1 | ###########################################################################
|
---|
2 | #
|
---|
3 | # OAIPlug.pm -- basic Open Archives Initiative (OAI) plugin
|
---|
4 | #
|
---|
5 | # A component of the Greenstone digital library software
|
---|
6 | # from the New Zealand Digital Library Project at the
|
---|
7 | # University of Waikato, New Zealand.
|
---|
8 | #
|
---|
9 | # Copyright (C) 1999 New Zealand Digital Library Project
|
---|
10 | #
|
---|
11 | # This program is free software; you can redistribute it and/or modify
|
---|
12 | # it under the terms of the GNU General Public License as published by
|
---|
13 | # the Free Software Foundation; either version 2 of the License, or
|
---|
14 | # (at your option) any later version.
|
---|
15 | #
|
---|
16 | # This program is distributed in the hope that it will be useful,
|
---|
17 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
18 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
19 | # GNU General Public License for more details.
|
---|
20 | #
|
---|
21 | # You should have received a copy of the GNU General Public License
|
---|
22 | # along with this program; if not, write to the Free Software
|
---|
23 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
24 | #
|
---|
25 | ###########################################################################
|
---|
26 |
|
---|
27 | package OAIPlugin;
|
---|
28 |
|
---|
29 | use extrametautil;
|
---|
30 | use unicode;
|
---|
31 | use util;
|
---|
32 |
|
---|
33 | use strict;
|
---|
34 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
---|
35 |
|
---|
36 | use ReadXMLFile;
|
---|
37 | use ReadTextFile; # needed for subroutine textcat_get_language_encoding
|
---|
38 | use metadatautil;
|
---|
39 | use MetadataRead;
|
---|
40 | use util;
|
---|
41 |
|
---|
42 | # methods with identical signatures take precedence in the order given in the ISA list.
|
---|
43 | sub BEGIN {
|
---|
44 | @OAIPlugin::ISA = ('MetadataRead', 'ReadXMLFile', 'ReadTextFile');
|
---|
45 | }
|
---|
46 |
|
---|
47 | my $set_list =
|
---|
48 | [ { 'name' => "auto",
|
---|
49 | 'desc' => "{OAIPlugin.metadata_set.auto}" },
|
---|
50 | { 'name' => "dc",
|
---|
51 | 'desc' => "{OAIPlugin.metadata_set.dc}" }
|
---|
52 | ];
|
---|
53 |
|
---|
54 | my $arguments =
|
---|
55 | [ { 'name' => "process_exp",
|
---|
56 | 'desc' => "{BasePlugin.process_exp}",
|
---|
57 | 'type' => "regexp",
|
---|
58 | 'reqd' => "no",
|
---|
59 | 'deft' => &get_default_process_exp() },
|
---|
60 | { 'name' => "metadata_set",
|
---|
61 | 'desc' => "{OAIPlugin.metadata_set}",
|
---|
62 | 'type' => "enumstring",
|
---|
63 | 'reqd' => "no",
|
---|
64 | 'list' => $set_list,
|
---|
65 | 'deft' => "dc" },
|
---|
66 | { 'name' => "document_field",
|
---|
67 | 'desc' => "{OAIPlugin.document_field}",
|
---|
68 | 'type' => "metadata",
|
---|
69 | 'reqd' => "no",
|
---|
70 | 'deft' => "gi.Sourcedoc" }
|
---|
71 | ];
|
---|
72 |
|
---|
73 | my $options = { 'name' => "OAIPlugin",
|
---|
74 | 'desc' => "{OAIPlugin.desc}",
|
---|
75 | 'abstract' => "no",
|
---|
76 | 'inherits' => "yes",
|
---|
77 | 'explodes' => "yes",
|
---|
78 | 'args' => $arguments };
|
---|
79 |
|
---|
80 |
|
---|
81 | sub new {
|
---|
82 | my ($class) = shift (@_);
|
---|
83 | my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
|
---|
84 | push(@$pluginlist, $class);
|
---|
85 |
|
---|
86 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
---|
87 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
---|
88 |
|
---|
89 | new ReadTextFile($pluginlist, $inputargs, $hashArgOptLists,1);
