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 | #
|
---|
12 | # Copyright (C) 2005 New Zealand Digital Library Project
|
---|
13 | #
|
---|
14 | # This program is free software; you can redistribute it and/or modify
|
---|
15 | # it under the terms of the GNU General Public License as published by
|
---|
16 | # the Free Software Foundation; either version 2 of the License, or
|
---|
17 | # (at your option) any later version.
|
---|
18 | #
|
---|
19 | # This program is distributed in the hope that it will be useful,
|
---|
20 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
21 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
22 | # GNU General Public License for more details.
|
---|
23 | #
|
---|
24 | # You should have received a copy of the GNU General Public License
|
---|
25 | # along with this program; if not, write to the Free Software
|
---|
26 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
27 | #
|
---|
28 | # TO DO: - Remove punctuation from metadata values before sorting.
|
---|
29 | # - Add an AZCompactList-style hlist option?
|
---|
30 | #
|
---|
31 | ###########################################################################
|
---|
32 |
|
---|
33 |
|
---|
34 | package List;
|
---|
35 |
|
---|
36 |
|
---|
37 | use BaseClassifier;
|
---|
38 | use Sort::Naturally;
|
---|
39 | use strict;
|
---|
40 |
|
---|
41 |
|
---|
42 | sub BEGIN {
|
---|
43 | @List::ISA = ('BaseClassifier');
|
---|
44 | }
|
---|
45 |
|
---|
46 |
|
---|
47 | my $metadata_selection_mode_list =
|
---|
48 | [
|
---|
49 | { 'name' => "firstvalue",
|
---|
50 | 'desc' => "{List.metadata_selection.firstvalue}"},
|
---|
51 | { 'name' => "firstvalidmetadata",
|
---|
52 | 'desc' => "{List.metadata_selection.firstvalidmetadata}"},
|
---|
53 | { 'name' => "allvalues",
|
---|
54 | 'desc' => "{List.metadata_selection.allvalues}"} ];
|
---|
55 |
|
---|
56 | my $metadata_selection_mode_default = "firstvalidmetadata";
|
---|
57 |
|
---|
58 | my $valid_metadata_selection_modes = { 'firstvalue' => 1,
|
---|
59 | 'firstvalidmetadata' => 1,
|
---|
60 | 'allvalues' => 1 };
|
---|
61 |
|
---|
62 | my $metadata_sort_mode_list =
|
---|
63 | [
|
---|
64 | { 'name' => "unicode",
|
---|
65 | 'desc' => "{List.metadata_sort.unicode}"},
|
---|
66 | { 'name' => "alphabetic",
|
---|
67 | 'desc' => "{List.metadata_sort.alphabetic}"},
|
---|
68 | { 'name' => "alphanumeric",
|
---|
69 | 'desc' => "{List.metadata_sort.alphanumeric}"} ];
|
---|
70 |
|
---|
71 | my $metadata_sort_mode_default = "alphanumeric";
|
---|
72 |
|
---|
73 | my $valid_metadata_sort_modes = { 'unicode' => 1,
|
---|
74 | 'alphabetic' => 1,
|
---|
75 | 'alphanumeric' => 1};
|
---|
76 | my $partition_type_list =
|
---|
77 | [ { 'name' => "per_letter",
|
---|
78 | 'desc' => "{List.level_partition.per_letter}" },
|
---|
79 | { 'name' => "approximate_size",
|
---|
80 | 'desc' => "{List.level_partition.approximate_size}"},
|
---|
81 | { 'name' => "constant_size",
|
---|
82 | 'desc' => "{List.level_partition.constant_size}" },
|
---|
83 | { 'name' => "all_values",
|
---|
84 | 'desc' => "{List.level_partition.all_values}" },
|
---|
85 | { 'name' => "none",
|
---|
86 | 'desc' => "{List.level_partition.none}" } ];
|
---|
87 |
|
---|
88 | my $partition_type_default = "per_letter";
|
---|
89 |
|
---|
90 | my $valid_partition_types = { 'per_letter' => 1,
|
---|
91 | 'constant_size' => 1,
|
---|
92 | 'approximate_size' => 1,
|
---|
93 | 'all_values' => 1,
|
---|
94 | 'none' => 1};
|
---|
95 |
|
---|
96 | my $partition_size_default = 30;
|
---|
97 |
|
---|
98 | my $numeric_partition_type_list =
|
---|
99 | [ { 'name' => "per_digit",
|
---|
100 | 'desc' => "{List.level_partition.per_digit}" },
|
---|
101 | { 'name' => "per_number",
|
---|
102 | 'desc' => "{List.level_partition.per_number}" },
|
---|
103 | { 'name' => "single_partition",
|
---|
104 | 'desc' => "{List.level_partition.single}" },
|
---|
105 | { 'name' => "approximate_size",
|
---|
106 | 'desc' => "{List.level_partition.approximate_size_numeric}"},
|
---|
107 | { 'name' => "constant_size",
|
---|
108 | 'desc' => "{List.level_partition.constant_size}" },
|
---|
109 | { 'name' => "all_values",
|
---|
110 | 'desc' => "{List.level_partition.all_values}" },
|
---|
111 | { 'name' => "none",
|
---|
112 | 'desc' => "{List.level_partition.none}" } ];
|
---|
113 |
|
---|
114 | my $numeric_partition_type_default = "single_partition";
|
---|
115 |
|
---|
116 | my $valid_numeric_partition_types = { 'per_digit' => 1,
|
---|
117 | 'per_number' => 1,
|
---|
118 | 'constant_size' => 1,
|
---|
119 | 'single_partition' => 1,
|
---|
120 | 'approximate_size' => 1,
|
---|
121 | 'all_values' => 1,
|
---|
122 | 'none' =>1 };
|
---|
123 |
|
---|
124 | my $numeric_partition_size_default = 30;
|
---|
125 |
|
---|
126 | my $numeric_partition_name_length_default = "-1"; # use the full number
|
---|
127 |
|
---|
128 | my $bookshelf_type_list =
|
---|
129 | [ { 'name' => "always",
|
---|
130 | 'desc' => "{List.bookshelf_type.always}" },
|
---|
131 | { 'name' => "duplicate_only",
|
---|
132 | 'desc' => "{List.bookshelf_type.duplicate_only}" },
|
---|
133 | { 'name' => "never",
|
---|
134 | 'desc' => "{List.bookshelf_type.never}" } ];
|
---|
135 |
|
---|
136 | my $bookshelf_type_default = "never";
|
---|
137 | my $sort_leaf_nodes_using_default = "Title";
|
---|
138 | my $arguments =
|
---|
139 | [ { 'name' => "metadata",
|
---|
140 | 'desc' => "{List.metadata}",
|
---|
141 | 'type' => "metadata",
|
---|
142 | 'reqd' => "yes" },
|
---|
143 |
|
---|
144 | { 'name' => "metadata_selection_mode_within_level",
|
---|
145 | 'desc' => "{List.metadata_selection_mode_within_level}",
|
---|
146 | 'type' => "enumstring", # Must be enumstring because multiple values can be specified (separated by '/')
|
---|
147 | 'list' => $metadata_selection_mode_list,
|
---|
148 | 'deft' => $metadata_selection_mode_default },
|
---|
149 |
|
---|
150 | { 'name' => "metadata_sort_mode_within_level",
|
---|
151 | 'desc' => "{List.metadata_sort_mode_within_level}",
|
---|
152 | 'type' => "enumstring", # Must be enumstring because multiple values can be specified (separated by '/')
|
---|
153 | 'list' => $metadata_sort_mode_list,
|
---|
154 | 'deft' => $metadata_sort_mode_default },
|
---|
155 |
|
---|
156 | { 'name' => "bookshelf_type",
|
---|
157 | 'desc' => "{List.bookshelf_type}",
|
---|
158 | 'type' => "enum",
|
---|
159 | 'list' => $bookshelf_type_list,
|
---|
160 | 'deft' => $bookshelf_type_default },
|
---|
161 |
|
---|
162 | { 'name' => "classify_sections",
|
---|
163 | 'desc' => "{List.classify_sections}",
|
---|
164 | 'type' => "flag" },
|
---|
165 |
|
---|
166 | { 'name' => "partition_type_within_level",
|
---|
167 | 'desc' => "{List.partition_type_within_level}",
|
---|
168 | 'type' => "enumstring", # Must be enumstring because multiple values can be specified (separated by '/')
|
---|
169 | 'list' => $partition_type_list,
|
---|
170 | 'deft' => $partition_type_default },
|
---|
171 |
|
---|
172 | { 'name' => "partition_size_within_level",
|
---|
173 | 'desc' => "{List.partition_size_within_level}",
|
---|
174 | 'type' => "string", # Must be string because multiple values can be specified (separated by '/')
|
---|
175 | 'deft' => $partition_size_default},
|
---|
176 |
|
---|
177 | { 'name' => "partition_name_length",
|
---|
178 | 'desc' => "{List.partition_name_length}",
|
---|
179 | 'type' => "string" },
|
---|
180 |
|
---|
181 | {'name' => "partition_sort_mode_within_level",
|
---|
182 | 'desc' => "{List.partition_sort_mode_within_level}",
|
---|
183 | 'type' => "enumstring", # Must be enumstring because multiple values can be specified (separated by '/')
|
---|
184 | 'list' => $metadata_sort_mode_list,
|
---|
185 | 'deft' => $metadata_sort_mode_default },
|
---|
186 |
|
---|
187 | { 'name' => "numeric_partition_type_within_level",
|
---|
188 | 'desc' => "{List.numeric_partition_type_within_level}",
|
---|
189 | 'type' => "enumstring", # Must be enumstring because multiple values can be specified (separated by '/')
|
---|
190 | 'list' => $numeric_partition_type_list,
|
---|
191 | 'deft' => $numeric_partition_type_default },
|
---|
192 |
|
---|
193 | {'name' => "numeric_partition_sort_mode_within_level",
|
---|
194 | 'desc' => "{List.numeric_partition_sort_mode_within_level}",
|
---|
195 | 'type' => "enumstring", # Must be enumstring because multiple values can be specified (separated by '/')
|
---|
196 | 'list' => $metadata_sort_mode_list,
|
---|
197 | 'deft' => $metadata_sort_mode_default },
|
---|
198 |
|
---|
199 | { 'name' => "numeric_partition_size_within_level",
|
---|
200 | 'desc' => "{List.numeric_partition_size_within_level}",
|
---|
201 | 'type' => "string", # Must be string because multiple values can be specified (separated by '/')
|
---|
202 | 'deft' => $numeric_partition_size_default},
|
---|
203 |
|
---|
204 | { 'name' => "numeric_partition_name_length_within_level",
|
---|
205 | 'desc' => "{List.numeric_partition_name_length_within_level}",
|
---|
206 | 'type' => "string",
|
---|
207 | 'deft' => $numeric_partition_name_length_default },
|
---|
208 |
|
---|
209 | { 'name' => "numbers_first",
|
---|
210 | 'desc' => "{List.numbers_first}",
|
---|
211 | 'type' => 'flag'},
|
---|
212 |
|
---|
213 | { 'name' => "sort_leaf_nodes_using",
|
---|
214 | 'desc' => "{List.sort_leaf_nodes_using}",
|
---|
215 | 'type' => "metadata",
|
---|
216 | 'deft' => $sort_leaf_nodes_using_default },
|
---|
217 |
|
---|
218 | { 'name' => "sort_leaf_nodes_sort_mode",
|
---|
219 | 'desc' => "{List.sort_leaf_nodes_sort_mode}",
|
---|
220 | 'type' => "enum",
|
---|
221 | 'list' => $metadata_sort_mode_list,
|
---|
222 | 'deft' => $metadata_sort_mode_default },
|
---|
223 |
|
---|
224 | { 'name' => "reverse_sort_leaf_nodes",
|
---|
225 | 'desc' => "{List.reverse_sort_leaf_nodes}",
|
---|
226 | 'type' => "flag"},
|
---|
227 |
|
---|
228 | { 'name' => "sort_using_unicode_collation",
|
---|
229 | 'desc' => "{List.metadata_sort.unicode} {List.sort_using_unicode_collation}",
