source: gsdl/trunk/perllib/classify.pm@ 14926

Last change on this file since 14926 was 14239, checked in by mdewsnip, 17 years ago

Fixed problem with loading custom plugins/classifiers when running pluginfo.pl -listall or classinfo.pl -listall.
  • Property svn:keywords set to Author Date Id Revision
File size: 15.0 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;
32use gsprintf;
33
34#use GDBM_File;
35use unbuildutil;
36
37
38sub gsprintf
39{
40 return &gsprintf::gsprintf(@_);
41}
42
43
44$next_classify_num = 1;
45$oid_to_clids = {};
46
47sub load_classifier_for_info {
48 my ($classifier) = shift @_;
49
50 # find the classifier
51 my $customclassname;
52 if (defined($ENV{'GSDLCOLLECTION'}))
53 {
54 $customclassname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "custom", $ENV{'GSDLCOLLECTION'},
55 "perllib", "classify", "${classifier}.pm");
56 }
57 my $colclassname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "perllib", "classify", "${classifier}.pm");
58 my $mainclassname = &util::filename_cat($ENV{'GSDLHOME'}, "perllib", "classify", "${classifier}.pm");
59
60 if (defined($customclassname) && -e $customclassname) { require $customclassname; }
61 elsif (-e $colclassname) { require $colclassname; }
62 elsif (-e $mainclassname) { require $mainclassname; }
63 else {
64 &gsprintf(STDERR, "{classify.could_not_find_classifier}\n", $classifier) && die "\n";
65 }
66 my ($classobj);
67 my $options = "-gsdlinfo";
68 eval ("\$classobj = new \$classifier([],[$options])");
69 die "$@" if $@;
70
71 return $classobj;
72}
73
74sub load_classifiers {
75 my ($classify_list, $build_dir, $outhandle) = @_;
76 my @classify_objects = ();
77 my $classify_number = 1;
78
79 my $colclassdir = &util::filename_cat($ENV{'GSDLCOLLECTDIR'},"perllib/classify");
80 unshift (@INC, $colclassdir);
81
82 foreach $classifyoption (@$classify_list) {
83
84 # get the classifier name
85 my $classname = shift @$classifyoption;
86 next unless defined $classname;
87
88 # find the classifier
89 my $customclassname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "custom", $ENV{'GSDLCOLLECTION'},
90 "perllib", "classify", "${classname}.pm");
91 my $colclassname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "perllib", "classify", "${classname}.pm");
92 my $mainclassname = &util::filename_cat($ENV{'GSDLHOME'}, "perllib", "classify", "${classname}.pm");
93
94 if (-e $customclassname) { require $customclassname; }
95 elsif (-e $colclassname) { require $colclassname; }
96 elsif (-e $mainclassname) { require $mainclassname; }
97 else { &gsprintf(STDERR, "{classify.could_not_find_classifier}\n", $classname) && die "\n";
98 # die "ERROR - couldn't find classifier \"$classname\"\n";
99 }
100
101 # create the classify object
102 my ($classobj);
103
104 my @newoptions;
105
106 # do these first so they can be overriden by user supplied options
107 push @newoptions, "-builddir", "$build_dir" if ($build_dir);
108 push @newoptions, "-outhandle", "$outhandle" if ($outhandle);
109 push @newoptions, "-verbosity", "2";
110
111 # backwards compatability hack: if the classifier options are
112 # in "x=y" format, convert them to parsearg ("-x y") format.
113 my ($opt, $key, $value);
114 foreach $opt (@$classifyoption) {
115 # if ($opt =~ /^(\w+)=(.*)$/) {
116 # push @newoptions, "-$1", $2;
117 # } else {
118 push @newoptions, $opt;
119 #}
120 }
121
122 map { $_ = "\"$_\""; } @newoptions;
123 my $options .= join (",", @newoptions);
124
125
126 eval ("\$classobj = new \$classname([],[$options])");
127 die "$@" if $@;
128
129 $classobj->set_number($classify_number);
130 $classify_number ++;
131
132 # add this object to the list
133 push (@classify_objects, $classobj);
134 }
135
136 my ($classobj);
137 eval ("\$classobj = new AllList()");
138 die "$@" if $@;
139 push (@classify_objects, $classobj);
140
141 return \@classify_objects;
142}
143
144# init_classifiers resets all the classifiers and readys them to process
145# the documents.
