| 1 | ###########################################################################
|
|---|
| 2 | #
|
|---|
| 3 | # List.pm -- A general and flexible list classifier with most of
|
|---|
| 4 | # the abilities of AZCompactList, and better Unicode,
|
|---|
| 5 | # metadata and sorting capabilities.
|
|---|
| 6 | #
|
|---|
| 7 | # A component of the Greenstone digital library software
|
|---|
| 8 | # from the New Zealand Digital Library Project at the
|
|---|
| 9 | # University of Waikato, New Zealand.
|
|---|
| 10 | #
|
|---|
| 11 | # Author: Michael Dewsnip, NZDL Project, University of Waikato, NZ
|
|---|
| 12 | #
|
|---|
| 13 | # Copyright (C) 2005 New Zealand Digital Library Project
|
|---|
| 14 | #
|
|---|
| 15 | # This program is free software; you can redistribute it and/or modify
|
|---|
| 16 | # it under the terms of the GNU General Public License as published by
|
|---|
| 17 | # the Free Software Foundation; either version 2 of the License, or
|
|---|
| 18 | # (at your option) any later version.
|
|---|
| 19 | #
|
|---|
| 20 | # This program is distributed in the hope that it will be useful,
|
|---|
| 21 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|---|
| 22 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|---|
| 23 | # GNU General Public License for more details.
|
|---|
| 24 | #
|
|---|
| 25 | # You should have received a copy of the GNU General Public License
|
|---|
| 26 | # along with this program; if not, write to the Free Software
|
|---|
| 27 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
|---|
| 28 | #
|
|---|
| 29 | # TO DO: - Remove punctuation from metadata values before sorting.
|
|---|
| 30 | # - Add an AZCompactList-style hlist option?
|
|---|
| 31 | #
|
|---|
| 32 | ###########################################################################
|
|---|
| 33 |
|
|---|
| 34 | package List;
|
|---|
| 35 |
|
|---|
| 36 |
|
|---|
| 37 | use BaseClassifier;
|
|---|
| 38 |
|
|---|
| 39 | use strict;
|
|---|
| 40 |
|
|---|
| 41 |
|
|---|
| 42 | sub BEGIN {
|
|---|
| 43 | @List::ISA = ('BaseClassifier');
|
|---|
| 44 | }
|
|---|
| 45 |
|
|---|
| 46 | my $partition_type_list =
|
|---|
| 47 | [ { 'name' => "per_letter",
|
|---|
| 48 | 'desc' => "{List.level_partition.per_letter}" },
|
|---|
| 49 | { 'name' => "constant_size",
|
|---|
| 50 | 'desc' => "{List.level_partition.constant_size}" },
|
|---|
| 51 | { 'name' => "per_letter_fixed_size",
|
|---|
| 52 | 'desc' => "{List.level_partition.per_letter_fixed_size}"},
|
|---|
| 53 | { 'name' => "none",
|
|---|
| 54 | 'desc' => "{List.level_partition.none}" } ];
|
|---|
| 55 |
|
|---|
| 56 | my $bookshelf_type_list =
|
|---|
| 57 | [ { 'name' => "always",
|
|---|
| 58 | 'desc' => "{List.bookshelf_type.always}" },
|
|---|
| 59 | { 'name' => "duplicate_only",
|
|---|
| 60 | 'desc' => "{List.bookshelf_type.duplicate_only}" },
|
|---|
| 61 | { 'name' => "never",
|
|---|
| 62 | 'desc' => "{List.bookshelf_type.never}" } ];
|
|---|
| 63 |
|
|---|
| 64 | my $arguments =
|
|---|
| 65 | [ { 'name' => "metadata",
|
|---|
| 66 | 'desc' => "{List.metadata}",
|
|---|
| 67 | 'type' => "metadata",
|
|---|
| 68 | 'reqd' => "yes" },
|
|---|
| 69 |
|
|---|
| 70 | # The interesting options
|
|---|
| 71 | { 'name' => "bookshelf_type",
|
|---|
| 72 | 'desc' => "{List.bookshelf_type}",
|
|---|
| 73 | 'type' => "enum",
|
|---|
| 74 | 'list' => $bookshelf_type_list,
|
|---|
| 75 | 'deft' => "never" },
|
|---|
| 76 | { 'name' => "classify_sections",
|
|---|
| 77 | 'desc' => "{List.classify_sections}",
|
|---|
| 78 | 'type' => "flag" },
|
|---|
| 79 | { 'name' => "partition_type_within_level",
|
|---|
| 80 | 'desc' => "{List.partition_type_within_level}",
|
|---|
| 81 | 'type' => "enum",
|
|---|
| 82 | 'list' => $partition_type_list,
|
|---|
| 83 | 'deft' => "per_letter" },
|
|---|
| 84 | { 'name' => "partition_size_within_level",
|
|---|
| 85 | 'desc' => "{List.partition_size_within_level}",
|
|---|
| 86 | 'type' => "int",
|
|---|
| 87 | 'range' => "1,",
|
|---|
| 88 | 'deft' => "39" },
|
|---|
| 89 | { 'name' => "partition_name_length",
|
|---|
| 90 | 'desc' => "{List.partition_name_length}",
|
|---|
| 91 | 'type' => "string" },
|
|---|
| 92 | { 'name' => "sort_leaf_nodes_using",
|
|---|
| 93 | 'desc' => "{List.sort_leaf_nodes_using}",
|
|---|
| 94 | 'type' => "metadata",
|
|---|
| 95 | 'deft' => "Title" },
|
|---|
| 96 | { 'name' => "sort_using_unicode_collation",
|
|---|
| 97 | 'desc' => "{List.sort_using_unicode_collation}",
|
|---|
| 98 | 'type' => "flag" },
|
|---|
| 99 | { 'name' => "use_hlist_for",
|
|---|
| 100 | 'desc' => "{List.use_hlist_for}",
|
|---|
| 101 | 'type' => "string" },
|
|---|
| 102 | { 'name' => "removeprefix",
|
|---|
| 103 | 'desc' => "{BasClas.removeprefix}",
|
|---|
| 104 | 'type' => "regexp" },
|
|---|
| 105 | { 'name' => "removesuffix",
|
|---|
| 106 | 'desc' => "{BasClas.removesuffix}",
|
|---|
| 107 | 'type' => "regexp" } ];
|
|---|
| 108 |
|
|---|
| 109 | my $options = { 'name' => "List",
|
|---|
| 110 | 'desc' => "{List.desc}",
|
|---|
| 111 | 'abstract' => "no",
|
|---|
| 112 | 'inherits' => "yes",
|
|---|
| 113 | 'args' => $arguments };
|
|---|
| 114 |
|
|---|
| 115 |
|
|---|
| 116 | sub new
|
|---|
| 117 | {
|
|---|
| 118 | my ($class) = shift(@_);
|
|---|
| 119 | my ($classifierslist, $inputargs, $hashArgOptLists) = @_;
|
|---|
| 120 | push(@$classifierslist, $class);
|
|---|
| 121 |
|
|---|
| 122 | push(@{$hashArgOptLists->{"ArgList"}}, @{$arguments});
|
|---|
| 123 | push(@{$hashArgOptLists->{"OptList"}}, $options);
|
|---|
| 124 |
|
|---|
| 125 | my $self = new BaseClassifier($classifierslist, $inputargs, $hashArgOptLists);
|
|---|
