| 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 |
|
|---|
| 31 | use BaseClassifier;
|
|---|
| 32 | use util;
|
|---|
| 33 | use cfgread;
|
|---|
| 34 | use sorttools;
|
|---|
| 35 |
|
|---|
| 36 | use strict;
|
|---|
| 37 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
|---|
| 38 |
|
|---|
| 39 | sub BEGIN {
|
|---|
| 40 | @HFileHierarchy::ISA = ('BaseClassifier');
|
|---|
| 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}",
|
|---|
| 63 | 'type' => "metadata",
|
|---|
| 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",
|
|---|
| 72 | 'reqd' => "no" },
|
|---|
| 73 | { 'name' => "documents_last",
|
|---|
| 74 | 'desc' => "{Hierarchy.documents_last}",
|
|---|
| 75 | 'type' => "flag",
|
|---|
| 76 | 'reqd' => "no"}
|
|---|
| 77 | ];
|
|---|
| 78 |
|
|---|
| 79 | my $options =
|
|---|
| 80 | { 'name' => "HFileHierarchy",
|
|---|
| 81 | 'desc' => "{HFileHierarchy.desc}",
|
|---|
| 82 | 'abstract' => "yes",
|
|---|
| 83 | 'inherits' => "yes",
|
|---|
| 84 | 'args' => $arguments };
|
|---|
| 85 |
|
|---|
| 86 |
|
|---|
| 87 | sub new {
|
|---|
| 88 | my ($class) = shift (@_);
|
|---|
| 89 | my ($classifierslist,$inputargs,$hashArgOptLists) = @_;
|
|---|
| 90 | push(@$classifierslist, $class);
|
|---|
| 91 |
|
|---|
| 92 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
|---|
| 93 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
|---|
| 94 |
|
|---|
| 95 | my $self = new BaseClassifier($classifierslist, $inputargs, $hashArgOptLists);
|
|---|
| 96 |
|
|---|
| 97 | if ($self->{'info_only'}) {
|
|---|
| 98 | # don't worry about any options etc
|
|---|
| 99 | return bless $self, $class;
|
|---|
| 100 | }
|
|---|
| 101 |
|
|---|
| 102 | my $metadata = $self->{'metadata'};
|
|---|
| 103 | if (!$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 |
|
|---|
| 110 | $self->{'buttonname'} = $self->generate_title_from_metadata($metadata) unless ($self->{'buttonname'});
|
|---|
| 111 | # strip ex from metadata
|
|---|
| 112 | $metadata =~ s/^ex\.//;
|
|---|
| 113 | $metadata =~ s/,ex\./,/g;
|
|---|
| 114 |
|
|---|
| 115 | my @meta_list = split(/,/, $metadata);
|
|---|
| 116 | $self->{'meta_list'} = \@meta_list;
|
|---|
| 117 |
|
|---|
| 118 | # sort = undef in this case is the same as sort=nosort
|
|---|
| 119 | if ($self->{'sort'} eq "nosort") {
|
|---|
| 120 | $self->{'sort'} = undef;
|
|---|
| 121 | }
|
|---|
| 122 | if (defined $self->{'sort'}) { # remove ex. namespace
|
|---|
| 123 | $self->{'sort'} =~ s/^ex\.//;
|
|---|
| 124 | }
|
|---|
| 125 |
|
|---|
| 126 | if ($self->{'hfile'}) {
|
|---|
| 127 | my $hfile = $self->{'hfile'};
|
|---|
| 128 | my $subjectfile;
|
|---|
| 129 | $subjectfile = &util::filename_cat($ENV{'GSDLCOLLECTDIR'},"etc", $hfile);
|
|---|
| 130 | if (!-e $subjectfile) {
|
|---|
| 131 | my $collfile = $subjectfile;
|
|---|
