source: main/trunk/greenstone2/perllib/classify.pm@ 31208

Last change on this file since 31208 was 29821, checked in by kjdon, 9 years ago

Detecting and handling the case where a 0 has been entered into hierarchy numbering in an hfile.

  • Property svn:keywords set to Author Date Id Revision
File size: 14.6 KB
Line 
1###########################################################################
2#
3# classify.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# functions to handle classifiers
27
28package classify;
29
30require util;
31use FileUtils;
32require AllList;
33
34use dbutil;
35use gsprintf;
36use strict; no strict 'subs';
37
38
39sub gsprintf
40{
41 return &gsprintf::gsprintf(@_);
42}
43
44
45sub load_classifier_for_info {
46 my ($classifier) = shift @_;
47
48 # find the classifier
49 # - used to have hardcoded list of places to load classifier from. We
50 # should, instead, try loading from all of the perllib places on the
51 # library path, as that improves support for extensions. Special cases
52 # needed for collection specific and custom classifier. [jmt12]
53 my @possible_class_paths;
54 if (defined($ENV{'GSDLCOLLECTION'}))
55 {
56 push(@possible_class_paths, &FileUtils::filenameConcatenate($ENV{'GSDLCOLLECTDIR'}, 'custom', $ENV{'GSDLCOLLECTION'}, 'perllib', 'classify', $classifier . '.pm'));
57 }
58 # (why does GSDLCOLLECTDIR get set to GSDLHOME for classinfo calls?)
59 if ($ENV{'GSDLCOLLECTDIR'} ne $ENV{'GSDLHOME'})
60 {
61 push(@possible_class_paths, &FileUtils::filenameConcatenate($ENV{'GSDLCOLLECTDIR'}, 'perllib', 'classify', $classifier . '.pm'));
62 }
63 foreach my $library_path (@INC)
64 {
65 # only interested in classify paths found in the library paths
66 if ($library_path =~ /classify$/)
67 {
68 push(@possible_class_paths, &FileUtils::filenameConcatenate($library_path, $classifier . '.pm'));
69 }
70 }
71 my $found_class = 0;
72 foreach my $possible_class_path (@possible_class_paths)
73 {
74 if (-e $possible_class_path)
75 {
76 require $possible_class_path;
77 $found_class = 1;
78 last;
79 }
80 }
81 if (!$found_class)
82 {
83 &gsprintf(STDERR, "{classify.could_not_find_classifier}\n", $classifier) && die "\n";
84 }
85
86 my ($classobj);
87 my $options = "-gsdlinfo";
88 eval ("\$classobj = new \$classifier([],[$options])");
89 die "$@" if $@;
90
91 return $classobj;
92}
93
94sub load_classifiers {
95 my ($classify_list, $build_dir, $outhandle) = @_;
96 my @classify_objects = ();
97 my $classify_number = 1;
98
99 # - ensure colclassdir doesn't already exist in INC before adding, other-
100 # wise we risk clobbering classifier inheritence implied by order of paths
101 # in INC [jmt12]
102 my $colclassdir = &FileUtils::filenameConcatenate($ENV{'GSDLCOLLECTDIR'},"perllib/classify");
103 &util::augmentINC($colclassdir);
104
105 foreach my $classifyoption (@$classify_list) {
106
107 # get the classifier name
108 my $classname = shift @$classifyoption;
109 next unless defined $classname;
110
111 # find the classifier
112 # - replaced as explained in load_classifier_for_info() [jmt12]
113 my @possible_class_paths;
114 if (defined($ENV{'GSDLCOLLECTION'}))
115 {
116 push(@possible_class_paths, &FileUtils::filenameConcatenate($ENV{'GSDLCOLLECTDIR'}, 'custom', $ENV{'GSDLCOLLECTION'}, 'perllib', 'classify', $classname . '.pm'));
117 }
118 # (why does GSDLCOLLECTDIR get set to GSDLHOME for classinfo calls?)
