root/main/trunk/greenstone2/perllib/classify.pm @ 27675

Revision 27675, 14.5 KB (checked in by ak19, 6 years ago)

Tidying up my previous 2 commits on this file.

  • Property svn:keywords set to Author Date Id Revision
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    map { $_ = "\"$_\""; } @newoptions;
171    my $options .= join (",", @newoptions);
172
173
174    eval ("\$classobj = new \$classname([],[$options])");
175    die "$@" if $@;
176
177    $classobj->set_number($classify_number);
178    $classify_number ++;
179
180        # add this object to the list
181    push (@classify_objects, $classobj);
182    }
183
184    my ($classobj);
185    eval ("\$classobj = new AllList()");
186    die "$@" if $@;
187    push (@classify_objects, $classobj);
188
189    return \@classify_objects;
190}
191
192# init_classifiers resets all the classifiers and readys them to process
193# the documents.
194sub init_classifiers {
195    my ($classifiers) = @_;
196   
197    foreach my $classobj (@$classifiers) {
198    $classobj->init();
199    }
200}
201
202
203
204# takes a hashref containing the metadata for an infodb entry, and extracts
205# the childrens numbers (from the 'contains' entry).   
206# assumes format is ".1;".2;".3
207sub get_children { 
208    my ($doc_db_hash) = @_;
209
210    my $children = undef;
211
212    my $contains = $doc_db_hash->{'contains'};
213    if (defined ($contains)) {
214    $contains =~ s/\@$//;  #remove trailing @
215    $contains =~ s/^\"\.//; #remove initial ".
216    @$children = split /\;\"\./, $contains;
217    }
218
219    return $children;
220}
221
222   
223sub recurse_sections {
224    my ($doc_obj, $children, $parentoid, $parentsection, $database_recs) = @_;
225
226    return if (!defined $children);
227
228    foreach my $child (sort { $a <=> $b} @$children) {
229    $doc_obj->create_named_section("$parentsection.$child");
230    my $doc_db_rec = $database_recs->{"$parentoid.$child"};
231    my $doc_db_hash = db_rec_to_hash($doc_db_rec);
232
233    # get child's children
234    my $newchildren = &get_children($doc_db_hash);
235
236    # add content for current section
237    add_section_content($doc_obj, "$parentsection.$child", $doc_db_hash);
238
239    # process all the children if there are any
240    if (defined ($newchildren))
241    {
242        recurse_sections($doc_obj, $newchildren, "$parentoid.$child",
243                 "$parentsection.$child", $database_recs);
244    }
245    }
246}                       
247
248
249sub add_section_content {
250    my ($doc_obj, $cursection, $doc_db_hash) = @_;
251 
252    foreach my $key (keys %$doc_db_hash) {
253    #don't need to store these metadata
254    next if $key =~ /(thistype|childtype|contains|docnum|doctype|classifytype)/i;
255    # but do want things like hastxt and archivedir
256    my @items = split /@/, $doc_db_hash->{$key};
257    # metadata is all from the info database so should already be in utf8
258    map {$doc_obj->add_utf8_metadata ($cursection, $key, $_); } @items;
259
260    }
261}
262
263
264# gets all the metadata from an infodb entry, and puts it into a hashref
265sub db_rec_to_hash {
266   
267    my ($infodb_str_ref) = @_;
268
269    my $hashref = {};
270
271    my @entries = split(/\n/, $infodb_str_ref);
272    foreach my $entry (@entries) {
273    my($key, $value) = ($entry =~ /^<([^>]*)>(.*?)$/ );
274    $hashref->{$key} .= '@' if defined $hashref->{$key};
275    $hashref->{$key} .= $value;
276   
277    }
278   
279    return $hashref;
280}                     
281
282
283sub reconstruct_doc_objs_metadata
284{
