[9206] | 1 | ###########################################################################
|
---|
| 2 | #
|
---|
| 3 | # HFileHierarchy.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 | # classifier plugin for generating hierarchical classifications
|
---|
| 27 |
|
---|
| 28 |
|
---|
| 29 | package HFileHierarchy;
|
---|
| 30 |
|
---|
[17209] | 31 | use BaseClassifier;
|
---|
[9206] | 32 | use util;
|
---|
| 33 | use cfgread;
|
---|
| 34 | use sorttools;
|
---|
| 35 |
|
---|
[10253] | 36 | use strict;
|
---|
| 37 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
---|
| 38 |
|
---|
[9206] | 39 | sub BEGIN {
|
---|
[17209] | 40 | @HFileHierarchy::ISA = ('BaseClassifier');
|
---|
[9206] | 41 | }
|
---|
| 42 |
|
---|
| 43 | my $arguments =
|
---|
| 44 | [ { 'name' => "metadata",
|
---|
| 45 | 'desc' => "{AZCompactList.metadata}",
|
---|
| 46 | 'type' => "metadata",
|
---|
| 47 | 'reqd' => "yes" },
|
---|
| 48 | { 'name' => "firstvalueonly",
|
---|
| 49 | 'desc' => "{AZCompactList.firstvalueonly}",
|
---|
| 50 | 'type' => "flag",
|
---|
| 51 | 'reqd' => "no" },
|
---|
| 52 | { 'name' => "allvalues",
|
---|
| 53 | 'desc' => "{AZCompactList.allvalues}",
|
---|
| 54 | 'type' => "flag",
|
---|
| 55 | 'reqd' => "no" },
|
---|
| 56 | { 'name' => "hfile",
|
---|
| 57 | 'desc' => "{Hierarchy.hfile}",
|
---|
| 58 | 'type' => "string",
|
---|
| 59 | 'deft' => "",
|
---|
| 60 | 'reqd' => "no" },
|
---|
| 61 | { 'name' => "sort",
|
---|
| 62 | 'desc' => "{Hierarchy.sort}",
|
---|
[11665] | 63 | 'type' => "metadata",
|
---|
[9206] | 64 | 'reqd' => "no" },
|
---|
| 65 | { 'name' => "reverse_sort",
|
---|
| 66 | 'desc' => "{Hierarchy.reverse_sort}",
|
---|
| 67 | 'type' => "flag",
|
---|
| 68 | 'reqd' => "no" },
|
---|
| 69 | { 'name' => "hlist_at_top",
|
---|
| 70 | 'desc' => "{Hierarchy.hlist_at_top}",
|
---|
| 71 | 'type' => "flag",
|
---|
[10663] | 72 | 'reqd' => "no" },
|
---|
| 73 | { 'name' => "documents_last",
|
---|
| 74 | 'desc' => "{Hierarchy.documents_last}",
|
---|
| 75 | 'type' => "flag",
|
---|
| 76 | 'reqd' => "no"}
|
---|
| 77 | ];
|
---|
[9206] | 78 |
|
---|
| 79 | my $options =
|
---|
| 80 | { 'name' => "HFileHierarchy",
|
---|
[10218] | 81 | 'desc' => "{HFileHierarchy.desc}",
|
---|
[9206] | 82 | 'abstract' => "yes",
|
---|
| 83 | 'inherits' => "yes",
|
---|
| 84 | 'args' => $arguments };
|
---|
| 85 |
|
---|
| 86 |
|
---|
| 87 | sub new {
|
---|
[10218] | 88 | my ($class) = shift (@_);
|
---|
| 89 | my ($classifierslist,$inputargs,$hashArgOptLists) = @_;
|
---|
| 90 | push(@$classifierslist, $class);
|
---|
[9206] | 91 |
|
---|
[17209] | 92 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
---|
| 93 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
---|
[9206] | 94 |
|
---|
[17209] | 95 | my $self = new BaseClassifier($classifierslist, $inputargs, $hashArgOptLists);
|
---|
[10218] | 96 |
|
---|
[10253] | 97 | if ($self->{'info_only'}) {
|
---|
| 98 | # don't worry about any options etc
|
---|
| 99 | return bless $self, $class;
|
---|
| 100 | }
|
---|
| 101 |
|
---|
[10218] | 102 | my $metadata = $self->{'metadata'};
|
---|
[9206] | 103 | if (!$metadata) {
|
---|
| 104 | print STDERR "$class Error: required option -metadata not supplied\n";
|
---|