119 if ($ENV{'GSDLCOLLECTDIR'} ne $ENV{'GSDLHOME'})
120 {
121 push(@possible_class_paths,&FileUtils::filenameConcatenate($ENV{'GSDLCOLLECTDIR'}, 'perllib', 'classify', $classname . '.pm'));
122 }
123 foreach my $library_path (@INC)
124 {
125 # only interested in classify paths found in the library paths
126 if ($library_path =~ /classify$/)
127 {
128 push(@possible_class_paths, &FileUtils::filenameConcatenate($library_path, $classname . '.pm'));
129 }
130 }
131 my $found_class = 0;
132 foreach my $possible_class_path (@possible_class_paths)
133 {
134 if (-e $possible_class_path)
135 {
136 require $possible_class_path;
137 $found_class = 1;
138 last;
139 }
140 }
141 if (!$found_class)
142 {
143 &gsprintf(STDERR, "{classify.could_not_find_classifier}\n", $classname) && die "\n";
144 }
145
146 # create the classify object
147 my ($classobj);
148
149 my @newoptions;
150
151 # do these first so they can be overriden by user supplied options
152 if ($build_dir) {
153 (my $build_dir_re = $build_dir) =~ s@\\@\\\\@g; # copy build_dir into build_dir_re and modify build_dir_re
154 push @newoptions, "-builddir", "$build_dir_re";
155 }
156 push @newoptions, "-outhandle", "$outhandle" if ($outhandle);
157 push @newoptions, "-verbosity", "2";
158
159 # backwards compatability hack: if the classifier options are
160 # in "x=y" format, convert them to parsearg ("-x y") format.
161 my ($opt, $key, $value);
162 foreach $opt (@$classifyoption) {
163 # if ($opt =~ /^(\w+)=(.*)$/) {
164 # push @newoptions, "-$1", $2;
165 # } else {
166 push @newoptions, $opt;
167 #}
168 }
169
170 eval ("\$classobj = new \$classname([],[\@newoptions])");
171 die "$@" if $@;
172
173 $classobj->set_number($classify_number);
174 $classify_number ++;
175
176 # add this object to the list
177 push (@classify_objects, $classobj);
178 }
179
180 my ($classobj);
181 eval ("\$classobj = new AllList()");
182 die "$@" if $@;
183 push (@classify_objects, $classobj);
184
185 return \@classify_objects;
186}
187
188# init_classifiers resets all the classifiers and readys them to process
189# the documents.
190sub init_classifiers {
191 my ($classifiers) = @_;
192
193 foreach my $classobj (@$classifiers) {
194 $classobj->init();
195 }
196}
197
198
199
200# takes a hashref containing the metadata for an infodb entry, and extracts
201# the childrens numbers (from the 'contains' entry).
202# assumes format is ".1;".2;".3
203sub get_children {
204 my ($doc_db_hash) = @_;
205
206 my $children = undef;
207
208 my $contains = $doc_db_hash->{'contains'};
209 if (defined ($contains)) {
210 $contains =~ s/\@$//; #remove trailing @
211 $contains =~ s/^\"\.//; #remove initial ".
212 @$children = split /\;\"\./, $contains;
213 }
214
215 return $children;
216}
217
218
219sub recurse_sections {
220 my ($doc_obj, $children, $parentoid, $parentsection, $database_recs) = @_;
221
222 return if (!defined $children);
223
224 foreach my $child (sort { $a <=> $b} @$children) {
225 $doc_obj->create_named_section("$parentsection.$child");
226 my $doc_db_rec = $database_recs->{"$parentoid.$child"};
227 my $doc_db_hash = db_rec_to_hash($doc_db_rec);
228
229 # get child's children
230 my $newchildren = &get_children($doc_db_hash);
231
232 # add content for current section
233 add_section_content($doc_obj, "$parentsection.$child", $doc_db_hash);
234
235 # process all the children if there are any
236 if (defined ($newchildren))
237 {
238 recurse_sections($doc_obj, $newchildren, "$parentoid.$child",
239 "$parentsection.$child", $database_recs);
240 }
241 }
242}
243
244
245sub add_section_content {
246 my ($doc_obj, $cursection, $doc_db_hash) = @_;
247
248 foreach my $key (keys %$doc_db_hash) {
249 #don't need to store these metadata
250 next if $key =~ /(thistype|childtype|contains|docnum|doctype|classifytype)/i;
