includes/language/converters/KkConverter.php

   1 <?php
   2 /**
   3  * This program is free software; you can redistribute it and/or modify
   4  * it under the terms of the GNU General Public License as published by
   5  * the Free Software Foundation; either version 2 of the License, or
   6  * (at your option) any later version.
   7  *
   8  * This program is distributed in the hope that it will be useful,
   9  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  11  * GNU General Public License for more details.
  12  *
  13  * You should have received a copy of the GNU General Public License along
  14  * with this program; if not, write to the Free Software Foundation, Inc.,
  15  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  16  * http://www.gnu.org/copyleft/gpl.html
  17  *
  18  * @file
  19  */
  20
  21 define( 'KK_C_UC', 'АӘБВГҒДЕЁЖЗИЙКҚЛМНҢОӨПРСТУҰҮФХҺЦЧШЩЪЫІЬЭЮЯ' ); # Kazakh Cyrillic uppercase
  22 define( 'KK_C_LC', 'аәбвгғдеёжзийкқлмнңоөпрстуұүфхһцчшщъыіьэюя' ); # Kazakh Cyrillic lowercase
  23 define( 'KK_L_UC', 'AÄBCÇDEÉFGĞHIİÏJKLMNÑOÖPQRSŞTUÜVWXYÝZ' ); # Kazakh Latin uppercase
  24 define( 'KK_L_LC', 'aäbcçdeéfgğhıiïjklmnñoöpqrsştuüvwxyýz' ); # Kazakh Latin lowercase
  25 // define( 'KK_A', 'ٴابپتجحدرزسشعفقكلمنڭەوۇۋۆىيچھ' ); # Kazakh Arabic
  26 define( 'H_HAMZA', 'ٴ' ); # U+0674 ARABIC LETTER HIGH HAMZA
  27 // define( 'ZWNJ', '‌' ); # U+200C ZERO WIDTH NON-JOINER
  28
  29 /**
  30  * Kazakh (Қазақша) converter routines
  31  *
  32  * @ingroup Languages
  33  */
  34 class KkConverter extends LanguageConverterSpecific {
  35
  36         public function getMainCode(): string {
  37                 return 'kk';
  38         }
  39
  40         public function getLanguageVariants(): array {
  41                 return [
  42                         'kk',
  43                         'kk-cyrl',
  44                         'kk-latn',
  45                         'kk-arab',
  46                         'kk-kz',
  47                         'kk-tr',
  48                         'kk-cn'
  49                 ];
  50         }
  51
  52         public function getVariantsFallbacks(): array {
  53                 return [
  54                         'kk' => 'kk-cyrl',
  55                         'kk-cyrl' => 'kk',
  56                         'kk-latn' => 'kk',
  57                         'kk-arab' => 'kk',
  58                         'kk-kz' => 'kk-cyrl',
  59                         'kk-tr' => 'kk-latn',
  60                         'kk-cn' => 'kk-arab'
  61                 ];
  62         }
  63
  64         protected function loadDefaultTables(): array {
  65                 // require __DIR__."/../../includes/KkConversion.php";
  66                 // Placeholder for future implementing. Remove the variables declarations
  67                 // after generating KkConversion.php
  68                 $kk2Cyrl = [];
  69                 $kk2Latn = [];
  70                 $kk2Arab = [];
  71                 $kk2KZ = [];
  72                 $kk2TR = [];
  73                 $kk2CN = [];
  74
  75                 return [
  76                         'kk-cyrl' => new ReplacementArray( $kk2Cyrl ),
  77                         'kk-latn' => new ReplacementArray( $kk2Latn ),
  78                         'kk-arab' => new ReplacementArray( $kk2Arab ),
  79                         'kk-kz' => new ReplacementArray( array_merge( $kk2Cyrl, $kk2KZ ) ),
  80                         'kk-tr' => new ReplacementArray( array_merge( $kk2Latn, $kk2TR ) ),
  81                         'kk-cn' => new ReplacementArray( array_merge( $kk2Arab, $kk2CN ) ),
  82                         'kk' => new ReplacementArray()
  83                 ];
  84         }
  85
  86         protected function postLoadTables( &$tables ) {
  87                 $tables['kk-kz']->merge( $tables['kk-cyrl'] );
  88                 $tables['kk-tr']->merge( $tables['kk-latn'] );
  89                 $tables['kk-cn']->merge( $tables['kk-arab'] );
  90         }
  91
  92         /**