| 126 |
|
|---|
| 127 | if ($self->{'info_only'}) {
|
|---|
| 128 | # don't worry about any options etc
|
|---|
| 129 | return bless $self, $class;
|
|---|
| 130 | }
|
|---|
| 131 |
|
|---|
| 132 | # The metadata elements to use (required)
|
|---|
| 133 | if (!$self->{'metadata'}) {
|
|---|
| 134 | die "Error: No metadata fields specified for List.\n";
|
|---|
| 135 | }
|
|---|
| 136 | my @metadata_groups = split(/\//, $self->{'metadata'});
|
|---|
| 137 | $self->{'metadata_groups'} = \@metadata_groups;
|
|---|
| 138 |
|
|---|
| 139 | # The classifier button name (default: the first metadata element specified)
|
|---|
| 140 | if (!$self->{'buttonname'}) {
|
|---|
| 141 | my $first_metadata_group = $metadata_groups[0];
|
|---|
| 142 | my $first_metadata_element = (split(/\;|,/, $first_metadata_group))[0];
|
|---|
| 143 | $self->{'buttonname'} = $self->generate_title_from_metadata($first_metadata_element);
|
|---|
| 144 | }
|
|---|
| 145 |
|
|---|
| 146 | # Whether to group items into a bookshelf, (must be 'always' for all metadata fields except the last)
|
|---|
| 147 | foreach my $metadata_group (@metadata_groups) {
|
|---|
| 148 | $self->{$metadata_group . ".bookshelf_type"} = "always";
|
|---|
| 149 | }
|
|---|
| 150 | my $last_metadata_group = $metadata_groups[$#metadata_groups];
|
|---|
| 151 | # Default: duplicate_only, ie. leave leaf nodes ungrouped (equivalent to AZCompactList -mingroup 2)
|
|---|
| 152 | $self->{$last_metadata_group . ".bookshelf_type"} = $self->{'bookshelf_type'};
|
|---|
| 153 |
|
|---|
| 154 | # Whether to use an hlist or a vlist for each level in the hierarchy (default: vlist)
|
|---|
| 155 | foreach my $metadata_group (@metadata_groups) {
|
|---|
| 156 | $self->{$metadata_group . ".list_type"} = "VList";
|
|---|
| 157 | }
|
|---|
| 158 | foreach my $metadata_group (split(/\,/, $self->{'use_hlist_for'})) {
|
|---|
| 159 | $self->{$metadata_group . ".list_type"} = "HList";
|
|---|
| 160 | }
|
|---|
| 161 |
|
|---|
| 162 | # How the items are grouped into partitions (default: no partition)
|
|---|
| 163 | # for each level (metadata group), separated by '/'
|
|---|
| 164 | if (!$self->{"partition_type_within_level"}) {
|
|---|
| 165 | foreach my $metadata_group (@metadata_groups) {
|
|---|
| 166 | $self->{$metadata_group . ".partition_type_within_level"} = "none";
|
|---|
| 167 | }
|
|---|
| 168 | } else {
|
|---|
| 169 | my @partition_type_within_levellist = split(/\//, $self->{'partition_type_within_level'});
|
|---|
| 170 | foreach my $metadata_group (@metadata_groups) {
|
|---|
| 171 | my $partition_type_within_levelelem = shift(@partition_type_within_levellist);
|
|---|
| 172 | if (defined($partition_type_within_levelelem)) {
|
|---|
| 173 | $self->{$metadata_group . ".partition_type_within_level"} = $partition_type_within_levelelem;
|
|---|
| 174 | }
|
|---|
| 175 | else {
|
|---|
| 176 | $self->{$metadata_group . ".partition_type_within_level"} = $self->{$metadata_groups[0] . ".partition_type_within_level"};
|
|---|
| 177 | }
|
|---|
| 178 | }
|
|---|
| 179 | }
|
|---|
| 180 |
|
|---|
| 181 | # The number of items in each partition
|
|---|
| 182 | if (!$self->{'partition_size_within_level'}) {
|
|---|
| 183 | # Default: 20
|
|---|
| 184 | foreach my $metadata_group (@metadata_groups) {
|
|---|
| 185 | $self->{$metadata_group . ".partition_size_within_level"} = 20;
|
|---|
| 186 | }
|
|---|
| 187 | }
|
|---|
| 188 | else {
|
|---|
| 189 | my @partition_size_within_levellist = split(/\//, $self->{'partition_size_within_level'});
|
|---|
| 190 |
|
|---|
| 191 | # Assign values based on the partition_size_within_level parameter
|
|---|
| 192 | foreach my $metadata_group (@metadata_groups) {
|
|---|
| 193 | my $partition_size_within_levelelem = shift(@partition_size_within_levellist);
|
|---|
| 194 | if (defined($partition_size_within_levelelem)) {
|
|---|
| 195 | $self->{$metadata_group . ".partition_size_within_level"} = $partition_size_within_levelelem;
|
|---|
| 196 | }
|
|---|
| 197 | else {
|
|---|
| 198 | $self->{$metadata_group . ".partition_size_within_level"} = $self->{$metadata_groups[0] . ".partition_size_within_level"};
|
|---|
| 199 | }
|
|---|
| 200 | }
|
|---|
| 201 | }
|
|---|
| 202 |
|
|---|
| 203 | # The removeprefix and removesuffix expressions
|
|---|
| 204 | if ($self->{'removeprefix'}) {
|
|---|
| 205 | # If there are more than one expressions, use '' to quote each experession and '/' to separate
|
|---|
| 206 | my @removeprefix_exprs_within_levellist = split(/'\/'/, $self->{'removeprefix'});
|
|---|
| 207 |
|
|---|
| 208 | foreach my $metadata_group (@metadata_groups) {
|
|---|
| 209 | my $removeprefix_expr_within_levelelem = shift(@removeprefix_exprs_within_levellist);
|
|---|
| 210 | if (defined($removeprefix_expr_within_levelelem) && $removeprefix_expr_within_levelelem ne "") {
|
|---|
| 211 | # Remove the other ' at the beginning and the end if there is any
|
|---|
| 212 | $removeprefix_expr_within_levelelem =~ s/^'//;
|
|---|
| 213 | $removeprefix_expr_within_levelelem =~ s/'$//;
|
|---|
| 214 | # Remove the extra ^ at the beginning
|
|---|
| 215 | $removeprefix_expr_within_levelelem =~ s/^\^//;
|
|---|
| 216 | $self->{$metadata_group . ".remove_prefix_expr"} = $removeprefix_expr_within_levelelem;
|
|---|
| 217 | } else {
|
|---|
| 218 | $self->{$metadata_group . ".remove_prefix_expr"} = $self->{$metadata_groups[0] . ".remove_prefix_expr"};
|
|---|
| 219 | }
|
|---|
| 220 | }
|
|---|
| 221 | }
|
|---|
| 222 | if ($self->{'removesuffix'}) {
|
|---|
| 223 | my @removesuffix_exprs_within_levellist = split(/'\/'/, $self->{'removesuffix'});
|
|---|
| 224 |
|
|---|