|
---|
230 | 'type' => "flag" },
|
---|
231 |
|
---|
232 | {'name' => "filter_metadata",
|
---|
233 | 'desc' => "{List.filter_metadata}",
|
---|
234 | 'type' => "metadata"},
|
---|
235 |
|
---|
236 | {'name' => "filter_regex",
|
---|
237 | 'desc' => "{List.filter_regex}",
|
---|
238 | 'type' => "regexp"},
|
---|
239 |
|
---|
240 | { 'name' => "standardize_capitalization",
|
---|
241 | 'desc' => "{List.standardize_capitalization}",
|
---|
242 | 'type' => "flag"},
|
---|
243 |
|
---|
244 | { 'name' => "removeprefix",
|
---|
245 | 'desc' => "{BasClas.removeprefix}",
|
---|
246 | 'type' => "regexp" },
|
---|
247 |
|
---|
248 | { 'name' => "removesuffix",
|
---|
249 | 'desc' => "{BasClas.removesuffix}",
|
---|
250 | 'type' => "regexp" } ];
|
---|
251 |
|
---|
252 | my $options = { 'name' => "List",
|
---|
253 | 'desc' => "{List.desc}",
|
---|
254 | 'abstract' => "no",
|
---|
255 | 'inherits' => "yes",
|
---|
256 | 'args' => $arguments };
|
---|
257 |
|
---|
258 |
|
---|
259 | sub new
|
---|
260 | {
|
---|
261 | my ($class) = shift(@_);
|
---|
262 | my ($classifierslist, $inputargs, $hashArgOptLists) = @_;
|
---|
263 | push(@$classifierslist, $class);
|
---|
264 |
|
---|
265 | push(@{$hashArgOptLists->{"ArgList"}}, @{$arguments});
|
---|
266 | push(@{$hashArgOptLists->{"OptList"}}, $options);
|
---|
267 |
|
---|
268 | my $self = new BaseClassifier($classifierslist, $inputargs, $hashArgOptLists);
|
---|
269 |
|
---|
270 | bless $self, $class;
|
---|
271 |
|
---|
272 | if ($self->{'info_only'}) {
|
---|
273 | # don't worry about any options etc
|
---|
274 | return $self;
|
---|
275 | }
|
---|
276 |
|
---|
277 | # The metadata elements to use (required)
|
---|
278 | if (!$self->{'metadata'}) {
|
---|
279 | die "Error: No metadata fields specified for List.\n";
|
---|
280 | }
|
---|
281 |
|
---|
282 | my @metadata_groups = split(/[\/]/, $self->{'metadata'});
|
---|
283 | $self->{'metadata_groups'} = \@metadata_groups;
|
---|
284 |
|
---|
285 | # The classifier button name (default: the first metadata element specified)
|
---|
286 | if (!$self->{'buttonname'}) {
|
---|
287 | my $first_metadata_group = $metadata_groups[0];
|
---|
288 | my $first_metadata_element = (split(/[\;|,\/]/, $first_metadata_group))[0];
|
---|
289 | $self->{'buttonname'} = $self->generate_title_from_metadata($first_metadata_element);
|
---|
290 | }
|
---|
291 |
|
---|
292 | # meta selection mode for each level
|
---|
293 | $self->set_metadata_groups_info_per_level("metadata_selection_mode_within_level", $metadata_selection_mode_default, $valid_metadata_selection_modes);
|
---|
294 |
|
---|
295 | # meta sort mode for each level
|
---|
296 | if ($self->{'sort_using_unicode_collation'}) {
|
---|
297 | print STDERR "WARNING: sort_using_unicode_collation is set, setting metadata_sort_mode_within_level to unicode for all levels, regardless of current setting\n";
|
---|
298 | $self->{'metadata_sort_mode_within_level'} = "unicode";
|
---|
299 | $metadata_sort_mode_default = "unicode";
|
---|
300 | } else {
|
---|
301 | if ($self->{'metadata_sort_mode_within_level'} =~ /unicode/) {
|
---|
302 | $self->{'sort_using_unicode_collation'} = 1;
|
---|
303 | }
|
---|
304 | }
|
---|
305 |
|
---|
306 |
|
---|
307 | $self->set_metadata_groups_info_per_level('metadata_sort_mode_within_level', $metadata_sort_mode_default, $valid_metadata_sort_modes);
|
---|
308 |
|
---|
309 | # Whether to group items into a bookshelf, (must be 'always' for all metadata fields except the last)
|
---|
310 | foreach my $metadata_group (@metadata_groups) {
|
---|
311 | $self->{$metadata_group . ".bookshelf_type"} = "always";
|
---|
312 | }
|
---|
313 | my $last_metadata_group = $metadata_groups[$#metadata_groups];
|
---|
314 | # Default: duplicate_only, ie. leave leaf nodes ungrouped (equivalent to AZCompactList -mingroup 2)
|
---|
315 | $self->{$last_metadata_group . ".bookshelf_type"} = $self->{'bookshelf_type'};
|
---|
316 |
|
---|
317 | # How the items are grouped into partitions (default: by letter)
|
---|
318 | # for each level (metadata group), separated by '/'
|
---|
319 | $self->set_metadata_groups_info_per_level("partition_type_within_level", $partition_type_default, $valid_partition_types);
|
---|
320 | $self->set_metadata_groups_info_per_level("numeric_partition_type_within_level", $numeric_partition_type_default, $valid_numeric_partition_types);
|
---|
321 |
|
---|
322 | # now check whether a level was none - need to set the equivalent level in the other half also to none
|
---|
323 | foreach my $metadata_group (@metadata_groups) {
|
---|
324 | if ($self->{$metadata_group . ".partition_type_within_level"} eq "none" || $self->{$metadata_group . ".numeric_partition_type_within_level"} eq "none") {
|
---|
325 |
|
---|
326 | print STDERR "WARNING: one of -partition_type_within_level or -numeric_partition_type_within_level was set to 'none' for level $metadata_group, overriding current value of both these options to 'none'\n";
|
---|
327 |
|
---|
328 | $self->{$metadata_group . ".partition_type_within_level"} = "none";
|
---|
329 | $self->{$metadata_group . ".numeric_partition_type_within_level"} = "none";
|
---|
330 | }
|
---|
331 | }
|
---|
332 |
|
---|
333 | $self->set_metadata_groups_info_per_level("partition_size_within_level", $partition_size_default);
|
---|
334 | $self->set_metadata_groups_info_per_level("numeric_partition_size_within_level", $numeric_partition_size_default);
|
---|
335 |
|
---|
336 | $self->set_metadata_groups_info_per_level('partition_sort_mode_within_level', $metadata_sort_mode_default, $valid_metadata_sort_modes);
|
---|
337 |
|
---|
338 | $self->set_metadata_groups_info_per_level('numeric_partition_sort_mode_within_level', $metadata_sort_mode_default, $valid_metadata_sort_modes);
|
---|
339 |
|
---|
340 | $self->set_metadata_groups_info_per_level("numeric_partition_name_length_within_level", $numeric_partition_name_length_default);
|
---|
341 |
|
---|
342 | # The removeprefix and removesuffix expressions
|
---|
343 | if ($self->{'removeprefix'}) {
|
---|
344 | # If there are more than one expressions, use '' to quote each experession and '/' to separate
|
---|
345 | my @removeprefix_exprs_within_levellist = split(/'\/'/, $self->{'removeprefix'});
|
---|
346 |
|
---|
347 | foreach my $metadata_group (@metadata_groups) {
|
---|
348 | my $removeprefix_expr_within_levelelem = shift(@removeprefix_exprs_within_levellist);
|
---|
349 | if (defined($removeprefix_expr_within_levelelem) && $removeprefix_expr_within_levelelem ne "") {
|
---|
350 | # Remove the other ' at the beginning and the end if there is any
|
---|
351 | $removeprefix_expr_within_levelelem =~ s/^'//;
|
---|
352 | $removeprefix_expr_within_levelelem =~ s/'$//;
|
---|
353 | # Remove the extra ^ at the beginning
|
---|
354 | $removeprefix_expr_within_levelelem =~ s/^\^//;
|
---|
355 | $self->{$metadata_group . ".remove_prefix_expr"} = $removeprefix_expr_within_levelelem;
|
---|
356 | } else {
|
---|
357 | $self->{$metadata_group . ".remove_prefix_expr"} = $self->{$metadata_groups[0] . ".remove_prefix_expr"};
|
---|
358 | }
|
---|
359 | }
|
---|
360 | }
|
---|
361 | if ($self->{'removesuffix'}) {
|
---|
362 | my @removesuffix_exprs_within_levellist = split(/'\/'/, $self->{'removesuffix'});
|
---|
363 |
|
---|
364 | foreach my $metadata_group (@metadata_groups) {
|
---|
365 | my $removesuffix_expr_within_levelelem = shift(@removesuffix_exprs_within_levellist);
|
---|
366 | if (defined($removesuffix_expr_within_levelelem) && $removesuffix_expr_within_levelelem ne "") {
|
---|
367 | $removesuffix_expr_within_levelelem =~ s/^'//;
|
---|
368 | $removesuffix_expr_within_levelelem =~ s/'$//;
|
---|
369 | # Remove the extra $ at the end
|
---|
370 | $removesuffix_expr_within_levelelem =~ s/\$$//;
|
---|
371 | $self->{$metadata_group . ".remove_suffix_expr"} = $removesuffix_expr_within_levelelem;
|
---|
372 | } else {
|
---|
373 | $self->{$metadata_group . ".remove_suffix_expr"} = $self->{$metadata_groups[0] . ".remove_suffix_expr"};
|
---|
374 | }
|
---|
375 | }
|
---|
376 | }
|
---|
377 |
|
---|
378 | # The metadata elements to use to sort the leaf nodes (default: Title)
|
---|
379 | my @sort_leaf_nodes_using_metadata_groups = split(/\|/, $self->{'sort_leaf_nodes_using'});
|
---|
380 | $self->{'sort_leaf_nodes_using_metadata_groups'} = \@sort_leaf_nodes_using_metadata_groups;
|
---|
381 |
|
---|
382 | foreach my $sort_group (@sort_leaf_nodes_using_metadata_groups) {
|
---|
383 | # set metadata_select_type, if not already set - might be already set if the same group was used in -metadata
|
---|
384 | if (!defined $self->{$sort_group . ".metadata_selection_mode_within_level"}) {
|
---|
385 | $self->{$sort_group . ".metadata_selection_mode_within_level"} = $metadata_selection_mode_default;
|
---|
386 | }
|
---|
387 |
|
---|
388 | }
|
---|
389 |
|
---|
390 | my @leaf_nodes_sort_modes = split (/\|/, $self->{'sort_leaf_nodes_sort_mode'});
|
---|
391 | foreach my $sort_group (@sort_leaf_nodes_using_metadata_groups) {
|
---|
392 | my $leaf_sort_mode = shift(@leaf_nodes_sort_modes);
|
---|
393 | if (!defined $self->{$sort_group . ".metadata_sort_mode_within_level"}) {
|
---|
394 | if (defined $leaf_sort_mode && defined $valid_metadata_sort_modes->{$leaf_sort_mode}) {
|
---|
395 | $self->{$sort_group . ".metadata_sort_mode_within_level"} = $leaf_sort_mode;
|
---|
396 | }
|
---|
397 | else {
|
---|
398 | $self->{$sort_group . ".metadata_sort_mode_within_level"} = $metadata_sort_mode_default;
|
---|
399 | }
|
---|
400 | }
|
---|
401 |
|
---|
402 | }
|
---|
403 |
|
---|
404 | # Create an instance of the Unicode::Collate object if better Unicode sorting is desired
|
---|
405 | if ($self->{'sort_using_unicode_collation'}) {
|
---|
406 | # To use this you first need to download the allkeys.txt file from
|
---|
407 | # http://www.unicode.org/Public/UCA/latest/allkeys.txt and put it in the Perl
|
---|
408 | # Unicode/Collate directory.