| 132 | $subjectfile = &util::filename_cat($ENV{'GSDLHOME'},"etc", $hfile);
|
|---|
| 133 | if (!-e $subjectfile) {
|
|---|
| 134 | my $outhandle = $self->{'outhandle'};
|
|---|
| 135 | print STDERR "\nHFileHierarchy Error: Can't locate subject file $hfile\n";
|
|---|
| 136 | print STDERR "This file should be in $collfile or $subjectfile\n";
|
|---|
| 137 | $self->print_txt_usage(""); # Use default resource bundle
|
|---|
| 138 | print STDERR "\nHFileHierarchy Error: Can't locate subject file $hfile\n";
|
|---|
| 139 | print STDERR "This file should be in $collfile or $subjectfile\n";
|
|---|
| 140 | die "\n";
|
|---|
| 141 | }
|
|---|
| 142 | }
|
|---|
| 143 | $self->{'descriptorlist'} = {}; # first field in subject file
|
|---|
| 144 | $self->{'locatorlist'} = {}; # second field in subject file
|
|---|
| 145 | $self->{'subjectfile'} = $subjectfile;
|
|---|
| 146 | }
|
|---|
| 147 |
|
|---|
| 148 |
|
|---|
| 149 | # $self->{'firstvalueonly'} = $firstvalueonly;
|
|---|
| 150 | # $self->{'allvalues'} = $allvalues;
|
|---|
| 151 |
|
|---|
| 152 | #$self->{'hlist_at_top'} = $hlist_at_top;
|
|---|
| 153 |
|
|---|
| 154 | # Clean out the unused keys
|
|---|
| 155 | delete $self->{'metadata'};
|
|---|
| 156 | delete $self->{'hfile'};
|
|---|
| 157 |
|
|---|
| 158 | return bless $self, $class;
|
|---|
| 159 | }
|
|---|
| 160 |
|
|---|
| 161 | sub init {
|
|---|
| 162 | my $self = shift (@_);
|
|---|
| 163 |
|
|---|
| 164 | my $subjectfile = $self->{'subjectfile'};
|
|---|
| 165 | if (defined $subjectfile) {
|
|---|
| 166 | # read in the subject file
|
|---|
| 167 | my $list = &cfgread::read_cfg_file ($self->{'subjectfile'}, undef, '^[^#]?\S');
|
|---|
| 168 | # $list is a hash that is indexed by the descriptor. The contents of this
|
|---|
| 169 | # hash is a list of two items. The first item is the OID and the second item
|
|---|
| 170 | # is the title
|
|---|
| 171 | foreach my $descriptor (keys (%$list)) {
|
|---|
| 172 | $self->{'descriptorlist'}->{$descriptor} = $list->{$descriptor}->[0];
|
|---|
| 173 | unless (defined $self->{'locatorlist'}->{$list->{$descriptor}->[0]}) {
|
|---|
| 174 | $self->{'locatorlist'}->{$list->{$descriptor}->[0]}->{'title'} = $list->{$descriptor}->[1];
|
|---|
| 175 | $self->{'locatorlist'}->{$list->{$descriptor}->[0]}->{'contents'} = [];
|
|---|
| 176 | }
|
|---|
| 177 | }
|
|---|
| 178 | }
|
|---|
| 179 | }
|
|---|
| 180 |
|
|---|
| 181 | sub hfile_classify
|
|---|
| 182 | {
|
|---|
| 183 | my $self = shift (@_);
|
|---|
| 184 | my ($doc_obj,$edit_mode,$sortmeta,$metavalues) = @_;
|
|---|
| 185 |
|
|---|
| 186 | my $outhandle = $self->{'outhandle'};
|
|---|
| 187 |
|
|---|
| 188 | my $doc_OID = $doc_obj->get_OID();
|
|---|
| 189 |
|
|---|
| 190 | foreach my $metaelement (@$metavalues) {
|
|---|
| 191 | if ((defined $self->{'descriptorlist'}->{$metaelement}) &&
|
|---|