| 225 | foreach my $metadata_group (@metadata_groups) {
|
|---|
| 226 | my $removesuffix_expr_within_levelelem = shift(@removesuffix_exprs_within_levellist);
|
|---|
| 227 | if (defined($removesuffix_expr_within_levelelem) && $removesuffix_expr_within_levelelem ne "") {
|
|---|
| 228 | $removesuffix_expr_within_levelelem =~ s/^'//;
|
|---|
| 229 | $removesuffix_expr_within_levelelem =~ s/'$//;
|
|---|
| 230 | # Remove the extra $ at the end
|
|---|
| 231 | $removesuffix_expr_within_levelelem =~ s/\$$//;
|
|---|
| 232 | $self->{$metadata_group . ".remove_suffix_expr"} = $removesuffix_expr_within_levelelem;
|
|---|
| 233 | } else {
|
|---|
| 234 | $self->{$metadata_group . ".remove_suffix_expr"} = $self->{$metadata_groups[0] . ".remove_suffix_expr"};
|
|---|
| 235 | }
|
|---|
| 236 | }
|
|---|
| 237 | }
|
|---|
| 238 |
|
|---|
| 239 | # The metadata elements to use to sort the leaf nodes (default: Title)
|
|---|
| 240 | my @sort_leaf_nodes_using_metadata_groups = ( "Title" );
|
|---|
| 241 | if ($self->{'sort_leaf_nodes_using'}) {
|
|---|
| 242 | @sort_leaf_nodes_using_metadata_groups = split(/\|/, $self->{'sort_leaf_nodes_using'});
|
|---|
| 243 | }
|
|---|
| 244 | $self->{'sort_leaf_nodes_using_metadata_groups'} = \@sort_leaf_nodes_using_metadata_groups;
|
|---|
| 245 |
|
|---|
| 246 | # Create an instance of the Unicode::Collate object if better Unicode sorting is desired
|
|---|
| 247 | if ($self->{'sort_using_unicode_collation'}) {
|
|---|
| 248 | # To use this you first need to download the allkeys.txt file from
|
|---|
| 249 | # http://www.unicode.org/Public/UCA/latest/allkeys.txt and put it in the Perl
|
|---|
| 250 | # Unicode/Collate directory.
|
|---|
| 251 | require Unicode::Collate;
|
|---|
| 252 | $self->{'unicode_collator'} = Unicode::Collate->new();
|
|---|
| 253 | }
|
|---|
| 254 |
|
|---|
| 255 | # An empty array for the document OIDs
|
|---|
| 256 | $self->{'OIDs'} = [];
|
|---|
| 257 |
|
|---|
| 258 | return bless $self, $class;
|
|---|
| 259 | }
|
|---|
| 260 |
|
|---|
| 261 |
|
|---|
| 262 | sub init
|
|---|
| 263 | {
|
|---|
| 264 | # Nothing to do...
|
|---|
| 265 | }
|
|---|
| 266 |
|
|---|
| 267 |
|
|---|
| 268 | # Called for each document in the collection
|
|---|
| 269 | sub classify
|
|---|
| 270 | {
|
|---|
| 271 | my $self = shift(@_);
|
|---|
| 272 | my ($doc_obj,$edit_mode) = @_;
|
|---|
| 273 |
|
|---|
| 274 | # If "-classify_sections" is set, classify every section of the document
|
|---|
| 275 | if ($self->{'classify_sections'}) {
|
|---|
| 276 | my $section = $doc_obj->get_next_section($doc_obj->get_top_section());
|
|---|
| 277 | while (defined $section) {
|
|---|
| 278 | $self->classify_section($doc_obj, $doc_obj->get_OID() . ".$section", $section, $edit_mode);
|
|---|
| 279 | $section = $doc_obj->get_next_section($section);
|
|---|
| 280 | }
|
|---|
| 281 | }
|
|---|
| 282 | # Otherwise just classify the top document section
|
|---|
| 283 | else {
|
|---|
| 284 | $self->classify_section($doc_obj, $doc_obj->get_OID(), $doc_obj->get_top_section(), $edit_mode);
|
|---|
| 285 | }
|
|---|
| 286 | }
|
|---|
| 287 |
|
|---|
| 288 |
|
|---|
| 289 | sub classify_section
|
|---|
| 290 | {
|
|---|
| 291 | my $self = shift(@_);
|
|---|
| 292 | my ($doc_obj,$section_OID,$section,$edit_mode) = @_;
|
|---|
| 293 |
|
|---|
| 294 | my @metadata_groups = @{$self->{'metadata_groups'}};
|
|---|
| 295 |
|
|---|
| 296 | # Only classify the section if it has a value for one of the metadata elements in the first group
|
|---|
| 297 | my $classify_section = 0;
|
|---|
| 298 | my $first_metadata_group = $metadata_groups[0];
|
|---|
| 299 | foreach my $first_metadata_group_element (split(/\;|,/, $first_metadata_group)) {
|
|---|
| 300 | my $real_first_metadata_group_element = $first_metadata_group_element;
|
|---|
| 301 | $real_first_metadata_group_element =~ s/^ex\.//;
|
|---|
| 302 | my $first_metadata_group_element_value = $doc_obj->get_metadata_element($section, $real_first_metadata_group_element);
|
|---|
| 303 |
|
|---|
| 304 | # Remove prefix/suffix if requested
|
|---|
| 305 | my $remove_prefix_expr = $self->{$first_metadata_group_element . ".remove_prefix_expr"};
|
|---|
| 306 | if (defined $remove_prefix_expr && $remove_prefix_expr ne "") {
|
|---|
| 307 | $first_metadata_group_element_value =~ s/^$remove_prefix_expr//;
|
|---|
| 308 | }
|
|---|
| 309 | my $remove_suffix_expr = $self->{$first_metadata_group_element . ".remove_suffix_expr"};
|
|---|
| 310 | if (defined $remove_suffix_expr && $remove_suffix_expr ne "") {
|
|---|
| 311 | $first_metadata_group_element_value =~ s/$remove_suffix_expr$//;
|
|---|
| 312 | }
|
|---|
| 313 |
|
|---|
| 314 | if (defined($first_metadata_group_element_value) && $first_metadata_group_element_value ne "") {
|
|---|
| 315 | # This section must be included in the classifier
|
|---|
| 316 | $classify_section = 1;
|
|---|
| 317 | last;
|
|---|
| 318 | }
|
|---|
| 319 | }
|
|---|
| 320 |
|
|---|
| 321 | # We're not classifying this section because it doesn't have the required metadata
|
|---|
| 322 | return if (!$classify_section);
|
|---|
| 323 |
|
|---|
| 324 | if (($edit_mode eq "delete") || ($edit_mode eq "update")) {
|
|---|
| 325 | $self->oid_array_delete($section_OID,'OIDs');
|
|---|
| 326 | if ($edit_mode eq "delete") {
|
|---|
| 327 | return;
|
|---|
| 328 | }
|
|---|
| 329 | }
|
|---|
| 330 |
|
|---|
| 331 | # Otherwise, include this section in the classifier
|
|---|
| 332 | push(@{$self->{'OIDs'}}, $section_OID);
|
|---|
| 333 |
|
|---|
| 334 | # Create a hash for the metadata values of each metadata element we're interested in
|
|---|
| 335 | my %metadata_groups_done = ();