|
---|
409 | require Unicode::Collate;
|
---|
410 | $self->{'unicode_collator'} = Unicode::Collate->new();
|
---|
411 | }
|
---|
412 |
|
---|
413 | # An empty array for the document/section OIDs that we are classifying
|
---|
414 | $self->{'OIDs'} = [];
|
---|
415 | # A hash for all the doc ids that we have seen, so we don't classify something twice
|
---|
416 | $self->{'all_doc_OIDs'} = {};
|
---|
417 | return $self;
|
---|
418 | }
|
---|
419 |
|
---|
420 |
|
---|
421 | sub init
|
---|
422 | {
|
---|
423 | # Nothing to do...
|
---|
424 | }
|
---|
425 |
|
---|
426 | sub set_metadata_groups_info_per_level
|
---|
427 | {
|
---|
428 | my $self = shift(@_);
|
---|
429 | my $info_name = shift(@_);
|
---|
430 | my $info_default = shift(@_);
|
---|
431 | my $info_valid_types_hash_ref = shift(@_);
|
---|
432 |
|
---|
433 | if (!defined $self->{$info_name}) {
|
---|
434 | print STDERR "List Error: no values were set for option $info_name\n";
|
---|
435 | }
|
---|
436 | my @info_list = split(/\//, $self->{$info_name});
|
---|
437 |
|
---|
438 | my $first = 1;
|
---|
439 | foreach my $metadata_group (@{$self->{'metadata_groups'}}) {
|
---|
440 | my $info_elem = shift(@info_list);
|
---|
441 | if (defined ($info_elem) && (!defined $info_valid_types_hash_ref || defined $info_valid_types_hash_ref->{$info_elem})) {
|
---|
442 | $self->{$metadata_group .".$info_name"} = $info_elem;
|
---|
443 | } else {
|
---|
444 | # its empty or an invalid entry
|
---|
445 | my $new_info_elem;
|
---|
446 | if ($first) {
|
---|
447 | $new_info_elem = $info_default;
|
---|
448 | } else {
|
---|
449 | # get the value we had at first
|
---|
450 | $new_info_elem = $self->{@{$self->{'metadata_groups'}}[0] . ".$info_name"};
|
---|
451 | }
|
---|
452 | $self->{$metadata_group .".$info_name"} = $new_info_elem;
|
---|
453 | if (defined $info_elem) {
|
---|
454 | print STDERR "List Error: $info_elem is not a valid value for $info_name, changing it to $new_info_elem\n";
|
---|
455 |
|
---|
456 | }
|
---|
457 | }
|
---|
458 | $first = 0;
|
---|
459 | }
|
---|
460 |
|
---|
461 | }
|
---|
462 |
|
---|
463 | # Called for each document in the collection
|
---|
464 | sub classify
|
---|
465 | {
|
---|
466 | my $self = shift(@_);
|
---|
467 | my ($doc_obj) = @_;
|
---|
468 |
|
---|
469 | if (defined $self->{'all_doc_OIDs'}->{$doc_obj->get_OID()}) {
|
---|
470 | print STDERR "Warning, List classifier has already seen document ".$doc_obj->get_OID().", not classifying again\n";
|
---|
471 | return;
|
---|
472 | }
|
---|
473 |
|
---|
474 | $self->{'all_doc_OIDs'}->{$doc_obj->get_OID()} = 1;
|
---|
475 |
|
---|
476 | # check against filter here
|
---|
477 | if ($self->{'filter_metadata'}) {
|
---|
478 | my $meta = $doc_obj->get_metadata_element($doc_obj->get_top_section(), $self->{'filter_metadata'});
|
---|
479 | return unless defined $meta;
|
---|
480 | if ($self->{'filter_regex'} ne "" && $meta !~ /$self->{'filter_regex'}/) {
|
---|
481 | print STDERR "doc $doc_obj doesn't pass filtering requirement\n" if ($self->{'verbosity'} > 3);
|
---|
482 | return;
|
---|
483 | }
|
---|
484 | }
|
---|
485 | # if we get here, we have passed the test for filtering
|
---|
486 | # If "-classify_sections" is set, classify every section of the document
|
---|
487 | if ($self->{'classify_sections'}) {
|
---|
488 | my $section = $doc_obj->get_next_section($doc_obj->get_top_section());
|
---|
489 | while (defined $section) {
|
---|
490 | $self->classify_section($doc_obj, $doc_obj->get_OID() . ".$section", $section);
|
---|
491 | $section = $doc_obj->get_next_section($section);
|
---|
492 | }
|
---|
493 | }
|
---|
494 | # Otherwise just classify the top document section
|
---|
495 | else {
|
---|
496 | $self->classify_section($doc_obj, $doc_obj->get_OID(), $doc_obj->get_top_section());
|
---|
497 | }
|
---|
498 |
|
---|
499 | }
|
---|
500 |
|
---|
501 | sub classify_section
|
---|
502 | {
|
---|
503 | my $self = shift(@_);
|
---|
504 | my ($doc_obj,$section_OID,$section) = @_;
|
---|
505 |
|
---|
506 | my @metadata_groups = @{$self->{'metadata_groups'}};
|
---|
507 |
|
---|
508 | # Only classify the section if it has a value for one of the metadata elements in the first group
|
---|
509 | my $classify_section = 0;
|
---|
510 | my $first_metadata_group = $metadata_groups[0];
|
---|
511 | my $remove_prefix_expr = $self->{$first_metadata_group . ".remove_prefix_expr"};
|
---|
512 | my $remove_suffix_expr = $self->{$first_metadata_group . ".remove_suffix_expr"};
|
---|
513 | foreach my $first_metadata_group_element (split(/\;|,/, $first_metadata_group)) {
|
---|
514 | my $real_first_metadata_group_element = $self->strip_ex_from_metadata($first_metadata_group_element);
|
---|
515 | my $first_metadata_group_element_value = $doc_obj->get_metadata_element($section, $real_first_metadata_group_element);
|
---|
516 | # Remove prefix/suffix if requested
|
---|
517 | if (defined ($first_metadata_group_element_value)) {
|
---|
518 | if (defined $remove_prefix_expr && $remove_prefix_expr ne "") {
|
---|
519 | $first_metadata_group_element_value =~ s/^$remove_prefix_expr//;
|
---|
520 | }
|
---|
521 |
|
---|
522 | if (defined $remove_suffix_expr && $remove_suffix_expr ne "") {
|
---|
523 | $first_metadata_group_element_value =~ s/$remove_suffix_expr$//;
|
---|
524 | }
|
---|
525 | }
|
---|
526 |
|
---|
527 | $first_metadata_group_element_value = &sorttools::format_metadata_for_sorting($first_metadata_group, $first_metadata_group_element_value, $doc_obj) unless $self->{'no_metadata_formatting'};
|
---|
528 | if (defined($first_metadata_group_element_value) && $first_metadata_group_element_value ne "") {
|
---|
529 | # This section must be included in the classifier as we have found a value
|
---|
530 | $classify_section = 1;
|
---|
531 | last;
|
---|
532 | }
|
---|
533 | }
|
---|
534 |
|
---|
535 | # We're not classifying this section because it doesn't have the required metadata
|
---|
536 | return if (!$classify_section);
|
---|
537 |
|
---|
538 | # Otherwise, include this section in the classifier
|
---|
539 |
|
---|
540 | push(@{$self->{'OIDs'}}, $section_OID);
|
---|
541 |
|
---|
542 | # Create a hash for the metadata values of each metadata element we're interested in
|
---|
543 | my %metadata_groups_done = ();
|
---|
544 | foreach my $metadata_group (@metadata_groups, @{$self->{'sort_leaf_nodes_using_metadata_groups'}}) {
|
---|
545 | # Take care not to do a metadata group more than once
|
---|
546 | unless ($metadata_groups_done{$metadata_group}) {
|
---|
547 | my $remove_prefix_expr = $self->{$metadata_group . ".remove_prefix_expr"};
|
---|
548 | my $remove_suffix_expr = $self->{$metadata_group . ".remove_suffix_expr"};
|
---|
549 | foreach my $metadata_element (split(/\;|,/, $metadata_group)) {
|
---|
550 | my $real_metadata_element = $self->strip_ex_from_metadata($metadata_element);
|
---|
551 |
|
---|
552 | my @metadata_values = @{$doc_obj->get_metadata($section, $real_metadata_element)};
|
---|
553 | foreach my $metadata_value (@metadata_values) {
|
---|
554 | #print STDERR "working with value $metadata_value\n";
|
---|
555 | # Strip leading and trailing whitespace
|
---|
556 | $metadata_value =~ s/^\s*//;
|
---|
557 | $metadata_value =~ s/\s*$//;
|
---|
558 |
|
---|
559 | # Remove prefix/suffix if requested
|
---|
560 | if (defined $remove_prefix_expr && $remove_prefix_expr ne "") {
|
---|
561 | $metadata_value =~ s/^$remove_prefix_expr//;
|
---|
562 | }
|
---|
563 | if (defined $remove_suffix_expr && $remove_suffix_expr ne "") {
|
---|
564 | $metadata_value =~ s/$remove_suffix_expr$//;
|
---|
565 | }
|
---|
566 |
|
---|
567 | # lowercase metadata both for sorting meta (d/D under D), and to allow CSS to be able to
|
---|
568 | # text-transform the stored lowercase values as capitalize or uppercase (can't CSS
|
---|
569 | # text-transform if stored uppercase). 2 CSS text-transforms have been added to core.css
|
---|
570 | my $lc_metadata_value = lc($metadata_value);
|
---|
