languages/classes/LanguageIu.php

   1 <?php
   2 /**
   3  * Inuktitut specific code.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup Language
  22  */
  23
  24 require_once __DIR__ . '/../LanguageConverter.php';
  25
  26 /**
  27  * Conversion script between Latin and Syllabics for Inuktitut.
  28  * - Syllabics -> lowercase Latin
  29  * - lowercase/uppercase Latin -> Syllabics
  30  *
  31  *
  32  * Based on:
  33  *   - http://commons.wikimedia.org/wiki/Image:Inuktitut.png
  34  *   - LanguageSr.php
  35  *
  36  * @ingroup Language
  37  */
  38 class IuConverter extends LanguageConverter {
  39         protected $mDoContentConvert;
  40
  41         public $mToLatin = array(
  42                 'ᐦ' => 'h', 'ᐃ' => 'i', 'ᐄ' => 'ii', 'ᐅ' => 'u', 'ᐆ' => 'uu', 'ᐊ' => 'a', 'ᐋ' => 'aa',
  43                 'ᑉ' => 'p', 'ᐱ' => 'pi', 'ᐲ' => 'pii', 'ᐳ' => 'pu', 'ᐴ' => 'puu', 'ᐸ' => 'pa', 'ᐹ' => 'paa',
  44                 'ᑦ' => 't', 'ᑎ' => 'ti', 'ᑏ' => 'tii', 'ᑐ' => 'tu', 'ᑑ' => 'tuu', 'ᑕ' => 'ta', 'ᑖ' => 'taa',
  45                 'ᒃ' => 'k', 'ᑭ' => 'ki', 'ᑮ' => 'kii', 'ᑯ' => 'ku', 'ᑰ' => 'kuu', 'ᑲ' => 'ka', 'ᑳ' => 'kaa',
  46                 'ᖅᒃ' => 'qq', 'ᖅᑭ' => 'qqi', 'ᖅᑮ' => 'qqii', 'ᖅᑯ' => 'qqu', 'ᖅᑰ' => 'ᖅqquu', 'ᖅᑲ' => 'qqa',
  47                 'ᖅᑳ' => 'qqaa', 'ᒡ' => 'g', 'ᒋ' => 'gi', 'ᒌ' => 'gii', 'ᒍ' => 'gu', 'ᒎ' => 'guu',
  48                 'ᒐ' => 'ga', 'ᒑ' => 'gaa', 'ᒻ' => 'm', 'ᒥ' => 'mi', 'ᒦ' => 'mii', 'ᒧ' => 'mu', 'ᒨ' => 'muu',
  49                 'ᒪ' => 'ma', 'ᒫ' => 'maa', 'ᓐ' => 'n', 'ᓂ' => 'ni', 'ᓃ' => 'nii', 'ᓄ' => 'nu', 'ᓅ' => 'nuu',
  50                 'ᓇ' => 'na', 'ᓈ' => 'naa', 'ᔅ' => 's', 'ᓯ' => 'si', 'ᓰ' => 'sii', 'ᓱ' => 'su', 'ᓲ' => 'suu',
  51                 'ᓴ' => 'sa', 'ᓵ' => 'saa', 'ᓪ' => 'l', 'ᓕ' => 'li', 'ᓖ' => 'lii', 'ᓗ' => 'lu', 'ᓘ' => 'luu',
  52                 'ᓚ' => 'la', 'ᓛ' => 'laa', 'ᔾ' => 'j', 'ᔨ' => 'ji', 'ᔩ' => 'jii', 'ᔪ' => 'ju', 'ᔫ' => 'juu',
  53                 'ᔭ' => 'ja', 'ᔮ' => 'jaa', 'ᕝ' => 'v', 'ᕕ' => 'vi', 'ᕖ' => 'vii', 'ᕗ' => 'vu', 'ᕘ' => 'vuu',
  54                 'ᕙ' => 'va', 'ᕚ' => 'vaa', 'ᕐ' => 'r', 'ᕆ' => 'ri', 'ᕇ' => 'rii', 'ᕈ' => 'ru', 'ᕉ' => 'ruu',
  55                 'ᕋ' => 'ra', 'ᕌ' => 'raa', 'ᖅ' => 'q', 'ᕿ' => 'qi', 'ᖀ' => 'qii', 'ᖁ' => 'qu', 'ᖂ' => 'quu',
  56                 'ᖃ' => 'qa', 'ᖄ' => 'qaa', 'ᖕ' => 'ng', 'ᖏ' => 'ngi', 'ᖐ' => 'ngii', 'ᖑ' => 'ngu',
  57                 'ᖒ' => 'nguu', 'ᖓ' => 'nga', 'ᖔ' => 'ngaa', 'ᖖ' => 'nng', 'ᙱ' => 'nngi', 'ᙲ' => 'nngii',
  58                 'ᙳ' => 'nngu', 'ᙴ' => 'nnguu', 'ᙵ' => 'nnga', 'ᙶ' => 'nngaa', 'ᖦ' => 'ɫ', 'ᖠ' => 'ɫi',
  59                 'ᖡ' => 'ɫii', 'ᖢ' => 'ɫu', 'ᖣ' => 'ɫuu', 'ᖤ' => 'ɫa', 'ᖥ' => 'ɫaa',
  60         );
  61
  62         public $mUpperToLowerCaseLatin = array(
  63                 'A' => 'a',     'B' => 'b',     'C' => 'c',     'D' => 'd',     'E' => 'e',
  64                 'F' => 'f',     'G' => 'g',     'H' => 'h',     'I' => 'i',     'J' => 'j',
  65                 'K' => 'k',     'L' => 'l',     'M' => 'm',     'N' => 'n',     'O' => 'o',
  66                 'P' => 'p',     'Q' => 'q',     'R' => 'r',     'S' => 's',     'T' => 't',
  67                 'U' => 'u',     'V' => 'v',     'W' => 'w',     'X' => 'x',     'Y' => 'y',
  68                 'Z' => 'z',
  69         );
  70
  71         public $mToSyllabics = array(
  72                 'h' => 'ᐦ', 'i' => 'ᐃ', 'ii' => 'ᐄ', 'u' => 'ᐅ', 'uu' => 'ᐆ', 'a' => 'ᐊ', 'aa' => 'ᐋ',
  73                 'p' => 'ᑉ', 'pi' => 'ᐱ', 'pii' => 'ᐲ', 'pu' => 'ᐳ', 'puu' => 'ᐴ', 'pa' => 'ᐸ', 'paa' => 'ᐹ',
