source: gsdl/trunk/perllib/classify.pm@ 15705

Last change on this file since 15705 was 15705, checked in by mdewsnip, 16 years ago

Merged the one lonely function in unbuildutil into dbutil, as it is GDBM-specific.
  • Property svn:keywords set to Author Date Id Revision
File size: 15.1 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;
31require AllList;
32
33use dbutil;
34use gsprintf;
35use strict; no strict 'subs';
36
37
38sub gsprintf
39{
40 return &gsprintf::gsprintf(@_);
41}
42
43
44my $oid_to_clids = {};
45
46sub load_classifier_for_info {
47 my ($classifier) = shift @_;
48
49 # find the classifier
50 my $customclassname;
51 if (defined($ENV{'GSDLCOLLECTION'}))
52 {
53 $customclassname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "custom", $ENV{'GSDLCOLLECTION'},
54 "perllib", "classify", "${classifier}.pm");
55 }
56 my $colclassname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "perllib", "classify", "${classifier}.pm");
57 my $mainclassname = &util::filename_cat($ENV{'GSDLHOME'}, "perllib", "classify", "${classifier}.pm");
58
59 if (defined($customclassname) && -e $customclassname) { require $customclassname; }
60 elsif (-e $colclassname) { require $colclassname; }
61 elsif (-e $mainclassname) { require $mainclassname; }
62 else {
63 &gsprintf(STDERR, "{classify.could_not_find_classifier}\n", $classifier) && die "\n";
64 }
65 my ($classobj);
66 my $options = "-gsdlinfo";
67 eval ("\$classobj = new \$classifier([],[$options])");
68 die "$@" if $@;
69
70 return $classobj;
71}
72
73sub load_classifiers {
74 my ($classify_list, $build_dir, $outhandle) = @_;
75 my @classify_objects = ();
76 my $classify_number = 1;
77
78 my $colclassdir = &util::filename_cat($ENV{'GSDLCOLLECTDIR'},"perllib/classify");
79 unshift (@INC, $colclassdir);
80
81 foreach my $classifyoption (@$classify_list) {
82
83 # get the classifier name
84 my $classname = shift @$classifyoption;
85 next unless defined $classname;
86
87 # find the classifier
88 my $customclassname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "custom", $ENV{'GSDLCOLLECTION'},
89 "perllib", "classify", "${classname}.pm");
90 my $colclassname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "perllib", "classify", "${classname}.pm");
91 my $mainclassname = &util::filename_cat($ENV{'GSDLHOME'}, "perllib", "classify", "${classname}.pm");
92
93 if (-e $customclassname) { require $customclassname; }
94 elsif (-e $colclassname) { require $colclassname; }
95 elsif (-e $mainclassname) { require $mainclassname; }
96 else { &gsprintf(STDERR, "{classify.could_not_find_classifier}\n", $classname) && die "\n";
97 # die "ERROR - couldn't find classifier \"$classname\"\n";
98 }
99
100 # create the classify object
101 my ($classobj);
102
103 my @newoptions;
104
105 # do these first so they can be overriden by user supplied options
106 push @newoptions, "-builddir", "$build_dir" if ($build_dir);
107 push @newoptions, "-outhandle", "$outhandle" if ($outhandle);
108 push @newoptions, "-verbosity", "2";
109
110 # backwards compatability hack: if the classifier options are
111 # in "x=y" format, convert them to parsearg ("-x y") format.
112 my ($opt, $key, $value);
113 foreach $opt (@$classifyoption) {
114 # if ($opt =~ /^(\w+)=(.*)$/) {
115 # push @newoptions, "-$1", $2;
116 # } else {
117 push @newoptions, $opt;
118 #}
119 }
120
121 map { $_ = "\"$_\""; } @newoptions;
122 my $options .= join (",", @newoptions);
123
124
125 eval ("\$classobj = new \$classname([],[$options])");
126 die "$@" if $@;
127
128 $classobj->set_number($classify_number);
129 $classify_number ++;
130
131 # add this object to the list
132 push (@classify_objects, $classobj);
133 }
134
135 my ($classobj);
136 eval ("\$classobj = new AllList()");
137 die "$@" if $@;
138 push (@classify_objects, $classobj);
139
140 return \@classify_objects;
141}
142
143# init_classifiers resets all the classifiers and readys them to process
144# the documents.