285    my $infodb_type = shift(@_);
286    my $infodb_file_path = shift(@_);
287    my $database_recs = shift(@_);
288
289    # dig out top level doc sections
290    my %top_sections = ();
291    my %top_docnums = ();
292    foreach my $key ( keys %$database_recs )
293    {
294    my $md_rec = $database_recs->{$key};
295    my $md_hash = db_rec_to_hash($md_rec);
296
297    if ((defined $md_hash->{'doctype'}) && ($md_hash->{'doctype'} eq "doc")) {
298        next if ($key =~ m/\./);
299        $top_sections{$key} = $md_hash;
300        $top_docnums{$key} = $md_hash->{'docnum'};
301    }
302    }
303
304    # for greenstone document objects based on metadata in database file
305    my @all_docs = ();
306    # we need to make sure the documents were processed in the same order as
307    # before, so sort based on their docnums
308    foreach my $oid ( sort { $top_docnums{$a} <=> $top_docnums{$b} } keys %top_sections )
309    {
310    my $doc_db_hash = $top_sections{$oid};
311
312    my $doc_obj = new doc();
313    $doc_obj->set_OID($oid);
314    my $top = $doc_obj->get_top_section();
315        add_section_content ($doc_obj, $top, $doc_db_hash);
316        my $children = &get_children($doc_db_hash);
317        recurse_sections($doc_obj, $children, $oid, $top, $database_recs);
318
319    push(@all_docs,$doc_obj);
320    }   
321
322    return \@all_docs;   
323}
324
325
326
327
328
329# classify_doc lets each of the classifiers classify a document
330sub classify_doc {
331    my ($classifiers, $doc_obj) = @_;
332
333    foreach my $classobj (@$classifiers) {
334    my $title = $classobj->{'title'};
335
336    $classobj->classify($doc_obj);
337    }
338}
339
340
341our $next_classify_num = 1;
342
343# output_classify_info outputs all the info needed for the classification
344# to the database
345sub output_classify_info
346{
347    my ($classifiers, $infodb_type, $infodb_handle, $remove_empty_classifications, $gli) = @_;
348
349    $gli = 0 unless defined $gli;
350
351    # create a classification containing all the info
352    my $classifyinfo = { 'classifyOID'=> 'browse',
353             'contains' => [] };
354
355    # get each of the classifications
356    foreach my $classifier (@$classifiers)
357    {
358    my $classifier_info = $classifier->get_classify_info($gli);
359    $classifier_info->{'classifyOID'} = "CL$next_classify_num" unless defined($classifier_info->{'classifyOID'});
360        print STDERR "*** outputting information for classifier: $classifier_info->{'classifyOID'}\n";
361
362    push(@{$classifyinfo->{'contains'}}, $classifier_info);
363    $next_classify_num++;
364    }
365
366    &print_classify_info($infodb_type, $infodb_handle, $classifyinfo, "", $remove_empty_classifications);
367}
368
369
370sub print_classify_info
371{
372    my ($infodb_type, $infodb_handle, $classifyinfo, $OID, $remove_empty_classifications) = @_;
373
374    $OID =~ s/^\.+//; # just for good luck
375
376    # book information is printed elsewhere
377    return if (defined ($classifyinfo->{'OID'}));
378 
379    # don't want empty classifications
380    return if (&check_contents ($classifyinfo, $remove_empty_classifications) == 0 && $remove_empty_classifications);
381   
382    $OID = $classifyinfo->{'classifyOID'} if defined ($classifyinfo->{'classifyOID'});
383
384    my %classify_infodb = ();
385    $classify_infodb{"doctype"} = [ "classify" ];
386    $classify_infodb{"hastxt"} = [ "0" ];
387    $classify_infodb{"childtype"} = [ $classifyinfo->{'childtype'} ]
388    if defined $classifyinfo->{'childtype'};
389    $classify_infodb{"Title"} = [ $classifyinfo->{'Title'} ]