146sub init_classifiers {
147 my ($classifiers) = @_;
148
149 foreach $classobj (@$classifiers) {
150 $classobj->init();
151 }
152}
153
154
155
156# takes a hashref containing the metadata for a gdbmfile entry, and extracts
157# the childrens numbers (from the 'contains' entry).
158# assumes format is ".1;".2;".3
159sub get_children {
160 my ($doc_db_hash) = @_;
161
162 my $children = undef;
163
164 $childs = $doc_db_hash->{'contains'};
165 if (defined ($childs)) {
166 $childs =~ s/\@$//; #remove trailing @
167 $childs =~ s/^\"\.//; #remove initial ".
168 @$children = split /\;\"\./, $childs;
169
170 }
171
172 return $children;
173}
174
175
176sub recurse_sections {
177 my ($doc_obj, $children, $parentoid, $parentsection, $gdbm_recs) = @_;
178
179 return if (!defined $children);
180
181 foreach my $child (sort { $a <=> $b} @$children) {
182 $doc_obj->create_named_section("$parentsection.$child");
183 my $doc_db_rec = $gdbm_recs->{"$parentoid.$child"};
184 my $doc_db_hash = db_rec_to_hash($doc_db_rec);
185
186 # get child's children
187 my $newchildren = &get_children($doc_db_hash);
188
189 # add content for current section
190 add_section_content($doc_obj, "$parentsection.$child", $doc_db_hash);
191
192 # process all the children if there are any
193 if (defined ($newchildren))
194 {
195 recurse_sections($doc_obj, $newchildren, "$parentoid.$child",
196 "$parentsection.$child", $gdbm_recs);
197 }
198 }
199}
200
201
202sub add_section_content {
203 my ($doc_obj, $cursection, $doc_db_hash) = @_;
204
205 foreach $key (keys %$doc_db_hash) {
206 #don't need to store these metadata
207 next if $key =~ /(thistype|childtype|contains|docnum|doctype|classifytype)/i;
208 # but do want things like hastxt and archivedir
209 my @items = split /@/, $doc_db_hash->{$key};
210 map {$doc_obj->add_metadata ($cursection, $key, $_); } @items;
211
212 }
213}
214
215
216# gets all the metadata from a gdbm file entry, and puts it into a hashref
217sub db_rec_to_hash {
218
219 my ($gdb_str_ref) = @_;
220
221 my $hashref = {};
222
223 my @entries = split(/\n/, $gdb_str_ref);
224 foreach $entry (@entries) {
225 my($key, $value) = ($entry =~ /^<([^>]*)>(.*?)$/ );
226 $hashref->{$key} .= '@' if defined $hashref->{$key};
227 $hashref->{$key} .= $value;
228
229 }
230
231 return $hashref;
232}
233
234
235sub reconstruct_doc_objs_metadata
236{
237 my ($fulldbname) = @_;
238
239# tie %gdbm_recs, 'GDBM_File', $fulldbname, &GDBM_WRCREAT, 0640;
240
241 my %gdbm_recs;
242 &unbuildutil::read_gdbm($fulldbname,\%gdbm_recs);
243
244
245 # dig out top level doc sections
246 my %top_sections = ();
247 my %top_docnums = ();
248 foreach my $key ( keys %gdbm_recs )
249 {
250 my $md_rec = $gdbm_recs{$key};
251 my $md_hash = db_rec_to_hash($md_rec);
252
253 if ((defined $md_hash->{'doctype'}) && ($md_hash->{'doctype'} eq "doc")) {
254 next if ($key =~ m/\./);
255 $top_sections{$key} = $md_hash;
256 $top_docnums{$key} = $md_hash->{'docnum'};
257 }
258 }
259
260 # for greenstone document objects based on metadata in gdbm file
261 my @all_docs = ();
262 # we need to make sure the documents were processed in the same order as
263 # before, so sort based on their docnums
264 foreach my $oid ( sort { $top_docnums{$a} <=> $top_docnums{$b} } keys %top_sections )
265 {
266 my $doc_db_hash = $top_sections{$oid};
267
268 my $doc_obj = new doc();
269 $doc_obj->set_OID($oid);
270 my $top = $doc_obj->get_top_section();
271 add_section_content ($doc_obj, $top, $doc_db_hash);