145sub init_classifiers {
146 my ($classifiers) = @_;
147
148 foreach my $classobj (@$classifiers) {
149 $classobj->init();
150 }
151}
152
153
154
155# takes a hashref containing the metadata for an infodb entry, and extracts
156# the childrens numbers (from the 'contains' entry).
157# assumes format is ".1;".2;".3
158sub get_children {
159 my ($doc_db_hash) = @_;
160
161 my $children = undef;
162
163 my $contains = $doc_db_hash->{'contains'};
164 if (defined ($contains)) {
165 $contains =~ s/\@$//; #remove trailing @
166 $contains =~ s/^\"\.//; #remove initial ".
167 @$children = split /\;\"\./, $contains;
168 }
169
170 return $children;
171}
172
173
174sub recurse_sections {
175 my ($doc_obj, $children, $parentoid, $parentsection, $database_recs) = @_;
176
177 return if (!defined $children);
178
179 foreach my $child (sort { $a <=> $b} @$children) {
180 $doc_obj->create_named_section("$parentsection.$child");
181 my $doc_db_rec = $database_recs->{"$parentoid.$child"};
182 my $doc_db_hash = db_rec_to_hash($doc_db_rec);
183
184 # get child's children
185 my $newchildren = &get_children($doc_db_hash);
186
187 # add content for current section
188 add_section_content($doc_obj, "$parentsection.$child", $doc_db_hash);
189
190 # process all the children if there are any
191 if (defined ($newchildren))
192 {
193 recurse_sections($doc_obj, $newchildren, "$parentoid.$child",
194 "$parentsection.$child", $database_recs);
195 }
196 }
197}
198
199
200sub add_section_content {
201 my ($doc_obj, $cursection, $doc_db_hash) = @_;
202
203 foreach my $key (keys %$doc_db_hash) {
204 #don't need to store these metadata
205 next if $key =~ /(thistype|childtype|contains|docnum|doctype|classifytype)/i;
206 # but do want things like hastxt and archivedir
207 my @items = split /@/, $doc_db_hash->{$key};
208 map {$doc_obj->add_metadata ($cursection, $key, $_); } @items;
209
210 }
211}
212
213
214# gets all the metadata from an infodb entry, and puts it into a hashref
215sub db_rec_to_hash {
216
217 my ($gdb_str_ref) = @_;
218
219 my $hashref = {};
220
221 my @entries = split(/\n/, $gdb_str_ref);
222 foreach my $entry (@entries) {
223 my($key, $value) = ($entry =~ /^<([^>]*)>(.*?)$/ );
224 $hashref->{$key} .= '@' if defined $hashref->{$key};
225 $hashref->{$key} .= $value;
226
227 }
228
229 return $hashref;
230}
231
232
233sub reconstruct_doc_objs_metadata
234{
235 my ($infodb_file_path) = @_;
236
237 my %database_recs;
238 &dbutil::read_infodb_file($infodb_file_path, \%database_recs);
239
240 # dig out top level doc sections
241 my %top_sections = ();
242 my %top_docnums = ();
243 foreach my $key ( keys %database_recs )
244 {
245 my $md_rec = $database_recs{$key};
246 my $md_hash = db_rec_to_hash($md_rec);
247
248 if ((defined $md_hash->{'doctype'}) && ($md_hash->{'doctype'} eq "doc")) {
249 next if ($key =~ m/\./);
250 $top_sections{$key} = $md_hash;
251 $top_docnums{$key} = $md_hash->{'docnum'};
252 }
253 }
254
255 # for greenstone document objects based on metadata in database file
256 my @all_docs = ();
257 # we need to make sure the documents were processed in the same order as
258 # before, so sort based on their docnums
259 foreach my $oid ( sort { $top_docnums{$a} <=> $top_docnums{$b} } keys %top_sections )
260 {
261 my $doc_db_hash = $top_sections{$oid};
262
263 my $doc_obj = new doc();
264 $doc_obj->set_OID($oid);
265 my $top = $doc_obj->get_top_section();
266 add_section_content ($doc_obj, $top, $doc_db_hash);