| 105 | $self->print_txt_usage(""); # Use default resource bundle
|
---|
| 106 |
|
---|
| 107 | die "$class Error: required option -metadata not supplied\n";
|
---|
| 108 | }
|
---|
| 109 |
|
---|
[10218] | 110 | $self->{'buttonname'} = $self->generate_title_from_metadata($metadata) unless ($self->{'buttonname'});
|
---|
[20422] | 111 | # strip ex from metadata
|
---|
[20425] | 112 | $metadata = $self->strip_ex_from_metadata($metadata);
|
---|
[9206] | 113 |
|
---|
| 114 | my @meta_list = split(/,/, $metadata);
|
---|
| 115 | $self->{'meta_list'} = \@meta_list;
|
---|
| 116 |
|
---|
[10218] | 117 | # sort = undef in this case is the same as sort=nosort
|
---|
| 118 | if ($self->{'sort'} eq "nosort") {
|
---|
| 119 | $self->{'sort'} = undef;
|
---|
| 120 | }
|
---|
[20422] | 121 | if (defined $self->{'sort'}) { # remove ex. namespace
|
---|
[20425] | 122 | $self->{'sort'} = $self->strip_ex_from_metadata($self->{'sort'});
|
---|
[20422] | 123 | }
|
---|
[9206] | 124 |
|
---|
[10218] | 125 | if ($self->{'hfile'}) {
|
---|
| 126 | my $hfile = $self->{'hfile'};
|
---|
[28564] | 127 | my $subjectfile;
|
---|
| 128 | $subjectfile = &FileUtils::filenameConcatenate($ENV{'GSDLCOLLECTDIR'},"etc", $hfile);
|
---|
[9206] | 129 | if (!-e $subjectfile) {
|
---|
| 130 | my $collfile = $subjectfile;
|
---|
[28564] | 131 | $subjectfile = &FileUtils::filenameConcatenate($ENV{'GSDLHOME'},"etc", $hfile);
|
---|
[9206] | 132 | if (!-e $subjectfile) {
|
---|
| 133 | my $outhandle = $self->{'outhandle'};
|
---|
| 134 | print STDERR "\nHFileHierarchy Error: Can't locate subject file $hfile\n";
|
---|
| 135 | print STDERR "This file should be in $collfile or $subjectfile\n";
|
---|
| 136 | $self->print_txt_usage(""); # Use default resource bundle
|
---|
| 137 | print STDERR "\nHFileHierarchy Error: Can't locate subject file $hfile\n";
|
---|
| 138 | print STDERR "This file should be in $collfile or $subjectfile\n";
|
---|
| 139 | die "\n";
|
---|
| 140 | }
|
---|
| 141 | }
|
---|
| 142 | $self->{'descriptorlist'} = {}; # first field in subject file
|
---|
| 143 | $self->{'locatorlist'} = {}; # second field in subject file
|
---|
| 144 | $self->{'subjectfile'} = $subjectfile;
|
---|
| 145 | }
|
---|
| 146 |
|
---|
| 147 |
|
---|
[10218] | 148 | # $self->{'firstvalueonly'} = $firstvalueonly;
|
---|
| 149 | # $self->{'allvalues'} = $allvalues;
|
---|
[9206] | 150 |
|
---|
[10218] | 151 | #$self->{'hlist_at_top'} = $hlist_at_top;
|
---|
[9206] | 152 |
|
---|
[10218] | 153 | # Clean out the unused keys
|
---|
| 154 | delete $self->{'metadata'};
|
---|
| 155 | delete $self->{'hfile'};
|
---|
[9206] | 156 |
|
---|
| 157 | return bless $self, $class;
|
---|
| 158 | }
|
---|
| 159 |
|
---|
| 160 | sub init {
|
---|
| 161 | my $self = shift (@_);
|
---|
| 162 |
|
---|
| 163 | my $subjectfile = $self->{'subjectfile'};
|
---|
| 164 | if (defined $subjectfile) {
|
---|
[24586] | 165 | # read in the subject file, but read in unicode mode to preserve special characters
|
---|
| 166 | my $list = &cfgread::read_cfg_file_unicode ($self->{'subjectfile'}, undef, '^[^#]?\S');
|
---|
[9206] | 167 | # $list is a hash that is indexed by the descriptor. The contents of this
|
---|
| 168 | # hash is a list of two items. The first item is the OID and the second item
|
