| 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 |
|
|---|
| 28 | package classify;
|
|---|
| 29 |
|
|---|
| 30 | require util;
|
|---|
| 31 | use gsprintf;
|
|---|
| 32 |
|
|---|
| 33 |
|
|---|
| 34 | sub gsprintf
|
|---|
| 35 | {
|
|---|
| 36 | return &gsprintf::gsprintf(@_);
|
|---|
| 37 | }
|
|---|
| 38 |
|
|---|
| 39 |
|
|---|
| 40 | $next_classify_num = 1;
|
|---|
| 41 | sub load_classifier_for_info {
|
|---|
| 42 | my ($classifier) = shift @_;
|
|---|
| 43 |
|
|---|
| 44 | # find the classifier
|
|---|
| 45 | my $colclassname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'},
|
|---|
| 46 | "perllib/classify",
|
|---|
| 47 | "${classifier}.pm");
|
|---|
| 48 | my $mainclassname = &util::filename_cat($ENV{'GSDLHOME'},
|
|---|
| 49 | "perllib/classify",
|
|---|
| 50 | "${classifier}.pm");
|
|---|
| 51 |
|
|---|
| 52 | if (-e $colclassname) { require $colclassname; }
|
|---|
| 53 | elsif (-e $mainclassname) { require $mainclassname; }
|
|---|
| 54 | else {
|
|---|
| 55 | &gsprintf(STDERR, "{classify.could_not_find_classifier}\n", $classifier) && die "\n";
|
|---|
| 56 | }
|
|---|
| 57 | my ($classobj);
|
|---|
| 58 | my $options = "-gsdlinfo";
|
|---|
| 59 |
|
|---|
| 60 | eval ("\$classobj = new \$classifier($options)");
|
|---|
| 61 | die "$@" if $@;
|
|---|
| 62 |
|
|---|
| 63 | return $classobj;
|
|---|
| 64 | }
|
|---|
| 65 |
|
|---|
| 66 | sub load_classifiers {
|
|---|
| 67 | my ($classify_list, $build_dir, $outhandle) = @_;
|
|---|
| 68 | my @classify_objects = ();
|
|---|
| 69 |
|
|---|
| 70 | foreach $classifyoption (@$classify_list) {
|
|---|
| 71 |
|
|---|
| 72 | # get the classifier name
|
|---|
| 73 | my $classname = shift @$classifyoption;
|
|---|
| 74 | next unless defined $classname;
|
|---|
| 75 |
|
|---|
| 76 | # find the classifier
|
|---|
| 77 | my $colclassname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'},"perllib/classify",
|
|---|
| 78 | "${classname}.pm");
|
|---|
| 79 | my $mainclassname = &util::filename_cat($ENV{'GSDLHOME'},"perllib/classify",
|
|---|
| 80 | "${classname}.pm");
|
|---|
| 81 |
|
|---|
| 82 | if (-e $colclassname) { require $colclassname; }
|
|---|
| 83 | elsif (-e $mainclassname) { require $mainclassname; }
|
|---|
| 84 | else { &gsprintf(STDERR, "{classify.could_not_find_classifier}\n", $classname) && die "\n";
|
|---|
| 85 | # die "ERROR - couldn't find classifier \"$classname\"\n";
|
|---|
| 86 | }
|
|---|
| 87 |
|
|---|
| 88 | # create the classify object
|
|---|
| 89 | my ($classobj);
|
|---|
| 90 |
|
|---|
| 91 | my @newoptions;
|
|---|
| 92 |
|
|---|
| 93 | # do these first so they can be overriden by user supplied options
|
|---|
| 94 | push @newoptions, "-builddir", "$build_dir" if ($build_dir);
|
|---|
| 95 | push @newoptions, "-outhandle", "$outhandle" if ($outhandle);
|
|---|
| 96 | push @newoptions, "-verbosity", "2";
|
|---|
| 97 |
|
|---|
| 98 | # backwards compatability hack: if the classifier options are
|
|---|
| 99 | # in "x=y" format, convert them to parsearg ("-x y") format.
|
|---|
| 100 | my ($opt, $key, $value);
|
|---|
| 101 | foreach $opt (@$classifyoption) {
|
|---|
| 102 | if ($opt =~ /^(\w+)=(.*)$/) {
|
|---|
| 103 | push @newoptions, "-$1", $2;
|
|---|
| 104 | } else {
|
|---|
| 105 | push @newoptions, $opt;
|
|---|
| 106 | }
|
|---|
| 107 | }
|
|---|
| 108 |
|
|---|
| 109 | map { $_ = "\"$_\""; } @newoptions;
|
|---|
| 110 | my $options .= join (",", @newoptions);
|
|---|
| 111 |
|
|---|
| 112 | eval ("\$classobj = new \$classname($options)");
|
|---|
| 113 | die "$@" if $@;
|
|---|
| 114 |
|
|---|
| 115 | # add this object to the list
|
|---|
| 116 | push (@classify_objects, $classobj);
|
|---|
| 117 | }
|
|---|
| 118 |
|
|---|
| 119 | return \@classify_objects;
|
|---|
| 120 | }
|
|---|
| 121 |
|
|---|
| 122 | # init_classifiers resets all the classifiers and readys them to process
|
|---|
| 123 | # the documents.