|
---|
90 | my $self = new ReadXMLFile($pluginlist, $inputargs, $hashArgOptLists);
|
---|
91 |
|
---|
92 | if ($self->{'info_only'}) {
|
---|
93 | # don't worry about modifying options
|
---|
94 | return bless $self, $class;
|
---|
95 | }
|
---|
96 | # trim any ex. from document field iff it's the only metadata namespace prefix
|
---|
97 | $self->{'document_field'} =~ s/^ex\.([^.]+)$/$1/;
|
---|
98 | return bless $self, $class;
|
---|
99 | }
|
---|
100 |
|
---|
101 | sub get_default_process_exp {
|
---|
102 | my $self = shift (@_);
|
---|
103 |
|
---|
104 | return q^(?i)(\.oai)$^;
|
---|
105 | }
|
---|
106 |
|
---|
107 | sub get_doctype {
|
---|
108 | my $self = shift(@_);
|
---|
109 |
|
---|
110 | return "OAI-PMH";
|
---|
111 | }
|
---|
112 |
|
---|
113 | sub xml_start_document {
|
---|
114 | my $self = shift (@_);
|
---|
115 | $self->{'in_metadata_node'} = 0;
|
---|
116 | $self->{'rawxml'} = "";
|
---|
117 | $self->{'saved_metadata'} = {};
|
---|
118 | }
|
---|
119 |
|
---|
120 | sub xml_end_document {
|
---|
121 | }
|
---|
122 |
|
---|
123 | sub xml_doctype {
|
---|
124 | my $self = shift(@_);
|
---|
125 |
|
---|
126 | my ($expat, $name, $sysid, $pubid, $internal) = @_;
|
---|
127 |
|
---|
128 | ##die "" if ($name !~ /^OAI-PMH$/);
|
---|
129 |
|
---|
130 | my $outhandle = $self->{'outhandle'};
|
---|
131 | print $outhandle "OAIPlugin: processing $self->{'file'}\n" if $self->{'verbosity'} > 1;
|
---|
132 | print STDERR "<Processing n='$self->{'file'}' p='OAIPlugin'>\n" if $self->{'gli'};
|
---|
133 |
|
---|
134 | }
|
---|
135 |
|
---|
136 |
|
---|
137 | sub xml_start_tag {
|
---|
138 | my $self = shift(@_);
|
---|
139 | my ($expat,$element) = @_;
|
---|
140 |
|
---|
141 | my %attr_hash = %_;
|
---|
142 |
|
---|
143 | my $attr = "";
|
---|
144 | map { $attr .= " $_=$attr_hash{$_}"; } keys %attr_hash;
|
---|
145 |
|
---|
146 | $self->{'rawxml'} .= "<$element$attr>";
|
---|
147 |
|
---|
148 | if ($element eq "metadata") {
|
---|
149 | $self->{'in_metadata_node'} = 1;
|
---|
150 | $self->{'metadata_xml'} = "";
|
---|
151 | }
|
---|
152 |
|
---|
153 | if ($self->{'in_metadata_node'}) {
|
---|
154 | $self->{'metadata_xml'} .= "<$element$attr>";
|
---|
155 | }
|
---|
156 | }
|
---|
157 |
|
---|
158 | sub xml_end_tag {
|
---|
159 | my $self = shift(@_);
|
---|
160 | my ($expat, $element) = @_;
|
---|
161 |
|
---|
162 | $self->{'rawxml'} .= "</$element>";
|
---|
163 |
|
---|
164 | if ($self->{'in_metadata_node'}) {
|
---|
165 | $self->{'metadata_xml'} .= "</$element>";
|
---|
166 | }
|
---|
167 |
|
---|
168 | if ($element eq "metadata") {
|
---|
169 | my $textref = \$self->{'metadata_xml'};
|
---|
170 | #my $metadata = $self->{'metadata'};
|
---|
171 | my $metadata = $self->{'saved_metadata'};
|
---|
172 | $self->extract_oai_metadata($textref,$metadata);
|
---|
173 |
|
---|
174 | $self->{'in_metadata_node'} = 0;
|
---|
175 | }
|
---|
176 |
|
---|
177 |
|
---|
178 | }
|
---|
179 |
|
---|
180 | sub xml_text {
|
---|
181 | my $self = shift(@_);
|
---|
182 | my ($expat) = @_;
|
---|
183 |
|
---|
184 | $self->{'rawxml'} .= $_;
|
---|
185 |
|
---|
186 | if ($self->{'in_metadata_node'}) {
|
---|
187 | $self->{'metadata_xml'} .= $_;
|
---|
188 | }
|
---|
189 | }
|
---|
190 |
|
---|
191 |
|
---|
192 | sub metadata_read {
|
---|
193 | my $self = shift (@_);
|
---|
194 |
|
---|
195 | my ($pluginfo, $base_dir, $file, $block_hash,
|
---|
196 | $extrametakeys, $extrametadata, $extrametafile,
|
---|
197 | $processor, $gli, $aux) = @_;
|
---|
198 |
|
---|
199 | # can we process this file??
