1 | ###########################################################################
|
---|
2 | #
|
---|
3 | # MARCXMLPlug.pm
|
---|
4 | # A component of the Greenstone digital library software
|
---|
5 | # from the New Zealand Digital Library Project at the
|
---|
6 | # University of Waikato, New Zealand.
|
---|
7 | #
|
---|
8 | # Copyright (C) 2001 New Zealand Digital Library Project
|
---|
9 | #
|
---|
10 | # This program is free software; you can redistribute it and/or modify
|
---|
11 | # it under the terms of the GNU General Public License as published by
|
---|
12 | # the Free Software Foundation; either version 2 of the License, or
|
---|
13 | # (at your option) any later version.
|
---|
14 | #
|
---|
15 | # This program is distributed in the hope that it will be useful,
|
---|
16 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
17 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
18 | # GNU General Public License for more details.
|
---|
19 | #
|
---|
20 | # You should have received a copy of the GNU General Public License
|
---|
21 | # along with this program; if not, write to the Free Software
|
---|
22 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
23 | #
|
---|
24 | ###########################################################################
|
---|
25 |
|
---|
26 | # Processes MARCXML documents. Note that this plugin does no
|
---|
27 | # syntax checking (though the XML::Parser module tests for
|
---|
28 | # well-formedness).
|
---|
29 |
|
---|
30 | package MARCXMLPlug;
|
---|
31 |
|
---|
32 | use XMLPlug;
|
---|
33 |
|
---|
34 | use strict;
|
---|
35 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
---|
36 |
|
---|
37 | sub BEGIN {
|
---|
38 | @MARCXMLPlug::ISA = ('XMLPlug');
|
---|
39 | }
|
---|
40 |
|
---|
41 | my $arguments = [{'name' => "metadata_mapping_file",
|
---|
42 | 'desc' => "{MARCXMLPlug.metadata_mapping_file}",
|
---|
43 | 'type' => "string",
|
---|
44 | 'reqd' => "no" }];
|
---|
45 |
|
---|
46 | my $options = { 'name' => "MARCXMLPlug",
|
---|
47 | 'desc' => "{MARCXMLPlug.desc}",
|
---|
48 | 'abstract' => "no",
|
---|
49 | 'inherits' => "yes",
|
---|
50 | 'args' => $arguments
|
---|
51 | };
|
---|
52 |
|
---|
53 | sub new {
|
---|
54 | my ($class) = shift (@_);
|
---|
55 | my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
|
---|
56 | push(@$pluginlist, $class);
|
---|
57 |
|
---|
58 | if(defined $arguments){ push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});}
|
---|
59 | if(defined $options) { push(@{$hashArgOptLists->{"OptList"}},$options)};
|
---|
60 |
|
---|
61 | my $self = new XMLPlug($pluginlist, $inputargs, $hashArgOptLists);
|
---|
62 |
|
---|
63 | $self->{'content'} = "";
|
---|
64 | $self->{'record_count'} = 1;
|
---|
65 | $self->{'language'} = "";
|
---|
66 | $self->{'encoding'} = "";
|
---|
67 | $self->{'marc_mapping'} = {};
|
---|
68 | $self->{'current_code'} = "";
|
---|
69 | $self->{'current_tag'} = "";
|
---|
70 | $self->{'current_element'} = "";
|
---|
71 | $self->{'metadata_mapping'} = undef;
|
---|
72 | $self->{'num_processed'} = 0;
|
---|
73 | $self->{'indent'} = 0;
|
---|
74 |
|
---|
75 | return bless $self, $class;
|
---|
76 | }
|
---|
77 |
|
---|
78 | sub get_doctype {
|
---|
79 | my $self = shift(@_);
|
---|
80 |
|
---|
81 | return "collection";
|
---|
82 | }
|
---|
83 |
|
---|
84 |
|
---|
85 | sub init {
|
---|
86 | my $self = shift (@_);
|
---|
87 | my ($verbosity, $outhandle, $failhandle) = @_;
|
---|
88 |
|
---|
89 | ## the mapping file has already been loaded
|
---|
90 | if (defined $self->{'metadata_mapping'} ){
|
---|
91 | $self->SUPER::init(@_);
|
---|
92 | return;
|
---|
93 | }
|
---|
94 |
|
---|
95 | my $metadata_mapping = {};
|
---|
96 |
|
---|
97 | # read in the metadata mapping file
|
---|
98 | my $mm_file = $self->{'metadata_mapping_file'};
|
---|
99 |
|
---|
100 | if (! defined $mm_file or $mm_file eq ""){
|
---|
101 | $mm_file = &util::filename_cat( $ENV{'GSDLHOME'}, "etc","marctodc.txt" );
|
---|
102 | $self->{'metadata_mapping_file'} = $mm_file;
|
---|
103 | }
|
---|
104 |
|
---|
105 | if (!-e $mm_file)
|
---|
106 | {
|
---|
107 | my $msg = "MARCXMLPlug ERROR: Can't locate mapping file \"" .
