| 1 | ###########################################################################
|
|---|
| 2 | #
|
|---|
| 3 | # AZList.pm --
|
|---|
| 4 | # A component of the Greenstone digital library software
|
|---|
| 5 | # from the New Zealand Digital Library Project at the
|
|---|
| 6 | # University of Waikato, New Zealand.
|
|---|
| 7 | #
|
|---|
| 8 | # Copyright (C) 1999 New Zealand Digital Library Project
|
|---|
| 9 | #
|
|---|
| 10 | # This program is free software; you can redistribute it and/or modify
|
|---|
| 11 | # it under the terms of the GNU General Public License as published by
|
|---|
| 12 | # the Free Software Foundation; either version 2 of the License, or
|
|---|
| 13 | # (at your option) any later version.
|
|---|
| 14 | #
|
|---|
| 15 | # This program is distributed in the hope that it will be useful,
|
|---|
| 16 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|---|
| 17 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|---|
| 18 | # GNU General Public License for more details.
|
|---|
| 19 | #
|
|---|
| 20 | # You should have received a copy of the GNU General Public License
|
|---|
| 21 | # along with this program; if not, write to the Free Software
|
|---|
| 22 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
|---|
| 23 | #
|
|---|
| 24 | ###########################################################################
|
|---|
| 25 |
|
|---|
| 26 | # classifier plugin for sorting alphabetically
|
|---|
| 27 |
|
|---|
| 28 | package AZList;
|
|---|
| 29 |
|
|---|
| 30 | use BasClas;
|
|---|
| 31 | use sorttools;
|
|---|
| 32 | use iso639;
|
|---|
| 33 |
|
|---|
| 34 | sub BEGIN {
|
|---|
| 35 | @ISA = ('BasClas');
|
|---|
| 36 | }
|
|---|
| 37 |
|
|---|
| 38 | use strict;
|
|---|
| 39 |
|
|---|
| 40 | my $arguments =
|
|---|
| 41 | [ { 'name' => "metadata",
|
|---|
| 42 | 'desc' => "{List.metadata}",
|
|---|
| 43 | 'type' => "metadata",
|
|---|
| 44 | 'reqd' => "yes" } ,
|
|---|
| 45 | { 'name' => "buttonname",
|
|---|
| 46 | 'desc' => "{AZList.buttonname}",
|
|---|
| 47 | 'type' => "string",
|
|---|
| 48 | 'deft' => "{BasClas.metadata.deft}",
|
|---|
| 49 | 'reqd' => "no" },
|
|---|
| 50 | { 'name' => "removeprefix",
|
|---|
| 51 | 'desc' => "{AZList.removeprefix}",
|
|---|
| 52 | 'type' => "regexp",
|
|---|
| 53 | 'deft' => "",
|
|---|
| 54 | 'reqd' => "no" } ,
|
|---|
| 55 | { 'name' => "removesuffix",
|
|---|
| 56 | 'desc' => "{AZList.removesuffix}",
|
|---|
| 57 | 'type' => "regexp",
|
|---|
| 58 | 'deft' => "",
|
|---|
| 59 | 'reqd' => "no" }
|
|---|
| 60 | ];
|
|---|
| 61 |
|
|---|
| 62 | my $options = { 'name' => "AZList",
|
|---|
| 63 | 'desc' => "{AZList.desc}",
|
|---|
| 64 | 'abstract' => "no",
|
|---|
| 65 | 'inherits' => "yes",
|
|---|
| 66 | 'args' => $arguments };
|
|---|
| 67 |
|
|---|
| 68 | # sub print_usage {
|
|---|
| 69 | # print STDERR "
|
|---|
| 70 | # usage: classify AZList [options]
|
|---|
| 71 | # options:
|
|---|
| 72 |
|
|---|
| 73 | # -metadata X (required) Metadata field used for classification.
|
|---|
| 74 | # List will be sorted by this element.
|
|---|
| 75 |
|
|---|
| 76 | # -buttonname X (optional) Button name for this classification.
|
|---|
| 77 | # defaults to metadata name.
|
|---|
| 78 |
|
|---|
| 79 | # -removeprefix regex (optional) A prefix to ignore in the Metadata values
|
|---|
| 80 | # for the field when sorting.