251 # but do want things like hastxt and archivedir
252 my @items = split /@/, $doc_db_hash->{$key};
253 # metadata is all from the info database so should already be in utf8
254 map {$doc_obj->add_utf8_metadata ($cursection, $key, $_); } @items;
255
256 }
257}
258
259
260# gets all the metadata from an infodb entry, and puts it into a hashref
261sub db_rec_to_hash {
262
263 my ($infodb_str_ref) = @_;
264
265 my $hashref = {};
266
267 my @entries = split(/\n/, $infodb_str_ref);
268 foreach my $entry (@entries) {
269 my($key, $value) = ($entry =~ /^<([^>]*)>(.*?)$/ );
270 $hashref->{$key} .= '@' if defined $hashref->{$key};
271 $hashref->{$key} .= $value;
272
273 }
274
275 return $hashref;
276}
277
278
279sub reconstruct_doc_objs_metadata
280{
281 my $infodb_type = shift(@_);
282 my $infodb_file_path = shift(@_);
283 my $database_recs = shift(@_);
284
285 # dig out top level doc sections
286 my %top_sections = ();
287 my %top_docnums = ();
288 foreach my $key ( keys %$database_recs )
289 {
290 my $md_rec = $database_recs->{$key};
291 my $md_hash = db_rec_to_hash($md_rec);
292
293 if ((defined $md_hash->{'doctype'}) && ($md_hash->{'doctype'} eq "doc")) {
294 next if ($key =~ m/\./);
295 $top_sections{$key} = $md_hash;
296 $top_docnums{$key} = $md_hash->{'docnum'};
297 }
298 }
299
300 # for greenstone document objects based on metadata in database file
301 my @all_docs = ();
302 # we need to make sure the documents were processed in the same order as
303 # before, so sort based on their docnums
304 foreach my $oid ( sort { $top_docnums{$a} <=> $top_docnums{$b} } keys %top_sections )
305 {
306 my $doc_db_hash = $top_sections{$oid};
307
308 my $doc_obj = new doc();
309 $doc_obj->set_OID($oid);
310 my $top = $doc_obj->get_top_section();
311 add_section_content ($doc_obj, $top, $doc_db_hash);
312 my $children = &get_children($doc_db_hash);
313 recurse_sections($doc_obj, $children, $oid, $top, $database_recs);
314
315 push(@all_docs,$doc_obj);
316 }
317
318 return \@all_docs;
319}
320
321
322
323
324
325# classify_doc lets each of the classifiers classify a document
326sub classify_doc {
327 my ($classifiers, $doc_obj) = @_;
328
329 foreach my $classobj (@$classifiers) {
330 my $title = $classobj->{'title'};
331
332 $classobj->classify($doc_obj);
333 }
334}
335
336
337our $next_classify_num = 1;
338
339# output_classify_info outputs all the info needed for the classification
340# to the database
341sub output_classify_info
342{
343 my ($classifiers, $infodb_type, $infodb_handle, $remove_empty_classifications, $gli) = @_;
344
345 $gli = 0 unless defined $gli;
346
347 # create a classification containing all the info
348 my $classifyinfo = { 'classifyOID'=> 'browse',
349 'contains' => [] };
350
351 # get each of the classifications
352 foreach my $classifier (@$classifiers)
353 {
354 my $classifier_info = $classifier->get_classify_info($gli);
355 if (defined $classifier_info) {
356 $classifier_info->{'classifyOID'} = "CL$next_classify_num" unless defined($classifier_info->{'classifyOID'});
357 print STDERR "*** outputting information for classifier: $classifier_info->{'classifyOID'}\n";
358
359 push(@{$classifyinfo->{'contains'}}, $classifier_info);
360 } else {
361 print STDERR "*** error with classifier CL$next_classify_num, not outputing it\n";
362 }
363 $next_classify_num++;
364
365 }
366
367 &print_classify_info($infodb_type, $infodb_handle, $classifyinfo, "", $remove_empty_classifications);
368}
369
370
371sub print_classify_info
372{
373 my ($infodb_type, $infodb_handle, $classifyinfo, $OID, $remove_empty_classifications) = @_;
374
375 $OID =~ s/^\.+//; # just for good luck
376
377 # book information is printed elsewhere
378 return if (defined ($classifyinfo->{'OID'}));
379