| 192 | (defined $self->{'locatorlist'}->{$self->{'descriptorlist'}->{$metaelement}})) {
|
|---|
| 193 |
|
|---|
| 194 | if ($edit_mode eq "delete") {
|
|---|
| 195 | # find it, and remove it
|
|---|
| 196 | my $existing_list = $self->{'locatorlist'}->{$self->{'descriptorlist'}->{$metaelement}}->{'contents'};
|
|---|
| 197 |
|
|---|
| 198 | my $filtered_list = [];
|
|---|
| 199 | foreach my $existing_oid_pair (@$existing_list) {
|
|---|
| 200 | if ($existing_oid_pair->[0] eq $doc_OID) {
|
|---|
| 201 | print $outhandle " Deleting $doc_OID for $metaelement in hierarchy\n";
|
|---|
| 202 | }
|
|---|
| 203 | else {
|
|---|
| 204 | push(@$filtered_list,$existing_oid_pair);
|
|---|
| 205 | }
|
|---|
| 206 | }
|
|---|
| 207 | $self->{'locatorlist'}->{$self->{'descriptorlist'}->{$metaelement}}->{'contents'} = $filtered_list;
|
|---|
| 208 | }
|
|---|
| 209 | else {
|
|---|
| 210 | push (@{$self->{'locatorlist'}->{$self->{'descriptorlist'}->{$metaelement}}->{'contents'}},
|
|---|
| 211 | [$doc_OID, $sortmeta]);
|
|---|
| 212 | my $localid = $self->{'descriptorlist'}->{$metaelement};
|
|---|
| 213 | my $classid = $self->get_number();
|
|---|
| 214 |
|
|---|
| 215 | $doc_obj->add_metadata($doc_obj->get_top_section(), "memberof", "CL$classid.$localid");
|
|---|
| 216 | }
|
|---|
| 217 | }
|
|---|
| 218 | }
|
|---|
| 219 | }
|
|---|
| 220 |
|
|---|
| 221 |
|
|---|
| 222 |
|
|---|
| 223 |
|
|---|
| 224 | sub hfile_get_classify_info {
|
|---|
| 225 | my $self = shift (@_);
|
|---|
| 226 |
|
|---|
| 227 | my ($classifyinfo) = @_;
|
|---|
| 228 |
|
|---|
| 229 | my $list = $self->{'locatorlist'};
|
|---|
| 230 |
|
|---|
| 231 | # sorted the keys - otherwise funny things happen - kjdon 03/01/03
|
|---|
| 232 | foreach my $OID (sort keys (%$list)) {
|
|---|
| 233 | my $tempinfo = $self->get_OID_entry ($OID, $classifyinfo, $list->{$OID}->{'title'}, "VList");
|
|---|
| 234 |
|
|---|
| 235 | if (defined $self->{'sort'}) {
|
|---|
| 236 | if ($self->{'reverse_sort'}) {
|
|---|
| 237 | foreach my $subOID (sort {$b->[1] cmp $a->[1];} @{$list->{$OID}->{'contents'}}) {
|
|---|
| 238 | push (@{$tempinfo->{'contains'}}, {'OID'=>$subOID->[0]});
|
|---|
| 239 | }
|
|---|
| 240 | }
|
|---|
| 241 | else {
|
|---|
| 242 | foreach my $subOID (sort {$a->[1] cmp $b->[1];} @{$list->{$OID}->{'contents'}}) {
|
|---|
| 243 | push (@{$tempinfo->{'contains'}}, {'OID'=>$subOID->[0]});
|
|---|
| 244 | }
|
|---|
| 245 | }
|
|---|
| 246 | }
|
|---|
| 247 | else {
|
|---|
| 248 | foreach my $subOID (@{$list->{$OID}->{'contents'}}) {
|
|---|
| 249 | push (@{$tempinfo->{'contains'}}, {'OID'=>$subOID->[0]});
|
|---|
| 250 | }
|
|---|
| 251 | }
|
|---|
| 252 | }
|
|---|
| 253 |
|
|---|
| 254 | return $classifyinfo;
|
|---|
| 255 | }
|
|---|
| 256 |
|
|---|
| 257 |
|
|---|
| 258 | sub supports_memberof {
|
|---|
| 259 | my $self = shift(@_);
