1 | ##########################################################################
|
---|
2 | #
|
---|
3 | # jenaTDBBuildproc.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) 1999 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 | # This document processor outputs a document for indexing (should be
|
---|
27 | # implemented by subclass) and storing in the database
|
---|
28 |
|
---|
29 | package jenaTDBBuildproc;
|
---|
30 |
|
---|
31 | use strict;
|
---|
32 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
---|
33 |
|
---|
34 | use docprint;
|
---|
35 | use util;
|
---|
36 | use FileUtils;
|
---|
37 |
|
---|
38 | use extrabuildproc;
|
---|
39 |
|
---|
40 |
|
---|
41 | BEGIN {
|
---|
42 | @jenaTDBBuildproc::ISA = ('extrabuildproc');
|
---|
43 | }
|
---|
44 |
|
---|
45 | sub new()
|
---|
46 | {
|
---|
47 | my $class = shift @_;
|
---|
48 |
|
---|
49 | my $self = new extrabuildproc (@_);
|
---|
50 |
|
---|
51 | # Do the following here so it doesn't keep checking (within the util.pm method)
|
---|
52 | # whether it needs to create the directory or not
|
---|
53 | my $tmp_dir = &util::get_collectlevel_tmp_dir();
|
---|
54 | $self->{'tmp_dir'} = $tmp_dir;
|
---|
55 |
|
---|
56 |
|
---|
57 | my $xslt_file_in = "gsdom2rdf.xsl";
|
---|
58 |
|
---|
59 | my $xslt_filename_in = &util::locate_config_file($xslt_file_in);
|
---|
60 | if (!defined $xslt_filename_in) {
|
---|
61 | print STDERR "Can not find $xslt_file_in, please make sure you have supplied the correct file path\n";
|
---|
62 | die "\n";
|
---|
63 | }
|
---|
64 |
|
---|
65 | my $xslt_filename_out = &FileUtils::filenameConcatenate($tmp_dir,$xslt_file_in);
|
---|
66 |
|
---|
67 | my $collection = $self->{'collection'};
|
---|
68 |
|
---|
69 | my $url_prefix = &util::get_full_greenstone_url_prefix();
|
---|
70 |
|
---|
71 | my $property_hashmap = { 'libraryurl' => $url_prefix,
|
---|
72 | 'collect' => $collection };
|
---|
73 |
|
---|
74 | file_copy_with_property_sub($xslt_filename_in,$xslt_filename_out,$property_hashmap);
|
---|
75 |
|
---|
76 | $self->{'xslt_file'} = $xslt_file_in;
|
---|
77 | $self->{'xslt_filename'} = $xslt_filename_out;
|
---|
78 |
|
---|
79 | return bless $self, $class;
|
---|
80 | }
|
---|
81 |
|
---|
82 |
|
---|
83 | sub property_lookup
|
---|
84 | {
|
---|
85 | my ($hashmap,$value) = @_;
|
---|
86 |
|
---|
87 | my $lookup = (defined $hashmap->{$value}) ? $hashmap->{$value} : "\@$value\@";
|
---|
88 |
|
---|
89 | return $lookup;
|
---|
90 | }
|
---|
91 |
|
---|
92 |
|
---|
93 | # Performs a text file copy, substituding substings of the form
|
---|
94 | # @xxx@ in the input file with the values set in hashmap
|
---|
95 | # passed in
|
---|
96 |
|
---|
97 | sub file_copy_with_property_sub
|
---|
98 | {
|
---|
99 | my ($filename_in,$filename_out,$property_hashmap) = @_;
|
---|
100 |
|
---|
101 | if (!open(FIN, "<$filename_in")) {
|
---|
102 | print STDERR "util::file_substitute_at_properteis failed to open $filename_in\n $!\n";
|
---|
103 | return;
|
---|
104 | }
|
---|
105 | binmode(FIN,":utf8");
|
---|
106 |
|
---|
107 | if (!open(FOUT, ">$filename_out")) {
|
---|
108 | print STDERR "util::file_substitute_at_properteis failed to open $filename_out\n $!\n";
|
---|
109 | return;
|
---|
110 | }
|
---|
111 | binmode(FOUT,":utf8");
|
---|
112 |