|
|---|
| 81 | # ";
|
|---|
| 82 | # }
|
|---|
| 83 |
|
|---|
| 84 | sub new {
|
|---|
| 85 | my $class = shift (@_);
|
|---|
| 86 | my $self = new BasClas($class, @_);
|
|---|
| 87 |
|
|---|
| 88 | # 14-05-02 To allow for proper inheritance of arguments - John Thompson
|
|---|
| 89 | my $option_list = $self->{'option_list'};
|
|---|
| 90 | push( @{$option_list}, $options );
|
|---|
| 91 |
|
|---|
| 92 | my ($metaname, $title, $removeprefix, $removesuffix);
|
|---|
| 93 |
|
|---|
| 94 | if (!parsargv::parse(\@_,
|
|---|
| 95 | q^metadata/.*/^, \$metaname,
|
|---|
| 96 | q^buttonname/.*/^, \$title,
|
|---|
| 97 | q^removeprefix/.*/^, \$removeprefix,
|
|---|
| 98 | q^removesuffix/.*/^, \$removesuffix,
|
|---|
| 99 | "allow_extra_options")) {
|
|---|
| 100 |
|
|---|
| 101 | print STDERR "\nIncorrect options passed to $class, check your collect.cfg file\n";
|
|---|
| 102 | $self->print_txt_usage(""); # Use default resource bundle
|
|---|
| 103 | die "\n";
|
|---|
| 104 | }
|
|---|
| 105 |
|
|---|
| 106 | if (!defined $metaname) {
|
|---|
| 107 | $self->print_txt_usage(""); # Use default resource bundle
|
|---|
| 108 | print STDERR "AZList used with no metadata name\n";
|
|---|
| 109 | die "\n";
|
|---|
| 110 | }
|
|---|
| 111 |
|
|---|
| 112 | $title = $metaname unless ($title);
|
|---|
| 113 |
|
|---|
| 114 | $self->{'list'} = {};
|
|---|
| 115 | $self->{'metaname'} = $metaname;
|
|---|
| 116 | $self->{'title'} = $title;
|
|---|
| 117 | if (defined($removeprefix) && $removeprefix) {
|
|---|
| 118 | $removeprefix =~ s/^\^//; # don't need a leading ^
|
|---|
| 119 | $self->{'removeprefix'} = $removeprefix;
|
|---|
| 120 | }
|
|---|
| 121 | if (defined($removesuffix) && $removesuffix) {
|
|---|
| 122 | $removesuffix =~ s/\$$//; # don't need a trailing $
|
|---|
| 123 | $self->{'removesuffix'} = $removesuffix;
|
|---|
| 124 | }
|
|---|
| 125 |
|
|---|
| 126 | return bless $self, $class;
|
|---|
| 127 | }
|
|---|
| 128 |
|
|---|
| 129 | sub init {
|
|---|
| 130 | my $self = shift (@_);
|
|---|
| 131 |
|
|---|
| 132 | $self->{'list'} = {};
|
|---|
| 133 | }
|
|---|
| 134 |
|
|---|
| 135 | sub classify {
|
|---|
| 136 | my $self = shift (@_);
|
|---|
| 137 | my ($doc_obj) = @_;
|
|---|
| 138 |
|
|---|
| 139 | my $doc_OID = $doc_obj->get_OID();
|
|---|
| 140 | my $metavalue = $doc_obj->get_metadata_element ($doc_obj->get_top_section(), $self->{'metaname'}, $self->{'ignore_namespace'});
|
|---|
| 141 |
|
|---|
| 142 | # if this document doesn't contain the metadata element we're
|
|---|
| 143 | # sorting by we won't include it in this classification
|
|---|
| 144 | if (defined $metavalue && $metavalue ne "") {
|
|---|
| 145 | if (defined($self->{'removeprefix'}) &&
|
|---|
| 146 | length($self->{'removeprefix'})) {
|
|---|
| 147 | $metavalue =~ s/^$self->{'removeprefix'}//;
|
|---|
