/[webpac2]/trunk/lib/WebPAC/Output/MARC.pm
This is repository of my old source code which isn't updated any more. Go to git.rot13.org for current projects!
ViewVC logotype

Annotation of /trunk/lib/WebPAC/Output/MARC.pm

Parent Directory Parent Directory | Revision Log Revision Log


Revision 626 - (hide annotations)
Mon Sep 4 16:15:07 2006 UTC (17 years, 9 months ago) by dpavlin
File size: 4177 byte(s)
fix MARC encoding whoes

1 dpavlin 578 package WebPAC::Output::MARC;
2    
3     use warnings;
4     use strict;
5    
6     use base qw/WebPAC::Common/;
7    
8     use MARC::Record 2.0; # need 2.0 for utf-8 encoding see marcpm.sf.net
9     use MARC::Lint;
10     use Data::Dump qw/dump/;
11 dpavlin 626 use Encode qw/from_to decode/;
12 dpavlin 578
13     =head1 NAME
14    
15     WebPAC::Output::MARC - Create MARC records from C<marc_*> normalisation rules
16    
17     =head1 VERSION
18    
19 dpavlin 626 Version 0.02
20 dpavlin 578
21     =cut
22    
23 dpavlin 626 our $VERSION = '0.02';
24 dpavlin 578
25     =head1 SYNOPSIS
26    
27     Create MARC records from C<marc_*> normalisation rules described in
28     L<WebPAC::Normalize>.
29    
30    
31     =head1 FUNCTIONS
32    
33     =head2 new
34    
35     my $marc = new WebPAC::Output::MARC(
36     path => '/path/to/output.marc',
37 dpavlin 626 native_encoding => 'iso-8859-2',
38     marc_encoding => 'utf-8',
39 dpavlin 578 lint => 1,
40     dump => 0,
41     )
42    
43     =cut
44    
45     sub new {
46     my $class = shift;
47     my $self = {@_};
48     bless($self, $class);
49    
50     my $log = $self->_get_logger;
51    
52     if ($self->{lint}) {
53     $self->{lint}= new MARC::Lint or
54     $log->warn("Can't create MARC::Lint object, linting is disabled");
55     }
56    
57     if (my $path = $self->{path}) {
58     open($self->{fh}, '>', $path) ||
59     $log->logdie("can't open MARC output $path: $!");
60    
61     $log->info("Creating MARC export file $path", $self->{lint} ? ' (with lint)' : '', "\n");
62     } else {
63     $log->logconfess("new called without path");
64     }
65    
66 dpavlin 626 $self->{native_encoding} ||= 'iso-8859-2';
67     $self->{marc_encoding} ||= 'utf-8';
68 dpavlin 578
69     $self ? return $self : return undef;
70     }
71    
72     =head2 add
73    
74     $marc->add(
75     id => $mfn,
76     fields => WebPAC::Normalize::_get_marc_fields(),
77     leader => WebPAC::Normalize::marc_leader(),
78 dpavlin 582 row => $row,
79 dpavlin 578 );
80    
81 dpavlin 582 C<row> is optional parametar which is used when dumping original row to
82     error log.
83    
84 dpavlin 578 =cut
85    
86     sub add {
87     my $self = shift;
88    
89     my $arg = {@_};
90    
91     my $log = $self->_get_logger;
92    
93     $log->logconfess("add needs fields and id arguments")
94     unless ($arg->{fields} && defined $arg->{id});
95    
96     my $marc = new MARC::Record;
97 dpavlin 626 $marc->encoding( $self->{marc_encoding} );
98 dpavlin 578
99 dpavlin 590 my $id = $arg->{id};
100 dpavlin 578
101     $log->logconfess("fields isn't array") unless (ref($arg->{fields}) eq 'ARRAY');
102    
103 dpavlin 626 my $fields = $arg->{fields};
104 dpavlin 578
105 dpavlin 626 $log->debug("original fields = ", sub { dump( $fields ) });
106    
107     # recode fields to marc_encoding
108     foreach my $j ( 0 .. $#$fields ) {
109     foreach my $i ( 0 .. ( ( $#{$fields->[$j]} - 3 ) / 2 ) ) {
110     my $f = $fields->[$j]->[ ($i * 2) + 4 ];
111     $f = decode( $self->{native_encoding}, $f );
112     $fields->[$j]->[ ($i * 2) + 4 ] = $f;
113     }
114     }
115    
116     $log->debug("recode fields = ", sub { dump( $fields ) });
117    
118     $marc->add_fields( @$fields );
119    
120 dpavlin 578 # tweak leader
121     if (my $new_l = $arg->{leader}) {
122    
123     my $leader = $marc->leader;
124    
125     foreach my $o ( keys %$new_l ) {
126     my $insert = $new_l->{$o};
127     $leader = substr($leader, 0, $o) .
128     $insert . substr($leader, $o+length($insert));
129     }
130     $marc->leader( $leader );
131     }
132    
133     if ($self->{lint}) {
134     $self->{lint}->check_record( $marc );
135 dpavlin 582 my @w = $self->{lint}->warnings;
136     if (@w) {
137     $log->error("MARC lint detected warning on record $id\n",
138 dpavlin 621 "<<<<< Original input row:\n",dump($arg->{row}), "\n",
139 dpavlin 582 ">>>>> Normalized MARC row: leader: [", $marc->leader(), "]\n", dump( $arg->{fields} ), "\n",
140     "!!!!! MARC lint warnings:\n",join("\n",@w),"\n"
141     );
142     map { $self->{_marc_lint_warnings}->{$_}++ } @w;
143     }
144 dpavlin 578 }
145    
146     if ($self->{dump}) {
147     $log->info("MARC record on record $id\n",
148 dpavlin 587 "<<<<< Original imput row:\n",dump($arg->{row}), "\n",
149 dpavlin 578 ">>>>> Normalized MARC row: leader: [", $marc->leader(), "]\n", dump( $arg->{fields} ), "\n",
150     );
151     }
152    
153 dpavlin 626 {
154     use bytes;
155     print {$self->{fh}} $marc->as_usmarc;
156     }
157 dpavlin 578
158     }
159    
160     =head2 finish
161    
162     Close MARC output file
163    
164     $marc->finish;
165    
166 dpavlin 582 It will also dump MARC lint warnings summary if called with C<lint>.
167    
168 dpavlin 578 =cut
169    
170     sub finish {
171     my $self = shift;
172    
173 dpavlin 582 my $log = $self->get_logger;
174    
175     close( $self->{fh} ) or $log->logdie("can't close ", $self->{path}, ": $!");
176    
177     if (my $w = $self->{_marc_lint_warnings}) {
178     $log->error("MARC lint warnings summary:\n",
179     join ("\n",
180     map { $w->{$_} . "\t" . $_ }
181     sort { $w->{$b} <=> $w->{$a} } keys %$w
182     )
183     );
184     }
185 dpavlin 578 }
186    
187     =head1 AUTHOR
188    
189     Dobrica Pavlinusic, C<< <dpavlin@rot13.org> >>
190    
191     =head1 COPYRIGHT & LICENSE
192    
193     Copyright 2006 Dobrica Pavlinusic, All Rights Reserved.
194    
195     This program is free software; you can redistribute it and/or modify it
196     under the same terms as Perl itself.
197    
198     =cut
199    
200     1; # End of WebPAC::Output::MARC

  ViewVC Help
Powered by ViewVC 1.1.26