  93          * Return cyrillic to latin reg conversion table
  94          * @since 1.36
  95          *
  96          * @return array
  97          */
  98         protected function getMCyrl2Latn(): array {
  99                 return [
 100                         # # Punctuation
 101                         '/№/u' => 'No.',
 102                         # # Е after vowels
 103                         '/([АӘЕЁИОӨҰҮЭЮЯЪЬ])Е/u' => '$1YE',
 104                         '/([АӘЕЁИОӨҰҮЭЮЯЪЬ])е/ui' => '$1ye',
 105                         # # leading ЁЮЯЩ
 106                         '/^Ё([' . KK_C_UC . ']|$)/u' => 'YO$1', '/^Ё([' . KK_C_LC . ']|$)/u' => 'Yo$1',
 107                         '/^Ю([' . KK_C_UC . ']|$)/u' => 'YU$1', '/^Ю([' . KK_C_LC . ']|$)/u' => 'Yu$1',
 108                         '/^Я([' . KK_C_UC . ']|$)/u' => 'YA$1', '/^Я([' . KK_C_LC . ']|$)/u' => 'Ya$1',
 109                         '/^Щ([' . KK_C_UC . ']|$)/u' => 'ŞÇ$1', '/^Щ([' . KK_C_LC . ']|$)/u' => 'Şç$1',
 110                         # # other ЁЮЯ
 111                         '/Ё/u' => 'YO', '/ё/u' => 'yo',
 112                         '/Ю/u' => 'YU', '/ю/u' => 'yu',
 113                         '/Я/u' => 'YA', '/я/u' => 'ya',
 114                         '/Щ/u' => 'ŞÇ', '/щ/u' => 'şç',
 115                         # # soft and hard signs
 116                         '/[ъЪ]/u' => 'ʺ', '/[ьЬ]/u' => 'ʹ',
 117                         # # other characters
 118                         '/А/u' => 'A', '/а/u' => 'a', '/Ә/u' => 'Ä', '/ә/u' => 'ä',
 119                         '/Б/u' => 'B', '/б/u' => 'b', '/В/u' => 'V', '/в/u' => 'v',
 120                         '/Г/u' => 'G', '/г/u' => 'g', '/Ғ/u' => 'Ğ', '/ғ/u' => 'ğ',
 121                         '/Д/u' => 'D', '/д/u' => 'd', '/Е/u' => 'E', '/е/u' => 'e',
 122                         '/Ж/u' => 'J', '/ж/u' => 'j', '/З/u' => 'Z', '/з/u' => 'z',
 123                         '/И/u' => 'Ï', '/и/u' => 'ï', '/Й/u' => 'Ý', '/й/u' => 'ý',
 124                         '/К/u' => 'K', '/к/u' => 'k', '/Қ/u' => 'Q', '/қ/u' => 'q',
 125                         '/Л/u' => 'L', '/л/u' => 'l', '/М/u' => 'M', '/м/u' => 'm',
 126                         '/Н/u' => 'N', '/н/u' => 'n', '/Ң/u' => 'Ñ', '/ң/u' => 'ñ',
 127                         '/О/u' => 'O', '/о/u' => 'o', '/Ө/u' => 'Ö', '/ө/u' => 'ö',
 128                         '/П/u' => 'P', '/п/u' => 'p', '/Р/u' => 'R', '/р/u' => 'r',
 129                         '/С/u' => 'S', '/с/u' => 's', '/Т/u' => 'T', '/т/u' => 't',
 130                         '/У/u' => 'W', '/у/u' => 'w', '/Ұ/u' => 'U', '/ұ/u' => 'u',
 131                         '/Ү/u' => 'Ü', '/ү/u' => 'ü', '/Ф/u' => 'F', '/ф/u' => 'f',
 132                         '/Х/u' => 'X', '/х/u' => 'x', '/Һ/u' => 'H', '/һ/u' => 'h',
 133                         '/Ц/u' => 'C', '/ц/u' => 'c', '/Ч/u' => 'Ç', '/ч/u' => 'ç',
 134                         '/Ш/u' => 'Ş', '/ш/u' => 'ş', '/Ы/u' => 'I', '/ы/u' => 'ı',
 135                         '/І/u' => 'İ', '/і/u' => 'i', '/Э/u' => 'É', '/э/u' => 'é',
 136                 ];
 137         }
 138
 139         /**
 140          * Return latin to cyrillic reg conversion table
 141          * @since 1.36
 142          *
 143          * @return array
 144          */
 145         protected function getMLatn2Cyrl(): array {
 146                 return [
 147                         # # Punctuation
 148                         '/#|No\./' => '№',
 149                         # # Şç
 150                         '/ŞÇʹ/u' => 'ЩЬ', '/Şçʹ/u' => 'Щь',
 151                         '/Ş[Çç]/u' => 'Щ', '/şç/u' => 'щ',
 152                         # # soft and hard signs
 153                         '/([' . KK_L_UC . '])ʺ([' . KK_L_UC . '])/u' => '$1Ъ$2',
 154                         '/ʺ([' . KK_L_LC . '])/u' => 'ъ$1',
 155                         '/([' . KK_L_UC . '])ʹ([' . KK_L_UC . '])/u' => '$1Ь$2',
 156                         '/ʹ([' . KK_L_LC . '])/u' => 'ь$1',
 157                         '/ʺ/u' => 'ъ',
 158                         '/ʹ/u' => 'ь',
 159                         # # Ye Yo Yu Ya.