571 | $lc_metadata_value = &sorttools::format_metadata_for_sorting($real_metadata_element, $lc_metadata_value, $doc_obj) unless $self->{'no_metadata_formatting'};
|
---|
572 |
|
---|
573 | # Add the metadata value into the list for this combination of metadata group
|
---|
574 | # and section - if we have some non-whitespace chars
|
---|
575 | # text that we have some non-whitespace chars
|
---|
576 | if ($lc_metadata_value =~ /\S/) {
|
---|
577 |
|
---|
578 | push(@{$self->{$metadata_group . ".list"}->{$section_OID}}, $lc_metadata_value);
|
---|
579 |
|
---|
580 | # add the actual value into the stored values so we can remember the case
|
---|
581 | if (!$self->{'standardize_capitalization'}) {
|
---|
582 | if (defined $self->{$metadata_group . ".actualvalues"}->{$lc_metadata_value}->{$metadata_value}) {
|
---|
583 | $self->{$metadata_group . ".actualvalues"}->{$lc_metadata_value}->{$metadata_value}++;
|
---|
584 | } else {
|
---|
585 | $self->{$metadata_group . ".actualvalues"}->{$lc_metadata_value}->{$metadata_value} = 1;
|
---|
586 | }
|
---|
587 | }
|
---|
588 | last if ($self->{$metadata_group . ".metadata_selection_mode_within_level"} eq "firstvalue");
|
---|
589 | }
|
---|
590 | } # foreach metadatavalue
|
---|
591 | last if ((@metadata_values > 0) && $self->{$metadata_group . ".metadata_selection_mode_within_level"} =~ /^(firstvalue|firstvalidmetadata)$/ );
|
---|
592 | } # foreach metadata element
|
---|
593 |
|
---|
594 | $metadata_groups_done{$metadata_group} = 1;
|
---|
595 | }
|
---|
596 | }
|
---|
597 | }
|
---|
598 |
|
---|
599 |
|
---|
600 | sub get_classify_info
|
---|
601 | {
|
---|
602 | my $self = shift(@_);
|
---|
603 |
|
---|
604 | # The metadata groups to classify by
|
---|
605 | my @metadata_groups = @{$self->{'metadata_groups'}};
|
---|
606 | my $first_metadata_group = $metadata_groups[0];
|
---|
607 |
|
---|
608 | # The OID values of the documents to include in the classifier
|
---|
609 | my @OIDs = @{$self->{'OIDs'}};
|
---|
610 |
|
---|
611 | # Create the root node of the classification hierarchy
|
---|
612 | my %classifier_node = ( 'thistype' => "Invisible",
|
---|
613 | #'childtype' => $self->{$first_metadata_group . ".list_type"},
|
---|
614 | 'Title' => $self->{'buttonname'},
|
---|
615 | 'contains' => [],
|
---|
616 | 'mdtype' => $first_metadata_group );
|
---|
617 |
|
---|
618 | # Recursively create the classification hierarchy, one level for each metadata group
|
---|
619 | $self->add_level(\@metadata_groups, \@OIDs, \%classifier_node);
|
---|
620 | return \%classifier_node;
|
---|
621 | }
|
---|
622 |
|
---|
623 |
|
---|
624 | sub add_level
|
---|
625 | {
|
---|
626 | my $self = shift(@_);
|
---|
627 | my @metadata_groups = @{shift(@_)};
|
---|
628 | my @OIDs = @{shift(@_)};
|
---|
629 | my $classifier_node = shift(@_);
|
---|
630 |
|
---|
631 | my $metadata_group = $metadata_groups[0];
|
---|
632 | if (!defined($self->{$metadata_group . ".list"})) {
|
---|
633 | print STDERR "Warning: No metadata values assigned to $metadata_group.\n";
|
---|
634 | return;
|
---|
635 | }
|
---|
636 |
|
---|
637 | # Create a mapping from metadata value to OID
|
---|
638 | my $OID_to_metadata_values_hash_ref = $self->{$metadata_group . ".list"};
|
---|
639 | my %metadata_value_to_OIDs_hash = ();
|
---|
640 | my %numeric_metadata_value_to_OIDs_hash = ();
|
---|
641 | foreach my $OID (@OIDs)
|
---|
642 | {
|
---|
643 | if ($OID_to_metadata_values_hash_ref->{$OID})
|
---|
644 | {
|
---|
645 | my @metadata_values = @{$OID_to_metadata_values_hash_ref->{$OID}};
|
---|
646 | foreach my $metadata_value (@metadata_values)
|
---|
647 | {
|
---|
648 | if ($metadata_value =~ /^[0-9]/) {
|
---|
649 | push(@{$numeric_metadata_value_to_OIDs_hash{$metadata_value}}, $OID);
|
---|
650 | } else {
|
---|
651 | push(@{$metadata_value_to_OIDs_hash{$metadata_value}}, $OID);
|
---|
652 | }
|
---|
653 | }
|
---|
654 | }
|
---|
655 | }
|
---|
656 | #print STDERR "Number of distinct values: " . scalar(keys %metadata_value_to_OIDs_hash) . "\n";
|
---|
657 | #print STDERR "Number of distinct numeric values: " . scalar(keys %numeric_metadata_value_to_OIDs_hash) . "\n";
|
---|
658 |
|
---|
659 | # Partition the values (if necessary)
|
---|
660 | my $partition_type_within_level = $self->{$metadata_group . ".partition_type_within_level"};
|
---|
661 | my $partition_size_within_level = $self->{$metadata_group . ".partition_size_within_level"};
|
---|
662 | my $partition_sort_mode_within_level = $self->{$metadata_group . ".partition_sort_mode_within_level"};
|
---|
663 | my $bookshelf_type_within_level = $self->{$metadata_group. ".bookshelf_type"};
|
---|
664 |
|
---|
665 |
|
---|
666 | #############################################
|
---|
667 | ### DO THE NUMBERS IF THEY ARE TO COME FIRST
|
---|
668 | #############################################
|
---|
669 | if ($self->{'numbers_first'} && keys(%numeric_metadata_value_to_OIDs_hash)) {
|
---|
670 | $self->partition_numeric_values(\@metadata_groups, $classifier_node, \%numeric_metadata_value_to_OIDs_hash);
|
---|
671 | }
|
---|
672 |
|
---|
673 | ############################################
|
---|
674 | # DO THE LETTERS
|
---|
675 | ############################################
|
---|
676 | if (keys(%metadata_value_to_OIDs_hash)){ # make sure we have some values
|
---|
677 | if ($partition_type_within_level =~ /^per_letter$/i) {
|
---|
678 | $self->split_per_letter_or_digit(\@metadata_groups, $classifier_node, $partition_sort_mode_within_level, \%metadata_value_to_OIDs_hash);
|
---|
679 | }
|
---|
680 | elsif ($partition_type_within_level =~ /^approximate_size$/i && scalar(keys %metadata_value_to_OIDs_hash) > $partition_size_within_level) {
|
---|
681 | $self->split_approximate_size(\@metadata_groups, $classifier_node, $partition_size_within_level, $partition_sort_mode_within_level, $bookshelf_type_within_level, \%metadata_value_to_OIDs_hash, $self->{'partition_name_length'});
|
---|
682 |
|
---|
683 | }
|
---|
684 | elsif ($partition_type_within_level =~ /^constant_size$/i && scalar(keys %metadata_value_to_OIDs_hash) > $partition_size_within_level) {
|
---|
685 | $self->split_constant_size(\@metadata_groups, $classifier_node, $partition_size_within_level, $partition_sort_mode_within_level, $bookshelf_type_within_level, \%metadata_value_to_OIDs_hash, $self->{'partition_name_length'});
|
---|
686 | }
|
---|
687 |
|
---|
688 | # Otherwise just add all the values to a VList
|
---|
689 | else {
|
---|
690 | if ($partition_type_within_level =~ /^all_values$/i) {
|
---|
691 | $classifier_node->{'childtype'} = "HList";
|
---|
692 | $self->{$metadata_group. ".bookshelf_type"} = "always";
|
---|
693 | }
|
---|
694 | $self->add_vlist(\@metadata_groups, $classifier_node, \%metadata_value_to_OIDs_hash);
|
---|
695 | }
|
---|
696 | }
|
---|
697 |
|
---|
698 | ###########################################
|
---|
699 | ### DO THE NUMBERS IF THEY ARE TO COME LAST
|
---|
700 | ###########################################
|
---|
701 | if (!$self->{'numbers_first'} && keys(%numeric_metadata_value_to_OIDs_hash)) {
|
---|
702 | $self->partition_numeric_values(\@metadata_groups, $classifier_node, \%numeric_metadata_value_to_OIDs_hash);
|
---|
703 | }
|
---|
704 | }
|
---|
705 |
|
---|
706 | sub partition_numeric_values
|
---|
707 | {
|
---|
708 | my $self = shift(@_);
|
---|
709 | my @metadata_groups = @{shift(@_)};
|
---|
710 | my $classifier_node = shift(@_);
|
---|
711 | my $numeric_metadata_value_to_OIDs_hash_ref = shift(@_);
|
---|
712 |
|
---|
713 | my $metadata_group = $metadata_groups[0];
|
---|
714 | my $numeric_partition_type_within_level = $self->{$metadata_group . ".numeric_partition_type_within_level"};
|
---|
715 | my $numeric_partition_size_within_level = $self->{$metadata_group . ".numeric_partition_size_within_level"};
|
---|
716 | my $numeric_partition_sort_mode_within_level = $self->{$metadata_group . ".numeric_partition_sort_mode_within_level"};
|
---|
717 | my $numeric_partition_name_length_within_level = $self->{$metadata_group . ".numeric_partition_name_length_within_level"};