|
|---|
| 336 | foreach my $metadata_group (@metadata_groups, @{$self->{'sort_leaf_nodes_using_metadata_groups'}}) {
|
|---|
| 337 | # Take care not to do a metadata group more than once
|
|---|
| 338 | unless ($metadata_groups_done{$metadata_group}) {
|
|---|
| 339 | foreach my $metadata_element (split(/\;|,/, $metadata_group)) {
|
|---|
| 340 | my $real_metadata_element = $metadata_element;
|
|---|
| 341 | $real_metadata_element =~ s/^ex\.//;
|
|---|
| 342 | my $remove_prefix_expr = $self->{$metadata_element . ".remove_prefix_expr"};
|
|---|
| 343 | my $remove_suffix_expr = $self->{$metadata_element . ".remove_suffix_expr"};
|
|---|
| 344 | my @metadata_values = @{$doc_obj->get_metadata($section, $real_metadata_element)};
|
|---|
| 345 | foreach my $metadata_value (@metadata_values) {
|
|---|
| 346 | # Strip leading and trailing whitespace
|
|---|
| 347 | $metadata_value =~ s/^\s*//;
|
|---|
| 348 | $metadata_value =~ s/\s*$//;
|
|---|
| 349 |
|
|---|
| 350 | # Remove prefix/suffix if requested
|
|---|
| 351 | if (defined $remove_prefix_expr && $remove_prefix_expr ne "") {
|
|---|
| 352 | $metadata_value =~ s/^$remove_prefix_expr//;
|
|---|
| 353 | }
|
|---|
| 354 | if (defined $remove_suffix_expr && $remove_suffix_expr ne "") {
|
|---|
| 355 | $metadata_value =~ s/$remove_suffix_expr$//;
|
|---|
| 356 | }
|
|---|
| 357 |
|
|---|
| 358 | # Convert the metadata value from a UTF-8 string to a Unicode string
|
|---|
| 359 | # This means that length() and substr() work properly
|
|---|
| 360 | # We need to be careful to convert classifier node title values back to UTF-8, however
|
|---|
| 361 | my $metadata_value_unicode_string = $self->convert_utf8_string_to_unicode_string($metadata_value);
|
|---|
| 362 |
|
|---|
| 363 | # Add the metadata value into the list for this combination of metadata group and section
|
|---|
| 364 | push(@{$self->{$metadata_group . ".list"}->{$section_OID}}, $metadata_value_unicode_string);
|
|---|
| 365 | }
|
|---|
| 366 | last if (@metadata_values > 0);
|
|---|
| 367 | }
|
|---|
| 368 |
|
|---|
| 369 | $metadata_groups_done{$metadata_group} = 1;
|
|---|
| 370 | }
|
|---|
| 371 | }
|
|---|
| 372 | }
|
|---|
| 373 |
|
|---|
| 374 |
|
|---|
| 375 | sub get_classify_info
|
|---|
| 376 | {
|
|---|
| 377 | my $self = shift(@_);
|
|---|
| 378 |
|
|---|
| 379 | # The metadata groups to classify by
|
|---|
| 380 | my @metadata_groups = @{$self->{'metadata_groups'}};
|
|---|
| 381 | my $first_metadata_group = $metadata_groups[0];
|
|---|
| 382 |
|
|---|
| 383 | # The OID values of the documents to include in the classifier
|
|---|
| 384 | my @OIDs = @{$self->{'OIDs'}};
|
|---|
| 385 |
|
|---|
| 386 | # Create the root node of the classification hierarchy
|
|---|
| 387 | my %classifier_node = ( 'thistype' => "Invisible",
|
|---|
| 388 | 'childtype' => $self->{$first_metadata_group . ".list_type"},
|
|---|
| 389 | 'Title' => $self->{'buttonname'},
|
|---|
| 390 | 'contains' => [],
|
|---|
| 391 | 'mdtype' => $first_metadata_group );
|
|---|
| 392 |
|
|---|
| 393 | # Recursively create the classification hierarchy, one level for each metadata group
|
|---|
| 394 | $self->add_level(\@metadata_groups, \@OIDs, \%classifier_node);
|
|---|
| 395 | return \%classifier_node;
|
|---|
| 396 | }
|
|---|
| 397 |
|
|---|
| 398 |
|
|---|
| 399 | sub add_level
|
|---|
| 400 | {
|
|---|
| 401 | my $self = shift(@_);
|
|---|
| 402 | my @metadata_groups = @{shift(@_)};
|
|---|
| 403 | my @OIDs = @{shift(@_)};
|
|---|
| 404 | my $classifier_node = shift(@_);
|
|---|
| 405 | # print STDERR "\nAdding AZ list for " . $classifier_node->{'Title'} . "\n";
|
|---|
| 406 |
|
|---|
| 407 | my $metadata_group = $metadata_groups[0];
|
|---|
| 408 | # print STDERR "Processing metadata group: " . $metadata_group . "\n";
|
|---|
| 409 | # print STDERR "Number of OID values: " . @OIDs . "\n";
|
|---|
| 410 |
|
|---|
| 411 | if (!defined($self->{$metadata_group . ".list"})) {
|
|---|
| 412 | print STDERR "Warning: No metadata values assigned to $metadata_group.\n";
|
|---|
| 413 | return;
|
|---|
| 414 | }
|
|---|
| 415 |
|
|---|
| 416 | # Create a mapping from metadata value to OID
|
|---|
| 417 | my $OID_to_metadata_values_hash_ref = $self->{$metadata_group . ".list"};
|
|---|
| 418 | my %metadata_value_to_OIDs_hash = ();
|
|---|
| 419 | foreach my $OID (@OIDs)
|
|---|
| 420 | {
|
|---|
| 421 | if ($OID_to_metadata_values_hash_ref->{$OID})
|
|---|
| 422 | {
|
|---|
| 423 | my @metadata_values = @{$OID_to_metadata_values_hash_ref->{$OID}};
|
|---|
| 424 | foreach my $metadata_value (@metadata_values)
|
|---|
| 425 | {
|
|---|
| 426 | push(@{$metadata_value_to_OIDs_hash{$metadata_value}}, $OID);
|
|---|
| 427 | }
|
|---|
| 428 | }
|
|---|
| 429 | }
|
|---|
| 430 | # print STDERR "Number of distinct values: " . scalar(keys %metadata_value_to_OIDs_hash) . "\n";
|
|---|
| 431 |
|
|---|
| 432 | # Partition the values (if necessary)
|
|---|
| 433 | my $partition_type_within_level = $self->{$metadata_group . ".partition_type_within_level"};
|
|---|
| 434 | if ($partition_type_within_level =~ /^per_letter$/i) {
|
|---|
| 435 | # Generate one hlist for each letter
|
|---|
| 436 | my @sortedmetadata_values = $self->sort_metadata_values_array(keys(%metadata_value_to_OIDs_hash));
|
|---|
| 437 | my %metadata_value_to_OIDs_subhash = ();
|
|---|
| 438 |
|
|---|
| 439 | my $lastpartition = substr($sortedmetadata_values[0], 0, 1);
|
|---|
| 440 | foreach my $metadata_value (@sortedmetadata_values) {
|
|---|
| 441 | my $metadata_valuepartition = substr($metadata_value, 0, 1);
|
|---|
| 442 |
|
|---|
| 443 | # Is this the start of a new partition?