 160                         '/Y[Ee]/u' => 'Е', '/ye/u' => 'е',
 161                         '/Y[Oo]/u' => 'Ё', '/yo/u' => 'ё',
 162                         '/Y[UWuw]/u' => 'Ю', '/y[uw]/u' => 'ю',
 163                         '/Y[Aa]/u' => 'Я', '/ya/u' => 'я',
 164                         # # other characters
 165                         '/A/u' => 'А', '/a/u' => 'а', '/Ä/u' => 'Ә', '/ä/u' => 'ә',
 166                         '/B/u' => 'Б', '/b/u' => 'б', '/C/u' => 'Ц', '/c/u' => 'ц',
 167                         '/Ç/u' => 'Ч', '/ç/u' => 'ч', '/D/u' => 'Д', '/d/u' => 'д',
 168                         '/E/u' => 'Е', '/e/u' => 'е', '/É/u' => 'Э', '/é/u' => 'э',
 169                         '/F/u' => 'Ф', '/f/u' => 'ф', '/G/u' => 'Г', '/g/u' => 'г',
 170                         '/Ğ/u' => 'Ғ', '/ğ/u' => 'ғ', '/H/u' => 'Һ', '/h/u' => 'һ',
 171                         '/I/u' => 'Ы', '/ı/u' => 'ы', '/İ/u' => 'І', '/i/u' => 'і',
 172                         '/Ï/u' => 'И', '/ï/u' => 'и', '/J/u' => 'Ж', '/j/u' => 'ж',
 173                         '/K/u' => 'К', '/k/u' => 'к', '/L/u' => 'Л', '/l/u' => 'л',
 174                         '/M/u' => 'М', '/m/u' => 'м', '/N/u' => 'Н', '/n/u' => 'н',
 175                         '/Ñ/u' => 'Ң', '/ñ/u' => 'ң', '/O/u' => 'О', '/o/u' => 'о',
 176                         '/Ö/u' => 'Ө', '/ö/u' => 'ө', '/P/u' => 'П', '/p/u' => 'п',
 177                         '/Q/u' => 'Қ', '/q/u' => 'қ', '/R/u' => 'Р', '/r/u' => 'р',
 178                         '/S/u' => 'С', '/s/u' => 'с', '/Ş/u' => 'Ш', '/ş/u' => 'ш',
 179                         '/T/u' => 'Т', '/t/u' => 'т', '/U/u' => 'Ұ', '/u/u' => 'ұ',
 180                         '/Ü/u' => 'Ү', '/ü/u' => 'ү', '/V/u' => 'В', '/v/u' => 'в',
 181                         '/W/u' => 'У', '/w/u' => 'у', '/Ý/u' => 'Й', '/ý/u' => 'й',
 182                         '/X/u' => 'Х', '/x/u' => 'х', '/Z/u' => 'З', '/z/u' => 'з',
 183                 ];
 184         }
 185
 186         /**
 187          * Return latin or cyrillic to arab reg conversion table.