|
---|
718 | my $bookshelf_type_within_level = $self->{$metadata_group. ".bookshelf_type"};
|
---|
719 |
|
---|
720 | if ($numeric_partition_type_within_level eq "single_partition") {
|
---|
721 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, "0-9", $numeric_metadata_value_to_OIDs_hash_ref);
|
---|
722 | }
|
---|
723 | elsif ($numeric_partition_type_within_level eq "per_digit") {
|
---|
724 | $self->split_per_letter_or_digit(\@metadata_groups, $classifier_node, $numeric_partition_sort_mode_within_level, $numeric_metadata_value_to_OIDs_hash_ref, 1, 1);
|
---|
725 | }
|
---|
726 | elsif ($numeric_partition_type_within_level eq "per_number") {
|
---|
727 | # each different number is a bucket
|
---|
728 | $self->split_per_letter_or_digit(\@metadata_groups, $classifier_node, $numeric_partition_sort_mode_within_level, $numeric_metadata_value_to_OIDs_hash_ref, 1, $numeric_partition_name_length_within_level);
|
---|
729 | }
|
---|
730 | elsif ($numeric_partition_type_within_level eq "constant_size" && scalar(keys %$numeric_metadata_value_to_OIDs_hash_ref) > $numeric_partition_size_within_level) {
|
---|
731 | # Generate hlists of a certain size
|
---|
732 |
|
---|
733 | $self->split_constant_size(\@metadata_groups, $classifier_node, $numeric_partition_size_within_level, $numeric_partition_sort_mode_within_level, $bookshelf_type_within_level, $numeric_metadata_value_to_OIDs_hash_ref, $numeric_partition_name_length_within_level, 1);
|
---|
734 | } elsif ($numeric_partition_type_within_level eq "approximate_size" && scalar(keys %$numeric_metadata_value_to_OIDs_hash_ref) > $numeric_partition_size_within_level) {
|
---|
735 | $self->split_approximate_size(\@metadata_groups, $classifier_node, $numeric_partition_size_within_level, $numeric_partition_sort_mode_within_level, $bookshelf_type_within_level, $numeric_metadata_value_to_OIDs_hash_ref, $numeric_partition_name_length_within_level, 1);
|
---|
736 | }
|
---|
737 | # Otherwise just add all the values to a VList
|
---|
738 | else {
|
---|
739 | if ($numeric_partition_type_within_level =~ /^all_values$/i) {
|
---|
740 | $classifier_node->{'childtype'} = "HList";
|
---|
741 | $self->{$metadata_group. ".bookshelf_type"} = "always";
|
---|
742 | }
|
---|
743 |
|
---|
744 | $self->add_vlist(\@metadata_groups, $classifier_node, $numeric_metadata_value_to_OIDs_hash_ref);
|
---|
745 | }
|
---|
746 |
|
---|
747 | }
|
---|
748 |
|
---|
749 | sub split_approximate_size
|
---|
750 | {
|
---|
751 |
|
---|
752 | my $self = shift(@_);
|
---|
753 | my @metadata_groups = @{shift(@_)};
|
---|
754 | my $classifier_node = shift(@_);
|
---|
755 | my $partition_size = shift(@_);
|
---|
756 | my $sort_mode = shift(@_);
|
---|
757 | my $bookshelf_type = shift(@_);
|
---|
758 | my $metadata_value_to_OIDs_hash_ref = shift(@_);
|
---|
759 | my $partition_name_length = shift(@_);
|
---|
760 | my $is_numeric = shift(@_);
|
---|
761 |
|
---|
762 | # Generate hlist based on the first letter of the metadata value (like per_letter), or based on
|
---|
763 | # numbers, but also with restriction on the partition size
|
---|
764 | # If a partition has fewer items than specified by the "partition_size_within_level", then group them together if possible
|
---|
765 | # If a partition has more items than specified, split into several hlists.
|
---|
766 | # Depends on the bookshelf_type, one item can be either a document (when bookshelf_type is "never") or a metadata value (otherwise)
|
---|
767 |
|
---|
768 | my @sortedmetadata_values = $self->sort_metadata_values_array($sort_mode, keys(%$metadata_value_to_OIDs_hash_ref));
|
---|
769 |
|
---|
770 | # Separate values by their first letter, each form a bucket, like the per_letter partition type
|
---|
771 | my $last_partition = $self->generate_partition_name($sortedmetadata_values[0], $partition_name_length, $is_numeric);
|
---|
772 |
|
---|
773 | my @partition_buckets = ();
|
---|
774 | my @metadata_values_in_bucket = ();
|
---|
775 | my $num_items_in_bucket = 0;
|
---|
776 |
|
---|
777 | foreach my $metadata_value (@sortedmetadata_values) {
|
---|
778 | my $metadata_valuepartition = $self->generate_partition_name($metadata_value, $partition_name_length, $is_numeric);
|
---|
779 | if ($metadata_valuepartition ne $last_partition) {
|
---|
780 | my @temp_array = @metadata_values_in_bucket;
|
---|
781 | # 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
|
---|
782 | my %partition_info = ();
|
---|
783 | $partition_info{'metadata_values'} = \@temp_array;
|
---|
784 | $partition_info{'size'} = $num_items_in_bucket;
|
---|
785 | $partition_info{'name'} = $last_partition;
|
---|
786 | push (@partition_buckets, \%partition_info);
|
---|
787 |
|
---|
788 | @metadata_values_in_bucket = ($metadata_value);
|
---|
789 | $num_items_in_bucket = ($bookshelf_type eq "never" ? scalar(@{$metadata_value_to_OIDs_hash_ref->{$metadata_value}}) : scalar(@metadata_values_in_bucket));
|
---|
790 | $last_partition = $metadata_valuepartition;
|
---|
791 | } else {
|
---|
792 | $num_items_in_bucket += ($bookshelf_type eq "never" ? scalar(@{$metadata_value_to_OIDs_hash_ref->{$metadata_value}}) : 1); #scalar(@metadata_values_in_bucket);
|
---|
793 | push (@metadata_values_in_bucket, $metadata_value);
|
---|
794 | }
|
---|
795 | }
|
---|
796 | # Last one
|
---|
797 | my %partition_info = ();
|
---|
798 | $partition_info{'metadata_values'} = \@metadata_values_in_bucket;
|
---|
799 | $partition_info{'size'} = $num_items_in_bucket;
|
---|
800 | $partition_info{'name'} = $last_partition;
|
---|
801 |
|
---|
802 | push (@partition_buckets, \%partition_info);
|
---|
803 |
|
---|
804 | # now go through the array of buckets, and merge small buckets
|
---|
805 | my @new_partition_buckets = ();
|
---|
806 | for (my $i = 0; $i < scalar(@partition_buckets) - 1; $i++) {
|
---|
807 |
|
---|
808 | my $this_bucket = $partition_buckets[$i];
|
---|
809 | my $next_bucket = $partition_buckets[$i+1];
|
---|
810 |
|
---|
811 | my $items_in_partition = $this_bucket->{'size'};
|
---|
812 |
|
---|
813 | if ($items_in_partition < $partition_size ) {
|
---|
814 | my $items_in_next_partition = $next_bucket->{'size'};
|
---|
815 | if ($items_in_partition + $items_in_next_partition <= $partition_size ) {
|
---|
816 | # merge this bucket into the next bucket
|
---|
817 | foreach my $metadata_value_to_merge (@{$this_bucket->{'metadata_values'}}) {
|
---|
818 | push(@{$next_bucket->{'metadata_values'}}, $metadata_value_to_merge);
|
---|
819 | }
|
---|
820 | $next_bucket->{'size'} += $items_in_partition;
|
---|
821 |
|
---|
822 | } else {
|
---|
823 | # remember this bucket
|
---|
824 | push (@new_partition_buckets, $this_bucket);
|
---|
825 | }
|
---|
826 | } else {
|
---|
827 | # remember this bucket
|
---|
828 | push (@new_partition_buckets, $this_bucket);
|
---|
829 | }
|
---|
830 | }
|
---|
831 | # add in the last bucket
|
---|
832 | my $last_bucket = $partition_buckets[scalar(@partition_buckets) - 1];
|
---|
833 | push (@new_partition_buckets, $last_bucket);
|
---|
834 |
|
---|
835 | # Add partitions to the main list, but divide big bucket into several
|
---|
836 | my $last_partition_end = "";
|
---|
837 | my $partition_start = "";
|
---|
838 | my $partition_end = "";
|
---|
839 | my $partition_name = "";
|
---|
840 | foreach my $partition (@new_partition_buckets) {
|
---|
841 | my @metadata_values = $self->sort_metadata_values_array($sort_mode, @{$partition->{'metadata_values'}});
|
---|
842 | my $items_in_partition = $partition->{'size'};
|
---|
843 | $partition_start = $self->generate_partition_start($metadata_values[0], $last_partition_end, $partition_name_length, $is_numeric);
|
---|
844 |
|
---|
845 | if ($items_in_partition <= $partition_size) {
|
---|
846 | # we can just add the partition as is
|
---|
847 | my %metadata_values_to_OIDs_subhashes = ();
|
---|
848 | for (my $i = 0; $i < scalar(@metadata_values); $i++) {
|
---|
849 | my $metadata_value = $metadata_values[$i];
