generateNormalizerDataMl.php

Go to the documentation of this file.

<?php
require_once __DIR__ . '/../Maintenance.php';
 
class GenerateNormalizerDataMl extends Maintenance {
    public function __construct() {
        parent::__construct();
        $this->addDescription( 'Generate the normalizer data file for Malayalam' );
    }
 
    public function getDbType() {
        return Maintenance::DB_NONE;
    }
 
    public function execute() {
        $hexPairs = [
            # From http://unicode.org/versions/Unicode5.1.0/#Malayalam_Chillu_Characters
            '0D23 0D4D 200D' => '0D7A',
            '0D28 0D4D 200D' => '0D7B',
            '0D30 0D4D 200D' => '0D7C',
            '0D32 0D4D 200D' => '0D7D',
            '0D33 0D4D 200D' => '0D7E',
 
            # From http://permalink.gmane.org/gmane.science.linguistics.wikipedia.technical/46413
            '0D15 0D4D 200D' => '0D7F',
        ];
 
        $pairs = [];
        foreach ( $hexPairs as $hexSource => $hexDest ) {
            $source = UtfNormal\Utils::hexSequenceToUtf8( $hexSource );
            $dest = UtfNormal\Utils::hexSequenceToUtf8( $hexDest );
            $pairs[$source] = $dest;
        }
 
        global $IP;
        file_put_contents( "$IP/serialized/normalize-ml.ser", serialize( $pairs ) );
        echo "ml: " . count( $pairs ) . " pairs written.\n";
    }
}
 
$maintClass = GenerateNormalizerDataMl::class;
require_once RUN_MAINTENANCE_IF_MAIN;