 188          * @since 1.36
 189          *
 190          * @return array
 191          */
 192         public function getMCyLa2Arab() {
 193                 return [
 194                         # # Punctuation -> Arabic
 195                         '/#|№|No\./u' => '؀', # U+0600
 196                         '/\,/' => '،', # U+060C
 197                         '/;/' => '؛', # U+061B
 198                         '/\?/' => '؟', # U+061F
 199                         '/%/' => '٪', # U+066A
 200                         '/\*/' => '٭', # U+066D
 201                         # # Digits -> Arabic
 202                         '/0/' => '۰', # U+06F0
 203                         '/1/' => '۱', # U+06F1
 204                         '/2/' => '۲', # U+06F2
 205                         '/3/' => '۳', # U+06F3
 206                         '/4/' => '۴', # U+06F4
 207                         '/5/' => '۵', # U+06F5
 208                         '/6/' => '۶', # U+06F6
 209                         '/7/' => '۷', # U+06F7
 210                         '/8/' => '۸', # U+06F8
 211                         '/9/' => '۹', # U+06F9
 212                         # # Cyrillic -> Arabic
 213                         '/Аллаһ/ui' => 'ﷲ',
 214                         '/([АӘЕЁИОӨҰҮЭЮЯЪЬ])е/ui' => '$1يە',
 215                         '/[еэ]/ui' => 'ە', '/[ъь]/ui' => '',
 216                         '/[аә]/ui' => 'ا', '/[оө]/ui' => 'و', '/[ұү]/ui' => 'ۇ', '/[ыі]/ui' => 'ى',
 217                         '/[и]/ui' => 'ىي', '/ё/ui' => 'يو', '/ю/ui' => 'يۋ', '/я/ui' => 'يا', '/[й]/ui' => 'ي',
 218                         '/ц/ui' => 'تس', '/щ/ui' => 'شش',
 219                         '/һ/ui' => 'ح', '/ч/ui' => 'تش',
 220                         # '/һ/ui' => 'ھ', '/ч/ui' => 'چ',
 221                         '/б/ui' => 'ب', '/в/ui' => 'ۆ', '/г/ui' => 'گ', '/ғ/ui' => 'ع',
 222                         '/д/ui' => 'د', '/ж/ui' => 'ج', '/з/ui' => 'ز', '/к/ui' => 'ك',
 223                         '/қ/ui' => 'ق', '/л/ui' => 'ل', '/м/ui' => 'م', '/н/ui' => 'ن',
 224                         '/ң/ui' => 'ڭ', '/п/ui' => 'پ', '/р/ui' => 'ر', '/с/ui' => 'س',
 225                         '/т/ui' => 'ت', '/у/ui' => 'ۋ', '/ф/ui' => 'ف', '/х/ui' => 'ح',
 226                         '/ш/ui' => 'ش',
 227                         # # Latin -> Arabic // commented for now...
 228                         /*'/Allah/ui' => 'ﷲ',
 229                         '/[eé]/ui' => 'ە', '/[yý]/ui' => 'ي', '/[ʺʹ]/ui' => '',
 230                         '/[aä]/ui' => 'ا', '/[oö]/ui' => 'و', '/[uü]/ui' => 'ۇ',
 231                         '/[ï]/ui' => 'ىي', '/[ıIiİ]/u' => 'ى',
 232                         '/c/ui' => 'تس',
 233                         '/ç/ui' => 'تش', '/h/ui' => 'ح',
 234                         #'/ç/ui' => 'چ', '/h/ui' => 'ھ',
 235                         '/b/ui' => 'ب','/d/ui' => 'د',
 236                         '/f/ui' => 'ف', '/g/ui' => 'گ', '/ğ/ui' => 'ع',
 237                         '/j/ui' => 'ج', '/k/ui' => 'ك', '/l/ui' => 'ل', '/m/ui' => 'م',
 238                         '/n/ui' => 'ن', '/ñ/ui' => 'ڭ', '/p/ui' => 'پ', '/q/ui' => 'ق',
 239                         '/r/ui' => 'ر', '/s/ui' => 'س', '/ş/ui' => 'ش', '/t/ui' => 'ت',
 240                         '/v/ui' => 'ۆ', '/w/ui' => 'ۋ', '/x/ui' => 'ح', '/z/ui' => 'ز',*/
 241                 ];
 242         }
 243
 244         public function translate( $text, $toVariant ) {
 245                 $text = parent::translate( $text, $toVariant );
 246
 247                 switch ( $toVariant ) {
 248                         case 'kk-cyrl':
 249                         case 'kk-kz':
 250                                 $letters = KK_L_UC . KK_L_LC . 'ʺʹ#0123456789';
 251                                 break;
 252
 253                         case 'kk-latn':
 254                         case 'kk-tr':
 255                                 $letters = KK_C_UC . KK_C_LC . '№0123456789';
 256                                 break;
 257
 258                         case 'kk-arab':
 259                         case 'kk-cn':
 260                                 $letters = KK_C_UC . KK_C_LC . /*KK_L_UC.KK_L_LC.'ʺʹ'.*/',;\?%\*№0123456789';
 261                                 break;
 262
 263                         default:
 264                                 return $text;
 265                 }
 266                 // disable conversion variables like $1, $2...