---|
| 169 | # is the title
|
---|
[10253] | 170 | foreach my $descriptor (keys (%$list)) {
|
---|
[9206] | 171 | $self->{'descriptorlist'}->{$descriptor} = $list->{$descriptor}->[0];
|
---|
| 172 | unless (defined $self->{'locatorlist'}->{$list->{$descriptor}->[0]}) {
|
---|
| 173 | $self->{'locatorlist'}->{$list->{$descriptor}->[0]}->{'title'} = $list->{$descriptor}->[1];
|
---|
| 174 | $self->{'locatorlist'}->{$list->{$descriptor}->[0]}->{'contents'} = [];
|
---|
| 175 | }
|
---|
| 176 | }
|
---|
| 177 | }
|
---|
| 178 | }
|
---|
| 179 |
|
---|
| 180 | sub hfile_classify
|
---|
| 181 | {
|
---|
| 182 | my $self = shift (@_);
|
---|
[23116] | 183 | my ($doc_obj,$sortmeta,$metavalues) = @_;
|
---|
[9206] | 184 |
|
---|
[18455] | 185 | my $outhandle = $self->{'outhandle'};
|
---|
| 186 |
|
---|
[9206] | 187 | my $doc_OID = $doc_obj->get_OID();
|
---|
| 188 |
|
---|
| 189 | foreach my $metaelement (@$metavalues) {
|
---|
| 190 | if ((defined $self->{'descriptorlist'}->{$metaelement}) &&
|
---|
| 191 | (defined $self->{'locatorlist'}->{$self->{'descriptorlist'}->{$metaelement}})) {
|
---|
| 192 |
|
---|
[23116] | 193 | push (@{$self->{'locatorlist'}->{$self->{'descriptorlist'}->{$metaelement}}->{'contents'}},
|
---|
| 194 | [$doc_OID, $sortmeta]);
|
---|
| 195 | my $localid = $self->{'descriptorlist'}->{$metaelement};
|
---|
| 196 | my $classid = $self->get_number();
|
---|
| 197 |
|
---|
| 198 | $doc_obj->add_metadata($doc_obj->get_top_section(), "memberof", "CL$classid.$localid");
|
---|
| 199 |
|
---|
[9206] | 200 | }
|
---|
| 201 | }
|
---|
| 202 | }
|
---|
| 203 |
|
---|
| 204 |
|
---|
| 205 |
|
---|
| 206 |
|
---|
| 207 | sub hfile_get_classify_info {
|
---|
| 208 | my $self = shift (@_);
|
---|
| 209 |
|
---|
| 210 | my ($classifyinfo) = @_;
|
---|
| 211 |
|
---|
| 212 | my $list = $self->{'locatorlist'};
|
---|
| 213 |
|
---|
[23081] | 214 | my $classifier_num = "CL".$self->get_number();
|
---|
[9206] | 215 | # sorted the keys - otherwise funny things happen - kjdon 03/01/03
|
---|
[10253] | 216 | foreach my $OID (sort keys (%$list)) {
|
---|
[23081] | 217 | my $tempinfo = $self->get_OID_entry ($OID, $classifyinfo, "$classifier_num.$OID", $list->{$OID}->{'title'}, "VList");
|
---|
[29821] | 218 | if (not defined ($tempinfo)) {
|
---|
| 219 | print STDERR "Error occurred for node $OID. Not creating the classifier \n";
|
---|
| 220 | return undef;
|
---|
| 221 | }
|
---|
[10218] | 222 | if (defined $self->{'sort'}) {
|
---|
[9206] | 223 | if ($self->{'reverse_sort'}) {
|
---|
[10253] | 224 | foreach my $subOID (sort {$b->[1] cmp $a->[1];} @{$list->{$OID}->{'contents'}}) {
|
---|
[9206] | 225 | push (@{$tempinfo->{'contains'}}, {'OID'=>$subOID->[0]});
|
---|
| 226 | }
|
---|
| 227 | }
|
---|
| 228 | else {
|
---|
[10253] | 229 | foreach my $subOID (sort {$a->[1] cmp $b->[1];} @{$list->{$OID}->{'contents'}}) {
|
---|
[9206] | 230 | push (@{$tempinfo->{'contains'}}, {'OID'=>$subOID->[0]});
|
---|
| 231 | }
|
---|
| 232 | }
|
---|
| 233 | }
|
---|
| 234 | else {
|
---|
[10253] | 235 | foreach my $subOID (@{$list->{$OID}->{'contents'}}) {
|
---|
[9206] | 236 | push (@{$tempinfo->{'contains'}}, {'OID'=>$subOID->[0]});
|
---|
| 237 | }
|
---|
| 238 | }
|
---|
| 239 | }
|
---|
| 240 |
|
---|