|
---|
200 | my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
|
---|
201 | return undef unless $self->can_process_this_file_for_metadata($filename_full_path);
|
---|
202 |
|
---|
203 | print STDERR "\n<Processing n='$file' p='OAIPlugin'>\n" if ($gli);
|
---|
204 | print STDERR "OAIPlugin: processing $file\n" if ($self->{'verbosity'}) > 1;
|
---|
205 |
|
---|
206 | if (!$self->parse_file($filename_full_path, $file, $gli)) {
|
---|
207 | $self->{'saved_metadata'} = undef;
|
---|
208 | return undef;
|
---|
209 | }
|
---|
210 |
|
---|
211 | my $verbosity = $self->{'verbosity'};
|
---|
212 | my $new_metadata = $self->{'saved_metadata'};
|
---|
213 | $self->{'saved_metadata'} = undef;
|
---|
214 |
|
---|
215 | # add the pretty metadata table as metadata
|
---|
216 | my $ppmd_table = $self->{'ppmd_table'};
|
---|
217 | $new_metadata->{'prettymd'} = $ppmd_table;
|
---|
218 | $self->{'ppmd_table'} = undef;
|
---|
219 |
|
---|
220 | my $document_metadata_field = $self->{'document_field'};
|
---|
221 | my $url_array = $new_metadata->{$document_metadata_field};
|
---|
222 | if (!defined $url_array) {
|
---|
223 | # try ex.
|
---|
224 | $url_array = $new_metadata->{"ex.$document_metadata_field"};
|
---|
225 | }
|
---|
226 | my $num_urls = (defined $url_array) ? scalar(@$url_array) : 0;
|
---|
227 | ##print STDERR "$num_urls urls for $file\n";
|
---|
228 | my $srcdoc_exists = 0;
|
---|
229 | my $srcdoc_pos = 0;
|
---|
230 | my $filename_dir = &util::filename_head($filename_full_path);
|
---|
231 |
|
---|
232 | # filenames in extrametadata must be relative to current dir, as
|
---|
233 | # DirectoryPlugin adds path info on itself
|
---|
234 | my ($filename_for_metadata) = $file =~ /([^\\\/]+)$/; # this assumes there will only be one record per oai file - is this always the case??
|
---|
235 | for (my $i=0; $i<$num_urls; $i++) {
|
---|
236 |
|
---|
237 | if ($url_array->[$i] !~ m/^(https?|ftp):/) {
|
---|
238 |
|
---|
239 | my $src_filename = &util::filename_cat($filename_dir, $url_array->[$i]);
|
---|
240 | if (-e $src_filename) {
|
---|
241 | $srcdoc_pos = $i;
|
---|
242 | $srcdoc_exists = 1;
|
---|
243 | # get the slashes the right way, use filename_cat
|
---|
244 | $filename_for_metadata = &util::filename_cat($url_array->[$i]);
|
---|
245 | last;
|
---|
246 | }
|
---|
247 | }
|
---|
248 | }
|
---|
249 |
|
---|
250 | if ($srcdoc_exists) {
|
---|
251 | $self->{'oai-files'}->{$file}->{'srcdoc_exists'} = 1;
|
---|
252 | }
|
---|
253 | else {
|
---|
254 | # save the rawxml for the source document
|
---|
255 | $self->{'oai-files'}->{$file}->{'srcdoc_exists'} = 0;
|
---|
256 | $self->{'oai-files'}->{$file}->{'rawxml'} = $self->{'rawxml'};
|
---|
257 | $self->{'rawxml'} = "";
|
---|
258 | }
|
---|
259 |
|
---|
260 | # return all the metadata we have extracted to the caller.
|
---|
261 | # Directory plug will pass it back in at read time, so we don't need to extract it again.
|
---|
262 |
|
---|
263 | # Extrametadata keys should be regular expressions
|
---|
264 | # Indexing into the extrameta data structures requires the filename's style of slashes to be in URL format
|
---|
265 | # Then need to convert the filename to a regex, no longer to protect windows directory chars \, but for
|
---|
266 | # protecting special characters like brackets in the filepath such as "C:\Program Files (x86)\Greenstone".
