source: trunk/gsdl/perllib/plugins/OAIPlug.pm@ 13886

Last change on this file since 13886 was 13886, checked in by shaoqun, 17 years ago

doesn't need to check the doctype for the oai files

  • Property svn:keywords set to Author Date Id Revision
File size: 9.5 KB
Line 
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
27package OAIPlug;
28
29use BasPlug;
30use unicode;
31use util;
32
33use strict;
34no strict 'refs'; # allow filehandles to be variables and viceversa
35
36use XMLPlug;
37
38sub BEGIN {
39 @OAIPlug::ISA = ('XMLPlug');
40}
41
42
43my $arguments =
44 [ { 'name' => "process_exp",
45 'desc' => "{BasPlug.process_exp}",
46 'type' => "regexp",
47 'reqd' => "no",
48 'deft' => &get_default_process_exp() },
49 ];
50
51my $options = { 'name' => "OAIPlug",
52 'desc' => "{OAIPlug.desc}",
53 'abstract' => "no",
54 'inherits' => "yes",
55 'args' => $arguments };
56
57
58sub new {
59 my ($class) = shift (@_);
60 my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
61 push(@$pluginlist, $class);
62
63 if(defined $arguments){ push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});}
64 if(defined $options) { push(@{$hashArgOptLists->{"OptList"}},$options)};
65
66 my $self = new XMLPlug($pluginlist, $inputargs, $hashArgOptLists);
67
68 return bless $self, $class;
69}
70
71sub get_default_process_exp {
72 my $self = shift (@_);
73
74 return q^(?i)(\.oai)$^;
75}
76
77sub get_doctype {
78 my $self = shift(@_);
79
80 return "OAI-PMH";
81}
82
83sub xml_start_document {
84 my $self = shift (@_);
85 $self->{'in_metadata_node'} = 0;
86 $self->{'rawxml'} = "";
87}
88
89sub xml_end_document {
90}
91
92sub xml_doctype {
93 my $self = shift(@_);
94
95 my ($expat, $name, $sysid, $pubid, $internal) = @_;
96
97 ##die "" if ($name !~ /^OAI-PMH$/);
98
99 my $outhandle = $self->{'outhandle'};
100 print $outhandle "OAIPlug: processing $self->{'file'}\n" if $self->{'verbosity'} > 1;
101 print STDERR "<Processing n='$self->{'file'}' p='OAIPlug'>\n" if $self->{'gli'};
102
103}
104
105
106sub xml_start_tag {
107 my $self = shift(@_);
108 my ($expat,$element) = @_;
109
110 my %attr_hash = %_;
111
112 my $attr = "";
113 map { $attr .= " $_=$attr_hash{$_}"; } keys %attr_hash;
114
115 $self->{'rawxml'} .= "<$element$attr>";
116
117 if ($element eq "metadata") {
118 $self->{'in_metadata_node'} = 1;
119 $self->{'metadata_xml'} = "";
120 }
121
122 if ($self->{'in_metadata_node'}) {
123 $self->{'metadata_xml'} .= "<$element$attr>";
124 }
125}
126
127sub xml_end_tag {
128 my $self = shift(@_);
129 my ($expat, $element) = @_;
130
131 $self->{'rawxml'} .= "</$element>";
132
133 if ($self->{'in_metadata_node'}) {
134 $self->{'metadata_xml'} .= "</$element>";
135 }
136
137 if ($element eq "metadata") {
138 my $textref = \$self->{'metadata_xml'};
139 my $metadata = $self->{'metadata'};
140 $self->extract_oai_metadata($textref,$metadata);
141
142 $self->{'in_metadata_node'} = 0;
143 }
144
145
146}
147
148sub xml_text {
149 my $self = shift(@_);
150 my ($expat) = @_;
151
152 $self->{'rawxml'} .= $_;
153
154 if ($self->{'in_metadata_node'}) {
155 $self->{'metadata_xml'} .= $_;
