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 | my $xslt_file = "gsdom2rdf.xsl";
|
---|
52 |
|
---|
53 | my $xslt_filename = &util::locate_config_file($xslt_file);
|
---|
54 | if (!defined $xslt_filename) {
|
---|
55 | print STDERR "Can not find $xslt_file, please make sure you have supplied the correct file path\n";
|
---|
56 | die "\n";
|
---|
57 | }
|
---|
58 |
|
---|
59 | $self->{'xslt_file'} = $xslt_file;
|
---|
60 | $self->{'xslt_filename'} = $xslt_filename;
|
---|
61 |
|
---|
62 | # Do the following here so it doesn't keep checking (within the util.pm method)
|
---|
63 | # whether it needs to create the directory or not
|
---|
64 | my $tmp_dir = &util::get_collectlevel_tmp_dir();
|
---|
65 | $self->{'tmp_dir'} = $tmp_dir;
|
---|
66 |
|
---|
67 | return bless $self, $class;
|
---|
68 | }
|
---|
69 |
|
---|
70 |
|
---|
71 |
|
---|
72 |
|
---|
73 | sub open_xslt_pipe
|
---|
74 | {
|
---|
75 | my $self = shift @_;
|
---|
76 | my ($output_file_name, $xslt_file)=@_;
|
---|
77 |
|
---|
78 | return unless defined $xslt_file and $xslt_file ne "" and &FileUtils::fileExists($xslt_file);
|
---|
79 |
|
---|
80 | my $apply_xslt_jar = &FileUtils::javaFilenameConcatenate($ENV{'GSDLHOME'},"bin","java","ApplyXSLT.jar");
|
---|
81 | my $xalan_jar = &FileUtils::javaFilenameConcatenate($ENV{'GSDLHOME'},"bin","java","xalan.jar");
|
---|
82 |
|
---|
83 | my $java_class_path = &util::javapathname_cat($apply_xslt_jar,$xalan_jar);
|
---|
84 |
|
---|
85 | $xslt_file = &util::makeFilenameJavaCygwinCompatible($xslt_file);
|
---|
86 |
|
---|
87 | my $mapping_file_path = "";
|
---|
88 |
|
---|
89 | my $cmd = "| java -cp \"$java_class_path\" org.nzdl.gsdl.ApplyXSLT -t \"$xslt_file\" ";
|
---|
90 |
|
---|
91 | if (defined $self->{'mapping_file'} and $self->{'mapping_file'} ne ""){
|
---|
92 | my $mapping_file_path = "\"".$self->{'mapping_file'}."\"";
|
---|
93 | $cmd .= "-m $mapping_file_path";
|
---|
94 | }
|
---|
95 |
|
---|
96 | print STDERR "*** cmd = $cmd\n";
|
---|
97 |
|
---|
98 | open(*XMLWRITER, $cmd)
|
---|
99 | or die "can't open pipe to xslt: $!";
|
---|
100 |
|
---|
101 |
|
---|
102 | $self->{'xslt_writer'} = *XMLWRITER;
|
---|
103 |
|
---|
104 | print XMLWRITER "<?DocStart?>\n";
|
---|
105 | print XMLWRITER "$output_file_name\n";
|
---|
106 |
|
---|
107 | }
|
---|
108 |
|
---|
109 |
|
---|
110 | sub close_xslt_pipe
|
---|
111 | {
|
---|
112 | my $self = shift @_;
|
---|
113 |
|
---|
114 |
|
---|
115 | return unless defined $self->{'xslt_writer'} ;
|
---|
116 |
|
---|
117 | my $xsltwriter = $self->{'xslt_writer'};
|
---|
118 |
|
---|
119 | print $xsltwriter "<?DocEnd?>\n";
|
---|
120 | close($xsltwriter);
|
---|
121 |
|
---|
122 | undef $self->{'xslt_writer'};
|
---|
123 |
|
---|
124 | }
|
---|
125 |
|
---|
126 |
|
---|
127 | sub textedit {
|
---|
128 | my $self = shift (@_);
|
---|
129 | my ($doc_obj) = @_;
|
---|
130 | my $handle = $self->{'output_handle'};
|
---|
131 |
|
---|
132 | my $doc_oid = $doc_obj->get_OID();
|
---|
133 |
|
---|
134 | my $tmp_dir = $self->{'tmp_dir'};
|
---|
135 | my $tmp_doc_filename = &FileUtils::filenameConcatenate($tmp_dir,"doc-$doc_oid.xml");
|
---|
136 | $tmp_doc_filename = &util::makeFilenameJavaCygwinCompatible($tmp_doc_filename);
|
---|
137 |
|
---|
138 | my $xslt_filename = $self->{'xslt_filename'};
|
---|
139 | $self->open_xslt_pipe($tmp_doc_filename, $xslt_filename); # stops with error if not able to open pipe
|
---|
140 |
|
---|
141 | my $outhandler = $self->{'xslt_writer'};
|
---|
142 | binmode($outhandler,":utf8");
|
---|
143 |
|
---|
144 | my $section_text = &docprint::get_section_xml($doc_obj,$doc_obj->get_top_section());
|
---|
145 | print $outhandler $section_text;
|
---|
146 |
|
---|
147 | $self->close_xslt_pipe();
|
---|
148 |
|
---|
149 |
|
---|
150 | # now feed to generated file to jena's (TDB) tripple store
|
---|
151 |
|
---|
152 |
|
---|
153 | }
|
---|
154 |
|
---|
155 |
|
---|
156 | sub texteditMG {
|
---|
157 | my $self = shift (@_);
|
---|
158 | my ($doc_obj) = @_;
|
---|