272 my $children = &get_children($doc_db_hash);
273 recurse_sections($doc_obj, $children, $oid, $top, \%gdbm_recs);
274
275 push(@all_docs,$doc_obj);
276 }
277
278# untie %gdbm_recs;
279
280 return \@all_docs;
281}
282
283
284
285
286
287# classify_doc lets each of the classifiers classify a document
288sub classify_doc {
289 my ($classifiers, $doc_obj) = @_;
290
291 foreach $classobj (@$classifiers) {
292 my $title = $classobj->{'title'};
293 $classobj->classify($doc_obj);
294 }
295}
296
297# output_classify_info outputs all the info needed for the classification
298# to the gdbm
299sub output_classify_info {
300 my ($classifiers, $handle, $remove_empty_classifications, $gli) = @_;
301# $handle = "main::STDOUT";
302
303 $gli = 0 unless defined $gli;
304
305 # create a classification containing all the info
306 my $classifyinfo = {'classifyOID'=>'browse',
307 'contains'=>[]};
308
309 # get each of the classifications
310 foreach $classobj (@$classifiers) {
311 my $tempinfo = $classobj->get_classify_info($gli);
312 my $classID = $tempinfo->{'classifyOID'};
313
314 $tempinfo->{'classifyOID'} = "CL$next_classify_num" unless defined($tempinfo->{'classifyOID'});
315 $next_classify_num++;
316
317 print STDERR "*** outputting information for classifier: $tempinfo->{'classifyOID'}\n";
318
319 push (@{$classifyinfo->{'contains'}}, $tempinfo);
320 }
321
322 &print_classify_info ($handle, $classifyinfo, "", $remove_empty_classifications);
323}
324
325sub print_classify_info {
326 my ($handle, $classifyinfo, $OID, $remove_empty_classifications) = @_;
327
328 $OID =~ s/^\.+//; # just for good luck
329
330 # book information is printed elsewhere
331 return if (defined ($classifyinfo->{'OID'}));
332
333 # don't want empty classifications
334 return if (&check_contents ($classifyinfo, $remove_empty_classifications) == 0 && $remove_empty_classifications);
335
336 $OID = $classifyinfo->{'classifyOID'} if defined ($classifyinfo->{'classifyOID'});
337
338 my $outputtext = "[$OID]\n";
339 $outputtext .= "<doctype>classify\n";
340 $outputtext .= "<hastxt>0\n";
341 $outputtext .= "<childtype>$classifyinfo->{'childtype'}\n"
342 if defined $classifyinfo->{'childtype'};
343 $outputtext .= "<Title>$classifyinfo->{'Title'}\n"
344 if defined $classifyinfo->{'Title'};
345 $outputtext .= "<numleafdocs>$classifyinfo->{'numleafdocs'}\n"
346 if defined $classifyinfo->{'numleafdocs'};
347 $outputtext .= "<thistype>$classifyinfo->{'thistype'}\n"
348 if defined $classifyinfo->{'thistype'};
349 $outputtext .= "<parameters>$classifyinfo->{'parameters'}\n"
350 if defined $classifyinfo->{'parameters'};
351 $outputtext .= "<supportsmemberof>$classifyinfo->{'supportsmemberof'}\n"
352 if defined $classifyinfo->{'supportsmemberof'};
353
354 my $contains_text = "<contains>";
355 my $mdoffset_text = "<mdoffset>";
356
357 my $next_subOID = 1;
358 my $first = 1;
359 foreach $tempinfo (@{$classifyinfo->{'contains'}}) {
360 # empty contents were made undefined by clean_contents()
361 next unless defined $tempinfo;
362
363 if (!defined ($tempinfo->{'classifyOID'}) ||
364 $tempinfo->{'classifyOID'} ne "oai") {
365 $contains_text .= ";" unless $first;
366 }
367 $mdoffset_text .= ";" unless $first;
368 $first = 0;
369
370 if (defined ($tempinfo->{'classifyOID'})) {
371 if ($tempinfo->{'classifyOID'} ne "oai") {
372 $contains_text .= $tempinfo->{'classifyOID'};
373 }
374
375 # Extra code for incremental building.