|
---|
113 | my $line;
|
---|
114 | while (defined($line = <FIN>)) {
|
---|
115 |
|
---|
116 | $line =~ s/\@([^@ ]+)\@/&property_lookup($property_hashmap,$1)/ige;
|
---|
117 |
|
---|
118 | print FOUT $line;
|
---|
119 | }
|
---|
120 |
|
---|
121 | close(FIN);
|
---|
122 | close(FOUT);
|
---|
123 | }
|
---|
124 |
|
---|
125 |
|
---|
126 | sub open_xslt_pipe
|
---|
127 | {
|
---|
128 | my $self = shift @_;
|
---|
129 | my ($output_file_name, $xslt_file)=@_;
|
---|
130 |
|
---|
131 | return unless defined $xslt_file and $xslt_file ne "" and &FileUtils::fileExists($xslt_file);
|
---|
132 |
|
---|
133 | my $apply_xslt_jar = &FileUtils::javaFilenameConcatenate($ENV{'GSDLHOME'},"bin","java","ApplyXSLT.jar");
|
---|
134 | my $xalan_jar = &FileUtils::javaFilenameConcatenate($ENV{'GSDLHOME'},"bin","java","xalan.jar");
|
---|
135 |
|
---|
136 | my $java_class_path = &util::javapathname_cat($apply_xslt_jar,$xalan_jar);
|
---|
137 |
|
---|
138 | $xslt_file = &util::makeFilenameJavaCygwinCompatible($xslt_file);
|
---|
139 |
|
---|
140 | my $mapping_file_path = "";
|
---|
141 |
|
---|
142 | my $cmd = "| java -cp \"$java_class_path\" org.nzdl.gsdl.ApplyXSLT -t \"$xslt_file\" ";
|
---|
143 |
|
---|
144 |
|
---|
145 | if (defined $self->{'mapping_file'} and $self->{'mapping_file'} ne ""){
|
---|
146 | my $mapping_file_path = "\"".$self->{'mapping_file'}."\"";
|
---|
147 | $cmd .= "-m $mapping_file_path";
|
---|
148 | }
|
---|
149 |
|
---|
150 | if (!open(*XMLWRITER, $cmd)) {
|
---|
151 | print STDERR "Can't open pipe to xslt: $!";
|
---|
152 | print STDERR "Command was:\n $cmd\n";
|
---|
153 | $self->{'xslt_writer'} = undef;
|
---|
154 | }
|
---|
155 | else {
|
---|
156 | $self->{'xslt_writer'} = *XMLWRITER;
|
---|
157 |
|
---|
158 | print XMLWRITER "<?DocStart?>\n";
|
---|
159 | print XMLWRITER "$output_file_name\n";
|
---|
160 | }
|
---|
161 | }
|
---|
162 |
|
---|
163 |
|
---|
164 | sub close_xslt_pipe
|
---|
165 | {
|
---|
166 | my $self = shift @_;
|
---|
167 |
|
---|
168 | return unless defined $self->{'xslt_writer'} ;
|
---|
169 |
|
---|
170 | my $xsltwriter = $self->{'xslt_writer'};
|
---|
171 |
|
---|
172 | print $xsltwriter "<?DocEnd?>\n";
|
---|
173 | close($xsltwriter);
|
---|
174 |
|
---|
175 | undef $self->{'xslt_writer'};
|
---|
176 |
|
---|
177 | }
|
---|
178 |
|
---|
179 | sub make_ttl_safe
|
---|
180 | {
|
---|
181 | my ($front,$str,$back) = @_;
|
---|
182 |
|
---|
183 | $str =~ s/\\/\\\\/g;
|
---|
184 |
|
---|
185 | $str =~ s/\&#x([0-9A-F]+);/chr(hex($1))/eig;
|
---|
186 | $str =~ s/\&#([0-9]+);/chr($1)/eig;
|
---|
187 |
|
---|
188 | $str =~ s/[\r\n]+/\\n/g;
|
---|
189 |
|
---|
190 | return "$front$str$back";
|
---|
191 | }
|
---|
192 |
|
---|
193 |
|
---|
194 | sub xml_to_ttl {
|
---|
195 | my $self = shift (@_);
|
---|
196 | my ($section_text,$output_root) = @_;
|
---|
197 | my $handle = $self->{'output_handle'};
|
---|
198 |
|
---|
199 | my $tmp_dir = $self->{'tmp_dir'};
|
---|
200 | my $tmp_doc_filename = &FileUtils::filenameConcatenate($tmp_dir,"$output_root.ttl");
|
---|
201 | my $tmp_doc_filename_cc = &util::makeFilenameJavaCygwinCompatible($tmp_doc_filename);
|
---|
202 |
|
---|
203 | my $xslt_filename = $self->{'xslt_filename'};
|
---|
204 | $self->open_xslt_pipe($tmp_doc_filename_cc, $xslt_filename); # stops with error if not able to open pipe