|
|---|
| 444 | if ($metadata_valuepartition ne $lastpartition) {
|
|---|
| 445 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $lastpartition, \%metadata_value_to_OIDs_subhash);
|
|---|
| 446 | %metadata_value_to_OIDs_subhash = ();
|
|---|
| 447 | $lastpartition = $metadata_valuepartition;
|
|---|
| 448 | }
|
|---|
| 449 |
|
|---|
| 450 | $metadata_value_to_OIDs_subhash{$metadata_value} = $metadata_value_to_OIDs_hash{$metadata_value};
|
|---|
| 451 | }
|
|---|
| 452 |
|
|---|
| 453 | # Don't forget to add the last partition
|
|---|
| 454 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $lastpartition, \%metadata_value_to_OIDs_subhash);
|
|---|
| 455 |
|
|---|
| 456 | # The partitions are stored in an HList
|
|---|
| 457 | $classifier_node->{'childtype'} = "HList";
|
|---|
| 458 | }
|
|---|
| 459 | elsif ($partition_type_within_level =~ /^per_letter_fixed_size$/i) {
|
|---|
| 460 | # Generate hlist based on the first letter of the metadata value (like per_letter) but with restriction on the partition size
|
|---|
| 461 | # If a letter has fewer items than specified by the "partition_size_within_level", then group them together if possible
|
|---|
| 462 | # If a letter has more items than specified, split into several hlists.
|
|---|
| 463 | # Depends on the bookshelf_type, one item can be either a document (when bookshelf_type is "never") or a metadata value (otherwise)
|
|---|
| 464 | my $partition_size_within_level = $self->{$metadata_group . ".partition_size_within_level"};
|
|---|
| 465 | my @sortedmetadata_values = $self->sort_metadata_values_array(keys(%metadata_value_to_OIDs_hash));
|
|---|
| 466 | my $bookshelf_type = $self->{$metadata_group . ".bookshelf_type"};
|
|---|
| 467 |
|
|---|
| 468 | # Separate values by their first letter, each form a bucket, like the per_letter partition type
|
|---|
| 469 | my $last_partition = substr($sortedmetadata_values[0], 0, 1);
|
|---|
| 470 | my %partition_buckets = ();
|
|---|
| 471 | my @metadata_values_in_bucket = ();
|
|---|
| 472 | my $num_items_in_bucket = 0;
|
|---|
| 473 | foreach my $metadata_value (@sortedmetadata_values) {
|
|---|
| 474 | my $metadata_valuepartition = substr($metadata_value, 0, 1);
|
|---|
| 475 | if ($metadata_valuepartition ne $last_partition) {
|
|---|
| 476 | my @temp_array = @metadata_values_in_bucket;
|
|---|
| 477 | # Cache the values that belong to this bucket, and the number of items in this bucket, not necessary to be the same number as the metadata values
|
|---|
| 478 | my %partition_info = ();
|
|---|
| 479 | $partition_info{'metadata_values'} = \@temp_array;
|
|---|
| 480 | $partition_info{'size'} = $num_items_in_bucket;
|
|---|
| 481 | $partition_buckets{$last_partition} = \%partition_info;
|
|---|
| 482 |
|
|---|
| 483 | @metadata_values_in_bucket = ($metadata_value);
|
|---|
| 484 | $num_items_in_bucket = $bookshelf_type eq "never" ? scalar(@{$metadata_value_to_OIDs_hash{$metadata_value}}) : scalar(@metadata_values_in_bucket);
|
|---|
| 485 | $last_partition = $metadata_valuepartition;
|
|---|
| 486 | } else {
|
|---|
| 487 | $num_items_in_bucket += $bookshelf_type eq "never" ? scalar(@{$metadata_value_to_OIDs_hash{$metadata_value}}) : scalar(@metadata_values_in_bucket);
|
|---|
| 488 | push (@metadata_values_in_bucket, $metadata_value);
|
|---|
| 489 | }
|
|---|
| 490 | }
|
|---|
| 491 | # Last one
|
|---|
| 492 | my %partition_info = ();
|
|---|
| 493 | $partition_info{'metadata_values'} = \@metadata_values_in_bucket;
|
|---|
| 494 | $partition_info{'size'} = $num_items_in_bucket;
|
|---|
| 495 | $partition_buckets{$last_partition} = \%partition_info;
|
|---|
| 496 |
|
|---|
| 497 | my @partition_keys = $self->sort_metadata_values_array(keys(%partition_buckets));
|
|---|
| 498 | for (my $i = 0; $i < scalar(@partition_keys) - 1; $i++) {
|
|---|
| 499 | my $partition = $partition_keys[$i];
|
|---|
| 500 | my $items_in_partition = $partition_buckets{$partition}->{'size'};
|
|---|
| 501 | # Merge small buckets together, but keep the numeric bucket apart
|
|---|
| 502 | if ($items_in_partition < $partition_size_within_level) {
|
|---|
| 503 | my $items_in_next_partition = $partition_buckets{$partition_keys[$i+1]}->{'size'};
|
|---|
| 504 | if ($items_in_partition + $items_in_next_partition <= $partition_size_within_level
|
|---|
| 505 | && !(($partition =~ /^[^0-9]/ && $partition_keys[$i+1] =~ /^[0-9]/)
|
|---|
| 506 | || ($partition =~ /^[0-9]/ && $partition_keys[$i+1] =~ /^[^0-9]/))) {
|
|---|
| 507 | foreach my $metadata_value_to_merge (@{$partition_buckets{$partition}->{'metadata_values'}}) {
|
|---|
| 508 | push(@{$partition_buckets{$partition_keys[$i+1]}->{'metadata_values'}}, $metadata_value_to_merge);
|
|---|
| 509 | }
|
|---|
| 510 | $partition_buckets{$partition_keys[$i+1]}->{'size'} += $items_in_partition;
|
|---|
| 511 | delete $partition_buckets{$partition};
|
|---|
| 512 | }
|
|---|
| 513 | }
|
|---|
| 514 | }
|
|---|
| 515 | @partition_keys = $self->sort_metadata_values_array(keys(%partition_buckets));
|
|---|
| 516 |
|
|---|
| 517 | # Add partitions, and divide big bucket into several
|
|---|
| 518 | my $last_partition_end = "";
|
|---|
| 519 | my $partition_start = "";
|
|---|
| 520 | foreach my $partition (@partition_keys) {
|
|---|
| 521 | my @metadata_values = $self->sort_metadata_values_array(@{$partition_buckets{$partition}->{'metadata_values'}});
|
|---|
| 522 | my $items_in_partition = $partition_buckets{$partition}->{'size'};
|
|---|
| 523 | $partition_start = $self->generate_partition_start($metadata_values[0], $last_partition_end, $self->{"partition_name_length"});
|
|---|
| 524 |
|
|---|
| 525 | if ($items_in_partition > $partition_size_within_level) {
|
|---|
| 526 | my $items_done = 0;
|
|---|
| 527 | my %metadata_values_to_OIDs_subhashes = ();
|
|---|
| 528 | for (my $i = 0; $i < scalar(@metadata_values); $i++) {
|
|---|
| 529 | my $metadata_value = $metadata_values[$i];
|
|---|