|
---|
108 | $self->{'metadata_mapping'} . "\".\n This file should be at $mm_file\n" .
|
---|
109 | " No marc files can be processed.\n";
|
---|
110 |
|
---|
111 | print $outhandle $msg;
|
---|
112 | print $failhandle $msg;
|
---|
113 | $self->{'metadata_mapping'} = undef;
|
---|
114 | # We pick up the error in process() if there is no $mm_file
|
---|
115 | # If we exit here, then pluginfo.pl will exit too!
|
---|
116 | }
|
---|
117 | elsif (open(MMIN, "<$mm_file"))
|
---|
118 | {
|
---|
119 | my $l=1;
|
---|
120 | my $line;
|
---|
121 | while (defined($line=<MMIN>))
|
---|
122 | {
|
---|
123 | chomp $line;
|
---|
124 | if ($line =~ m/^(\d+\w?)\s*->\s*([\w\^]+)$/)
|
---|
125 | {
|
---|
126 | my $marc_info = $1;
|
---|
127 | my $gsdl_info = $2;
|
---|
128 | $metadata_mapping->{$marc_info} = $gsdl_info;
|
---|
129 | }
|
---|
130 | elsif ($line !~ m/^\#/ # allow comments (# in first column)
|
---|
131 | && $line !~ m/^\s*$/) # allow blank lines
|
---|
132 | {
|
---|
133 | print $outhandle "Parse error on line $l of $mm_file:\n";
|
---|
134 | print $outhandle " \"$line\"\n";
|
---|
135 | }
|
---|
136 | $l++
|
---|
137 | }
|
---|
138 | close(MMIN);
|
---|
139 | }
|
---|
140 | else
|
---|
141 | {
|
---|
142 | print STDERR "Unable to open $mm_file: $!\n";
|
---|
143 | }
|
---|
144 |
|
---|
145 | $self->{'metadata_mapping'} = $metadata_mapping;
|
---|
146 |
|
---|
147 | ##map { print STDERR $_."=>".$metadata_mapping->{$_}."\n"; } keys %$metadata_mapping;
|
---|
148 |
|
---|
149 | $self->SUPER::init(@_);
|
---|
150 | }
|
---|
151 |
|
---|
152 | # Called for DOCTYPE declarations - use die to bail out if this doctype
|
---|
153 | # is not meant for this plugin
|
---|
154 | sub xml_doctype {
|
---|
155 | my $self = shift(@_);
|
---|
156 |
|
---|
157 | my ($expat, $name, $sysid, $pubid, $internal) = @_;
|
---|
158 | return;
|
---|
159 |
|
---|
160 | }
|
---|
161 |
|
---|
162 |
|
---|
163 | sub xml_start_document {
|
---|
164 | my $self = shift(@_);
|
---|
165 |
|
---|
166 | my ($expat, $name, $sysid, $pubid, $internal) = @_;
|
---|
167 |
|
---|
168 |
|
---|
169 | my $file = $self->{'file'};
|
---|
170 | my $filename = $self->{'filename'};
|
---|
171 |
|
---|
172 | my ($language, $encoding) = $self->textcat_get_language_encoding ($filename);
|
---|
173 |
|
---|
174 | $self->{'language'} = $language;
|
---|
175 | $self->{'encoding'} = $encoding;
|
---|
176 | $self->{'element_count'} = 1;
|
---|
177 | $self->{'indent'} = 0;
|
---|
178 | my $outhandle = $self->{'outhandle'};
|
---|
179 | print $outhandle "MARCXMLPlug: processing $self->{'file'}\n" if $self->{'verbosity'} > 1;
|
---|
180 | print STDERR "<Processing n='$self->{'file'}' p='MARCXMLPlug'>\n" if $self->{'gli'};
|
---|
181 |