|
---|
850 | $metadata_values_to_OIDs_subhashes{$metadata_value} = $metadata_value_to_OIDs_hash_ref->{$metadata_value};
|
---|
851 | }
|
---|
852 | my $last_metadata_value = $metadata_values[scalar(@metadata_values)-1];
|
---|
853 | $partition_end = $self->generate_partition_end($last_metadata_value, $partition_start, $partition_name_length, $is_numeric);
|
---|
854 | $partition_name = $partition_start;
|
---|
855 | if ($partition_end ne $partition_start) {
|
---|
856 | $partition_name = $partition_name . "-" . $partition_end;
|
---|
857 | }
|
---|
858 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $partition_name, \%metadata_values_to_OIDs_subhashes);
|
---|
859 | $last_partition_end = $partition_end;
|
---|
860 | } else {
|
---|
861 | # we have too many items, need to split the partition
|
---|
862 | my $items_done = 0;
|
---|
863 | my %metadata_values_to_OIDs_subhashes = ();
|
---|
864 | for (my $i = 0; $i < scalar(@metadata_values); $i++) {
|
---|
865 | my $metadata_value = $metadata_values[$i];
|
---|
866 | # If the bookshelf_type is "never", count the documents, otherwise count the distinct metadata values
|
---|
867 | my $items_for_this_md_value = $bookshelf_type eq "never" ? scalar(@{$metadata_value_to_OIDs_hash_ref->{$metadata_value}}) : 1;
|
---|
868 |
|
---|
869 | if ($items_done + $items_for_this_md_value > $partition_size && $items_done != 0) {
|
---|
870 | # Save the stored items into a partition
|
---|
871 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $partition_name, \%metadata_values_to_OIDs_subhashes);
|
---|
872 | $last_partition_end = $partition_end;
|
---|
873 | $partition_start = $self->generate_partition_start($metadata_value, $last_partition_end, $partition_name_length, $is_numeric);
|
---|
874 | $items_done = 0;
|
---|
875 | %metadata_values_to_OIDs_subhashes = ();
|
---|
876 | }
|
---|
877 |
|
---|
878 | # If bookshelf_type is "never" and the current metadata value holds too many items, need to split into several partitions
|
---|
879 | if ($bookshelf_type eq "never" && $items_for_this_md_value > $partition_size) {
|
---|
880 |
|
---|
881 | my $partitionname_for_this_value = $self->generate_partition_start($metadata_value, $last_partition_end, $partition_name_length, $is_numeric);
|
---|
882 | # Get the number of partitions needed for this value
|
---|
883 | my $num_splits = int($items_for_this_md_value / $partition_size);
|
---|
884 | $num_splits++ if ($items_for_this_md_value / $partition_size > $num_splits);
|
---|
885 | my @OIDs_for_this_value = @{$metadata_value_to_OIDs_hash_ref->{$metadata_value}};
|
---|
886 | for (my $i = 0; $i < $num_splits; $i++) {
|
---|
887 | my %OIDs_subhashes_for_this_value = ();
|
---|
888 | my @OIDs_for_this_partition = ();
|
---|
889 | for (my $d = $i * $partition_size; $d < (($i+1) * $partition_size > $items_for_this_md_value ? $items_for_this_md_value : ($i+1) * $partition_size); $d++) {
|
---|
890 | push (@OIDs_for_this_partition, $OIDs_for_this_value[$d]);
|
---|
891 | }
|
---|
892 |
|
---|
893 | # The last bucket might have only a few items and need to be merged with buckets for subsequent metadata values
|
---|
894 | if ($i == $num_splits - 1 && scalar(@OIDs_for_this_partition) < $partition_size) {
|
---|
895 | $partition_start = $partitionname_for_this_value;
|
---|
896 | $partition_name = $partition_start;
|
---|
897 | $metadata_values_to_OIDs_subhashes{$metadata_value} = \@OIDs_for_this_partition;
|
---|
898 | $items_done += scalar(@OIDs_for_this_partition);
|
---|
899 | $last_partition_end = $partitionname_for_this_value
|
---|
900 | } else {
|
---|
901 |
|
---|
902 | # Add an HList for this bucket
|
---|
903 | $OIDs_subhashes_for_this_value{$metadata_value} = \@OIDs_for_this_partition;
|
---|
904 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $partitionname_for_this_value, \%OIDs_subhashes_for_this_value);
|
---|
905 | $last_partition_end = $partitionname_for_this_value;
|
---|
906 | }
|
---|
907 | }
|
---|
908 | } else {
|
---|
909 |
|
---|
910 | $metadata_values_to_OIDs_subhashes{$metadata_value} = $metadata_value_to_OIDs_hash_ref->{$metadata_value};
|
---|
911 | $items_done += $bookshelf_type eq "never" ? scalar(@{$metadata_values_to_OIDs_subhashes{$metadata_value}}) : 1;
|
---|
912 | $partition_end = $self->generate_partition_end($metadata_value, $partition_start, $partition_name_length, $is_numeric);
|
---|
913 | $partition_name = $partition_start;
|
---|
914 | if ($partition_end ne $partition_start) {
|
---|
915 | $partition_name = $partition_name . "-" . $partition_end;
|
---|
916 | }
|
---|
917 |
|
---|
918 | }
|
---|
919 |
|
---|
920 | # The last partition
|
---|
921 | if($i == scalar(@metadata_values) - 1 && $items_done >0) {
|
---|
922 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $partition_name, \%metadata_values_to_OIDs_subhashes);
|
---|
923 | }
|
---|
924 |
|
---|
925 | }
|
---|
926 | } # end if items in partition > partition size
|
---|
927 |
|
---|
928 | }
|
---|
929 |
|
---|
930 | # The partitions are stored in an HList
|
---|
931 | $classifier_node->{'childtype'} = "HList";
|
---|
932 |
|
---|
933 | }
|
---|
934 |
|
---|
935 | sub split_constant_size
|
---|
936 | {
|
---|
937 | my $self = shift(@_);
|
---|
938 | my @metadata_groups = @{shift(@_)};
|
---|
939 | my $classifier_node = shift(@_);
|
---|
940 | my $partition_size = shift(@_);
|
---|
941 | my $sort_mode = shift(@_);
|
---|
942 | my $bookshelf_type = shift(@_);
|
---|
943 | my $metadata_value_to_OIDs_hash_ref = shift(@_);
|
---|
944 | my $partition_name_length = shift(@_);
|
---|
945 | my $is_numeric = shift(@_);
|
---|
946 |
|
---|
947 | my @sortedmetadata_values = $self->sort_metadata_values_array($sort_mode, keys(%$metadata_value_to_OIDs_hash_ref));
|
---|
948 | my $items_in_partition = 0;
|
---|
949 | my %metadata_value_to_OIDs_subhash = ();
|
---|
950 | my $lastpartitionend = "";
|
---|
951 | my $partitionstart;
|
---|
952 |
|
---|
953 | foreach my $metadata_value (@sortedmetadata_values) {
|
---|
954 | if ($items_in_partition == 0) {
|
---|
955 | # a new partition, set the name
|
---|
956 | $partitionstart = $self->generate_partition_start($metadata_value, $lastpartitionend, $partition_name_length, $is_numeric);
|
---|
957 | }
|
---|
958 | my $numitems_for_this_value = ($bookshelf_type eq "never" ? scalar(@{$metadata_value_to_OIDs_hash_ref->{$metadata_value}}) : 1);
|
---|
959 | if ($items_in_partition + $numitems_for_this_value <= $partition_size) {
|
---|
960 | # add all the current values into the temporary list
|
---|
961 | $metadata_value_to_OIDs_subhash{$metadata_value} = $metadata_value_to_OIDs_hash_ref->{$metadata_value};
|
---|
962 | $items_in_partition += $numitems_for_this_value;
|
---|
963 | } elsif ($items_in_partition < $partition_size) {
|
---|
964 | # only want to add some of the values into temporary list
|
---|
965 | # note, we only get here if bookshelf type is never
|
---|
966 | my @OIDs = @{$metadata_value_to_OIDs_hash_ref->{$metadata_value}};
|
---|
967 | @OIDs = $self->sort_leaf_items(\@OIDs);
|
---|
968 | my $num_items_needed = $partition_size - $items_in_partition;
|
---|
969 | my @slice = splice(@OIDs, 0, $num_items_needed);
|
---|
970 | $metadata_value_to_OIDs_subhash{$metadata_value} = \@slice;
|
---|
971 |
|
---|
972 | # now we have filled up the partition
|
---|
973 | my $partitionend = $self->generate_partition_end($metadata_value, $partitionstart, $partition_name_length, $is_numeric);
|
---|
974 | my $partitionname = $partitionstart;
|
---|
975 | if ($partitionend ne $partitionstart) {
|
---|
976 | $partitionname = $partitionname . "-" . $partitionend;
|
---|
977 | }
|
---|
978 |
|
---|
979 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $partitionname, \%metadata_value_to_OIDs_subhash);
|
---|
980 | %metadata_value_to_OIDs_subhash = ();
|
---|
981 | $items_in_partition = 0;
|
---|
982 | $lastpartitionend = $partitionend;
|
---|
983 |
|
---|
984 | # can we get more partitions from this metadata value?
