WebPAC/Normalize/XML.pm

package WebPAC::Normalize::XML;

use warnings;
use strict;

use base qw/WebPAC::Common WebPAC::Normalize/;
use XML::Simple;
use Data::Dumper;
use Text::Iconv;

=head1 NAME

WebPAC::Normalize::XML - apply XML normalisaton rules

=head1 VERSION

Version 0.01

=cut

our $VERSION = '0.01';

=head1 SYNOPSIS

This module uses C<conf/normalize/*.xml> files to perform normalisation
from input records

=cut

=head1 FUNCTIONS

=head2 open

Read normalisation rules defined using XML from C<conf/normalize/*.xml> and
parse it.

 my $n = new WebPAC::Normalize::XML;
 $n->open(
        tag => 'isis',
        xml_file => '/path/to/conf/normalize/isis.xml',
 );

C<tag> defines tag to use within C<xml_file>

C<xml_file> defines path to normalize XML.

=cut

sub open {
        my $self = shift;

        my $arg = {@_};

        my $log = $self->_get_logger();

        foreach my $req (qw/tag xml_file/) {
                $log->logconfess("need argument $req") unless $arg->{$req};
        }

        $self->{'tag'} = $arg->{'tag'};
        my $xml_file = $arg->{'xml_file'};

        $log->info("using $xml_file tag <",$self->{'tag'},">");

        $log->logdie("normalisation xml file '$xml_file' doesn't exist!") if (! -e $xml_file);

        $self->{'import_xml_file'} = $xml_file;

        $self->{'import_xml'} = XMLin($xml_file,
                ForceArray => [ $self->{'tag'}, $self->{'tags'}, 'config', 'format' ],
        );

        $log->debug("import xml is ",sub { Dumper($self->{'import_xml'}) });

        return $self;
}


=head2 _x

Convert string from XML UTF-8 encoding to code page defined in C<xml_file>.

 my $text = $n->_x('utf8 text');

Default application code page is C<ISO-8859-2>. You will probably want to
change that when creating new instance of object based on this one.

=cut

sub _x {
        my $self = shift;
        my $utf8 = shift || return;

        # create UTF-8 convertor for import_xml files
        $self->{'utf2cp'} ||= Text::Iconv->new('UTF-8' ,$self->{'code_page'} || 'ISO-8859-2');

        return $self->{'utf2cp'}->convert($utf8) ||
                $self->_get_logger()->logwarn("can't convert '$utf8'");
}


=head1 AUTHOR

Dobrica Pavlinusic, C<< <dpavlin@rot13.org> >>

=head1 COPYRIGHT & LICENSE

Copyright 2005 Dobrica Pavlinusic, All Rights Reserved.

This program is free software; you can redistribute it and/or modify it
under the same terms as Perl itself.

=cut

1; # End of WebPAC::Normalize::XML
1	package WebPAC::Normalize::XML;
2
3	use warnings;
4	use strict;
5
6	use base qw/WebPAC::Common WebPAC::Normalize/;
7	use XML::Simple;
8	use Data::Dumper;
9	use Text::Iconv;
10
11	=head1 NAME
12
13	WebPAC::Normalize::XML - apply XML normalisaton rules
14
15	=head1 VERSION
16
17	Version 0.01
18
19	=cut
20
21	our $VERSION = '0.01';
22
23	=head1 SYNOPSIS
24
25	This module uses C<conf/normalize/*.xml> files to perform normalisation
26	from input records
27
28	=cut
29
30	=head1 FUNCTIONS
31
32	=head2 open
33
34	Read normalisation rules defined using XML from C<conf/normalize/*.xml> and
35	parse it.
36
37	my $n = new WebPAC::Normalize::XML;
38	$n->open(
39	tag => 'isis',
40	xml_file => '/path/to/conf/normalize/isis.xml',
41	);
42
43	C<tag> defines tag to use within C<xml_file>
44
45	C<xml_file> defines path to normalize XML.
46
47	=cut
48
49	sub open {
50	my $self = shift;
51
52	my $arg = {@_};
53
54	my $log = $self->_get_logger();
55
56	foreach my $req (qw/tag xml_file/) {
57	$log->logconfess("need argument $req") unless $arg->{$req};
58	}
59
60	$self->{'tag'} = $arg->{'tag'};
61	my $xml_file = $arg->{'xml_file'};
62
63	$log->info("using $xml_file tag <",$self->{'tag'},">");
64
65	$log->logdie("normalisation xml file '$xml_file' doesn't exist!") if (! -e $xml_file);
66
67	$self->{'import_xml_file'} = $xml_file;
68
69	$self->{'import_xml'} = XMLin($xml_file,
70	ForceArray => [ $self->{'tag'}, $self->{'tags'}, 'config', 'format' ],
71	);
72
73	$log->debug("import xml is ",sub { Dumper($self->{'import_xml'}) });
74
75	return $self;
76	}
77
78
79	=head2 _x
80
81	Convert string from XML UTF-8 encoding to code page defined in C<xml_file>.
82
83	my $text = $n->_x('utf8 text');
84
85	Default application code page is C<ISO-8859-2>. You will probably want to
86	change that when creating new instance of object based on this one.
87
88	=cut
89
90	sub _x {
91	my $self = shift;
92	my $utf8 = shift \|\| return;
93
94	# create UTF-8 convertor for import_xml files
95	$self->{'utf2cp'} \|\|= Text::Iconv->new('UTF-8' ,$self->{'code_page'} \|\| 'ISO-8859-2');
96
97	return $self->{'utf2cp'}->convert($utf8) \|\|
98	$self->_get_logger()->logwarn("can't convert '$utf8'");
99	}
100
101
102	=head1 AUTHOR
103
104	Dobrica Pavlinusic, C<< <dpavlin@rot13.org> >>
105
106	=head1 COPYRIGHT & LICENSE
107
108	Copyright 2005 Dobrica Pavlinusic, All Rights Reserved.
109
110	This program is free software; you can redistribute it and/or modify it
111	under the same terms as Perl itself.
112
113	=cut
114
115	1; # End of WebPAC::Normalize::XML