| 530 | # If the bookshelf_type is "never", count the documents, otherwise count the distinct metadata values
|
|---|
| 531 | my $items_for_this_md_value = $bookshelf_type eq "never" ? scalar(@{$metadata_value_to_OIDs_hash{$metadata_value}}) : 1;
|
|---|
| 532 |
|
|---|
| 533 | my $partitionend = $self->generate_partition_end($metadata_value, $partition_start, $self->{"partition_name_length"});
|
|---|
| 534 | my $partitionname = $partition_start;
|
|---|
| 535 | if ($partitionend ne $partition_start) {
|
|---|
| 536 | $partitionname = $partitionname . "-" . $partitionend;
|
|---|
| 537 | }
|
|---|
| 538 |
|
|---|
| 539 | # Start a new partition
|
|---|
| 540 | if ($items_done + $items_for_this_md_value > $partition_size_within_level && $items_done != 0) {
|
|---|
| 541 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $partitionname, \%metadata_values_to_OIDs_subhashes);
|
|---|
| 542 | $last_partition_end = $partitionend;
|
|---|
| 543 | $partition_start = $self->generate_partition_start($metadata_value, $last_partition_end, $self->{"partition_name_length"});
|
|---|
| 544 | $items_done = 0;
|
|---|
| 545 | %metadata_values_to_OIDs_subhashes = ();
|
|---|
| 546 | }
|
|---|
| 547 |
|
|---|
| 548 | # If bookshelf_type is "never" and the current metadata value holds too many items, need to split into several partitions
|
|---|
| 549 | if ($bookshelf_type eq "never" && $items_for_this_md_value > $partition_size_within_level) {
|
|---|
| 550 | my $partitionname_for_this_value = $self->generate_partition_start($metadata_value, $last_partition_end, $self->{"partition_name_length"});
|
|---|
| 551 | # Get the number of partitions needed for this value
|
|---|
| 552 | my $num_splits = int($items_for_this_md_value / $partition_size_within_level);
|
|---|
| 553 | $num_splits++ if ($items_for_this_md_value / $partition_size_within_level > $num_splits);
|
|---|
| 554 |
|
|---|
| 555 | my @OIDs_for_this_value = @{$metadata_value_to_OIDs_hash{$metadata_value}};
|
|---|
| 556 | for (my $i = 0; $i < $num_splits; $i++) {
|
|---|
| 557 | my %OIDs_subhashes_for_this_value = ();
|
|---|
| 558 | my @OIDs_for_this_partition = ();
|
|---|
| 559 | for (my $d = $i * $partition_size_within_level; $d < (($i+1) * $partition_size_within_level > $items_for_this_md_value ? $items_for_this_md_value : ($i+1) * $partition_size_within_level); $d++) {
|
|---|
| 560 | push (@OIDs_for_this_partition, $OIDs_for_this_value[$d]);
|
|---|
| 561 | }
|
|---|
| 562 |
|
|---|
| 563 | # The last bucket might have only a few items and need to be merged with buckets for subsequent metadata values
|
|---|
| 564 | if ($i == $num_splits - 1 && scalar(@OIDs_for_this_partition) < $partition_size_within_level) {
|
|---|
| 565 | $metadata_values_to_OIDs_subhashes{$metadata_value} = \@OIDs_for_this_partition;
|
|---|
| 566 | $items_done += scalar(@OIDs_for_this_partition);
|
|---|
| 567 | next;
|
|---|
| 568 | }
|
|---|
| 569 |
|
|---|
| 570 | # Add an HList for this bucket
|
|---|
| 571 | $OIDs_subhashes_for_this_value{$metadata_value} = \@OIDs_for_this_partition;
|
|---|
| 572 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $partitionname_for_this_value, \%OIDs_subhashes_for_this_value);
|
|---|
| 573 | $last_partition_end = $partitionname_for_this_value;
|
|---|
| 574 | }
|
|---|
| 575 | next;
|
|---|
| 576 | }
|
|---|
| 577 |
|
|---|
| 578 | $metadata_values_to_OIDs_subhashes{$metadata_value} = $metadata_value_to_OIDs_hash{$metadata_value};
|
|---|
| 579 | $items_done += $bookshelf_type eq "never" ? scalar(@{$metadata_values_to_OIDs_subhashes{$metadata_value}}) : 1;
|
|---|
| 580 |
|
|---|
| 581 | # The last partition
|
|---|
| 582 | if($i == scalar(@metadata_values) - 1) {
|
|---|
| 583 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $partitionname, \%metadata_values_to_OIDs_subhashes);
|
|---|
| 584 | }
|
|---|
| 585 | }
|
|---|
| 586 | }
|
|---|
| 587 | else {
|
|---|
| 588 | # The easier case, just add a partition
|
|---|
| 589 | my %metadata_values_to_OIDs_subhashes = ();
|
|---|
| 590 | for (my $i = 0; $i < scalar(@metadata_values); $i++) {
|
|---|
| 591 | my $metadata_value = $metadata_values[$i];
|
|---|
| 592 | $metadata_values_to_OIDs_subhashes{$metadata_value} = $metadata_value_to_OIDs_hash{$metadata_value};
|
|---|
| 593 | }
|
|---|
| 594 | my $last_metadata_value = $metadata_values[scalar(@metadata_values)-1];
|
|---|
| 595 | my $partitionend = $self->generate_partition_end($last_metadata_value, $partition_start, $self->{"partition_name_length"});
|
|---|
| 596 | my $partitionname = $partition_start;
|
|---|
| 597 | if ($partitionend ne $partition_start) {
|
|---|
| 598 | $partitionname = $partitionname . "-" . $partitionend;
|
|---|
| 599 | }
|
|---|
| 600 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $partitionname, \%metadata_values_to_OIDs_subhashes);
|
|---|
| 601 | $last_partition_end = $partitionend;
|
|---|
| 602 | }
|
|---|
| 603 | }
|
|---|
| 604 | }
|
|---|
| 605 | else {
|
|---|
| 606 | # Generate hlists of a certain size
|
|---|
| 607 | my $partition_size_within_level = $self->{$metadata_group . ".partition_size_within_level"};
|
|---|
| 608 | if ($partition_type_within_level =~ /^constant_size$/i && scalar(keys %metadata_value_to_OIDs_hash) > $partition_size_within_level) {
|
|---|
| 609 | my @sortedmetadata_values = $self->sort_metadata_values_array(keys(%metadata_value_to_OIDs_hash));
|
|---|
| 610 | my $itemsdone = 0;
|
|---|
| 611 | my %metadata_value_to_OIDs_subhash = ();
|
|---|
| 612 | my $lastpartitionend = "";
|
|---|
| 613 | my $partitionstart;
|
|---|
| 614 | foreach my $metadata_value (@sortedmetadata_values) {
|
|---|
| 615 | $metadata_value_to_OIDs_subhash{$metadata_value} = $metadata_value_to_OIDs_hash{$metadata_value};
|
|---|
| 616 | $itemsdone++;
|
|---|
| 617 | my $itemsinpartition = scalar(keys %metadata_value_to_OIDs_subhash);
|
|---|
| 618 |
|
|---|
| 619 | # Is this the start of a new partition?
