source: main/trunk/greenstone2/perllib/plugins/GreenstoneMETSPlugin.pm@ 22597

Last change on this file since 22597 was 21803, checked in by kjdon, 14 years ago

set file_id to null if ID doesn't match FILE.* (previously it was ending up with mets: from the previous match
  • Property svn:keywords set to Author Date Id Revision
File size: 9.3 KB
Line 
1###########################################################################
2#
3# GreenstoneMETSPlugin.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 GreenstoneArchive METS documents. Note that this plugin does no
27# syntax checking (though the XML::Parser module tests for
28# well-formedness). It's assumed that the GreenstoneArchive files conform
29# to their DTD.
30
31
32package GreenstoneMETSPlugin;
33
34use ghtml;
35
36use strict;
37no strict 'refs'; # allow filehandles to be variables and viceversa
38
39use ReadXMLFile;
40use XML::XPath;
41use XML::XPath::XMLParser;
42
43sub BEGIN {
44 @GreenstoneMETSPlugin::ISA = ('ReadXMLFile');
45}
46
47my $arguments = [
48 ];
49my $options = { 'name' => "GreenstoneMETSPlugin",
50 'desc' => "{GreenstoneMETSPlugin.desc}",
51 'abstract' => "no",
52 'inherits' => "yes" };
53
54
55sub get_default_process_exp {
56 my $self = shift (@_);
57
58 return q^(?i)docmets\.xml$^;
59}
60
61sub new {
62 my ($class) = shift (@_);
63 my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
64 push(@$pluginlist, $class);
65
66 # have no args - do we still want this?
67 #push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
68 push(@{$hashArgOptLists->{"OptList"}},$options);
69
70 my $self = new ReadXMLFile($pluginlist, $inputargs, $hashArgOptLists);
71
72 $self->{'section'} = "";
73 $self->{'section_level'} = 0;
74 $self->{'metadata_name'} = "";
75 $self->{'metadata_value'} = "";
76 $self->{'content'} = "";
77
78 return bless $self, $class;
79}
80
81sub xml_start_document {
82 my $self = shift (@_);
83 my ($expat, $element) = @_;
84
85 $self->{'section'} = "";
86 $self->{'section_level'} = 0;
87 $self->{'metadata_name'} = "";
88 $self->{'metadata_value'} = "";
89 $self->{'content'} = "";
90
91 #**defined a dmdSection Table
92 $self->{'dmdSec_table'}={};
93
94 #**defined a fileSection Table
95 $self->{'fileSec_table'}={};
96
97 #***open doctxt.xml and read the data in
98 my $filename = $self->{'filename'};
99
100 $filename =~ s/docmets.xml$/doctxt.xml/;
101
102 if (!open (FILEIN,"<$filename")){
103 print STDERR "Warning: unable to open the $filename\n";
104 $self->{'xmltxt'} = "";
105 }
106 else {
107 my $xml_text = "";
108 while (defined (my $line = <FILEIN>)) {
109 if ($line !~ m/^<!DOCTYPE.*>/) {
110 $xml_text .= $line;
111 }
112 }
113 my $xml_parser = XML::XPath->new (xml=> $xml_text);
114 #my $xml_tree = $xml_parser->parse ($xml_text);
115
116 #eval {$self->{'parser_text'}->parse};
117 $self->{'parsed_xml'} = $xml_parser;
118 }
119 my $outhandle = $self->{'outhandle'};
120 print $outhandle "GreenstoneMETSPlugin: processing $self->{'file'}\n" if $self->{'verbosity'} > 1;
121 print STDERR "<Processing n='$self->{'file'}' p='GreenstoneMETSPlugin'>\n" if ($self->{'gli'});
122
123}