376 # We need to store a listing of the classifiers each DOI is in
377 my $clids = [];
378 #rint STDERR "==1. Recording reverse lookup for " . $tempinfo->{'classifyOID'} . "==\n";
379 if(defined($oid_to_clids->{$tempinfo->{'classifyOID'}})) {
380 #rint STDERR "Found existing array!\n";
381 $clids = $oid_to_clids->{$tempinfo->{'classifyOID'}};
382 }
383 #rint STDERR "Appended $OID to \"" . join(";", @{$clids}) . "\"\n";
384 push(@{$clids}, $OID);
385 $oid_to_clids->{$tempinfo->{'classifyOID'}} = $clids;
386 #rint STDERR "Result: \"" . join(";", @{$clids}) . "\"\n";
387
388 &print_classify_info ($handle, $tempinfo, $tempinfo->{'classifyOID'},
389 $remove_empty_classifications);
390 } elsif (defined ($tempinfo->{'OID'})) {
391 $contains_text .= $tempinfo->{'OID'};
392 $mdoffset_text .= $tempinfo->{'offset'} if (defined ($tempinfo->{'offset'}));
393
394
395 # note: we don't want to print the contents of the books
396 # Extra code for incremental building.
397 # We need to store a listing of the classifiers each DOI is in
398 my $clids = [];
399 #rint STDERR "==2. Recording reverse lookup for " . $tempinfo->{'OID'} . "==\n";
400 if(defined($oid_to_clids->{$tempinfo->{'OID'}})) {
401 #rint STDERR "Found existing array!\n";
402 $clids = $oid_to_clids->{$tempinfo->{'OID'}};
403 }
404 #rint STDERR "Appended $OID to \"" . join(";", @{$clids}) . "\"\n";
405 push(@{$clids}, $OID);
406 $oid_to_clids->{$tempinfo->{'OID'}} = $clids;
407 #rint STDERR "Result: \"" . join(";", @{$clids}) . "\"\n";
408
409
410 } else {
411
412 # Supress having top-level node in Collage classifier
413 # so no bookshelf icon appears, top-level, along with the
414 # applet
415
416 if (!defined ($tempinfo->{'Title'}) || $tempinfo->{'Title'} ne "Collage") {
417 $contains_text .= "\".$next_subOID";
418 }
419
420 # Extra code for incremental building.
421 # We need to store a listing of the classifiers each DOI is in
422 my $clids = [];
423 #rint STDERR "==3. Recording reverse lookup for $OID.$next_subOID==\n";
424 if(defined($oid_to_clids->{$OID . "." . $next_subOID})) {
425 #rint STDERR "Found existing array!\n";
426 $clids = $oid_to_clids->{$OID . "." . $next_subOID};
427 }
428 #rint STDERR "Appended $OID to \"" . join(";", @{$clids}) . "\"\n";
429 push(@{$clids}, $OID);
430 $oid_to_clids->{$OID . "." . $next_subOID} = $clids;
431 #rint STDERR "Result: \"" . join(";", @{$clids}) . "\"\n";
432
433 &print_classify_info ($handle, $tempinfo, "$OID.$next_subOID",
434 $remove_empty_classifications);
435 $next_subOID++;
436 }
437 }
438
439 $outputtext .= "$contains_text\n";
440 $outputtext .= "<mdtype>$classifyinfo->{'mdtype'}\n"
441 if defined $classifyinfo->{'mdtype'};
442 $outputtext .= "$mdoffset_text\n"
443 if ($mdoffset_text !~ m/^<mdoffset>;+$/);
444
445 $outputtext .= '-' x 70 . "\n";
446
447 print $handle $outputtext;
448
449}
450
451sub check_contents {
452 my ($classifyinfo,$remove_empty_classifications) = @_;
453 $remove_empty_classifications = 0 unless ($remove_empty_classifications);
454 my $num_leaf_docs = 0;
455 my $sub_num_leaf_docs = 0;
456
457 return $classifyinfo->{'numleafdocs'} if (defined $classifyinfo->{'numleafdocs'});
458
459 foreach $content (@{$classifyinfo->{'contains'}}) {
460 if (defined $content->{'OID'}) {
461 # found a book
462 $num_leaf_docs ++;
463 } elsif (($sub_num_leaf_docs = &check_contents ($content,$remove_empty_classifications)) > 0) {
464 # there's a book somewhere below
465 $num_leaf_docs += $sub_num_leaf_docs;
466 } else {
467 if ($remove_empty_classifications){
468 # section contains no books so we want to remove
469 # it from its parents contents
470 $content = undef;
471 }
472 }
473 }
474
475 $classifyinfo->{'numleafdocs'} = $num_leaf_docs;
476 return $num_leaf_docs;
477}
478
4791;
Note: See TracBrowser for help on using the repository browser.