С' => 'S', 'с' => 's',
  98                 'Т' => 'T', 'т' => 't', 'Ф' => 'F', 'ф' => 'f',
  99                 'Х' => 'H', 'х' => 'h', 'Ч' => 'Ç', 'ч' => 'ç',
 100                 'Ш' => 'Ş', 'ш' => 'ş', 'Ы' => 'I', 'ы' => 'ı',
 101                 'Э' => 'E', 'э' => 'e', 'Е' => 'E', 'е' => 'e',
 102                 'Я' => 'Â', 'я' => 'â', 'У' => 'U', 'у' => 'u',
 103                 'О' => 'O', 'о' => 'o',
 104
 105                 'Ё' => 'Yo', 'ё' => 'yo', 'Ю' => 'Yu', 'ю' => 'yu',
 106                 'Ц' => 'Ts', 'ц' => 'ts', 'Щ' => 'Şç', 'щ' => 'şç',
 107                 'Ь' => '', 'ь' => '', 'Ъ' => '', 'ъ' => '',
 108
 109         ];
 110
 111         public $mLatinToCyrillic = [
 112                 'Â' => 'Я', 'â' => 'я', 'B' => 'Б', 'b' => 'б',
 113                 'Ç' => 'Ч', 'ç' => 'ч', 'D' => 'Д', 'd' => 'д',
 114                 'F' => 'Ф', 'f' => 'ф', 'G' => 'Г', 'g' => 'г',
 115                 'H' => 'Х', 'h' => 'х', 'I' => 'Ы', 'ı' => 'ы',
 116                 'İ' => 'И', 'i' => 'и', 'J' => 'Ж', 'j' => 'ж',
 117                 'K' => 'К', 'k' => 'к', 'L' => 'Л', 'l' => 'л',
 118                 'M' => 'М', 'm' => 'м', 'N' => 'Н', 'n' => 'н',
 119                 'O' => 'О', 'o' => 'о', 'P' => 'П', 'p' => 'п',
 120                 'R' => 'Р', 'r' => 'р', 'S' => 'С', 's' => 'с',
 121                 'Ş' => 'Ш', 'ş' => 'ш', 'T' => 'Т', 't' => 'т',
 122                 'V' => 'В', 'v' => 'в', 'Z' => 'З', 'z' => 'з',
 123
 124                 'ya' => 'я', 'Ya' => 'Я', 'YA' => 'Я',
 125                 'ye' => 'е', 'YE' => 'Е', 'Ye' => 'Е',
 126
 127                 // hack, hack, hack
 128                 'A' => 'А', 'a' => 'а', 'E' => 'Е', 'e' => 'е',
 129                 'Ö' => 'Ё', 'ö' => 'ё', 'U' => 'У', 'u' => 'у',
 130                 'Ü' => 'Ю', 'ü' => 'ю', 'Y' => 'Й', 'y' => 'й',
 131                 'C' => 'Дж', 'c' => 'дж', 'Ğ' => 'Гъ', 'ğ' => 'гъ',
 132                 'Ñ' => 'Нъ', 'ñ' => 'нъ', 'Q' => 'Къ', 'q' => 'къ',
 133
 134                 ];
 135
 136         public $mExceptions = [];
 137         public $mCyrl2LatnPatterns = [];
 138         public $mLatn2CyrlPatterns = [];
 139         public $mCyrlCleanUpRegexes = [];
 140
 141         public $mExceptionsLoaded = false;
 142
 143         function loadDefaultTables() {
 144                 $this->mTables = [
 145                         'crh-latn' => new ReplacementArray( $this->mCyrillicToLatin ),
 146                         'crh-cyrl' => new ReplacementArray( $this->mLatinToCyrillic ),
 147                         'crh' => new ReplacementArray()
 148                 ];
 149         }
 150
 151         function loadExceptions() {
 152                 if ( $this->mExceptionsLoaded ) {
 153                         return;
 154                 }
 155
 156                 $this->mExceptionsLoaded = true;
 157                 $crhExceptions = new MediaWiki\Languages\Data\CrhExceptions();
 158                 list( $this->mExceptions, $this->mCyrl2LatnPatterns, $this->mLatn2CyrlPatterns,
 159                         $this->mCyrlCleanUpRegexes ) = $crhExceptions->loadExceptions( self::L_LC . self::C_LC,
 160                         self::L_UC . self::C_UC );
 161         }
 162
 163         /**
 164          * A function wrapper:
 165          *   - if there is no selected variant, leave the link
 166          *     names as they were
 167          *   - do not try to find variants for usernames
 168          *
 169          * @param string &$link
 170          * @param Title &$nt
 171          * @param bool $ignoreOtherCond
 172          */
 173         function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
 174                 // check for user namespace
 175                 if ( is_object( $nt ) ) {
 176                         $ns = $nt->getNamespace();
 177                         if ( $ns == NS_USER || $ns == NS_USER_TALK ) {
 178                                 return;
 179                         }
 180                 }
 181
 182                 $oldlink = $link;
 183                 parent::findVariantLink( $link, $nt, $ignoreOtherCond );
 184                 if ( $this->getPreferredVariant() == $this->mMainLanguageCode ) {
 185                         $link = $oldlink;
 186                 }
 187         }
 188
 189         /**
 190          *  It translates text into variant, specials:
 191          *    - ommiting roman numbers
 192          *
 193          * @param string $text
 194          * @param bool $toVariant
 195          *
 196          * @throws MWException
 197          * @return string
 198          */
 199         function translate( $text, $toVariant ) {
 200                 $letters = '';
 201                 switch ( $toVariant ) {
 202                         case 'crh-cyrl':
 203                                 $letters = self::L_UC . self::L_LC . "\'";
 204                                 break;
 205                         case 'crh-latn':
 206                                 $letters = self::C_UC . self::C_LC . "";
 207                                 break;
 208                         default:
 209                                 return $text;
 210                                 break;
 211                 }
 212
 213                 if ( !$this->mTablesLoaded ) {
 214                         $this->loadTables();
 215                 }
 216
 217                 if ( !isset( $this->mTables[$toVariant] ) ) {
 218                         throw new MWException( "Broken variant table: " . implode( ',', array_keys( $this->mTables ) ) );
 219                 }
 220
 221                 // check for roman numbers like VII, XIX...