|
---|
182 | }
|
---|
183 |
|
---|
184 | sub xml_end_document {
|
---|
185 |
|
---|
186 | }
|
---|
187 |
|
---|
188 | sub xml_start_tag {
|
---|
189 | my $self = shift;
|
---|
190 | my $expat = shift;
|
---|
191 | my $element = shift;
|
---|
192 |
|
---|
193 | my $text = $self->escape_text($_);
|
---|
194 |
|
---|
195 | $self->{'current_element'} = $element;
|
---|
196 |
|
---|
197 | ##get all atributes of this element and store it in a map name=>value
|
---|
198 | my %attr_map = ();
|
---|
199 | my $attrstring = $_;
|
---|
200 | while ($attrstring =~ /(\w+)=\"(\w+)\"/){
|
---|
201 | $attr_map{$1}=$2;
|
---|
202 | $attrstring = $'; #'
|
---|
203 | }
|
---|
204 |
|
---|
205 |
|
---|
206 | my $processor = $self->{'processor'};
|
---|
207 |
|
---|
208 | ##create a new document for each record
|
---|
209 | if ($element eq "record") {
|
---|
210 | my $filename = $self->{'filename'};
|
---|
211 | my $language = $self->{'language'};
|
---|
212 | my $encoding = $self->{'encoding'};
|
---|
213 | my $file = $self->{'file'};
|
---|
214 | my $doc_obj = new doc($filename);
|
---|
215 | $doc_obj->set_OIDtype ($processor->{'OIDtype'}, $processor->{'OIDmetadata'});
|
---|
216 | $doc_obj->add_utf8_metadata($doc_obj->get_top_section(), "Language", $language);
|
---|
217 | $doc_obj->add_utf8_metadata($doc_obj->get_top_section(), "Encoding", $encoding);
|
---|
218 | my ($filemeta) = $file =~ /([^\\\/]+)$/;
|
---|
219 | $doc_obj->add_utf8_metadata($doc_obj->get_top_section(), "Source", &ghtml::dmsafe($filemeta));
|
---|
220 | $doc_obj->add_utf8_metadata($doc_obj->get_top_section(), "SourceSegment", "$self->{'record_count'}");
|
---|
221 | if ($self->{'cover_image'}) {
|
---|
222 | $self->associate_cover_image($doc_obj, $filename);
|
---|
223 | }
|
---|
224 | $doc_obj->add_utf8_metadata($doc_obj->get_top_section(), "Plugin", "$self->{'plugin_type'}");
|
---|
225 | $doc_obj->add_metadata($doc_obj->get_top_section(), "FileFormat", "XML");
|
---|
226 |
|
---|
227 | $doc_obj->set_OID();
|
---|
228 | $self->set_OID($doc_obj, $doc_obj->get_OID() , $self->{'record_count'});
|
---|
229 |
|
---|
230 | my $outhandle = $self->{'outhandle'};
|
---|
231 | print $outhandle "Record $self->{'record_count'} - MARCXMLPlug: processing $self->{'file'}\n" if $self->{'verbosity'} > 1;
|
---|
232 |
|
---|
233 | $self->{'record_count'}++;
|
---|
234 | $self->{'doc_obj'} = $doc_obj;
|
---|
235 | $self->{'num_processed'}++;
|
---|
236 |
|
---|
237 | }
|
---|
238 |
|
---|
239 | ## get the marc code, for example 520
|
---|
240 | if ($element eq "datafield") {
|
---|
241 | if (defined $attr_map{'tag'} and $attr_map{'tag'} ne ""){
|
---|
242 | $self->{'current_tag'} = $attr_map{tag};
|
---|
243 | }
|
---|
244 | }