|
---|
267 | $filename_for_metadata = &util::filepath_to_url_format($filename_for_metadata);
|
---|
268 | $filename_for_metadata = &util::filename_to_regex($filename_for_metadata);
|
---|
269 |
|
---|
270 | # Check that we haven't already got some metadata
|
---|
271 | if (defined &extrametautil::getmetadata($extrametadata, $filename_for_metadata)) {
|
---|
272 | print STDERR "\n**** OAIPlugin: Need to merge new metadata with existing stored metadata: file = $filename_for_metadata\n" if $verbosity > 3;
|
---|
273 |
|
---|
274 | my $file_metadata_table = &extrametautil::getmetadata($extrametadata, $filename_for_metadata);
|
---|
275 |
|
---|
276 | foreach my $metaname (keys %{$new_metadata}) {
|
---|
277 | # will create new entry if one does not already exist
|
---|
278 | push(@{$file_metadata_table->{$metaname}}, @{$new_metadata->{$metaname}});
|
---|
279 | }
|
---|
280 |
|
---|
281 | } else {
|
---|
282 | &extrametautil::setmetadata($extrametadata, $filename_for_metadata, $new_metadata);
|
---|
283 | &extrametautil::addmetakey($extrametakeys, $filename_for_metadata);
|
---|
284 | }
|
---|
285 |
|
---|
286 | if ($srcdoc_exists) {
|
---|
287 | if (!defined &extrametautil::getmetafile($extrametafile, $filename_for_metadata)) {
|
---|
288 | &extrametautil::setmetafile($extrametafile, $filename_for_metadata, {});
|
---|
289 | }
|
---|
290 | #maps the file to full path
|
---|
291 | &extrametautil::setmetafile_for_named_file($extrametafile, $filename_for_metadata, $file, $filename_full_path);
|
---|
292 |
|
---|
293 | }
|
---|
294 | return 1;
|
---|
295 |
|
---|
296 | }
|
---|
297 |
|
---|
298 |
|
---|
299 | sub read {
|
---|
300 | my $self = shift (@_);
|
---|
301 |
|
---|
302 | my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
|
---|
303 |
|
---|
304 | if (!defined $self->{'oai-files'}->{$file}) {
|
---|
305 | return undef;
|
---|
306 | }
|
---|
307 |
|
---|
308 | my $srcdoc_exists = $self->{'oai-files'}->{$file}->{'srcdoc_exists'};
|
---|
309 | if ($srcdoc_exists) {
|
---|
310 | # do nothing more - all the metadata has been extracted and associated with the srcdoc
|
---|
311 | # no more need to access details of this $file => tidy up as you go
|
---|
312 | delete $self->{'oai-files'}->{$file};
|
---|
313 | return 0; # not processed here, but don't pass on to rest of plugins
|
---|
314 | }
|
---|
315 |
|
---|
316 | my $filename = $file;
|
---|
317 | $filename = &util::filename_cat ($base_dir, $file) if $base_dir =~ /\w/;
|
---|
318 |
|
---|
319 | # Do encoding stuff on metadata
|
---|
320 | my ($language, $encoding) = $self->textcat_get_language_encoding ($filename);
|
---|
321 |
|
---|
322 | # create a new document
|
---|
323 | my $doc_obj = new doc ($filename, "indexed_doc", $self->{'file_rename_method'});
|
---|
324 | my $top_section = $doc_obj->get_top_section;
|
---|
325 | my $plugin_type = $self->{'plugin_type'};
|
---|
326 |
|
---|
327 | my ($filemeta) = $file =~ /([^\\\/]+)$/;
|
---|
328 | my $plugin_filename_encoding = $self->{'filename_encoding'};
|
---|
329 | my $filename_encoding = $self->deduce_filename_encoding($file,$metadata,$plugin_filename_encoding);
|
---|
330 | $self->set_Source_metadata($doc_obj, $filename, $filename_encoding);
|
---|
331 |
|
---|
332 | $doc_obj->add_utf8_metadata($top_section, "Language", $language);
|
---|
333 | $doc_obj->add_utf8_metadata($top_section, "Encoding", $encoding);
|
---|
334 | $doc_obj->add_utf8_metadata($top_section, "Plugin", $plugin_type);
|
---|
335 | $doc_obj->add_metadata($top_section, "FileFormat", "OAI");
|