 222                 // Lookahead assertion ensures $roman doesn't match the empty string
 223                 $roman = '/^(?=[MDCLXVI])M{0,4}(C[DM]|D?C{0,3})(X[LC]|L?X{0,3})(I[VX]|V?I{0,3})$/u';
 224
 225                 # match any sub-string of the relevant letters and convert it
 226                 $matches = preg_split( '/(\b|^)[^' . $letters . ']+(\b|$)/u',
 227                         $text, -1, PREG_SPLIT_OFFSET_CAPTURE );
 228                 $mstart = 0;
 229                 $ret = '';
 230                 foreach ( $matches as $m ) {
 231                         # copy over the non-matching bit
 232                         $ret .= substr( $text, $mstart, $m[1] - $mstart );
 233                         # skip certain classes of strings
 234
 235                         if ( array_key_exists( $m[0], $this->mExceptions ) ) {
 236                                 # if it's an exception, just copy down the right answer
 237                                 $ret .= $this->mExceptions[$m[0]];
 238                         } elseif ( ! $m[0] || # empty strings
 239                                          preg_match( $roman, $m[0] ) || # roman numerals
 240                                          preg_match( '/[^' . $letters . ']/', $m[0] ) # mixed orthography
 241                                         ) {
 242                                 $ret .= $m[0];
 243                         } else {
 244                                 # convert according to the rules
 245                                 $token = $this->regsConverter( $m[0], $toVariant );
 246                                 $ret .= parent::translate( $token, $toVariant );
 247                         }
 248                         $mstart = $m[1] + strlen( $m[0] );
 249                 }
 250
 251                 # pick up stray quote marks
 252                 switch ( $toVariant ) {
 253                         case 'crh-cyrl':
 254                                 $ret = strtr( $ret, [ '“' => '«', '”' => '»', ] );
 255                                 $ret = $this->regsConverter( $ret, 'cyrl-cleanup' );
 256                                 break;
 257                         case 'crh-latn':
 258                                 $ret = strtr( $ret, [ '«' => '"', '»' => '"', ] );
 259                                 break;
 260                 }
 261
 262                 return $ret;
 263         }
 264
 265         private function regsConverter( $text, $toVariant ) {
 266                 if ( $text == '' ) return $text;
 267
 268                 $pat = [];
 269                 $rep = [];
 270                 switch ( $toVariant ) {
 271                         case 'crh-latn':
 272                                 foreach ( $this->mCyrl2LatnPatterns as $pat => $rep ) {
 273                                         $text = preg_replace( $pat, $rep, $text );
 274                                 }
 275                                 return $text;
 276                         case 'crh-cyrl':
 277                                 foreach ( $this->mLatn2CyrlPatterns as $pat => $rep ) {
 278                                         $text = preg_replace( $pat, $rep, $text );
 279                                 }
 280                                 return $text;
 281                         case 'cyrl-cleanup':
 282                                 foreach ( $this->mCyrlCleanUpRegexes as $pat => $rep ) {
 283                                         $text = preg_replace( $pat, $rep, $text );
 284                                 }
 285                                 return $text;
 286                         default:
 287                                 return $text;
 288                 }
 289         }
 290
 291 }
 292
 293 /**
 294  * Crimean Tatar (Qırımtatarca)
 295  *
 296  * @ingroup Language
 297  */
 298 class LanguageCrh extends Language {
 299
 300         function __construct() {
 301                 parent::__construct();
 302
 303                 $variants = [ 'crh', 'crh-cyrl', 'crh-latn' ];
 304                 $variantfallbacks = [
 305                         'crh' => 'crh-latn',
 306                         'crh-cyrl' => 'crh-latn',
 307                         'crh-latn' => 'crh-cyrl',
 308                 ];
 309
 310                 $this->mConverter = new CrhConverter( $this, 'crh', $variants, $variantfallbacks );
 311         }
 312 }