|
---|
985 | while (scalar(@OIDs) >= $partition_size) {
|
---|
986 | my @slice = splice(@OIDs, 0, $partition_size);
|
---|
987 | $metadata_value_to_OIDs_subhash{$metadata_value} = \@slice;
|
---|
988 | $partitionstart = $self->generate_partition_start($metadata_value, $lastpartitionend, $partition_name_length, $is_numeric);
|
---|
989 | my $partitionend = $self->generate_partition_end($metadata_value, $partitionstart, $partition_name_length, $is_numeric);
|
---|
990 | my $partitionname = $partitionstart;
|
---|
991 | if ($partitionend ne $partitionstart) {
|
---|
992 | $partitionname = $partitionname . "-" . $partitionend;
|
---|
993 | }
|
---|
994 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $partitionname, \%metadata_value_to_OIDs_subhash);
|
---|
995 | %metadata_value_to_OIDs_subhash = ();
|
---|
996 | $items_in_partition = 0;
|
---|
997 | $lastpartitionend = $partitionend;
|
---|
998 |
|
---|
999 | }
|
---|
1000 | if (scalar(@OIDs) > 0) {
|
---|
1001 | $metadata_value_to_OIDs_subhash{$metadata_value} = \@OIDs;
|
---|
1002 | $items_in_partition = scalar(@OIDs);
|
---|
1003 | $partitionstart = $self->generate_partition_start($metadata_value, $lastpartitionend, $partition_name_length, $is_numeric);
|
---|
1004 | }
|
---|
1005 |
|
---|
1006 |
|
---|
1007 | }
|
---|
1008 |
|
---|
1009 | if ($items_in_partition == $partition_size) {
|
---|
1010 | # its the end of a partition
|
---|
1011 | my $partitionend = $self->generate_partition_end($metadata_value, $partitionstart, $partition_name_length, $is_numeric);
|
---|
1012 | my $partitionname = $partitionstart;
|
---|
1013 | if ($partitionend ne $partitionstart) {
|
---|
1014 | $partitionname = $partitionname . "-" . $partitionend;
|
---|
1015 | }
|
---|
1016 |
|
---|
1017 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $partitionname, \%metadata_value_to_OIDs_subhash);
|
---|
1018 | %metadata_value_to_OIDs_subhash = ();
|
---|
1019 | $items_in_partition = 0;
|
---|
1020 | $lastpartitionend = $partitionend;
|
---|
1021 | }
|
---|
1022 | } # foreach metadata value
|
---|
1023 |
|
---|
1024 | if ($items_in_partition > 0) {
|
---|
1025 | # we have to add the last partition
|
---|
1026 | my $partitionend = $self->generate_partition_end(@sortedmetadata_values[@sortedmetadata_values-1], $partitionstart, $partition_name_length, $is_numeric);
|
---|
1027 | my $partitionname = $partitionstart;
|
---|
1028 | if ($partitionend ne $partitionstart) {
|
---|
1029 | $partitionname = $partitionname . "-" . $partitionend;
|
---|
1030 | }
|
---|
1031 |
|
---|
1032 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $partitionname, \%metadata_value_to_OIDs_subhash);
|
---|
1033 | }
|
---|
1034 |
|
---|
1035 | # The partitions are stored in an HList
|
---|
1036 | $classifier_node->{'childtype'} = "HList";
|
---|
1037 |
|
---|
1038 | }
|
---|
1039 |
|
---|
1040 | sub split_per_letter_or_digit
|
---|
1041 | {
|
---|
1042 | my $self = shift(@_);
|
---|
1043 | my @metadata_groups = @{shift(@_)};
|
---|
1044 | my $classifier_node = shift(@_);
|
---|
1045 | my $sort_mode = shift(@_);
|
---|
1046 | my $metadata_value_to_OIDs_hash_ref = shift(@_);
|
---|
1047 | my $is_numeric = shift(@_);
|
---|
1048 | my $numeric_partition_length = shift(@_);
|
---|
1049 |
|
---|
1050 | if (not defined $is_numeric) {
|
---|
1051 | $is_numeric = 0;
|
---|
1052 | }
|
---|
1053 | if ($is_numeric && not defined($numeric_partition_length)) {
|
---|
1054 | $numeric_partition_length = 1;
|
---|
1055 | }
|
---|
1056 | # Generate one hlist for each letter
|
---|
1057 | my @sortedmetadata_values = $self->sort_metadata_values_array($sort_mode, keys(%$metadata_value_to_OIDs_hash_ref));
|
---|
1058 | my %metadata_value_to_OIDs_subhash = ();
|
---|
1059 |
|
---|
1060 | my $lastpartition = $self->generate_partition_name($sortedmetadata_values[0], $numeric_partition_length, $is_numeric);
|
---|
1061 | foreach my $metadata_value (@sortedmetadata_values) {
|
---|
1062 |
|
---|
1063 | my $metadata_valuepartition = $self->generate_partition_name($metadata_value, $numeric_partition_length, $is_numeric);
|
---|
1064 |
|
---|
1065 | # Is this the start of a new partition?
|
---|
1066 | if ($metadata_valuepartition ne $lastpartition) {
|
---|
1067 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $lastpartition, \%metadata_value_to_OIDs_subhash);
|
---|
1068 | %metadata_value_to_OIDs_subhash = ();
|
---|
1069 | $lastpartition = $metadata_valuepartition;
|
---|
1070 | }
|
---|
1071 |
|
---|
1072 | $metadata_value_to_OIDs_subhash{$metadata_value} = %$metadata_value_to_OIDs_hash_ref{$metadata_value};
|
---|
1073 | }
|
---|
1074 |
|
---|
1075 | # Don't forget to add the last partition
|
---|
1076 | $self->add_hlist_partition(\@metadata_groups, $classifier_node, $lastpartition, \%metadata_value_to_OIDs_subhash);
|
---|
1077 |
|
---|
1078 | # The partitions are stored in an HList
|
---|
1079 | $classifier_node->{'childtype'} = "HList";
|
---|
1080 |
|
---|
1081 | }
|
---|
1082 |
|
---|
1083 | sub generate_partition_name
|
---|
1084 | {
|
---|
1085 | my $self = shift(@_);
|
---|
1086 | my $mvalue = shift(@_);
|
---|
1087 | my $numeric_partition_length = shift(@_);
|
---|
1088 | my $is_numeric = shift(@_);
|
---|
1089 |
|
---|
1090 | if (!$is_numeric || $numeric_partition_length == 1 ) {
|
---|
1091 | return substr($mvalue, 0, 1);
|
---|
1092 | }
|
---|
1093 | if ($numeric_partition_length == -1) {
|
---|
1094 | my ($all_digits) = $mvalue =~ /^([0-9]+)/;
|
---|
1095 | return $all_digits;
|
---|
1096 | }
|
---|
1097 | my ($some_digits) = $mvalue =~ /^([0-9]{1,$numeric_partition_length})/;
|
---|
1098 | return $some_digits;
|
---|
1099 | }
|
---|
1100 |
|
---|
1101 | sub generate_partition_start
|
---|
1102 | {
|
---|
1103 | my $self = shift(@_);
|
---|
1104 | my $metadata_value = shift(@_);
|
---|
1105 | my $lastpartitionend = shift(@_);
|
---|
1106 | my $partition_name_length = shift(@_);
|
---|
1107 | my $is_numeric = shift(@_);
|
---|
1108 |
|
---|
1109 | if ($is_numeric) {
|
---|
1110 | return $self->generate_partition_name($metadata_value, $partition_name_length, $is_numeric);
|
---|
1111 | }
|
---|
1112 |
|
---|
1113 | if ($partition_name_length) {
|
---|
1114 | return substr($metadata_value, 0, $partition_name_length);
|
---|
1115 | }
|
---|
1116 |
|
---|
1117 | my $partitionstart = substr($metadata_value, 0, 1);
|
---|
1118 | if ($partitionstart le $lastpartitionend) {
|
---|
1119 | $partitionstart = substr($metadata_value, 0, 2);
|
---|
1120 | # Give up after three characters
|
---|
1121 | if ($partitionstart le $lastpartitionend) {
|
---|
1122 | $partitionstart = substr($metadata_value, 0, 3);
|
---|
1123 | }
|
---|
1124 | }
|
---|
1125 |
|
---|
1126 | return $partitionstart;
|
---|
1127 | }
|
---|
1128 |
|
---|
1129 |
|
---|
1130 | sub generate_partition_end
|
---|
1131 | {
|
---|
1132 | my $self = shift(@_);
|
---|
1133 | my $metadata_value = shift(@_);
|
---|
1134 | my $partitionstart = shift(@_);
|
---|
1135 | my $partition_name_length = shift(@_);
|
---|
1136 | my $is_numeric = shift(@_);
|
---|
1137 |
|
---|
1138 | if ($is_numeric) {
|
---|
1139 | return $self->generate_partition_name($metadata_value, $partition_name_length, $is_numeric);
|
---|
1140 | }
|
---|
1141 | if ($partition_name_length) {
|
---|
1142 | return substr($metadata_value, 0, $partition_name_length);
|
---|
1143 | }
|
---|
1144 |
|
---|
1145 | my $partitionend = substr($metadata_value, 0, length($partitionstart));
|
---|
1146 | if ($partitionend gt $partitionstart) {
|
---|
1147 | $partitionend = substr($metadata_value, 0, 1);
|
---|
1148 | if ($partitionend le $partitionstart) {
|
---|
1149 | $partitionend = substr($metadata_value, 0, 2);
|
---|
1150 | # Give up after three characters
|
---|
1151 | if ($partitionend le $partitionstart) {
|
---|
1152 | $partitionend = substr($metadata_value, 0, 3);
|
---|
1153 | }
|
---|
1154 | }
|
---|
1155 | }
|
---|
1156 |
|
---|
1157 | return $partitionend;
|
---|
1158 | }
|
---|
1159 |
|
---|
1160 |
|
---|
1161 | sub add_hlist_partition
|
---|
1162 | {
|
---|
1163 | my $self = shift(@_);
|
---|
1164 | my @metadata_groups = @{shift(@_)};
|
---|
1165 | my $classifier_node = shift(@_);
|
---|
1166 | my $partitionname = shift(@_);
|
---|
1167 | my $metadata_value_to_OIDs_hash_ref = shift(@_);
|
---|
1168 |
|
---|
1169 | # Create an hlist partition
|
---|
1170 | # Note that we don't need to convert from unicode-aware strings
|
---|
1171 | # to utf8 here, as that is handled elsewhere in the code
|
---|
1172 | my %child_classifier_node = ( 'Title' => $partitionname, #'Title' => $self->convert_unicode_string_to_utf8_string($partitionname),
|
---|
1173 | 'childtype' => "VList",
|
---|
1174 | 'contains' => [] );
|
---|
1175 |
|
---|
1176 | # Add the children to the hlist partition
|
---|
1177 | $self->add_vlist(\@metadata_groups, \%child_classifier_node, $metadata_value_to_OIDs_hash_ref);
|
---|
1178 | push(@{$classifier_node->{'contains'}}, \%child_classifier_node);
|
---|
1179 | }
|
---|
1180 |
|
---|
1181 |
|
---|
1182 | sub add_vlist
|
---|
1183 | {
|
---|