156 }
157}
158
159
160
161
162sub read {
163 my $self = shift (@_);
164
165 my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
166
167 my $outhandle = $self->{'outhandle'};
168
169 my $filename = $file;
170 $filename = &util::filename_cat ($base_dir, $file) if $base_dir =~ /\w/;
171
172 return 0 if ((-d $filename) && ($filename =~ m/srcdocs$/));
173
174 if ($self->SUPER::read(@_)) {
175
176 # Do encoding stuff
177 my ($language, $encoding) = $self->textcat_get_language_encoding ($filename);
178
179 my $url_array = $metadata->{'URL'};
180 my $num_urls = (defined $url_array) ? scalar(@$url_array) : 0;
181
182 my $srcdoc_exists = 0;
183 my $srcdoc_pos = 0;
184 my $filename_dir = &util::filename_head($filename);
185
186 for (my $i=0; $i<$num_urls; $i++) {
187
188 if ($url_array->[$i] !~ m/^(http|ftp):/) {
189
190 my $src_filename = &util::filename_cat($filename_dir, $url_array->[$i]);
191
192 if (-e $src_filename) {
193 $srcdoc_pos = $i;
194 $srcdoc_exists = 1;
195 }
196 }
197 }
198
199 if ($srcdoc_exists)
200 {
201 print $outhandle "OAIPlug: passing metadata on to $url_array->[0]\n"
202 if ($self->{'verbosity'}>1);
203
204
205 # Make pretty print metadata table stick with src filename
206 my $ppmd_table = $self->{'ppmd_table'};
207 $metadata->{'prettymd'} = [ $ppmd_table ];
208 $self->{'ppmd_table'} = undef;
209
210 return &plugin::read ($pluginfo, $filename_dir, $url_array->[0],
211 $metadata, $processor, $maxdocs, $total_count, $gli);
212 }
213 else
214 {
215 # create a new document
216 my $doc_obj = new doc ($filename, "indexed_doc");
217 my $top_section = $doc_obj->get_top_section;
218 my $plugin_type = $self->{'plugin_type'};
219
220 $doc_obj->add_utf8_metadata($top_section, "Language", $language);
221 $doc_obj->add_utf8_metadata($top_section, "Encoding", $encoding);
222 $doc_obj->add_utf8_metadata($top_section, "Plugin", $plugin_type);
223 $doc_obj->add_metadata($top_section, "FileFormat", "OAI");
224 $doc_obj->add_metadata($top_section, "FileSize", (-s $filename));
225
226 # include any metadata passed in from previous plugins
227 # note that this metadata is associated with the top level section
228 $self->extra_metadata ($doc_obj, $doc_obj->get_top_section(), $metadata);
229
230 # do plugin specific processing of doc_obj
231 my $textref = \$self->{'rawxml'};
232 unless (defined ($self->process($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj))) {
233 print STDERR "<ProcessingError n='$file'>\n" if ($gli);
234 return -1;
235 }
236
237 # do any automatic metadata extraction
238 $self->auto_extract_metadata ($doc_obj);
239
240 # add an OID
241 $doc_obj->set_OID();
242
243 my $ppmd_table = $self->{'ppmd_table'};
244 $doc_obj->add_utf8_metadata($top_section,"prettymd",$ppmd_table);
245 $self->{'ppmd_table'} = undef;
246
247 # process the document
248 $processor->process($doc_obj);
249
250 $self->{'num_processed'} ++;
251
252 return 1; # processed the file
253 }
254 }
255 else {
256 return undef;
257 }
258}
259
260
261# do plugin specific processing of doc_obj
262sub process {
263 my $self = shift (@_);
264 my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