| 148 | }
|
|---|
| 149 | if (defined($self->{'removesuffix'}) &&
|
|---|
| 150 | length($self->{'removesuffix'})) {
|
|---|
| 151 | $metavalue =~ s/$self->{'removesuffix'}$//;
|
|---|
| 152 | }
|
|---|
| 153 |
|
|---|
| 154 | if ($self->{'metaname'} eq 'Language') {
|
|---|
| 155 | $metavalue = $iso639::fromiso639{$metavalue};
|
|---|
| 156 | } elsif ($self->{'metaname'} eq 'Creator') {
|
|---|
| 157 | &sorttools::format_string_name_english (\$metavalue);
|
|---|
| 158 | } else {
|
|---|
| 159 | &sorttools::format_string_english (\$metavalue);
|
|---|
| 160 | }
|
|---|
| 161 | if (defined $self->{'list'}->{$doc_OID}) {
|
|---|
| 162 | my $outhandle = $self->{'outhandle'};
|
|---|
| 163 | print $outhandle "WARNING: AZList::classify called multiple times for $doc_OID\n";
|
|---|
| 164 | }
|
|---|
| 165 | if ($metavalue) {
|
|---|
| 166 | $self->{'list'}->{$doc_OID} = $metavalue;
|
|---|
| 167 | } else {
|
|---|
| 168 | my $outhandle = $self->{'outhandle'};
|
|---|
| 169 | print $outhandle "WARNING: AZList: $doc_OID metadata is empty - not classifying\n";
|
|---|
| 170 | }
|
|---|
| 171 | }
|
|---|
| 172 | }
|
|---|
| 173 |
|
|---|
| 174 | sub alpha_numeric_cmp
|
|---|
| 175 | {
|
|---|
| 176 | my ($self,$a,$b) = @_;
|
|---|
| 177 |
|
|---|
| 178 | my $title_a = $self->{'list'}->{$a};
|
|---|
| 179 | my $title_b = $self->{'list'}->{$b};
|
|---|
| 180 |
|
|---|
| 181 | if ($title_a =~ m/^(\d+(\.\d+)?)/)
|
|---|
| 182 | {
|
|---|
| 183 | my $val_a = $1;
|
|---|
| 184 | if ($title_b =~ m/^(\d+(\.\d+)?)/)
|
|---|
| 185 | {
|
|---|
| 186 | my $val_b = $1;
|
|---|
| 187 | if ($val_a != $val_b)
|
|---|
| 188 | {
|
|---|
| 189 | return ($val_a <=> $val_b);
|
|---|
| 190 | }
|
|---|
| 191 | }
|
|---|
| 192 | }
|
|---|
| 193 |
|
|---|
| 194 | return ($title_a cmp $title_b);
|
|---|
| 195 | }
|
|---|
| 196 |
|
|---|
| 197 | sub get_classify_info {
|
|---|
| 198 | my $self = shift (@_);
|
|---|
| 199 |
|
|---|
| 200 | my @classlist
|
|---|
| 201 | = sort { $self->alpha_numeric_cmp($a,$b) } keys %{$self->{'list'}};
|
|---|
| 202 |
|
|---|
| 203 | return $self->splitlist (\@classlist);
|
|---|
| 204 | }
|
|---|
| 205 |
|
|---|
| 206 | sub get_entry {
|
|---|
| 207 | my $self = shift (@_);
|
|---|
| 208 | my ($title, $childtype, $thistype) = @_;
|
|---|
| 209 |
|
|---|
| 210 | # organise into classification structure
|
|---|
| 211 | my %classifyinfo = ('childtype'=>$childtype,
|
|---|
| 212 | 'Title'=>$title,
|
|---|
| 213 | 'contains'=>[]);
|
|---|
| 214 | $classifyinfo{'thistype'} = $thistype
|
|---|
| 215 | if defined $thistype && $thistype =~ /\w/;
|
|---|
| 216 |
|
|---|
| 217 | return \%classifyinfo;
|
|---|
| 218 | }
|
|---|
| 219 |
|
|---|
| 220 | # splitlist takes an ordered list of classifications (@$classlistref) and splits it
|
|---|
| 221 | # up into alphabetical sub-sections.