124
125sub xml_end_document {
126}
127
128sub xml_doctype {
129}
130
131sub xml_start_tag {
132 my $self = shift(@_);
133 my ($expat, $element) = @_;
134 $self->{'element'} = $element;
135 #**deal with dmdSection
136 if ($element =~ /^(mets:)?dmdSec$/ || $element =~ /(gsdl3:)?Metadata$/){
137 $self->xml_dmd_start_tag (@_);
138 } elsif ($element =~ /^(mets:)?file$/) {
139 # only store the file_id for sections with text. Not for default ids (assoc files)
140 if ($_{'ID'} =~ m/FILE(.*)/) {
141 $self->{'file_Id'} = $1;
142 }
143 else {
144 undef $self->{'file_Id'};
145 }
146 } elsif ($element =~ /^(mets:)?FLocat$/){
147 #***deal with fileSection
148 $self->xml_fileloc_start_tag (@_);
149 } elsif ($element =~ /^(mets:)?div$/){
150 #***deal with StrucMap Section
151 $self->xml_strucMap_start_tag (@_);
152 }
153}
154
155sub xml_dmd_start_tag {
156 my $self = shift (@_);
157 my ($expat, $element) = @_;
158
159 if ($element =~ /^(mets:)?dmdSec$/){
160 my ($section_num) = ($_{'ID'} =~ m/DM(.*)/);
161 $self->{'dmdSec_table'}->{"$section_num"}=[];
162 $self->{'dmdSec_table'}->{'section_num'}=$section_num;
163 } elsif ($element =~ /^(gsdl3:)?Metadata$/) {
164 $self->{'metadata_name'} = $_{'name'};
165 }
166}
167
168sub xml_fileloc_start_tag {
169 my $self = shift (@_);
170 my ($expat, $element) = @_;
171
172 my $xlink = $_{'xlink:href'};
173 if (!defined $xlink) {
174 # try without namespace
175 $xlink = $_{'href'};
176 }
177 #my ($section_num) = ($_{'ID'} =~ m/^FLOCAT(.*)$/);
178 my $section_num = $self->{'file_Id'};
179 return if (!defined $section_num);
180
181 $self->{'fileSec_table'}->{"$section_num"}=[];
182 $self->{'fileSec_table'}->{'section_num'}=$section_num;
183
184 my ($filename,$xpath_expr)=($xlink =~ m/^file:(.*)\#xpointer\((.*)\)$/);
185 my $nodeset = $self->{'parsed_xml'}->findnodes ($xpath_expr);
186 my $node_size= $nodeset->size;
187
188 if ($node_size==0) {
189 print STDERR "Warning: no text associated with XPATH $xpath_expr\n";
190 }
191 else {
192 foreach my $node ($nodeset->get_nodelist) {
193 my $xml_content = XML::XPath::XMLParser::as_string($node);
194 my $unescaped_xml_content = &ghtml::unescape_html($xml_content);
195 my $section_content={'section_content'=> $unescaped_xml_content};
196
197 my $content_list = $self->{'fileSec_table'}->{"$section_num"};
198 push (@$content_list, $section_content);
199 }
200 }
201}
202
203sub xml_strucMap_start_tag {
204 my $self = shift (@_);
205 my ($expat, $element) = @_;
206
207
208 my ($section_num) = ($_{'ID'} =~ m/DS(.*)/);
209
210 if ($_{'ID'} ne "DSAll"){
211 if ($self->{'section_level'}==0) {
212 $self->open_document();
213 } else {
214 my $doc_obj = $self->{'doc_obj'};
215 $self->{'section'}=
216 $doc_obj->insert_section($doc_obj->get_end_child($self->{'section'}));
217 }
218 $self->{'section_level'}++;
219
220 #***Add metadata from dmdSection
221 my $md_list = $self->{'dmdSec_table'}->{"$section_num"};
222
223 foreach my $md_pair (@$md_list){
224 my $metadata_name = $md_pair->{'metadata_name'};
225 my $metadata_value = $md_pair->{'metadata_value'};
226 $self->{'doc_obj'}->add_utf8_metadata($self->{'section'}, $metadata_name, $metadata_value);