|
---|
245 |
|
---|
246 |
|
---|
247 | ## append the subcode to the marc code for example 520a or 520b
|
---|
248 | if ($element eq "subfield"){
|
---|
249 | if (defined $attr_map{'code'} and $attr_map{'code'} ne "" and $self->{'current_tag'} ne ""){
|
---|
250 | $self->{'current_code'} = $attr_map{'code'};
|
---|
251 | }
|
---|
252 | }
|
---|
253 |
|
---|
254 | if ($element eq "record"){
|
---|
255 | $self->{'indent'} = 0;
|
---|
256 | }
|
---|
257 | else {
|
---|
258 | if ($element ne "subfield"){
|
---|
259 | $self->{'indent'} = 1;
|
---|
260 | }
|
---|
261 | else{
|
---|
262 | $self->{'indent'} = 2;
|
---|
263 | }
|
---|
264 | }
|
---|
265 |
|
---|
266 |
|
---|
267 | if ($element ne "collection"){
|
---|
268 | $self->{'content'} .= "<br/>".$self->calculate_indent($self->{'indent'}).$text;
|
---|
269 | }
|
---|
270 | }
|
---|
271 |
|
---|
272 |
|
---|
273 |
|
---|
274 | sub xml_end_tag {
|
---|
275 | my $self = shift(@_);
|
---|
276 | my ($expat, $element) = @_;
|
---|
277 | my $text = $self->escape_text($_);
|
---|
278 |
|
---|
279 | if ($element eq "record" and defined $self->{'doc_obj'}) {
|
---|
280 | # process the document
|
---|
281 | my $processor = $self->{'processor'};
|
---|
282 | my $doc_obj = $self->{'doc_obj'};
|
---|
283 | $self->{'content'} .= "<br/>".$text;
|
---|
284 |
|
---|
285 | $doc_obj->add_utf8_text($doc_obj->get_top_section(),$self->{'content'});
|
---|
286 | $processor->process($doc_obj);
|
---|
287 |
|
---|
288 | ##clean up
|
---|
289 | $self->{'content'} = "";
|
---|
290 | $self->{'doc_obj'} = undef;
|
---|
291 | return;
|
---|
292 | }
|
---|
293 |
|
---|
294 | ## map the xmlmarc to gsdl metadata
|
---|
295 | if ($element eq "datafield" and defined $self->{'doc_obj'} and defined $self->{'marc_mapping'}){
|
---|
296 | my $metadata_mapping = $self->{'metadata_mapping'};
|
---|
297 | my $marc_mapping = $self->{'marc_mapping'};
|
---|
298 | my $doc_obj = $self->{'doc_obj'};
|
---|
299 |
|
---|
300 | ##map { print STDERR $_."=>".$marc_mapping->{$_}."\n"; } keys %$marc_mapping;
|
---|
301 | ##map { print STDERR $_."=>".$metadata_mapping->{$_}."\n"; } keys %$metadata_mapping;
|
---|
302 |
|
---|
303 | foreach my $marc_field (keys %$metadata_mapping){
|
---|
304 | my ($meta_name,$meta_value);
|
---|
305 | my $matched_field = $marc_mapping->{$marc_field};
|
---|
306 | my $subfield = undef;
|
---|
307 | if (defined $matched_field){
|
---|
308 | ## test whether this field has subfield
|
---|
309 | if ($marc_field =~ /\d\d\d(\w)/){
|
---|
310 | $subfield = $1;
|
---|
311 | }
|
---|
312 | $meta_name = $metadata_mapping->{$marc_field};
|
---|
313 |
|
---|
314 | if (defined $subfield){
|
---|
315 | my %mapped_subfield = {@$matched_field};
|
---|