267 my $children = &get_children($doc_db_hash);
268 recurse_sections($doc_obj, $children, $oid, $top, \%database_recs);
269
270 push(@all_docs,$doc_obj);
271 }
272
273 return \@all_docs;
274}
275
276
277
278
279
280# classify_doc lets each of the classifiers classify a document
281sub classify_doc {
282 my ($classifiers, $doc_obj) = @_;
283
284 foreach my $classobj (@$classifiers) {
285 my $title = $classobj->{'title'};
286 $classobj->classify($doc_obj);
287 }
288}
289
290
291# output_classify_info outputs all the info needed for the classification
292# to the database
293sub output_classify_info
294{
295 my ($classifiers, $infodb_handle, $remove_empty_classifications, $gli) = @_;
296
297 $gli = 0 unless defined $gli;
298
299 # create a classification containing all the info
300 my $classifyinfo = { 'classifyOID'=> 'browse',
301 'contains' => [] };
302
303 # get each of the classifications
304 my $next_classify_num = 1;
305 foreach my $classifier (@$classifiers)
306 {
307 my $classifier_info = $classifier->get_classify_info($gli);
308 $classifier_info->{'classifyOID'} = "CL$next_classify_num" unless defined($classifier_info->{'classifyOID'});
309 print STDERR "*** outputting information for classifier: $classifier_info->{'classifyOID'}\n";
310
311 push(@{$classifyinfo->{'contains'}}, $classifier_info);
312 $next_classify_num++;
313 }
314
315 &print_classify_info ($infodb_handle, $classifyinfo, "", $remove_empty_classifications);
316}
317
318
319sub print_classify_info
320{
321 my ($infodb_handle, $classifyinfo, $OID, $remove_empty_classifications) = @_;
322
323 $OID =~ s/^\.+//; # just for good luck
324
325 # book information is printed elsewhere
326 return if (defined ($classifyinfo->{'OID'}));
327
328 # don't want empty classifications
329 return if (&check_contents ($classifyinfo, $remove_empty_classifications) == 0 && $remove_empty_classifications);
330
331 $OID = $classifyinfo->{'classifyOID'} if defined ($classifyinfo->{'classifyOID'});
332
333 my %classify_infodb = ();
334 $classify_infodb{"doctype"} = [ "classify" ];
335 $classify_infodb{"hastxt"} = [ "0" ];
336 $classify_infodb{"childtype"} = [ $classifyinfo->{'childtype'} ]
337 if defined $classifyinfo->{'childtype'};
338 $classify_infodb{"Title"} = [ $classifyinfo->{'Title'} ]
339 if defined $classifyinfo->{'Title'};
340 $classify_infodb{"numleafdocs"} = [ $classifyinfo->{'numleafdocs'} ]
341 if defined $classifyinfo->{'numleafdocs'};
342 $classify_infodb{"thistype"} = [ $classifyinfo->{'thistype'} ]
343 if defined $classifyinfo->{'thistype'};
344 $classify_infodb{"parameters"} = [ $classifyinfo->{'parameters'} ]
345 if defined $classifyinfo->{'parameters'};
346 $classify_infodb{"supportsmemberof"} = [ $classifyinfo->{'supportsmemberof'} ]
347 if defined $classifyinfo->{'supportsmemberof'};
348
349 my $contains_text = "";
350 my $mdoffset_text = "";
351
352 my $next_subOID = 1;
353 my $first = 1;
354 foreach my $tempinfo (@{$classifyinfo->{'contains'}}) {
355 # empty contents were made undefined by clean_contents()
356 next unless defined $tempinfo;
357
358 if (!defined ($tempinfo->{'classifyOID'}) ||
359 $tempinfo->{'classifyOID'} ne "oai") {
360 $contains_text .= ";" unless $first;
361 }
362 $mdoffset_text .= ";" unless $first;
363 $first = 0;
364
365 if (defined ($tempinfo->{'classifyOID'})) {
366 if ($tempinfo->{'classifyOID'} ne "oai") {
367 $contains_text .= $tempinfo->{'classifyOID'};
368 }
369
370 # Extra code for incremental building.