227 }
228
229 #*** Add content from fileSection
230 my $content_list = $self->{'fileSec_table'}->{"$section_num"};
231
232 foreach my $section_content (@$content_list){
233 my $content = $section_content->{'section_content'};
234 $self->{'doc_obj'}->add_utf8_text($self->{'section'},$content);
235 }
236 }
237}
238
239sub get_doctype {
240 my $self = shift(@_);
241
242 return "mets:mets";
243}
244
245sub xml_end_tag {
246 my $self = shift(@_);
247 my ($expat, $element) = @_;
248
249 if ($element =~ /^(gsdl3:)?Metadata$/) {
250 my $section_num = $self->{'dmdSec_table'}->{'section_num'};
251 my $metadata_name=$self->{'metadata_name'};
252 my $metadata_value=$self->{'metadata_value'};
253
254 my $md_pair={'metadata_name' => $metadata_name,
255 'metadata_value'=> $metadata_value};
256
257 my $md_list = $self->{'dmdSec_table'}->{"$section_num"};
258
259 push(@$md_list,$md_pair);
260
261 $self->{'metadata_name'} = "";
262 $self->{'metadata_value'} = "";
263 } elsif ($element =~ /^(mets:)?file$/){
264 $self->{'file_id'} = "";
265 }
266
267
268 #*** StrucMap Section
269 if ($element =~ /^(mets:)?div$/) {
270 $self->{'section_level'}--;
271 $self->{'section'} = $self->{'doc_obj'}->get_parent_section($self->{'section'});
272 $self->close_document() if $self->{'section_level'}==0;
273 }
274 $self->{'element'} = "";
275}
276
277sub xml_text {
278 my $self = shift(@_);
279 my ($expat) = @_;
280
281 if ($self->{'element'} =~ /^(gsdl3:)?Metadata$/) {
282 $self->{'metadata_value'} .= $_;
283 }
284}
285
286sub open_document {
287 my $self = shift(@_);
288
289 # create a new document
290 $self->{'doc_obj'} = new doc ();
291 $self->{'section'} = "";
292}
293
294sub close_document {
295 my $self = shift(@_);
296
297 # add the associated files
298 my $assoc_files =
299 $self->{'doc_obj'}->get_metadata($self->{'doc_obj'}->get_top_section(), "gsdlassocfile");
300
301 # for when "assocfilepath" isn't the same directory that doc.xml is in...
302 my $assoc_filepath_list= $self->{'doc_obj'}->get_metadata($self->{'doc_obj'}->get_top_section(), "assocfilepath");
303
304 my $assoc_filepath=shift (@$assoc_filepath_list);
305 if (defined ($assoc_filepath)) {
306 # make absolute rather than relative...
307 $self->{'filename'} =~ m@^(.*[\\/]archives)@;
308 $assoc_filepath = "$1/$assoc_filepath/";
309 } else {
310 $assoc_filepath = $self->{'filename'};
311 $assoc_filepath =~ s/[^\\\/]*$//;
312 }
313
314 foreach my $assoc_file_info (@$assoc_files) {
315 my ($assoc_file, $mime_type, $dir) = split (":", $assoc_file_info);
316 my $real_dir = &util::filename_cat($assoc_filepath, $assoc_file),
317 my $assoc_dir = (defined $dir && $dir ne "")
318 ? &util::filename_cat($dir, $assoc_file) : $assoc_file;
319 $self->{'doc_obj'}->associate_file($real_dir, $assoc_dir, $mime_type);
320 }
321 $self->{'doc_obj'}->delete_metadata($self->{'doc_obj'}->get_top_section(), "gsdlassocfile");
322
323 # process the document
324 $self->{'processor'}->process($self->{'doc_obj'}, $self->{'file'});
325}
326
327
3281;
329
Note: See TracBrowser for help on using the repository browser.