316 | $meta_value = $mapped_subfield{$subfield};
|
---|
317 | }
|
---|
318 | else{ ## get all values
|
---|
319 | my $i =0;
|
---|
320 | foreach my $value (@$matched_field){
|
---|
321 | if ($i%2 != 0){
|
---|
322 | $meta_value .= $value." ";
|
---|
323 | }
|
---|
324 | $i++;
|
---|
325 | }
|
---|
326 | }
|
---|
327 |
|
---|
328 | ## escape [ and ]
|
---|
329 | $meta_value =~ s/\[/\\\[/g;
|
---|
330 | $meta_value =~ s/\]/\\\]/g;
|
---|
331 | ##print STDERR "$meta_name=$meta_value\n";
|
---|
332 | $doc_obj->add_utf8_metadata($doc_obj->get_top_section(),$meta_name, $meta_value);
|
---|
333 |
|
---|
334 | }
|
---|
335 |
|
---|
336 | }
|
---|
337 |
|
---|
338 | ##clean up
|
---|
339 | $self->{'marc_mapping'} = undef;
|
---|
340 | $self->{'current_tag'} = "";
|
---|
341 | }
|
---|
342 |
|
---|
343 | if ($element eq "datafield"){
|
---|
344 | $self->{'indent'} = 1;
|
---|
345 | $self->{'content'} .= "<br/>".$self->calculate_indent($self->{'indent'}).$text;
|
---|
346 | }
|
---|
347 | else{
|
---|
348 | $self->{'content'} .= $text;
|
---|
349 | }
|
---|
350 |
|
---|
351 | }
|
---|
352 |
|
---|
353 |
|
---|
354 | sub set_OID {
|
---|
355 | my $self = shift (@_);
|
---|
356 | my ($doc_obj, $id, $record_number) = @_;
|
---|
357 |
|
---|
358 | $doc_obj->set_OID($id . "r" . $record_number);
|
---|
359 | }
|
---|
360 |
|
---|
361 | sub xml_text {
|
---|
362 | my $self = shift(@_);
|
---|
363 | my ($expat) = @_;
|
---|
364 |
|
---|
365 |
|
---|
366 | ## store the text of a marc code, for exapmle 520a=>A poem about....
|
---|
367 | if ($self->{'current_element'} eq "subfield" and $self->{'current_code'} ne "" and $_ ne "" ){
|
---|
368 | ##stored it in the marc_mapping
|
---|
369 | push(@{$self->{'marc_mapping'}->{$self->{'current_tag'}}},$self->{'current_code'});
|
---|
370 | push(@{$self->{'marc_mapping'}->{$self->{'current_tag'}}},$_);
|
---|
371 | $self->{'current_code'} = "";
|
---|
372 | }
|
---|
373 |
|
---|
374 | $self->{'content'} .=$self->escape_text($_);
|
---|
375 |
|
---|
376 | }
|
---|
377 |
|
---|
378 | sub calculate_indent{
|
---|
379 | my ($self,$num) = @_;
|
---|
380 |
|
---|
381 | my $indent ="";
|
---|
382 |
|
---|
383 | for (my $i=0; $i<$num;$i++){
|
---|
384 | $indent .= " ";
|
---|
385 | }
|
---|
386 |
|
---|
387 | return $indent;
|
---|
388 |
|
---|
389 | }
|
---|
390 |
|
---|
391 | sub escape_text {
|
---|
392 | my ($self,$text) = @_;
|
---|
393 | # special characters in the xml encoding
|
---|
394 | $text =~ s/&/&/g; # this has to be first...
|
---|
395 | $text =~ s/</</g;
|
---|
396 | $text =~ s/>/>/g;
|
---|
397 | $text =~ s/\"/"/g;
|
---|
398 |
|
---|
399 | return $text;
|
---|
400 | }
|
---|
401 |
|
---|
402 |
|
---|
403 | 1;
|
---|
404 |
|
---|
405 |
|
---|