view release on metacpan or  search on metacpan

lib/Lingua/Han/Cantonese.pm  view on Meta::CPAN

package Lingua::Han::Cantonese;

use warnings;
use strict;
use vars qw($VERSION);
$VERSION = '0.12';

use File::Spec;
use Lingua::Han::Utils qw/Unihan_value/;

sub new {
	my $class = shift;
	my $dir = __FILE__; $dir =~ s/\.pm//o;
	-d $dir or die "Directory $dir nonexistent!";
	my $self = { @_ };
	my %ct;
	my $file = File::Spec->catfile($dir, 'Cantonese.dat');
	open(FH, $file)	or die "$file: $!";
	while(<FH>) {
		my ($uni, $ct) = split(/\s+/);
		$ct{$uni} = $ct;
	}
	close(FH);
	$self->{'ct'} = \%ct;
	return bless $self => $class;
}

sub han2Cantonese {
	my ($self, $hanzi) = @_;

	my @code = Unihan_value($hanzi);

	my @result;
	foreach my $code (@code) {
		my $value = $self->{'ct'}->{$code};
		if (defined $value) {
			$value =~ s/\d//isg unless ($self->{'tone'});
		} else {
			# if it's not a Chinese, return original word
			$value = pack("U*", hex $code);
		}
		push @result, lc $value;
	}

	return wantarray ? @result : join('', @result);

}

1;
__END__
=encoding utf8

=head1 NAME

Lingua::Han::Cantonese - Retrieve the Cantonese(GuangDongHua) of Chinese character(HanZi).

=head1 SYNOPSIS

  use Lingua::Han::Cantonese;

  my $h2p = new Lingua::Han::Cantonese();
  print $h2p->han2Cantonese("我"); # ngo
  my @result = $h2p->han2Cantonese("爱你"); # @result = ('ngoi', 'nei');

  # we can set the tone up
  my $h2p = new Lingua::Han::Cantonese(tone => 1);
  print $h2p->han2Cantonese("我"); #ngo5
  my @result = $h2p->han2Cantonese("爱你"); # @result = ('ngoi3', 'nei5');
  print $h2p->han2Cantonese("林道"); #lam4dou3
  print $h2p->han2Cantonese("I love 余瑞华 a"); #i love jyu4seoi6waa4 a

=head1 DESCRIPTION

Retrieve the Cantonese(GuangDongHua) of Chinese character(HanZi).

=head1 RETURN VALUE

Usually, it returns its Cantonese/spell. It includes more than 20,000 words (from Unicode.org Unihan.txt, version 4.1.0).

if not(I mean it's not a Chinese character), returns the original word;

=head1 OPTION

=over 4