  74                 't' => 'ᑦ', 'ti' => 'ᑎ', 'tii' => 'ᑏ', 'tu' => 'ᑐ', 'tuu' => 'ᑑ', 'ta' => 'ᑕ', 'taa' => 'ᑖ',
  75                 'k' => 'ᒃ', 'ki' => 'ᑭ', 'kii' => 'ᑮ', 'ku' => 'ᑯ', 'kuu' => 'ᑰ', 'ka' => 'ᑲ', 'kaa' => 'ᑳ',
  76                 'g' => 'ᒡ', 'gi' => 'ᒋ', 'gii' => 'ᒌ', 'gu' => 'ᒍ', 'guu' => 'ᒎ', 'ga' => 'ᒐ', 'gaa' => 'ᒑ',
  77                 'm' => 'ᒻ', 'mi' => 'ᒥ', 'mii' => 'ᒦ', 'mu' => 'ᒧ', 'muu' => 'ᒨ', 'ma' => 'ᒪ', 'maa' => 'ᒫ',
  78                 'n' => 'ᓐ', 'ni' => 'ᓂ', 'nii' => 'ᓃ', 'nu' => 'ᓄ', 'nuu' => 'ᓅ', 'na' => 'ᓇ', 'naa' => 'ᓈ',
  79                 's' => 'ᔅ', 'si' => 'ᓯ', 'sii' => 'ᓰ', 'su' => 'ᓱ', 'suu' => 'ᓲ', 'sa' => 'ᓴ', 'saa' => 'ᓵ',
  80                 'l' => 'ᓪ', 'li' => 'ᓕ', 'lii' => 'ᓖ', 'lu' => 'ᓗ', 'luu' => 'ᓘ', 'la' => 'ᓚ', 'laa' => 'ᓛ',
  81                 'j' => 'ᔾ', 'ji' => 'ᔨ', 'jii' => 'ᔩ', 'ju' => 'ᔪ', 'juu' => 'ᔫ', 'ja' => 'ᔭ', 'jaa' => 'ᔮ',
  82                 'v' => 'ᕝ', 'vi' => 'ᕕ', 'vii' => 'ᕖ', 'vu' => 'ᕗ', 'vuu' => 'ᕘ', 'va' => 'ᕙ', 'vaa' => 'ᕚ',
  83                 'r' => 'ᕐ', 'ri' => 'ᕆ', 'rii' => 'ᕇ', 'ru' => 'ᕈ', 'ruu' => 'ᕉ', 'ra' => 'ᕋ', 'raa' => 'ᕌ',
  84                 'qq' => 'ᖅᒃ', 'qqi' => 'ᖅᑭ', 'qqii' => 'ᖅᑮ', 'qqu' => 'ᖅᑯ', 'qquu' => 'ᖅᑰ', 'qqa' => 'ᖅᑲ',
  85                 'qqaa' => 'ᖅᑳ', 'q' => 'ᖅ', 'qi' => 'ᕿ', 'qii' => 'ᖀ', 'qu' => 'ᖁ', 'quu' => 'ᖂ',
  86                 'qa' => 'ᖃ', 'qaa' => 'ᖄ', 'ng' => 'ᖕ', 'ngi' => 'ᖏ', 'ngii' => 'ᖐ', 'ngu' => 'ᖑ',
  87                 'nguu' => 'ᖒ', 'nga' => 'ᖓ', 'ngaa' => 'ᖔ', 'nng' => 'ᖖ', 'nngi' => 'ᙱ', 'nngii' => 'ᙲ',
  88                 'nngu' => 'ᙳ', 'nnguu' => 'ᙴ', 'nnga' => 'ᙵ', 'nngaa' => 'ᙶ', 'ɫ' => 'ᖦ', 'ɫi' => 'ᖠ',
  89                 'ɫii' => 'ᖡ', 'ɫu' => 'ᖢ', 'ɫuu' => 'ᖣ', 'ɫa' => 'ᖤ', 'ɫaa' => 'ᖥ',
  90         );
  91
  92         function loadDefaultTables() {
  93                 $this->mTables = array(
  94                         'lowercase' => new ReplacementArray( $this->mUpperToLowerCaseLatin ),
  95                         'ike-cans' => new ReplacementArray( $this->mToSyllabics ),
  96                         'ike-latn' => new ReplacementArray( $this->mToLatin ),
  97                         'iu' => new ReplacementArray()
  98                 );
  99         }
 100
 101         /**
 102          * rules should be defined as -{Syllabic | Latin-} -or-
 103          * -{code:text | code:text | ...}-
 104          * update: delete all rule parsing because it's not used
 105          * currently, and just produces a couple of bugs
 106          *
 107          * @param string $rule
 108          * @param array $flags
 109          * @return array
 110          */
 111         function parseManualRule( $rule, $flags = array() ) {
 112                 if ( in_array( 'T', $flags ) ) {
 113                         return parent::parseManualRule( $rule, $flags );
 114                 }
 115
 116                 $carray = array();
 117                 // otherwise ignore all formatting
 118                 foreach ( $this->mVariants as $v ) {
 119                         $carray[$v] = $rule;
 120                 }
 121
 122                 return $carray;
 123         }
 124
 125         /**
 126          * Do not convert content on talk pages
 127          *
 128          * @param string $text
 129          * @param Parser $parser
 130          * @return string
 131          */
 132         function parserConvert( $text, &$parser ) {
 133                 $this->mDoContentConvert = !( is_object( $parser->getTitle() )
 134                         && $parser->getTitle()->isTalkPage() );
 135
 136                 return parent::parserConvert( $text, $parser );
 137         }