| 241 | return $classifyinfo;
|
---|
| 242 | }
|
---|
| 243 |
|
---|
| 244 |
|
---|
| 245 | sub supports_memberof {
|
---|
| 246 | my $self = shift(@_);
|
---|
| 247 |
|
---|
| 248 | return "true";
|
---|
| 249 | }
|
---|
| 250 |
|
---|
| 251 | sub get_OID_entry {
|
---|
| 252 | my $self = shift (@_);
|
---|
[23081] | 253 | my ($OID, $classifyinfo, $classifyOID, $title, $classifytype) = @_;
|
---|
[9206] | 254 |
|
---|
| 255 | $OID = "" unless defined $OID;
|
---|
| 256 | $OID =~ s/^\.+//;
|
---|
| 257 |
|
---|
| 258 | my ($headOID, $tailOID) = $OID =~ /^(\d+)(.*)$/;
|
---|
| 259 | $tailOID = "" unless defined $tailOID;
|
---|
| 260 |
|
---|
| 261 | if (!defined $headOID) {
|
---|
| 262 | $classifyinfo->{'Title'} = $title;
|
---|
[23081] | 263 | $classifyinfo->{'classifyOID'} = $classifyOID;
|
---|
[9206] | 264 | $classifyinfo->{'classifytype'} = $classifytype;
|
---|
| 265 | return $classifyinfo;
|
---|
| 266 | }
|
---|
[29821] | 267 | if ($headOID eq "0") {
|
---|
| 268 | print STDERR "Error: Hierarchy numbering must not contain 0\n";
|
---|
| 269 | return undef;
|
---|
| 270 | }
|
---|
[9206] | 271 | $classifyinfo->{'contains'} = [] unless defined $classifyinfo->{'contains'};
|
---|
[10663] | 272 | if ($self->{'documents_last'}) {
|
---|
| 273 | # documents should come after nodes in the classifier
|
---|
| 274 |
|
---|
| 275 | my $doc_pos = 0;
|
---|
| 276 | foreach my $thing (@{$classifyinfo->{'contains'}}) {
|
---|
| 277 | last if defined $thing->{'OID'};
|
---|
| 278 | $doc_pos++;
|
---|
| 279 | }
|
---|
| 280 |
|
---|
| 281 | while ($doc_pos < $headOID) {
|
---|
| 282 | splice(@{$classifyinfo->{'contains'}}, $doc_pos, 0, $self->get_entry("", $classifytype));
|
---|
| 283 | $doc_pos++;
|
---|
| 284 | }
|
---|
| 285 |
|
---|
[23081] | 286 | return $self->get_OID_entry ($tailOID, $classifyinfo->{'contains'}->[($headOID-1)], $classifyOID, $title, $classifytype);
|
---|
[10663] | 287 |
|
---|
| 288 | }
|
---|
| 289 |
|
---|
| 290 | # else, documents come before nodes
|
---|
[9206] | 291 | my $offset = 0;
|
---|
| 292 | foreach my $thing (@{$classifyinfo->{'contains'}}) {
|
---|
| 293 | $offset ++ if defined $thing->{'OID'};
|
---|
| 294 | }
|
---|
| 295 |
|
---|
| 296 | while (scalar(@{$classifyinfo->{'contains'}}) < ($headOID+$offset)) {
|
---|
| 297 | push (@{$classifyinfo->{'contains'}}, $self->get_entry("", $classifytype));
|
---|
| 298 | }
|
---|
| 299 |
|
---|
[23081] | 300 | return $self->get_OID_entry ($tailOID, $classifyinfo->{'contains'}->[($headOID+$offset-1)], $classifyOID, $title, $classifytype);
|
---|
[9206] | 301 | }
|
---|
| 302 |
|
---|
| 303 | sub get_entry {
|
---|
| 304 | my $self = shift (@_);
|
---|
| 305 | my ($title, $childtype, $thistype) = @_;
|
---|
| 306 | my $memberof = &supports_memberof();
|
---|
| 307 |
|
---|
| 308 | # organise into classification structure
|
---|
| 309 | my %classifyinfo = ('childtype'=>$childtype,
|
---|
| 310 | 'Title'=>$title,
|
---|
| 311 | 'supportsmemberof'=>$memberof,
|
---|
| 312 | 'contains'=>[]);
|
---|
| 313 | $classifyinfo{'thistype'} = $thistype
|
---|
| 314 | if defined $thistype && $thistype =~ /\w/;
|
---|
| 315 |
|
---|
| 316 | return \%classifyinfo;
|
---|
| 317 | }
|
---|
| 318 |
|
---|
| 319 |
|
---|
| 320 | 1;
|
---|