---|
336 | $doc_obj->add_metadata($top_section, "FileSize", (-s $filename));
|
---|
337 |
|
---|
338 | # include any metadata passed in from previous plugins
|
---|
339 | # note that this metadata is associated with the top level section
|
---|
340 | # this will include all the metadata from the oai file that we extracted
|
---|
341 | # during metadata_read
|
---|
342 | $self->extra_metadata ($doc_obj, $doc_obj->get_top_section(), $metadata);
|
---|
343 |
|
---|
344 | # do plugin specific processing of doc_obj
|
---|
345 | my $text = $self->{'oai-files'}->{$file}->{'rawxml'};
|
---|
346 | delete $self->{'oai-files'}->{$file};
|
---|
347 |
|
---|
348 | unless (defined ($self->process(\$text, $pluginfo, $base_dir, $file, $metadata, $doc_obj))) {
|
---|
349 | print STDERR "<ProcessingError n='$file'>\n" if ($gli);
|
---|
350 | return -1;
|
---|
351 | }
|
---|
352 |
|
---|
353 | # do any automatic metadata extraction
|
---|
354 | $self->auto_extract_metadata ($doc_obj);
|
---|
355 |
|
---|
356 | # add an OID
|
---|
357 | $self->add_OID($doc_obj);
|
---|
358 |
|
---|
359 | # process the document
|
---|
360 | $processor->process($doc_obj);
|
---|
361 |
|
---|
362 | $self->{'num_processed'} ++;
|
---|
363 |
|
---|
364 | return 1; # processed the file
|
---|
365 | }
|
---|
366 |
|
---|
367 |
|
---|
368 | # do plugin specific processing of doc_obj
|
---|
369 | sub process {
|
---|
370 | my $self = shift (@_);
|
---|
371 | my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
|
---|
372 | my $outhandle = $self->{'outhandle'};
|
---|
373 |
|
---|
374 | print STDERR "<Processing n='$file' p='OAIPlugin'>\n" if ($gli);
|
---|
375 | print $outhandle "OAIPlugin: processing $file\n"
|
---|
376 | if $self->{'verbosity'} > 1;
|
---|
377 |
|
---|
378 | my $cursection = $doc_obj->get_top_section();
|
---|
379 |
|
---|
380 | ## $self->extract_metadata ($textref, $metadata, $doc_obj, $cursection);
|
---|
381 |
|
---|
382 | # add text to document object
|
---|
383 |
|
---|
384 | # $$textref =~ s/<(.*?)>/$1 /g;
|
---|
385 | $$textref =~ s/</</g;
|
---|
386 | $$textref =~ s/>/>/g;
|
---|
387 | $$textref =~ s/\[/[/g;
|
---|
388 | $$textref =~ s/\]/]/g;
|
---|
389 |
|
---|
390 | $doc_obj->add_utf8_text($cursection, $$textref);
|
---|
391 |
|
---|
392 | return 1;
|
---|
393 | }
|
---|
394 |
|
---|
395 |
|
---|
396 | # Improvement is to merge this with newer version in MetadataPass
|
---|
397 |
|
---|
398 | sub open_prettyprint_metadata_table
|
---|
399 | {
|
---|
400 | my $self = shift(@_);
|
---|
401 |
|
---|
402 | my $att = "width=100% cellspacing=2";
|
---|
403 | my $style = "style=\'border-bottom: 4px solid #000080\'";
|
---|
404 |
|
---|
405 | $self->{'ppmd_table'} = "\n<table $att $style>";
|
---|
406 | }
|
---|
407 |
|
---|
408 | sub add_prettyprint_metadata_line
|
---|
409 | {
|
---|
410 | my $self = shift(@_);
|
---|
411 | my ($metaname, $metavalue_utf8) = @_;
|
---|
412 |
|
---|
413 | $metavalue_utf8 = &util::hyperlink_text($metavalue_utf8);
|
---|
414 |
|
---|
415 | $self->{'ppmd_table'} .= " <tr bgcolor=#b5d3cd>\n";
|
---|
416 | $self->{'ppmd_table'} .= " <td width=30%>\n";
|
---|
417 | $self->{'ppmd_table'} .= " $metaname\n";
|
---|
418 | $self->{'ppmd_table'} .= " </td>\n";
|
---|
419 | $self->{'ppmd_table'} .= " <td>\n";
|
---|
420 | $self->{'ppmd_table'} .= " $metavalue_utf8\n";
|
---|
421 | $self->{'ppmd_table'} .= " </td>\n";
|
---|
422 | $self->{'ppmd_table'} .= " </tr>\n";
|
---|
423 |
|
---|
424 | }
|
---|
425 |
|
---|
426 | sub close_prettyprint_metadata_table
|
---|
427 | {
|