 138
 139         /**
 140          * A function wrapper:
 141          *   - if there is no selected variant, leave the link
 142          *     names as they were
 143          *   - do not try to find variants for usernames
 144          *
 145          * @param string $link
 146          * @param Title $nt
 147          * @param bool $ignoreOtherCond
 148          */
 149         function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
 150                 // check for user namespace
 151                 if ( is_object( $nt ) ) {
 152                         $ns = $nt->getNamespace();
 153                         if ( $ns == NS_USER || $ns == NS_USER_TALK ) {
 154                                 return;
 155                         }
 156                 }
 157
 158                 $oldlink = $link;
 159                 parent::findVariantLink( $link, $nt, $ignoreOtherCond );
 160                 if ( $this->getPreferredVariant() == $this->mMainLanguageCode ) {
 161                         $link = $oldlink;
 162                 }
 163         }
 164
 165         /**
 166          * An ugly function wrapper for parsing Image titles
 167          * (to prevent image name conversion)
 168          *
 169          * @param string $text
 170          * @param bool $toVariant
 171          *
 172          * @return string
 173          */
 174         function autoConvert( $text, $toVariant = false ) {
 175                 global $wgTitle;
 176                 if ( is_object( $wgTitle ) && $wgTitle->getNamespace() == NS_FILE ) {
 177                         $imagename = $wgTitle->getNsText();
 178                         if ( preg_match( "/^$imagename:/", $text ) ) {
 179                                 return $text;
 180                         }
 181                 }
 182                 return parent::autoConvert( $text, $toVariant );
 183         }
 184
 185         /**
 186          * It translates text into variant
 187          *
 188          * @param string $text
 189          * @param bool $toVariant
 190          *
 191          * @return string
 192          */
 193         function translate( $text, $toVariant ) {
 194                 // If $text is empty or only includes spaces, do nothing
 195                 // Otherwise translate it
 196                 if ( trim( $text ) ) {
 197                         $this->loadTables();
 198                         // To syllabics, first translate uppercase to lowercase Latin
 199                         if ( $toVariant == 'ike-cans' ) {
 200                                 $text = $this->mTables['lowercase']->replace( $text );
 201                         }
 202                         $text = $this->mTables[$toVariant]->replace( $text );
 203                 }
 204                 return $text;
 205         }
 206 }
 207
 208 /**
 209  * Inuktitut
 210  *
 211  * @ingroup Language
 212  */
 213 class LanguageIu extends Language {
 214         function __construct() {
 215                 global $wgHooks;
 216
 217                 parent::__construct();
 218
 219                 $variants = array( 'iu', 'ike-cans', 'ike-latn' );
 220                 $variantfallbacks = array(
 221                         'iu' => 'ike-cans',
 222                         'ike-cans' => 'iu',
 223                         'ike-latn' => 'iu',
 224                 );
 225
 226                 $flags = array();
 227                 $this->mConverter = new IuConverter( $this, 'iu', $variants, $variantfallbacks, $flags );
 228                 $wgHooks['PageContentSaveComplete'][] = $this->mConverter;
 229         }
 230 }