265 my $outhandle = $self->{'outhandle'};
266
267 print STDERR "<Processing n='$file' p='OAIPlug'>\n" if ($gli);
268 print $outhandle "OAIPlug: processing $file\n"
269 if $self->{'verbosity'} > 1;
270
271 my $cursection = $doc_obj->get_top_section();
272
273## $self->extract_metadata ($textref, $metadata, $doc_obj, $cursection);
274
275 # add text to document object
276
277# $$textref =~ s/<(.*?)>/$1 /g;
278 $$textref =~ s/</&lt;/g;
279 $$textref =~ s/>/&gt;/g;
280
281## print STDERR "*** adding text: $$textref\n";
282
283 $doc_obj->add_utf8_text($cursection, $$textref);
284
285 return 1;
286}
287
288
289# Improvement is to merge this with newer version in MetadataPass
290
291sub open_prettyprint_metadata_table
292{
293 my $self = shift(@_);
294
295 my $att = "width=100% cellspacing=2";
296 my $style = "style=\'border-bottom: 4px solid #000080\'";
297
298 $self->{'ppmd_table'} = "\n<table $att $style>";
299}
300
301sub add_prettyprint_metadata_line
302{
303 my $self = shift(@_);
304 my ($metaname, $metavalue_utf8) = @_;
305
306 $metavalue_utf8 =~ s/hdl\.handle\.net/mcgonagall.cs.waikato.ac.nz:8080\/dspace\/handle/;
307 $metavalue_utf8 = &util::hyperlink_text($metavalue_utf8);
308
309 $self->{'ppmd_table'} .= " <tr bgcolor=#b5d3cd>\n";
310 $self->{'ppmd_table'} .= " <td width=30%>\n";
311 $self->{'ppmd_table'} .= " $metaname\n";
312 $self->{'ppmd_table'} .= " </td>\n";
313 $self->{'ppmd_table'} .= " <td>\n";
314 $self->{'ppmd_table'} .= " $metavalue_utf8\n";
315 $self->{'ppmd_table'} .= " </td>\n";
316 $self->{'ppmd_table'} .= " </tr>\n";
317
318}
319
320sub close_prettyprint_metadata_table
321{
322 my $self = shift(@_);
323
324 $self->{'ppmd_table'} .= "</table>\n";
325}
326
327
328
329
330sub extract_oai_metadata {
331 my $self = shift (@_);
332 my ($textref, $metadata) = @_;
333 my $outhandle = $self->{'outhandle'};
334
335 # Only handles DC metadata
336
337 $self->open_prettyprint_metadata_table();
338
339 if ($$textref =~ m/<metadata\s*>(.*?)<\/metadata\s*>/s)
340 {
341 my $metadata_text = $1;
342 $metadata_text =~ s/^.*?<(oai_dc:)?dc.*?>(.*?)<\/(oai_dc:)?dc>.*?/$2/s;
343
344 while ($metadata_text =~ m/<(.*?)>(.*?)<\/(.*?)>(.*)/s)
345 {
346 # if URL given for document as identifier metadata, store it ...
347 # $doc_obj->add_utf8_metadata($cursection, "URL", $web_url);
348
349 my $metaname = $1;
350 my $metavalue = $2;
351 $metadata_text = $4;
352
353 $metaname =~ s/^(dc:)?(.)/\u$2/;
354
355 if ($metaname eq "Identifier")
356 {
357 # name clashes with GSDL reserved metadata name for hash id
358 $metaname = "URL";
359 }
360
361 if (defined $metadata->{$metaname})
362 {
363 push(@{$metadata->{$metaname}},$metavalue);
364
365 }
366 else
367 {
368 $metadata->{$metaname} = [ $metavalue ];
369 }
370
371 $self->add_prettyprint_metadata_line($metaname, $metavalue);
372
373 }
374 }
375
376 $self->close_prettyprint_metadata_table();
377}
378
379## we know from the file extension, so doesn't need to check the doctype
380sub check_doctype {
381
382 return 1;
383}
384
3851;
Note: See TracBrowser for help on using the repository browser.