|
|---|
| 620 | if ($itemsinpartition == 1) {
|
|---|
| 621 | $partitionstart = $self->generate_partition_start($metadata_value, $lastpartitionend, $self->{"partition_name_length"});
|
|---|
| 622 | }
|
|---|
| 623 |
|
|---|
| 624 | # Is this the end of the partition?
|
|---|
| 625 | if ($itemsinpartition == $partition_size_within_level || $itemsdone == @sortedmetadata_values) {
|
|---|
| 626 | my $partitionend = $self->generate_partition_end($metadata_value, $partitionstart, $self->{"partition_name_length"});
|
|---|
| 627 | my $partitionname = $partitionstart;
|
|---|
| 628 | if ($partitionend ne $partitionstart) {
|
|---|
| 629 | $partitionname = $partitionname . "-" . $partitionend;
|
|---|
| 630 | }
|
|---|
| 631 |
|
|---|
| 632 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $partitionname, \%metadata_value_to_OIDs_subhash);
|
|---|
| 633 | %metadata_value_to_OIDs_subhash = ();
|
|---|
| 634 | $lastpartitionend = $partitionend;
|
|---|
| 635 | }
|
|---|
| 636 | }
|
|---|
| 637 |
|
|---|
| 638 | # The partitions are stored in an HList
|
|---|
| 639 | $classifier_node->{'childtype'} = "HList";
|
|---|
| 640 | }
|
|---|
| 641 |
|
|---|
| 642 | # Otherwise just add all the values to a VList
|
|---|
| 643 | else {
|
|---|
| 644 | $self->add_vlist(\@metadata_groups, $classifier_node, \%metadata_value_to_OIDs_hash);
|
|---|
| 645 | }
|
|---|
| 646 | }
|
|---|
| 647 | }
|
|---|
| 648 |
|
|---|
| 649 |
|
|---|
| 650 | sub convert_utf8_string_to_unicode_string
|
|---|
| 651 | {
|
|---|
| 652 | my $self = shift(@_);
|
|---|
| 653 | my $utf8_string = shift(@_);
|
|---|
| 654 |
|
|---|
| 655 | my $unicode_string = "";
|
|---|
| 656 | foreach my $unicode_value (@{&unicode::utf82unicode($utf8_string)}) {
|
|---|
| 657 | $unicode_string .= chr($unicode_value);
|
|---|
| 658 | }
|
|---|
| 659 | return $unicode_string;
|
|---|
| 660 | }
|
|---|
| 661 |
|
|---|
| 662 |
|
|---|
| 663 | sub convert_unicode_string_to_utf8_string
|
|---|
| 664 | {
|
|---|
| 665 | my $self = shift(@_);
|
|---|
| 666 | my $unicode_string = shift(@_);
|
|---|
| 667 |
|
|---|
| 668 | my @unicode_array;
|
|---|
| 669 | for (my $i = 0; $i < length($unicode_string); $i++) {
|
|---|
| 670 | push(@unicode_array, ord(substr($unicode_string, $i, 1)));
|
|---|
| 671 | }
|
|---|
| 672 | return &unicode::unicode2utf8(\@unicode_array);
|
|---|
| 673 | }
|
|---|
| 674 |
|
|---|
| 675 |
|
|---|
| 676 | sub generate_partition_start
|
|---|
| 677 | {
|
|---|
| 678 | my $self = shift(@_);
|
|---|
| 679 | my $metadata_value = shift(@_);
|
|---|
| 680 | my $lastpartitionend = shift(@_);
|
|---|
| 681 | my $partition_name_length = shift(@_);
|
|---|
| 682 |
|
|---|
| 683 | if ($partition_name_length) {
|
|---|
| 684 | return substr($metadata_value, 0, $partition_name_length);
|
|---|
| 685 | }
|
|---|
| 686 |
|
|---|
| 687 | my $partitionstart = substr($metadata_value, 0, 1);
|
|---|
| 688 | if ($partitionstart le $lastpartitionend) {
|
|---|
| 689 | $partitionstart = substr($metadata_value, 0, 2);
|
|---|
| 690 | # Give up after three characters
|
|---|
| 691 | if ($partitionstart le $lastpartitionend) {
|
|---|
| 692 | $partitionstart = substr($metadata_value, 0, 3);
|
|---|
| 693 | }
|
|---|
| 694 | }
|
|---|
| 695 |
|
|---|
| 696 | return $partitionstart;
|
|---|
| 697 | }
|
|---|
| 698 |
|
|---|
| 699 |
|
|---|
| 700 | sub generate_partition_end
|
|---|
| 701 | {
|
|---|
| 702 | my $self = shift(@_);
|
|---|
| 703 | my $metadata_value = shift(@_);
|
|---|
| 704 | my $partitionstart = shift(@_);
|
|---|
| 705 | my $partition_name_length = shift(@_);
|
|---|
| 706 |
|
|---|
| 707 | if ($partition_name_length) {
|
|---|
| 708 | return substr($metadata_value, 0, $partition_name_length);
|
|---|
| 709 | }
|
|---|
| 710 |
|
|---|
| 711 | my $partitionend = substr($metadata_value, 0, length($partitionstart));
|
|---|
| 712 | if ($partitionend gt $partitionstart) {
|
|---|
| 713 | $partitionend = substr($metadata_value, 0, 1);
|
|---|
| 714 | if ($partitionend le $partitionstart) {
|
|---|
| 715 | $partitionend = substr($metadata_value, 0, 2);
|
|---|
| 716 | # Give up after three characters
|
|---|
| 717 | if ($partitionend le $partitionstart) {
|
|---|
| 718 | $partitionend = substr($metadata_value, 0, 3);
|
|---|
| 719 | }
|
|---|
| 720 | }
|
|---|
| 721 | }
|
|---|
| 722 |
|
|---|
| 723 | return $partitionend;
|
|---|
| 724 | }
|
|---|
| 725 |
|
|---|
| 726 |
|
|---|
| 727 | sub add_hlist_partition
|
|---|
| 728 | {
|
|---|
| 729 | my $self = shift(@_);
|
|---|
| 730 | my @metadata_groups = @{shift(@_)};
|
|---|
| 731 | my $classifier_node = shift(@_);
|
|---|
| 732 | my $partitionname = shift(@_);
|
|---|
| 733 | my $metadata_value_to_OIDs_hash_ref = shift(@_);
|
|---|
| 734 |
|
|---|
| 735 | # Create an hlist partition
|
|---|
| 736 | my %child_classifier_node = ( 'Title' => $self->convert_unicode_string_to_utf8_string($partitionname),
|
|---|
| 737 | 'childtype' => "VList",
|
|---|