---|
428 | my $self = shift(@_);
|
---|
429 |
|
---|
430 | $self->{'ppmd_table'} .= "</table>\n";
|
---|
431 | }
|
---|
432 |
|
---|
433 | my $qualified_dc_mapping = {
|
---|
434 | "alternative" => "dc.title",
|
---|
435 | "tableOfContents" => "dc.description",
|
---|
436 | "abstract" => "dc.description",
|
---|
437 | "created" => "dc.date",
|
---|
438 | "valid" => "dc.date",
|
---|
439 | "available" => "dc.date",
|
---|
440 | "issued" => "dc.date",
|
---|
441 | "modified" => "dc.date",
|
---|
442 | "dateAccepted" => "dc.date",
|
---|
443 | "dateCopyrighted" => "dc.date",
|
---|
444 | "dateSubmitted" => "dc.date",
|
---|
445 | "extent" => "dc.format",
|
---|
446 | "medium" => "dc.format",
|
---|
447 | "isVersionOf" => "dc.relation",
|
---|
448 | "hasVersion" => "dc.relation",
|
---|
449 | "isReplacedBy" => "dc.relation",
|
---|
450 | "replaces" => "dc.relation",
|
---|
451 | "isRequiredBy" => "dc.relation",
|
---|
452 | "requires" => "dc.relation",
|
---|
453 | "isPartOf" => "dc.relation",
|
---|
454 | "hasPart" => "dc.relation",
|
---|
455 | "isReferencedBy" => "dc.relation",
|
---|
456 | "references" => "dc.relation",
|
---|
457 | "isFormatOf" => "dc.relation",
|
---|
458 | "hasFormat" => "dc.relation",
|
---|
459 | "conformsTo" => "dc.relation",
|
---|
460 | "spatial" => "dc.coverage",
|
---|
461 | "temporal" => "dc.coverage",
|
---|
462 | # these are now top level elements in our qualified dc metadata set
|
---|
463 | # "audience" => "dc.any",
|
---|
464 | # "accrualMethod" => "dc.any",
|
---|
465 | # "accrualPeriodicity" => "dc.any",
|
---|
466 | # "accrualPolicy" => "dc.any",
|
---|
467 | # "instructionalMethod" => "dc.any",
|
---|
468 | # "provenance" => "dc.any",
|
---|
469 | # "rightsHolder" => "dc.any",
|
---|
470 | "mediator" => "dc.audience",
|
---|
471 | "educationLevel" => "dc.audience",
|
---|
472 | "accessRights" => "dc.rights",
|
---|
473 | "license" => "dc.rights",
|
---|
474 | "bibliographicCitation" => "dc.identifier"
|
---|
475 | };
|
---|
476 |
|
---|
477 | sub remap_dc_metadata
|
---|
478 | {
|
---|
479 | my $self = shift(@_);
|
---|
480 |
|
---|
481 | my ($metaname) = @_;
|
---|
482 |
|
---|
483 | my ($prefix,$name) = ($metaname =~ m/^(.*?)\.(.*?)$/);
|
---|
484 |
|
---|
485 | if (defined $qualified_dc_mapping->{$name}) {
|
---|
486 |
|
---|
487 | return $qualified_dc_mapping->{$name}."^".$name;
|
---|
488 | }
|
---|
489 |
|
---|
490 |
|
---|
491 | return $metaname; # didn't get a match, return param passed in unchanged
|
---|
492 | }
|
---|
493 |
|
---|
494 |
|
---|
495 | sub extract_oai_metadata {
|
---|
496 | my $self = shift (@_);
|
---|
497 | my ($textref, $metadata) = @_;
|
---|
498 | my $outhandle = $self->{'outhandle'};
|
---|
499 |
|
---|
500 | $self->open_prettyprint_metadata_table();
|
---|
501 |
|
---|
502 | if ($$textref =~ m/<metadata\s*>(.*?)<\/metadata\s*>/s)
|
---|
503 | {
|
---|
504 | my $metadata_text = $1;
|
---|
505 |
|
---|
506 | # locate and remove outermost tag (ignoring any attribute information in top-level tag)
|
---|
507 | my ($outer_tagname,$inner_metadata_text) = ($metadata_text =~ m/<([^ >]+).*?>(.*?)<\/\1>/s);
|
---|
508 | # split tag into namespace and tag name
|
---|
509 | my($namespace,$top_level_prefix) = ($outer_tagname =~ m/^(.*?):(.*?)$/);
|
---|
510 | # sometimes, the dc namespace is not specified as the prefix in each element (like <dc:title>)
|
---|
511 | # but is rather defined in the wrapper element containing the various dc meta elements,
|
---|
512 | # like <dc><title></title><creator></creator></dc>.