|
|---|
| 124 | sub init_classifiers {
|
|---|
| 125 | my ($classifiers) = @_;
|
|---|
| 126 |
|
|---|
| 127 | foreach $classobj (@$classifiers) {
|
|---|
| 128 | $classobj->init();
|
|---|
| 129 | }
|
|---|
| 130 | }
|
|---|
| 131 |
|
|---|
| 132 | # classify_doc lets each of the classifiers classify a document
|
|---|
| 133 | sub classify_doc {
|
|---|
| 134 | my ($classifiers, $doc_obj) = @_;
|
|---|
| 135 |
|
|---|
| 136 | foreach $classobj (@$classifiers) {
|
|---|
| 137 | $classobj->classify($doc_obj);
|
|---|
| 138 | }
|
|---|
| 139 | }
|
|---|
| 140 |
|
|---|
| 141 | # output_classify_info outputs all the info needed for the classification
|
|---|
| 142 | # to the gdbm
|
|---|
| 143 | sub output_classify_info {
|
|---|
| 144 | my ($classifiers, $handle, $allclassifications, $gli) = @_;
|
|---|
| 145 | # $handle = "main::STDOUT";
|
|---|
| 146 |
|
|---|
| 147 | $gli = 0 unless defined $gli;
|
|---|
| 148 |
|
|---|
| 149 | # create a classification containing all the info
|
|---|
| 150 | my $classifyinfo = {'classifyOID'=>'browse',
|
|---|
| 151 | 'contains'=>[]};
|
|---|
| 152 |
|
|---|
| 153 | # get each of the classifications
|
|---|
| 154 | foreach $classobj (@$classifiers) {
|
|---|
| 155 | my $tempinfo = $classobj->get_classify_info($gli);
|
|---|
| 156 | $tempinfo->{'classifyOID'} = "CL$next_classify_num";
|
|---|
| 157 | $next_classify_num++;
|
|---|
| 158 | push (@{$classifyinfo->{'contains'}}, $tempinfo);
|
|---|
| 159 | }
|
|---|
| 160 |
|
|---|
| 161 | &print_classify_info ($handle, $classifyinfo, "", $allclassifications);
|
|---|
| 162 | }
|
|---|
| 163 |
|
|---|
| 164 | sub print_classify_info {
|
|---|
| 165 | my ($handle, $classifyinfo, $OID, $allclassifications) = @_;
|
|---|
| 166 |
|
|---|
| 167 | $OID =~ s/^\.+//; # just for good luck
|
|---|
| 168 |
|
|---|
| 169 | # book information is printed elsewhere
|
|---|
| 170 | return if (defined ($classifyinfo->{'OID'}));
|
|---|
| 171 |
|
|---|
| 172 | # don't want empty classifications
|
|---|
| 173 | if ($allclassifications || &check_contents ($classifyinfo) > 0) {
|
|---|
| 174 |
|
|---|
| 175 | $OID = $classifyinfo->{'classifyOID'} if defined ($classifyinfo->{'classifyOID'});
|
|---|
| 176 |
|
|---|
| 177 | my $outputtext = "[$OID]\n";
|
|---|
| 178 | $outputtext .= "<doctype>classify\n";
|
|---|
| 179 | $outputtext .= "<hastxt>0\n";
|
|---|
| 180 | $outputtext .= "<childtype>$classifyinfo->{'childtype'}\n"
|
|---|
| 181 | if defined $classifyinfo->{'childtype'};
|
|---|
| 182 | $outputtext .= "<Title>$classifyinfo->{'Title'}\n"
|
|---|
| 183 | if defined $classifyinfo->{'Title'};
|
|---|
| 184 | $outputtext .= "<numleafdocs>$classifyinfo->{'numleafdocs'}\n"
|
|---|
| 185 | if defined $classifyinfo->{'numleafdocs'};
|
|---|
| 186 | $outputtext .= "<thistype>$classifyinfo->{'thistype'}\n"
|
|---|
| 187 | if defined $classifyinfo->{'thistype'};
|
|---|
| 188 | $outputtext .= "<parameters>$classifyinfo->{'parameters'}\n"
|
|---|
| 189 | if defined $classifyinfo->{'parameters'};
|
|---|
| 190 |
|
|---|
| 191 | my $contains_text = "<contains>";
|
|---|
| 192 | my $mdoffset_text = "<mdoffset>";
