maintenance/language/generateNormalizerData.php

   1 <?php
   2 /**
   3  * Generates normalizer data files for Arabic and Malayalam.
   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 MaintenanceLanguage
  22  */
  23
  24 require_once( __DIR__ . '/../../includes/normal/UtfNormalUtil.php' );
  25
  26 require_once( __DIR__ . '/../Maintenance.php' );
  27
  28 /**
  29  * Generates normalizer data files for Arabic and Malayalam.
  30  * For NFC see includes/normal.
  31  *
  32  * @ingroup MaintenanceLanguage
  33  */
  34 class GenerateNormalizerData extends Maintenance {
  35         public $dataFile;
  36
  37         public function __construct() {
  38                 parent::__construct();
  39                 $this->addOption( 'unicode-data-file', 'The local location of the data file ' .
  40                         'from http://unicode.org/Public/UNIDATA/UnicodeData.txt', false, true );
  41         }
  42
  43         public function execute() {
  44                 if ( !$this->hasOption( 'unicode-data-file' ) ) {
  45                         $this->dataFile = 'UnicodeData.txt';
  46                         if ( !file_exists( $this->dataFile ) ) {
  47                                 $this->error( "Unable to find UnicodeData.txt. Please specify its location with --unicode-data-file=<FILE>" );
  48                                 exit( 1 );
  49                         }
  50                 } else {
  51                         $this->dataFile = $this->getOption( 'unicode-data-file' );
  52                         if ( !file_exists( $this->dataFile ) ) {
  53                                 $this->error( 'Unable to find the specified data file.' );
  54                                 exit( 1 );
  55                         }
  56                 }
  57
  58                 $this->generateArabic();
  59                 $this->generateMalayalam();
  60         }
  61
  62         function generateArabic() {
  63                 $file = fopen( $this->dataFile, 'r' );
  64                 if ( !$file ) {
  65                         $this->error( 'Unable to open the data file.' );
  66                         exit( 1 );
  67                 }
  68
  69                 // For the file format, see http://www.unicode.org/reports/tr44/
  70                 $fieldNames = array(
  71                         'Code',
  72                         'Name',
  73                         'General_Category',
  74                         'Canonical_Combining_Class',
  75                         'Bidi_Class',
  76                         'Decomposition_Type_Mapping',
  77                         'Numeric_Type_Value',
  78                         'Bidi_Mirrored',
  79                         'Unicode_1_Name',
  80                         'ISO_Comment',
  81                         'Simple_Uppercase_Mapping',
  82                         'Simple_Lowercase_Mapping',
  83                         'Simple_Titlecase_Mapping'
  84                 );
  85
  86                 $pairs = array();
  87
  88                 $lineNum = 0;
  89                 while ( false !== ( $line = fgets( $file ) ) ) {
  90                         ++$lineNum;
  91
  92                         # Strip comments
  93                         $line = trim( substr( $line, 0, strcspn( $line, '#' ) ) );
  94                         if ( $line === '' ) {
  95                                 continue;
  96                         }
  97
  98                         # Split fields
  99                         $numberedData = explode( ';', $line );
 100                         $data = array();
 101                         foreach ( $fieldNames as $number => $name ) {
 102                                 $data[$name] = $numberedData[$number];
 103                         }
 104
 105                         $code = base_convert( $data['Code'], 16, 10 );
 106                         if ( ( $code >= 0xFB50 && $code <= 0xFDFF ) # Arabic presentation forms A
 107                                 || ( $code >= 0xFE70 && $code <= 0xFEFF ) ) # Arabic presentation forms B
 108                         {
 109                                 if ( $data['Decomposition_Type_Mapping'] === '' ) {
 110                                         // No decomposition
 111                                         continue;
 112                                 }
 113                                 if ( !preg_match( '/^ *(<\w*>) +([0-9A-F ]*)$/',
 114                                         $data['Decomposition_Type_Mapping'], $m ) )
 115                                 {
 116                                         $this->error( "Can't parse Decomposition_Type/Mapping on line $lineNum" );
 117                                         $this->error( $line );
 118                                         continue;
 119                                 }
 120
 121                                 $source = hexSequenceToUtf8( $data['Code'] );
 122                                 $dest = hexSequenceToUtf8( $m[2] );
 123                                 $pairs[$source] = $dest;
 124                         }
 125                 }
 126
 127                 global $IP;
 128                 file_put_contents( "$IP/serialized/normalize-ar.ser", serialize( $pairs ) );
 129                 echo "ar: " . count( $pairs ) . " pairs written.\n";
 130         }
 131
 132         function generateMalayalam() {
 133                 $hexPairs = array(
 134                         # From http://unicode.org/versions/Unicode5.1.0/#Malayalam_Chillu_Characters
 135                         '0D23 0D4D 200D' => '0D7A',
 136                         '0D28 0D4D 200D' => '0D7B',
 137                         '0D30 0D4D 200D' => '0D7C',
 138                         '0D32 0D4D 200D' => '0D7D',
 139                         '0D33 0D4D 200D' => '0D7E',
 140
 141                         # From http://permalink.gmane.org/gmane.science.linguistics.wikipedia.technical/46413
 142                         '0D15 0D4D 200D' => '0D7F',
 143                 );
 144
 145                 $pairs = array();
 146                 foreach ( $hexPairs as $hexSource => $hexDest ) {
 147                         $source = hexSequenceToUtf8( $hexSource );
 148                         $dest = hexSequenceToUtf8( $hexDest );
 149                         $pairs[$source] = $dest;
 150                 }
 151
 152                 global $IP;
 153                 file_put_contents( "$IP/serialized/normalize-ml.ser", serialize( $pairs ) );
 154                 echo "ml: " . count( $pairs ) . " pairs written.\n";
 155         }
 156 }
 157
 158 $maintClass = 'GenerateNormalizerData';
 159 require_once( RUN_MAINTENANCE_IF_MAIN );