|
---|
205 |
|
---|
206 | my $xml_outhandler = $self->{'xslt_writer'};
|
---|
207 |
|
---|
208 | if (defined $xml_outhandler) {
|
---|
209 | binmode($xml_outhandler,":utf8");
|
---|
210 |
|
---|
211 | ### my $section_text = &docprint::get_section_xml($doc_obj);
|
---|
212 |
|
---|
213 | $section_text =~ s/(<Metadata[^>]*>)(.*?)(<\/Metadata>)/&make_ttl_safe($1,$2,$3)/gse;
|
---|
214 | ## $1&make_ttl_safe($2)$3
|
---|
215 |
|
---|
216 | ## print STDERR "*** st = $section_text\n\n";
|
---|
217 | ## $self->debug_section_text($section_text);
|
---|
218 |
|
---|
219 | print $xml_outhandler $section_text;
|
---|
220 | }
|
---|
221 |
|
---|
222 | $self->close_xslt_pipe();
|
---|
223 |
|
---|
224 | # now feed the generated file to jena's (TDB) triple store
|
---|
225 |
|
---|
226 | my $outhandle = $self->{'outhandle'};
|
---|
227 | print $outhandle " Inserting triples for $output_root\n";
|
---|
228 |
|
---|
229 | my $collection = $self->{'collection'};
|
---|
230 |
|
---|
231 | if (-f $tmp_doc_filename) {
|
---|
232 |
|
---|
233 | # my $cmd = "gs-triplestore-add $collection \"$tmp_doc_filename\"";
|
---|
234 | my $cmd = "gs-triplestore-add3 $collection \"$tmp_doc_filename\"";
|
---|
235 |
|
---|
236 | my $status = system($cmd);
|
---|
237 | if ($status != 0) {
|
---|
238 | print STDERR "Error: failed to run:\n $cmd\n$!\n";
|
---|
239 | }
|
---|
240 |
|
---|
241 | # print STDERR "**** temporarily supressing deletion of: $tmp_doc_filename\n";
|
---|
242 | unlink $tmp_doc_filename;
|
---|
243 | }
|
---|
244 | else {
|
---|
245 | print STDERR "*** Failed to generate: $tmp_doc_filename\n";
|
---|
246 | }
|
---|
247 |
|
---|
248 | }
|
---|
249 |
|
---|
250 | sub textedit {
|
---|
251 | my $self = shift (@_);
|
---|
252 | my ($doc_obj) = @_;
|
---|
253 | my $handle = $self->{'output_handle'};
|
---|
254 |
|
---|
255 | # print STDERR "**** jenaTDBBuildproc::textedit()\n";
|
---|
256 |
|
---|
257 | my $doc_oid = $doc_obj->get_OID();
|
---|
258 | my $ttl_output_root_file = "doc-$doc_oid";
|
---|
259 |
|
---|
260 | my $section_text = &docprint::get_section_xml($doc_obj);
|
---|
261 | $self->xml_to_ttl($section_text,$ttl_output_root_file);
|
---|
262 |
|
---|
263 | # my $tmp_dir = $self->{'tmp_dir'};
|
---|
264 | # my $tmp_doc_filename = &FileUtils::filenameConcatenate($tmp_dir,"doc-$doc_oid.ttl");
|
---|
265 | # my $tmp_doc_filename_cc = &util::makeFilenameJavaCygwinCompatible($tmp_doc_filename);
|
---|
266 |
|
---|
267 | # my $xslt_filename = $self->{'xslt_filename'};
|
---|
268 | # $self->open_xslt_pipe($tmp_doc_filename_cc, $xslt_filename); # stops with error if not able to open pipe
|
---|
269 |
|
---|
270 | # my $xml_outhandler = $self->{'xslt_writer'};
|
---|
271 |
|
---|
272 | # if (defined $xml_outhandler) {
|
---|
273 | # binmode($xml_outhandler,":utf8");
|
---|
274 |
|
---|
275 | # my $section_text = &docprint::get_section_xml($doc_obj);
|
---|
276 |
|
---|
277 | # $section_text =~ s/(<Metadata[^>]*>)(.*?)(<\/Metadata>)/&make_ttl_safe($1,$2,$3)/gse;
|
---|
278 | # ## $1&make_ttl_safe($2)$3
|
---|
279 |
|
---|
280 | # ## print STDERR "*** st = $section_text\n\n";
|
---|
281 |
|
---|
282 | # print $xml_outhandler $section_text;
|
---|
283 | # }
|
---|
284 |
|
---|
285 | # $self->close_xslt_pipe();