380 # don't want empty classifications
381 return if (&check_contents ($classifyinfo, $remove_empty_classifications) == 0 && $remove_empty_classifications);
382
383 $OID = $classifyinfo->{'classifyOID'} if defined ($classifyinfo->{'classifyOID'});
384
385 my %classify_infodb = ();
386 $classify_infodb{"doctype"} = [ "classify" ];
387 $classify_infodb{"hastxt"} = [ "0" ];
388 $classify_infodb{"childtype"} = [ $classifyinfo->{'childtype'} ]
389 if defined $classifyinfo->{'childtype'};
390 $classify_infodb{"Title"} = [ $classifyinfo->{'Title'} ]
391 if defined $classifyinfo->{'Title'};
392 $classify_infodb{"numleafdocs"} = [ $classifyinfo->{'numleafdocs'} ]
393 if defined $classifyinfo->{'numleafdocs'};
394 $classify_infodb{"thistype"} = [ $classifyinfo->{'thistype'} ]
395 if defined $classifyinfo->{'thistype'};
396 $classify_infodb{"parameters"} = [ $classifyinfo->{'parameters'} ]
397 if defined $classifyinfo->{'parameters'};
398 $classify_infodb{"supportsmemberof"} = [ $classifyinfo->{'supportsmemberof'} ]
399 if defined $classifyinfo->{'supportsmemberof'};
400
401 my $contains_text = "";
402 my $mdoffset_text = "";
403
404 my $next_subOID = 1;
405 my $first = 1;
406 foreach my $tempinfo (@{$classifyinfo->{'contains'}}) {
407 # empty contents were made undefined by clean_contents()
408 next unless defined $tempinfo;
409 if (!defined ($tempinfo->{'classifyOID'}) ||
410 $tempinfo->{'classifyOID'} ne "oai") {
411 $contains_text .= ";" unless $first;
412 }
413 $mdoffset_text .= ";" unless $first;
414 $first = 0;
415
416 if (defined ($tempinfo->{'classifyOID'}))
417 {
418 if ($tempinfo->{'classifyOID'} ne "oai")
419 {
420 $contains_text .= $tempinfo->{'classifyOID'};
421 }
422
423 &print_classify_info ($infodb_type, $infodb_handle, $tempinfo, $tempinfo->{'classifyOID'},
424 $remove_empty_classifications);
425 }
426 elsif (defined ($tempinfo->{'OID'}))
427 {
428 $contains_text .= $tempinfo->{'OID'};
429 $mdoffset_text .= $tempinfo->{'offset'} if (defined ($tempinfo->{'offset'}));
430 }
431 else
432 {
433 # Supress having top-level node in Collage classifier
434 # so no bookshelf icon appears, top-level, along with the
435 # applet
436 if (!defined ($tempinfo->{'Title'}) || $tempinfo->{'Title'} ne "Collage")
437 {
438 $contains_text .= "\".$next_subOID";
439 }
440
441 &print_classify_info ($infodb_type, $infodb_handle, $tempinfo, "$OID.$next_subOID",
442 $remove_empty_classifications);
443 $next_subOID++;
444 }
445 }
446
447 $classify_infodb{"contains"} = [ $contains_text ];
448 $classify_infodb{"mdtype"} = [ $classifyinfo->{'mdtype'} ]
449 if defined $classifyinfo->{'mdtype'};
450 $classify_infodb{"mdoffset"} = [ $mdoffset_text ]
451 if ($mdoffset_text !~ m/^;+$/);
452
453 &dbutil::write_infodb_entry($infodb_type, $infodb_handle, $OID, \%classify_infodb);
454}
455
456
457sub check_contents {
458 my ($classifyinfo,$remove_empty_classifications) = @_;
459 $remove_empty_classifications = 0 unless ($remove_empty_classifications);
460 my $num_leaf_docs = 0;
461 my $sub_num_leaf_docs = 0;
462
463 return $classifyinfo->{'numleafdocs'} if (defined $classifyinfo->{'numleafdocs'});
464
465 foreach my $content (@{$classifyinfo->{'contains'}}) {
466 if (defined $content->{'OID'}) {
467 # found a book
468 $num_leaf_docs ++;
469 } elsif (($sub_num_leaf_docs = &check_contents ($content,$remove_empty_classifications)) > 0) {
470 # there's a book somewhere below
471 $num_leaf_docs += $sub_num_leaf_docs;
472 } else {
473 if ($remove_empty_classifications){
474 # section contains no books so we want to remove
475 # it from its parents contents
476 $content = undef;
477 }
478 }
479 }
480
481 $classifyinfo->{'numleafdocs'} = $num_leaf_docs;
482 return $num_leaf_docs;
483}
484
4851;
Note: See TracBrowser for help on using the repository browser.