| 738 | 'contains' => [] );
|
|---|
| 739 |
|
|---|
| 740 | # Add the children to the hlist partition
|
|---|
| 741 | $self->add_vlist(\@metadata_groups, \%child_classifier_node, $metadata_value_to_OIDs_hash_ref);
|
|---|
| 742 | push(@{$classifier_node->{'contains'}}, \%child_classifier_node);
|
|---|
| 743 | }
|
|---|
| 744 |
|
|---|
| 745 |
|
|---|
| 746 | sub add_vlist
|
|---|
| 747 | {
|
|---|
| 748 | my $self = shift(@_);
|
|---|
| 749 | my @metadata_groups = @{shift(@_)};
|
|---|
| 750 | my $classifier_node = shift(@_);
|
|---|
| 751 | my $metadata_value_to_OIDs_hash_ref = shift(@_);
|
|---|
| 752 |
|
|---|
| 753 | my $metadata_group = shift(@metadata_groups);
|
|---|
| 754 | $classifier_node->{'mdtype'} = $metadata_group;
|
|---|
| 755 |
|
|---|
| 756 | # Create an entry in the vlist for each value
|
|---|
| 757 | foreach my $metadata_value ($self->sort_metadata_values_array(keys(%{$metadata_value_to_OIDs_hash_ref})))
|
|---|
| 758 | {
|
|---|
| 759 | my @OIDs = @{$metadata_value_to_OIDs_hash_ref->{$metadata_value}};
|
|---|
| 760 |
|
|---|
| 761 | # If there is only one item and 'bookshelf_type' is not always (ie. never or duplicate_only), add the item to the list
|
|---|
| 762 | if (@OIDs == 1 && $self->{$metadata_group . ".bookshelf_type"} ne "always") {
|
|---|
| 763 | my $OID = $OIDs[0];
|
|---|
| 764 |
|
|---|
| 765 | # Find the offset of this metadata value
|
|---|
| 766 | my $offset = 0;
|
|---|
| 767 | my $OID_to_metadata_values_hash_ref = $self->{$metadata_group . ".list"};
|
|---|
| 768 | my @metadata_values = @{$OID_to_metadata_values_hash_ref->{$OID}};
|
|---|
| 769 | for (my $i = 0; $i < scalar(@metadata_values); $i++) {
|
|---|
| 770 | if ($metadata_value eq $metadata_values[$i]) {
|
|---|
| 771 | $offset = $i;
|
|---|
| 772 | last;
|
|---|
| 773 | }
|
|---|
| 774 | }
|
|---|
| 775 | push(@{$classifier_node->{'contains'}}, { 'OID' => $OID, 'offset' => $offset });
|
|---|
| 776 | }
|
|---|
| 777 | # If 'bookshelf_type' is 'never', list all the items even if there are duplicated values
|
|---|
| 778 | elsif ($self->{$metadata_group . ".bookshelf_type"} eq "never") {
|
|---|
| 779 | $self->add_sorted_leaf_items(\@OIDs, $classifier_node);
|
|---|
| 780 | }
|
|---|
| 781 | # Otherwise create a sublist (bookshelf) for the metadata value
|
|---|
| 782 | else {
|
|---|
| 783 | my %child_classifier_node = ( 'Title' => $self->convert_unicode_string_to_utf8_string($metadata_value),
|
|---|
| 784 | 'childtype' => "VList",
|
|---|
| 785 | 'contains' => [] );
|
|---|
| 786 |
|
|---|
| 787 | # If there are metadata elements remaining, recursively apply the process
|
|---|
| 788 | if (@metadata_groups > 0) {
|
|---|
| 789 | my $next_metadata_group = $metadata_groups[0];
|
|---|
| 790 | $child_classifier_node{'childtype'} = $self->{$next_metadata_group . ".list_type"};
|
|---|
| 791 | $self->add_level(\@metadata_groups, \@OIDs, \%child_classifier_node);
|
|---|
| 792 | }
|
|---|
| 793 | # Otherwise just add the documents as children of this list
|
|---|
| 794 | else {
|
|---|
| 795 | $self->add_sorted_leaf_items(\@OIDs, \%child_classifier_node);
|
|---|
| 796 | }
|
|---|
| 797 |
|
|---|
| 798 | # Add the sublist to the list
|
|---|
| 799 | push(@{$classifier_node->{'contains'}}, \%child_classifier_node);
|
|---|
| 800 | }
|
|---|
| 801 | }
|
|---|
| 802 | }
|
|---|
| 803 |
|
|---|
| 804 | sub add_sorted_leaf_items
|
|---|
| 805 | {
|
|---|
| 806 | my $self = shift(@_);
|
|---|
| 807 | my @OIDs = @{shift(@_)};
|
|---|
| 808 | my $classifier_node = shift(@_);
|
|---|
| 809 |
|
|---|
| 810 | # Sort leaf nodes and add to list
|
|---|
| 811 | my @sort_leaf_nodes_usingmetadata_groups = @{$self->{'sort_leaf_nodes_using_metadata_groups'}};
|
|---|
| 812 | foreach my $sort_leaf_nodes_usingmetaelem (reverse @sort_leaf_nodes_usingmetadata_groups) {
|
|---|
| 813 | my $OID_to_metadata_values_hash_ref = $self->{$sort_leaf_nodes_usingmetaelem . ".list"};
|
|---|
| 814 | # Force a stable sort (Perl 5.6's sort isn't stable)
|
|---|
| 815 | # !! The [0] bits aren't ideal (multiple metadata values) !!
|
|---|
| 816 | @OIDs = @OIDs[ sort { $OID_to_metadata_values_hash_ref->{$OIDs[$a]}[0] cmp $OID_to_metadata_values_hash_ref->{$OIDs[$b]}[0] || $a <=> $b; } 0..$#OIDs ];
|
|---|
| 817 | }
|
|---|
| 818 |
|
|---|
| 819 | foreach my $OID (@OIDs) {
|
|---|
| 820 | push(@{$classifier_node->{'contains'}}, { 'OID' => $OID });
|
|---|
| 821 | }
|
|---|
| 822 | }
|
|---|
| 823 |
|
|---|
| 824 |
|
|---|
| 825 | sub sort_metadata_values_array
|
|---|
| 826 | {
|
|---|
| 827 | my $self = shift(@_);
|
|---|
| 828 | my @metadata_values = @_;
|
|---|
| 829 |
|
|---|
| 830 | if ($self->{'unicode_collator'}) {
|
|---|
| 831 | return $self->{'unicode_collator'}->sort(@metadata_values);
|
|---|
| 832 | }
|
|---|
| 833 | else {
|
|---|
| 834 | return sort(@metadata_values);
|
|---|
| 835 | }
|
|---|
| 836 | }
|
|---|
| 837 |
|
|---|
| 838 |
|
|---|
| 839 | 1;
|
|---|