159 | my $handle = $self->{'output_handle'};
|
---|
160 |
|
---|
161 | my $doc_oid = $doc_obj->get_OID();
|
---|
162 |
|
---|
163 |
|
---|
164 | my $doc_section = 0; # just for this document
|
---|
165 |
|
---|
166 |
|
---|
167 | my $text = "";
|
---|
168 | my $text_extra = "";
|
---|
169 |
|
---|
170 | # get the text for this document
|
---|
171 | my $section = $doc_obj->get_top_section();
|
---|
172 | while (defined $section) {
|
---|
173 | # update a few statistics
|
---|
174 | $doc_section++;
|
---|
175 |
|
---|
176 | my $title = $doc_obj->get_metadata_element($section, "Title");
|
---|
177 |
|
---|
178 | if (defined $title && ($title =~ m/\S/)) {
|
---|
179 | print "$doc_oid: Title = $title\n";
|
---|
180 | }
|
---|
181 |
|
---|
182 | my $dc_title = $doc_obj->get_metadata_element($section, "dc.Title");
|
---|
183 |
|
---|
184 | if (defined $dc_title && ($dc_title =~ m/\S/)) {
|
---|
185 | print "$doc_oid: dc.Title = $dc_title\n";
|
---|
186 | }
|
---|
187 |
|
---|
188 |
|
---|
189 | my $id3_title = $doc_obj->get_metadata_element($section, "ex.ID3.Title");
|
---|
190 |
|
---|
191 | if (defined $id3_title && ($id3_title =~ m/\S/)) {
|
---|
192 | print "$doc_oid: id3.Title = $id3_title\n";
|
---|
193 | }
|
---|
194 |
|
---|
195 | $section = $doc_obj->get_next_section($section);
|
---|
196 | }
|
---|
197 |
|
---|
198 | print $handle "$text$text_extra";
|
---|
199 | }
|
---|
200 |
|
---|
201 |
|
---|
202 |
|
---|
203 | sub texteditADB {
|
---|
204 | my $self = shift (@_);
|
---|
205 | my ($doc_obj,$file,$mode) = @_;
|
---|
206 |
|
---|
207 | # Code written on the assumption that that jenaTDB does a replace
|
---|
208 | # operation when presented with a docid that already extis.
|
---|
209 | # => don't need to do anything special to distinguish between
|
---|
210 | # a mode of "add" and "update"
|
---|
211 |
|
---|
212 | my $outhandle = $self->{'outhandle'};
|
---|
213 |
|
---|
214 |
|
---|
215 | my $source_dir = $self->{'source_dir'}; # typically the archives dir
|
---|
216 | my $build_dir = $self->{'build_dir'};
|
---|
217 |
|
---|
218 | # full path to adb database
|
---|
219 | my $adb_filename
|
---|
220 | = &util::filename_cat($build_dir, "jenaTDB", "lsh-features.adb");
|
---|
221 |
|
---|
222 | # get doc id
|
---|
223 | my $doc_oid = $doc_obj->get_OID();
|
---|
224 |
|
---|
225 | # map to assoc dir
|
---|
226 | my $top_section = $doc_obj->get_top_section();
|
---|
227 | my $assoc_file
|
---|
228 | = $doc_obj->get_metadata_element ($top_section,"assocfilepath");
|
---|
229 | my $assoc_filename = &util::filename_cat($source_dir,$assoc_file);
|
---|
230 |
|
---|
231 | my $chr12_filename = &util::filename_cat($assoc_filename,"doc.chr12");
|
---|
232 | my $powerlog_filename = &util::filename_cat($assoc_filename,"doc.power");
|
---|
233 |
|
---|
234 | print $outhandle " Inserting tripples for $doc_oid\n";
|
---|
235 |
|
---|
236 | # my $cmd = "jenaTDB -d \"$adb_filename\" -I -k \"$doc_oid\" -f \"$chr12_filename\" -w \"$powerlog_filename\"";
|
---|
237 |
|
---|
238 | # my $status = system($cmd);
|
---|
239 | # if ($status != 0) {
|
---|
240 | # print STDERR "Error: failed to run:\n $cmd\n$!\n";
|
---|
241 | # }
|
---|
242 |
|
---|
243 | }
|
---|
244 |
|
---|
245 | sub text {
|
---|
246 | my $self = shift (@_);
|
---|
247 | my ($doc_obj,$file) = @_;
|
---|
248 |
|
---|
249 | $self->textedit($doc_obj,$file,"add");
|
---|
250 | }
|
---|
251 |
|
---|
252 | sub textreindex
|
---|
253 | {
|
---|
254 | my $self = shift @_;
|
---|
255 | my ($doc_obj,$file) = @_;
|
---|
256 |
|
---|
257 | $self->textedit($doc_obj,$file,"update");
|
---|
258 | }
|
---|
259 |
|
---|
260 | sub textdelete
|
---|
261 | {
|
---|
262 | my $self = shift @_;
|
---|
263 |
|
---|
264 | my ($doc_obj,$file) = @_;
|
---|
265 |
|
---|
266 | print STDERR "Warning: jenaTDB command-line does not currently support delete operation\n";
|
---|
267 |
|
---|
268 | # $self->textedit($doc_obj,$file,"delete");
|
---|
269 | }
|
---|
270 |
|
---|
271 |
|
---|
272 |
|
---|
273 |
|
---|
274 |
|
---|
275 | 1;
|
---|