 267                 $varsfix = '\$[0-9]';
 268
 269                 $matches = preg_split(
 270                         '/' . $varsfix . '[^' . $letters . ']+/u',
 271                         $text,
 272                         -1,
 273                         PREG_SPLIT_OFFSET_CAPTURE
 274                 );
 275
 276                 $mstart = 0;
 277                 $ret = '';
 278
 279                 foreach ( $matches as $m ) {
 280                         $ret .= substr( $text, $mstart, (int)$m[1] - $mstart );
 281                         $ret .= $this->regsConverter( $m[0], $toVariant );
 282                         $mstart = (int)$m[1] + strlen( $m[0] );
 283                 }
 284
 285                 return $ret;
 286         }
 287
 288         /**
 289          * @param string $text
 290          * @param string $toVariant
 291          * @return mixed|string
 292          */
 293         private function regsConverter( $text, $toVariant ) {
 294                 if ( $text == '' ) {
 295                         return $text;
 296                 }
 297
 298                 switch ( $toVariant ) {
 299                         case 'kk-arab':
 300                         case 'kk-cn':
 301                                 $letters = KK_C_LC . KK_C_UC; /*.KK_L_LC.KK_L_UC*/
 302                                 $front = 'әөүіӘӨҮІ'; /*.'äöüiÄÖÜİ'*/
 303                                 $excludes = 'еэгғкқЕЭГҒКҚ'; /*.'eégğkqEÉGĞKQ'*/
 304                                 // split text to words
 305                                 $matches = preg_split( '/[\b\s\-\.:]+/', $text, -1, PREG_SPLIT_OFFSET_CAPTURE );
 306                                 $mstart = 0;
 307                                 $ret = '';
 308                                 foreach ( $matches as $m ) {
 309                                         $ret .= substr( $text, $mstart, (int)$m[1] - $mstart );
 310                                         // is matched the word to front vowels?
 311                                         // exclude a words matched to е, э, г, к, к, қ,
 312                                         // them should be without hamza
 313                                         if ( preg_match( '/[' . $front . ']/u', $m[0] ) &&
 314                                                 !preg_match( '/[' . $excludes . ']/u', $m[0] )
 315                                         ) {
 316                                                 $ret .= preg_replace( '/[' . $letters . ']+/u', H_HAMZA . '$0', $m[0] );
 317                                         } else {
 318                                                 $ret .= $m[0];
 319                                         }
 320                                         $mstart = (int)$m[1] + strlen( $m[0] );
 321                                 }
 322                                 $text =& $ret;
 323                                 $mCyLa2Arab = $this->getMCyLa2Arab();
 324                                 foreach ( $mCyLa2Arab as $pat => $rep ) {
 325                                         $text = preg_replace( $pat, $rep, $text );
 326                                 }
 327                                 return $text;
 328
 329                         case 'kk-latn':
 330                         case 'kk-tr':
 331                                 $mCyrl2Latn = $this->getMCyrl2Latn();
 332                                 foreach ( $mCyrl2Latn as $pat => $rep ) {
 333                                         $text = preg_replace( $pat, $rep, $text );
 334                                 }
 335                                 return $text;
 336
 337                         case 'kk-cyrl':
 338                         case 'kk-kz':
 339                                 $mLatn2Cyrl = $this->getMLatn2Cyrl();
 340                                 foreach ( $mLatn2Cyrl as $pat => $rep ) {
 341                                         $text = preg_replace( $pat, $rep, $text );
 342                                 }
 343                                 return $text;
 344
 345                         default:
 346                                 return $text;
 347                 }
 348         }
 349
 350         public function convertCategoryKey( $key ) {
 351                 return $this->autoConvert( $key, 'kk' );
 352         }
 353 }