|
|---|
| 260 |
|
|---|
| 261 | return "true";
|
|---|
| 262 | }
|
|---|
| 263 |
|
|---|
| 264 | sub get_OID_entry {
|
|---|
| 265 | my $self = shift (@_);
|
|---|
| 266 | my ($OID, $classifyinfo, $title, $classifytype) = @_;
|
|---|
| 267 |
|
|---|
| 268 | $OID = "" unless defined $OID;
|
|---|
| 269 | $OID =~ s/^\.+//;
|
|---|
| 270 |
|
|---|
| 271 | my ($headOID, $tailOID) = $OID =~ /^(\d+)(.*)$/;
|
|---|
| 272 | $tailOID = "" unless defined $tailOID;
|
|---|
| 273 |
|
|---|
| 274 | if (!defined $headOID) {
|
|---|
| 275 | $classifyinfo->{'Title'} = $title;
|
|---|
| 276 | $classifyinfo->{'classifytype'} = $classifytype;
|
|---|
| 277 | return $classifyinfo;
|
|---|
| 278 | }
|
|---|
| 279 |
|
|---|
| 280 | $classifyinfo->{'contains'} = [] unless defined $classifyinfo->{'contains'};
|
|---|
| 281 | if ($self->{'documents_last'}) {
|
|---|
| 282 | # documents should come after nodes in the classifier
|
|---|
| 283 |
|
|---|
| 284 | my $doc_pos = 0;
|
|---|
| 285 | foreach my $thing (@{$classifyinfo->{'contains'}}) {
|
|---|
| 286 | last if defined $thing->{'OID'};
|
|---|
| 287 | $doc_pos++;
|
|---|
| 288 | }
|
|---|
| 289 |
|
|---|
| 290 | while ($doc_pos < $headOID) {
|
|---|
| 291 | splice(@{$classifyinfo->{'contains'}}, $doc_pos, 0, $self->get_entry("", $classifytype));
|
|---|
| 292 | $doc_pos++;
|
|---|
| 293 | }
|
|---|
| 294 |
|
|---|
| 295 | return $self->get_OID_entry ($tailOID, $classifyinfo->{'contains'}->[($headOID-1)], $title, $classifytype);
|
|---|
| 296 |
|
|---|
| 297 | }
|
|---|
| 298 |
|
|---|
| 299 | # else, documents come before nodes
|
|---|
| 300 | my $offset = 0;
|
|---|
| 301 | foreach my $thing (@{$classifyinfo->{'contains'}}) {
|
|---|
| 302 | $offset ++ if defined $thing->{'OID'};
|
|---|
| 303 | }
|
|---|
| 304 |
|
|---|
| 305 | while (scalar(@{$classifyinfo->{'contains'}}) < ($headOID+$offset)) {
|
|---|
| 306 | push (@{$classifyinfo->{'contains'}}, $self->get_entry("", $classifytype));
|
|---|
| 307 | }
|
|---|
| 308 |
|
|---|
| 309 | return $self->get_OID_entry ($tailOID, $classifyinfo->{'contains'}->[($headOID+$offset-1)], $title, $classifytype);
|
|---|
| 310 | }
|
|---|
| 311 |
|
|---|
| 312 | sub get_entry {
|
|---|
| 313 | my $self = shift (@_);
|
|---|
| 314 | my ($title, $childtype, $thistype) = @_;
|
|---|
| 315 | my $memberof = &supports_memberof();
|
|---|
| 316 |
|
|---|
| 317 | # organise into classification structure
|
|---|
| 318 | my %classifyinfo = ('childtype'=>$childtype,
|
|---|
| 319 | 'Title'=>$title,
|
|---|
| 320 | 'supportsmemberof'=>$memberof,
|
|---|
| 321 | 'contains'=>[]);
|
|---|
| 322 | $classifyinfo{'thistype'} = $thistype
|
|---|
| 323 | if defined $thistype && $thistype =~ /\w/;
|
|---|
| 324 |
|
|---|
| 325 | return \%classifyinfo;
|
|---|
| 326 | }
|
|---|
| 327 |
|
|---|
| 328 |
|
|---|
| 329 | 1;
|
|---|