|
---|
513 | # In such a case, we use this wrapper element as the top_level_prefix
|
---|
514 |
|
---|
515 | # if there was no prefix, then the tag itself becomes the top_level_prefix
|
---|
516 | if(!defined $top_level_prefix && defined $outer_tagname) {
|
---|
517 | $top_level_prefix = $outer_tagname;
|
---|
518 | }
|
---|
519 |
|
---|
520 | #process each element one by one
|
---|
521 | while ($inner_metadata_text =~ m/<([^ >]+).*?>(.*?)<\/\1>(.*)/s)
|
---|
522 | {
|
---|
523 |
|
---|
524 | my $metaname = $1;
|
---|
525 | my $metavalue = $2;
|
---|
526 | $inner_metadata_text = $3;
|
---|
527 |
|
---|
528 | # greenstone uses . for namespace, while oai uses :
|
---|
529 | $metaname =~ s/:/\./;
|
---|
530 | # if there is no namespace, then we use the outer tag name or
|
---|
531 | # namespace for this element
|
---|
532 | if ($metaname !~ m/\./)
|
---|
533 | {
|
---|
534 | $metaname = "$top_level_prefix.$metaname";
|
---|
535 | }
|
---|
536 |
|
---|
537 | # if metadata set is auto, leave as is, otherwise convert to
|
---|
538 | # specified namespace
|
---|
539 | if ($self->{'metadata_set'} ne "auto") {
|
---|
540 | if ($metaname !~ /^gi\./) { # hack to not overwrite gi metadata
|
---|
541 | $metaname =~ s/^([^\.]*)\./$self->{'metadata_set'}\./;
|
---|
542 | if ($self->{'metadata_set'} eq "dc") {
|
---|
543 | # convert qualified dc terms to gs version, e.g.
|
---|
544 | # spatial becomes coverage^spatial
|
---|
545 | $metaname = $self->remap_dc_metadata($metaname);
|
---|
546 | }
|
---|
547 | }
|
---|
548 | }
|
---|
549 |
|
---|
550 | # uppercase the first char of the name
|
---|
551 | $metaname =~ s/\.(.)/\.\u$1/;
|
---|
552 | $metavalue =~ s/\[/[/g;
|
---|
553 | $metavalue =~ s/\]/]/g;
|
---|
554 |
|
---|
555 | # so that GLI can see this metadata, store here as ex.dc.Title etc
|
---|
556 | my $ex_metaname = $metaname;
|
---|
557 | $ex_metaname =~ s/^ex\.//; # remove any pre-existing ex. prefix
|
---|
558 | $ex_metaname = "ex.$ex_metaname"; # at last can prefix ex.
|
---|
559 |
|
---|
560 | if (defined $metadata->{$ex_metaname})
|
---|
561 | {
|
---|
562 | push(@{$metadata->{$ex_metaname}},$metavalue);
|
---|
563 |
|
---|
564 | }
|
---|
565 | else
|
---|
566 | {
|
---|
567 | $metadata->{$ex_metaname} = [ $metavalue ];
|
---|
568 | }
|
---|
569 |
|
---|
570 | # but don't add ex to the pretty print line
|
---|
571 | $self->add_prettyprint_metadata_line($metaname, $metavalue);
|
---|
572 |
|
---|
573 | }
|
---|
574 | }
|
---|
575 |
|
---|
576 | $self->close_prettyprint_metadata_table();
|
---|
577 | }
|
---|
578 |
|
---|
579 | ## we know from the file extension, so doesn't need to check the doctype
|
---|
580 | sub check_doctype {
|
---|
581 |
|
---|
582 | return 1;
|
---|
583 | }
|
---|
584 |
|
---|
585 | 1;
|
---|