371 # We need to store a listing of the classifiers each DOI is in
372 my $clids = [];
373 #rint STDERR "==1. Recording reverse lookup for " . $tempinfo->{'classifyOID'} . "==\n";
374 if(defined($oid_to_clids->{$tempinfo->{'classifyOID'}})) {
375 #rint STDERR "Found existing array!\n";
376 $clids = $oid_to_clids->{$tempinfo->{'classifyOID'}};
377 }
378 #rint STDERR "Appended $OID to \"" . join(";", @{$clids}) . "\"\n";
379 push(@{$clids}, $OID);
380 $oid_to_clids->{$tempinfo->{'classifyOID'}} = $clids;
381 #rint STDERR "Result: \"" . join(";", @{$clids}) . "\"\n";
382
383 &print_classify_info ($infodb_handle, $tempinfo, $tempinfo->{'classifyOID'},
384 $remove_empty_classifications);
385 } elsif (defined ($tempinfo->{'OID'})) {
386 $contains_text .= $tempinfo->{'OID'};
387 $mdoffset_text .= $tempinfo->{'offset'} if (defined ($tempinfo->{'offset'}));
388
389
390 # note: we don't want to print the contents of the books
391 # Extra code for incremental building.
392 # We need to store a listing of the classifiers each DOI is in
393 my $clids = [];
394 #rint STDERR "==2. Recording reverse lookup for " . $tempinfo->{'OID'} . "==\n";
395 if(defined($oid_to_clids->{$tempinfo->{'OID'}})) {
396 #rint STDERR "Found existing array!\n";
397 $clids = $oid_to_clids->{$tempinfo->{'OID'}};
398 }
399 #rint STDERR "Appended $OID to \"" . join(";", @{$clids}) . "\"\n";
400 push(@{$clids}, $OID);
401 $oid_to_clids->{$tempinfo->{'OID'}} = $clids;
402 #rint STDERR "Result: \"" . join(";", @{$clids}) . "\"\n";
403
404
405 } else {
406
407 # Supress having top-level node in Collage classifier
408 # so no bookshelf icon appears, top-level, along with the
409 # applet
410
411 if (!defined ($tempinfo->{'Title'}) || $tempinfo->{'Title'} ne "Collage") {
412 $contains_text .= "\".$next_subOID";
413 }
414
415 # Extra code for incremental building.
416 # We need to store a listing of the classifiers each DOI is in
417 my $clids = [];
418 #rint STDERR "==3. Recording reverse lookup for $OID.$next_subOID==\n";
419 if(defined($oid_to_clids->{$OID . "." . $next_subOID})) {
420 #rint STDERR "Found existing array!\n";
421 $clids = $oid_to_clids->{$OID . "." . $next_subOID};
422 }
423 #rint STDERR "Appended $OID to \"" . join(";", @{$clids}) . "\"\n";
424 push(@{$clids}, $OID);
425 $oid_to_clids->{$OID . "." . $next_subOID} = $clids;
426 #rint STDERR "Result: \"" . join(";", @{$clids}) . "\"\n";
427
428 &print_classify_info ($infodb_handle, $tempinfo, "$OID.$next_subOID",
429 $remove_empty_classifications);
430 $next_subOID++;
431 }
432 }
433
434 $classify_infodb{"contains"} = [ $contains_text ];
435 $classify_infodb{"mdtype"} = [ $classifyinfo->{'mdtype'} ]
436 if defined $classifyinfo->{'mdtype'};
437 $classify_infodb{"mdoffset"} = [ $mdoffset_text ]
438 if ($mdoffset_text !~ m/^;+$/);
439
440 &dbutil::write_infodb_entry($infodb_handle, $OID, \%classify_infodb);
441}
442
443
444sub check_contents {
445 my ($classifyinfo,$remove_empty_classifications) = @_;
446 $remove_empty_classifications = 0 unless ($remove_empty_classifications);
447 my $num_leaf_docs = 0;
448 my $sub_num_leaf_docs = 0;
449
450 return $classifyinfo->{'numleafdocs'} if (defined $classifyinfo->{'numleafdocs'});
451
452 foreach my $content (@{$classifyinfo->{'contains'}}) {
453 if (defined $content->{'OID'}) {
454 # found a book
455 $num_leaf_docs ++;
456 } elsif (($sub_num_leaf_docs = &check_contents ($content,$remove_empty_classifications)) > 0) {
457 # there's a book somewhere below
458 $num_leaf_docs += $sub_num_leaf_docs;
459 } else {
460 if ($remove_empty_classifications){
461 # section contains no books so we want to remove
462 # it from its parents contents
463 $content = undef;
464 }
465 }
466 }
467
468 $classifyinfo->{'numleafdocs'} = $num_leaf_docs;
469 return $num_leaf_docs;
470}
471
4721;
Note: See TracBrowser for help on using the repository browser.