File Coverage

blib/lib/Lingua/ZH/CEDICT/Textfile.pm

Criterion	Covered	Total	%
statement	18	37	48.6
branch	0	10	0.0
condition	3	12	25.0
subroutine	5	6	83.3
pod	1	2	50.0
total	27	67	40.3

package Lingua::ZH::CEDICT::Textfile;

# Copyright (c) 2002-2005 Christian Renz
# This module is free software; you can redistribute it and/or modify
# it under the same terms as Perl itself.

# $Id: Textfile.pm,v 1.3 2002/08/13 20:57:45 crenz Exp $

use bytes;
use strict;
use warnings;
use vars qw($VERSION @ISA);

$VERSION = '0.03';
@ISA = qw(Lingua::ZH::CEDICT);

sub new {
    my $class = shift;
    my $self = +{@_};

    $self->{filename}      ||= "cedict_ts.u8";
    $self->{sourceCharSet} ||= "UTF-8";
    $self->{targetCharSet} ||= "UTF-8";

    bless $self, $class;
}

sub init {
    my ($self) = @_;
    my $fh;

    my $iconv;
    if ($self->{sourceCharset} && $self->{targetCharset} &&
        ($self->{sourceCharset} ne $self->{targetCharset})) {
        require Text::Iconv;
        Text::Iconv->raise_error(1);
        $iconv = Text::Iconv->new($self->{sourceCharset}, $self->{targetCharset});
    }

    $self->{entry} = [];
    open $fh, "<" . $self->{filename}
        or die "Can't open $self->{filename}: $!\n";
    $self->{version} = <$fh>;
    while (<$fh>) {
        next unless /\w/;
        $_ = $iconv->convert($_) if defined $iconv;

        m|^(\S+)\s(\S+)\s\[([a-z0-9: ]+)\]\s/(.*)/\s*$| or
#        m|^(\S+\|\S*)\s\[([a-z0-9: ]+)\]\s/(.*)/\s*$| or
            die "Line $.: Invalid entry '$_'\n";
#        my @zi = split /\|/, $1;
#        $zi[1] ||= '';
#        $zi[1] = '' if (index($zi[1], '?') >= 0);
        my $tonelessPinyin = $self->removePinyinTones($3);
#        print "@zi // $2 // $3\n";
        push @{$self->{entry}}, [ $1, $2, $3, $tonelessPinyin, $4 ];
#        push @{$self->{entry}}, [ $zi[0], $zi[1], $2, $tonelessPinyin, $3 ];
    }
    close $fh;

    $self->{numEntries} = scalar @{$self->{entry}};
}

1;

__END__