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