|
|---|
| 193 |
|
|---|
| 194 | my $next_subOID = 1;
|
|---|
| 195 | my $first = 1;
|
|---|
| 196 | foreach $tempinfo (@{$classifyinfo->{'contains'}}) {
|
|---|
| 197 | # empty contents were made undefined by clean_contents()
|
|---|
| 198 | next unless defined $tempinfo;
|
|---|
| 199 |
|
|---|
| 200 | $contains_text .= ";" unless $first;
|
|---|
| 201 | $mdoffset_text .= ";" unless $first;
|
|---|
| 202 | $first = 0;
|
|---|
| 203 |
|
|---|
| 204 | if (defined ($tempinfo->{'classifyOID'})) {
|
|---|
| 205 | $contains_text .= $tempinfo->{'classifyOID'};
|
|---|
| 206 | &print_classify_info ($handle, $tempinfo, $tempinfo->{'classifyOID'},
|
|---|
| 207 | $allclassifications);
|
|---|
| 208 | } elsif (defined ($tempinfo->{'OID'})) {
|
|---|
| 209 | $contains_text .= $tempinfo->{'OID'};
|
|---|
| 210 | $mdoffset_text .= $tempinfo->{'offset'}
|
|---|
| 211 | if (defined ($tempinfo->{'offset'}))
|
|---|
| 212 | # note: we don't want to print the contents of the books
|
|---|
| 213 | } else {
|
|---|
| 214 |
|
|---|
| 215 | # Supress having top-level node in Collage classifier
|
|---|
| 216 | # so no bookshelf icon appears, top-level, along with the
|
|---|
| 217 | # applet
|
|---|
| 218 |
|
|---|
| 219 | if (!defined ($tempinfo->{'Title'}) || $tempinfo->{'Title'} ne "Collage") {
|
|---|
| 220 | $contains_text .= "\".$next_subOID";
|
|---|
| 221 | }
|
|---|
| 222 |
|
|---|
| 223 | &print_classify_info ($handle, $tempinfo, "$OID.$next_subOID",
|
|---|
| 224 | $allclassifications);
|
|---|
| 225 | $next_subOID++;
|
|---|
| 226 | }
|
|---|
| 227 | }
|
|---|
| 228 |
|
|---|
| 229 | $outputtext .= "$contains_text\n";
|
|---|
| 230 | $outputtext .= "<mdtype>$classifyinfo->{'mdtype'}\n"
|
|---|
| 231 | if defined $classifyinfo->{'mdtype'};
|
|---|
| 232 | $outputtext .= "$mdoffset_text\n"
|
|---|
| 233 | if ($mdoffset_text !~ m/^<mdoffset>;+$/);
|
|---|
| 234 |
|
|---|
| 235 | $outputtext .= '-' x 70 . "\n";
|
|---|
| 236 |
|
|---|
| 237 | print $handle $outputtext;
|
|---|
| 238 | }
|
|---|
| 239 | }
|
|---|
| 240 |
|
|---|
| 241 | sub check_contents {
|
|---|
| 242 | my ($classifyinfo) = @_;
|
|---|
| 243 | my $num_leaf_docs = 0;
|
|---|
| 244 | my $sub_num_leaf_docs = 0;
|
|---|
| 245 |
|
|---|
| 246 | return $classifyinfo->{'numleafdocs'} if (defined $classifyinfo->{'numleafdocs'});
|
|---|
| 247 |
|
|---|
| 248 | foreach $content (@{$classifyinfo->{'contains'}}) {
|
|---|
| 249 | if (defined $content->{'OID'}) {
|
|---|
| 250 | # found a book
|
|---|
| 251 | $num_leaf_docs ++;
|
|---|
| 252 | } elsif (($sub_num_leaf_docs = &check_contents ($content)) > 0) {
|
|---|
| 253 | # there's a book somewhere below
|
|---|
| 254 | $num_leaf_docs += $sub_num_leaf_docs;
|
|---|
| 255 | } else {
|
|---|
| 256 | # section contains no books so we want to remove
|
|---|
| 257 | # it from its parents contents
|
|---|
| 258 | $content = undef;
|
|---|
| 259 | }
|
|---|
| 260 | }
|
|---|
| 261 |
|
|---|
| 262 | $classifyinfo->{'numleafdocs'} = $num_leaf_docs;
|
|---|
| 263 | return $num_leaf_docs;
|
|---|
| 264 | }
|
|---|
| 265 |
|
|---|
| 266 | 1;
|
|---|