390    if defined $classifyinfo->{'Title'};
391    $classify_infodb{"numleafdocs"} = [ $classifyinfo->{'numleafdocs'} ]
392    if defined $classifyinfo->{'numleafdocs'};
393    $classify_infodb{"thistype"} = [ $classifyinfo->{'thistype'} ]
394    if defined $classifyinfo->{'thistype'};
395    $classify_infodb{"parameters"} = [ $classifyinfo->{'parameters'} ]
396    if defined $classifyinfo->{'parameters'};
397    $classify_infodb{"supportsmemberof"} = [ $classifyinfo->{'supportsmemberof'} ]
398    if defined $classifyinfo->{'supportsmemberof'};
399   
400    my $contains_text = "";
401    my $mdoffset_text = "";
402   
403    my $next_subOID = 1;
404    my $first = 1;
405    foreach my $tempinfo (@{$classifyinfo->{'contains'}}) {
406    # empty contents were made undefined by clean_contents()
407    next unless defined $tempinfo;
408    if (!defined ($tempinfo->{'classifyOID'}) ||
409        $tempinfo->{'classifyOID'} ne "oai") {
410        $contains_text .= ";" unless $first;
411    }
412    $mdoffset_text .= ";" unless $first;
413    $first = 0;
414   
415    if (defined ($tempinfo->{'classifyOID'}))
416    {
417        if ($tempinfo->{'classifyOID'} ne "oai")
418        {
419        $contains_text .= $tempinfo->{'classifyOID'};
420        }
421
422        &print_classify_info ($infodb_type, $infodb_handle, $tempinfo, $tempinfo->{'classifyOID'},
423                  $remove_empty_classifications);
424    }
425    elsif (defined ($tempinfo->{'OID'}))
426    {
427        $contains_text .= $tempinfo->{'OID'};
428        $mdoffset_text .= $tempinfo->{'offset'} if (defined ($tempinfo->{'offset'}));
429    }
430    else
431    {
432        # Supress having top-level node in Collage classifier
433        # so no bookshelf icon appears, top-level, along with the
434        # applet
435        if (!defined ($tempinfo->{'Title'}) || $tempinfo->{'Title'} ne "Collage")
436        {
437        $contains_text .= "\".$next_subOID";
438        }
439
440        &print_classify_info ($infodb_type, $infodb_handle, $tempinfo, "$OID.$next_subOID",
441                  $remove_empty_classifications);
442        $next_subOID++;
443    }
444    }
445   
446    $classify_infodb{"contains"} = [ $contains_text ];
447    $classify_infodb{"mdtype"} = [ $classifyinfo->{'mdtype'} ]
448    if defined $classifyinfo->{'mdtype'};
449    $classify_infodb{"mdoffset"} = [ $mdoffset_text ]
450    if ($mdoffset_text !~ m/^;+$/);
451   
452    &dbutil::write_infodb_entry($infodb_type, $infodb_handle, $OID, \%classify_infodb);
453}
454
455
456sub check_contents {
457    my ($classifyinfo,$remove_empty_classifications) = @_;
458    $remove_empty_classifications = 0 unless ($remove_empty_classifications);
459    my $num_leaf_docs = 0;
460    my $sub_num_leaf_docs = 0;
461
462    return $classifyinfo->{'numleafdocs'} if (defined $classifyinfo->{'numleafdocs'});
463
464    foreach my $content (@{$classifyinfo->{'contains'}}) {
465    if (defined $content->{'OID'}) {
466        # found a book
467        $num_leaf_docs ++;
468    } elsif (($sub_num_leaf_docs = &check_contents ($content,$remove_empty_classifications)) > 0) {
469        # there's a book somewhere below
470        $num_leaf_docs += $sub_num_leaf_docs;
471    } else {
472        if ($remove_empty_classifications){
473        # section contains no books so we want to remove
474        # it from its parents contents
475        $content = undef;
476        }
477    }
478    }
479
480    $classifyinfo->{'numleafdocs'} = $num_leaf_docs;
481    return $num_leaf_docs;
482}
483
4841;
Note: See TracBrowser for help on using the browser.