|
---|
286 |
|
---|
287 | # # now feed the generated file to jena's (TDB) tripple store
|
---|
288 |
|
---|
289 | # my $outhandle = $self->{'outhandle'};
|
---|
290 | # print $outhandle " Inserting tripples for $doc_oid\n";
|
---|
291 |
|
---|
292 | # my $collection = $self->{'collection'};
|
---|
293 |
|
---|
294 | # if (-f $tmp_doc_filename) {
|
---|
295 |
|
---|
296 | # # my $cmd = "gs-triplestore-add $collection \"$tmp_doc_filename\"";
|
---|
297 | # my $cmd = "gs-triplestore-add3 $collection \"$tmp_doc_filename\"";
|
---|
298 |
|
---|
299 | # my $status = system($cmd);
|
---|
300 | # if ($status != 0) {
|
---|
301 | # print STDERR "Error: failed to run:\n $cmd\n$!\n";
|
---|
302 | # }
|
---|
303 |
|
---|
304 | # # print STDERR "**** temporarily supressing deletion of: $tmp_doc_filename\n";
|
---|
305 | # unlink $tmp_doc_filename;
|
---|
306 | # }
|
---|
307 | # else {
|
---|
308 | # print STDERR "*** Failed to generate: $tmp_doc_filename\n";
|
---|
309 | # }
|
---|
310 |
|
---|
311 | }
|
---|
312 |
|
---|
313 |
|
---|
314 | sub text {
|
---|
315 | my $self = shift (@_);
|
---|
316 | my ($doc_obj,$file) = @_;
|
---|
317 |
|
---|
318 | $self->textedit($doc_obj,$file,"add");
|
---|
319 | }
|
---|
320 |
|
---|
321 | sub textreindex
|
---|
322 | {
|
---|
323 | my $self = shift @_;
|
---|
324 | my ($doc_obj,$file) = @_;
|
---|
325 |
|
---|
326 | $self->textedit($doc_obj,$file,"update");
|
---|
327 | }
|
---|
328 |
|
---|
329 | sub textdelete
|
---|
330 | {
|
---|
331 | my $self = shift @_;
|
---|
332 |
|
---|
333 | my ($doc_obj,$file) = @_;
|
---|
334 |
|
---|
335 | print STDERR "Warning: jenaTDB command-line does not currently support delete operation\n";
|
---|
336 |
|
---|
337 | # $self->textedit($doc_obj,$file,"delete");
|
---|
338 | }
|
---|
339 |
|
---|
340 |
|
---|
341 | sub infodbedit
|
---|
342 | {
|
---|
343 | my $self = shift (@_);
|
---|
344 | my ($doc_obj, $filename, $edit_mode) = @_;
|
---|
345 |
|
---|
346 | # print STDERR "**** jenaTDBBuidproc::infodbedit(): $filename, $edit_mode\n";
|
---|
347 |
|
---|
348 | # only output this document if it is a "indexed_doc" or "info_doc" (database only) document
|
---|
349 | my $doctype = $doc_obj->get_doc_type();
|
---|
350 | return if ($doctype ne "indexed_doc" && $doctype ne "info_doc");
|
---|
351 |
|
---|
352 |
|
---|
353 | #
|
---|
354 | # The following is done in basebuildproc, consider if it makes sense to do here
|
---|
355 | #
|
---|
356 |
|
---|
357 | # #add this document to the browse structure
|
---|
358 | # push(@{$self->{'doclist'}},$doc_obj->get_OID())
|
---|
359 | # unless ($doctype eq "classification");
|
---|
360 | # $self->{'num_docs'} += 1 unless ($doctype eq "classification");
|
---|
361 |
|
---|
362 | # if (!defined $filename) {
|
---|
363 | # # a reconstructed doc
|
---|
364 | # my $num_reconstructed_bytes = $doc_obj->get_metadata_element ($doc_obj->get_top_section (), "total_numbytes");
|
---|
365 | # if (defined $num_reconstructed_bytes) {
|
---|
366 | # $self->{'num_bytes'} += $num_reconstructed_bytes;
|
---|
367 | # }
|
---|
368 | # }
|
---|
369 |
|
---|
370 | # classify the document
|
---|
371 | &classify::classify_doc ($self->{'classifiers'}, $doc_obj);
|
---|
372 |
|
---|
373 |
|
---|
374 | }
|
---|
375 |
|
---|
376 |
|
---|
377 | 1;
|
---|