1184 | my $self = shift(@_);
|
---|
1185 | my @metadata_groups = @{shift(@_)};
|
---|
1186 | my $classifier_node = shift(@_);
|
---|
1187 | my $metadata_value_to_OIDs_hash_ref = shift(@_);
|
---|
1188 | my $metadata_group = shift(@metadata_groups);
|
---|
1189 | $classifier_node->{'mdtype'} = $metadata_group;
|
---|
1190 |
|
---|
1191 | my $sort_type = $self->{$metadata_group .".metadata_sort_mode_within_level"};
|
---|
1192 | # Create an entry in the vlist for each value
|
---|
1193 | foreach my $metadata_value ($self->sort_metadata_values_array($sort_type, keys(%{$metadata_value_to_OIDs_hash_ref})))
|
---|
1194 | {
|
---|
1195 | my @OIDs = @{$metadata_value_to_OIDs_hash_ref->{$metadata_value}};
|
---|
1196 |
|
---|
1197 | # If there is only one item and 'bookshelf_type' is not always (ie. never or duplicate_only), add the item to the list
|
---|
1198 | if (@OIDs == 1 && $self->{$metadata_group . ".bookshelf_type"} ne "always") {
|
---|
1199 | my $OID = $OIDs[0];
|
---|
1200 | my $offset = $self->metadata_offset($metadata_group, $OID, $metadata_value);
|
---|
1201 | push(@{$classifier_node->{'contains'}}, { 'OID' => $OID, 'offset' => $offset });
|
---|
1202 | }
|
---|
1203 | # If 'bookshelf_type' is 'never', list all the items even if there are duplicated values
|
---|
1204 | elsif ($self->{$metadata_group . ".bookshelf_type"} eq "never") {
|
---|
1205 |
|
---|
1206 | @OIDs = $self->sort_leaf_items(\@OIDs);
|
---|
1207 | foreach my $OID (@OIDs) {
|
---|
1208 | my $offset = $self->metadata_offset($metadata_group, $OID, $metadata_value);
|
---|
1209 | push(@{$classifier_node->{'contains'}}, { 'OID' => $OID , 'offset' => $offset });
|
---|
1210 | }
|
---|
1211 |
|
---|
1212 | }
|
---|
1213 | # Otherwise create a sublist (bookshelf) for the metadata value
|
---|
1214 | else {
|
---|
1215 | my $metadata_value_display = $self->get_metadata_value_display($metadata_group, $metadata_value);
|
---|
1216 | my %child_classifier_node = ( 'Title' => $metadata_value_display,
|
---|
1217 | 'childtype' => "VList",
|
---|
1218 | 'mdtype' => $metadata_group,
|
---|
1219 | 'contains' => [] );
|
---|
1220 |
|
---|
1221 | # If there are metadata elements remaining, recursively apply the process
|
---|
1222 | if (@metadata_groups > 0) {
|
---|
1223 | my $next_metadata_group = $metadata_groups[0];
|
---|
1224 |
|
---|
1225 | # separate metadata into those that belong in the next/sub-metadata_group
|
---|
1226 | # and those that belong at the current level's metadata_group
|
---|
1227 |
|
---|
1228 | my $OID_to_metadata_values_hash_ref = $self->{$next_metadata_group . ".list"};
|
---|
1229 | my @current_level_OIDs = ();
|
---|
1230 | my @next_level_OIDs = ();
|
---|
1231 | foreach my $OID (@OIDs)
|
---|
1232 | {
|
---|
1233 | if ($OID_to_metadata_values_hash_ref->{$OID}) {
|
---|
1234 | push(@next_level_OIDs, $OID);
|
---|
1235 | } else {
|
---|
1236 | push(@current_level_OIDs, $OID);
|
---|
1237 | }
|
---|
1238 | }
|
---|
1239 | # recursively process those docs belonging to the sub-metadata_group
|
---|
1240 | $self->add_level(\@metadata_groups, \@next_level_OIDs, \%child_classifier_node);
|
---|
1241 |
|
---|
1242 | # For those docs that don't belong in the sub/next_metadata_group, but which belong
|
---|
1243 | # at this level, just add the documents as children of this list at the current level
|
---|
1244 | @current_level_OIDs = $self->sort_leaf_items(\@current_level_OIDs);
|
---|
1245 | foreach my $current_level_OID (@current_level_OIDs) {
|
---|
1246 | my $offset = $self->metadata_offset($metadata_group, $current_level_OID, $metadata_value);
|
---|
1247 | push(@{$child_classifier_node{'contains'}}, { 'OID' => $current_level_OID , 'offset' => $offset });
|
---|
1248 | }
|
---|
1249 | }
|
---|
1250 | # Otherwise just add the documents as children of this list
|
---|
1251 | else {
|
---|
1252 | @OIDs = $self->sort_leaf_items(\@OIDs);
|
---|
1253 | foreach my $OID (@OIDs) {
|
---|
1254 | my $offset = $self->metadata_offset($metadata_group, $OID, $metadata_value);
|
---|
1255 | push(@{$child_classifier_node{'contains'}}, { 'OID' => $OID , 'offset' => $offset });
|
---|
1256 | }
|
---|
1257 |
|
---|
1258 | }
|
---|
1259 |
|
---|
1260 | # Add the sublist to the list
|
---|
1261 | push(@{$classifier_node->{'contains'}}, \%child_classifier_node);
|
---|
1262 | }
|
---|
1263 | }
|
---|
1264 | }
|
---|
1265 |
|
---|
1266 | sub metadata_offset
|
---|
1267 | {
|
---|
1268 | my $self = shift(@_);
|
---|
1269 | my $metadata_group = shift(@_);
|
---|
1270 | my $OID = shift(@_);
|
---|
1271 | my $metadata_value = shift(@_);
|
---|
1272 |
|
---|
1273 | my $OID_to_metadata_values_hash_ref = $self->{$metadata_group . ".list"};
|
---|
1274 | my @metadata_values = @{$OID_to_metadata_values_hash_ref->{$OID}};
|
---|
1275 | for (my $i = 0; $i < scalar(@metadata_values); $i++) {
|
---|
1276 | if ($metadata_value eq $metadata_values[$i]) {
|
---|
1277 | return $i;
|
---|
1278 | }
|
---|
1279 | }
|
---|
1280 |
|
---|
1281 | return 0;
|
---|
1282 | }
|
---|
1283 |
|
---|
1284 | sub sort_leaf_items
|
---|
1285 | {
|
---|
1286 | my $self = shift(@_);
|
---|
1287 | my @OIDs = @{shift(@_)};
|
---|
1288 |
|
---|
1289 | # Sort leaf nodes and add to list
|
---|
1290 | my @sort_leaf_nodes_using_metadata_groups = @{$self->{'sort_leaf_nodes_using_metadata_groups'}};
|
---|
1291 | foreach my $sort_leaf_nodes_usingmetaelem (reverse @sort_leaf_nodes_using_metadata_groups) {
|
---|
1292 | my $OID_to_metadata_values_hash_ref = $self->{$sort_leaf_nodes_usingmetaelem . ".list"};
|
---|
1293 | my $sort_type = $self->{$sort_leaf_nodes_usingmetaelem . ".metadata_sort_mode_within_level"};
|
---|
1294 | # Force a stable sort (Perl 5.6's sort isn't stable)
|
---|
1295 | # !! The [0] bits aren't ideal (multiple metadata values) !!
|
---|
1296 | @OIDs = @OIDs[ sort {
|
---|
1297 | if (defined($OID_to_metadata_values_hash_ref->{$OIDs[$a]}) && defined($OID_to_metadata_values_hash_ref->{$OIDs[$b]}))
|
---|
1298 | {
|
---|
1299 | if ($sort_type eq "numeric") {
|
---|
1300 | $OID_to_metadata_values_hash_ref->{$OIDs[$a]}[0] <=> $OID_to_metadata_values_hash_ref->{$OIDs[$b]}[0];
|
---|
1301 | } elsif ($sort_type eq "alphabetic") {
|
---|
1302 | $OID_to_metadata_values_hash_ref->{$OIDs[$a]}[0] cmp $OID_to_metadata_values_hash_ref->{$OIDs[$b]}[0];
|
---|
1303 | } else {
|
---|
1304 | ncmp($OID_to_metadata_values_hash_ref->{$OIDs[$a]}[0], $OID_to_metadata_values_hash_ref->{$OIDs[$b]}[0]);
|
---|
1305 | }
|
---|
1306 | }
|
---|
1307 | else
|
---|
1308 | {
|
---|
1309 | $a <=> $b;
|
---|
1310 | }
|
---|
1311 | } 0..$#OIDs ];
|
---|
1312 | }
|
---|
1313 | if ($self->{'reverse_sort_leaf_nodes'}) {
|
---|
1314 | return reverse @OIDs;
|
---|
1315 | }
|
---|
1316 | return @OIDs;
|
---|
1317 | }
|
---|
1318 |
|
---|
1319 |
|
---|
1320 | sub sort_metadata_values_array
|
---|
1321 | {
|
---|
1322 | my $self = shift(@_);
|
---|
1323 | my ($sort_mode) = shift(@_);
|
---|
1324 | my @metadata_values = @_;
|
---|
1325 |
|
---|
1326 | if ($sort_mode eq "unicode") {
|
---|
1327 | if ($self->{'unicode_collator'}) {
|
---|
1328 | return $self->{'unicode_collator'}->sort(@metadata_values);
|
---|
1329 | }
|
---|
1330 | # the collator wasn't loaded, fall back on default
|
---|
1331 | $sort_mode = "alphanumeric";
|
---|
1332 | }
|
---|
1333 | if ($sort_mode eq "numeric") {
|
---|
1334 | return sort {$a <=> $b} @metadata_values;
|
---|
1335 | }
|
---|
1336 | if ($sort_mode eq "alphabetic") {
|
---|
1337 | return sort {$a cmp $b} @metadata_values;
|
---|
1338 | }
|
---|
1339 | # natural sort
|
---|
1340 | return nsort(@metadata_values);
|
---|
1341 | }
|
---|
1342 |
|
---|
1343 |
|
---|
1344 | # we are not using this any more. Using nsort instead
|
---|
1345 | # $a and $b args automatically passed in and shouldn't be declared
|
---|
1346 | sub alpha_numeric_cmp
|
---|
1347 | {
|
---|
1348 | my $self = shift (@_);
|
---|
1349 | my ($aStr, $bStr) = @_;
|
---|
1350 | if ($aStr =~ m/^(\d+(\.\d+)?)/)
|
---|
1351 | {
|
---|
1352 | my $val_a = $1;
|
---|
1353 | if ($bStr =~ m/^(\d+(\.\d+)?)/)
|
---|
1354 | {
|
---|
1355 | my $val_b = $1;
|
---|
1356 | if ($val_a != $val_b)
|
---|
1357 | {
|
---|
1358 | return ($val_a <=> $val_b);
|
---|
1359 | }
|
---|
1360 | }
|
---|
1361 | }
|
---|
1362 |
|
---|
1363 | return ($aStr cmp $bStr);
|
---|
1364 | }
|
---|
1365 |
|
---|
1366 |
|
---|
1367 | sub get_metadata_value_display {
|
---|
1368 | my $self = shift(@_);
|
---|
1369 | my ($metadata_group, $metadata_value) = @_;
|
---|
1370 | return $metadata_value if $self->{'standardize_capitalization'};
|
---|
1371 | my $actual_values_hash = $self->{$metadata_group . ".actualvalues"}->{$metadata_value};
|
---|
1372 | my $display_value ="";
|
---|
1373 | my $max_count=0;
|
---|
1374 | foreach my $v (keys %$actual_values_hash) {
|
---|
1375 | if ($actual_values_hash->{$v} > $max_count) {
|
---|
1376 | $display_value = $v;
|
---|
1377 | $max_count = $actual_values_hash->{$v};
|
---|
1378 | }
|
---|
1379 | }
|
---|
1380 | return $display_value;
|
---|
1381 | }
|
---|
1382 | 1;
|
---|