|
|---|
| 222 | sub splitlist {
|
|---|
| 223 | my $self = shift (@_);
|
|---|
| 224 | my ($classlistref) = @_;
|
|---|
| 225 | my $classhash = {};
|
|---|
| 226 |
|
|---|
| 227 | # top level
|
|---|
| 228 | my $childtype = "HList";
|
|---|
| 229 | if (scalar (@$classlistref) <= 39) {$childtype = "VList";}
|
|---|
| 230 | my $classifyinfo = $self->get_entry ($self->{'title'}, $childtype, "Invisible");
|
|---|
| 231 |
|
|---|
| 232 | # don't need to do any splitting if there are less than 39 (max + min -1) classifications
|
|---|
| 233 | if ((scalar @$classlistref) <= 39) {
|
|---|
| 234 | foreach my $subOID (@$classlistref) {
|
|---|
| 235 | push (@{$classifyinfo->{'contains'}}, {'OID'=>$subOID});
|
|---|
| 236 | }
|
|---|
| 237 | return $classifyinfo;
|
|---|
| 238 | }
|
|---|
| 239 |
|
|---|
| 240 | # first split up the list into separate A-Z and 0-9 classifications
|
|---|
| 241 | foreach my $classification (@$classlistref) {
|
|---|
| 242 | my $title = $self->{'list'}->{$classification};
|
|---|
| 243 |
|
|---|
| 244 | $title =~ s/^(&.{1,6};|<[^>]>|[^a-zA-Z0-9])//g; # remove any unwanted stuff
|
|---|
| 245 | # only need first char for classification
|
|---|
| 246 | $title =~ m/^(.)/; $title=$1;
|
|---|
| 247 | $title =~ tr/[a-z]/[A-Z]/;
|
|---|
| 248 | if ($title =~ /^[0-9]$/) {$title = '0-9';}
|
|---|
| 249 | elsif ($title !~ /^[A-Z]$/) {
|
|---|
| 250 | my $outhandle = $self->{'outhandle'};
|
|---|
| 251 | print $outhandle "AZList: WARNING $classification has badly formatted title ($title)\n";
|
|---|
| 252 | }
|
|---|
| 253 | $classhash->{$title} = [] unless defined $classhash->{$title};
|
|---|
| 254 | push (@{$classhash->{$title}}, $classification);
|
|---|
| 255 | }
|
|---|
| 256 | $classhash = $self->compactlist ($classhash);
|
|---|
| 257 |
|
|---|
| 258 | my @tmparr = ();
|
|---|
| 259 | foreach my $subsection (sort keys (%$classhash)) {
|
|---|
| 260 | push (@tmparr, $subsection);
|
|---|
| 261 | }
|
|---|
| 262 |
|
|---|
| 263 | # if there's a 0-9 section it will have been sorted to the beginning
|
|---|
| 264 | # but we want it at the end
|
|---|
| 265 | if ($tmparr[0] eq '0-9') {
|
|---|
| 266 | shift @tmparr;
|
|---|
| 267 | push (@tmparr, '0-9');
|
|---|
| 268 | }
|
|---|
| 269 |
|
|---|
| 270 | foreach my $subclass (@tmparr) {
|
|---|
| 271 | my $tempclassify = $self->get_entry($subclass, "VList");
|
|---|
| 272 | foreach my $subsubOID (@{$classhash->{$subclass}}) {
|
|---|
| 273 | push (@{$tempclassify->{'contains'}}, {'OID'=>$subsubOID});
|
|---|
| 274 | }
|
|---|
| 275 | push (@{$classifyinfo->{'contains'}}, $tempclassify);
|
|---|
| 276 | }
|
|---|
| 277 |
|
|---|
| 278 | return $classifyinfo;
|
|---|
| 279 | }
|
|---|
| 280 |
|
|---|
| 281 | sub compactlist {
|
|---|
| 282 | my $self = shift (@_);
|
|---|
| 283 | my ($classhashref) = @_;
|
|---|
| 284 | my $compactedhash = {};
|
|---|
| 285 | my @currentOIDs = ();
|
|---|
| 286 | my $currentfirstletter = ""; # start of working bin
|
|---|
| 287 | my $currentlastletter = ""; # end of working bin
|
|---|
| 288 | my $lastkey = ""; # the name of the last completed key
|
|---|
| 289 |
|
|---|
| 290 | # minimum and maximum documents to be displayed per page.
|
|---|
| 291 | # the actual maximum will be max + (min-1).
|
|---|
| 292 | # the smallest sub-section is a single letter at present
|
|---|
| 293 | # so in this case there may be many times max documents
|
|---|
| 294 | # displayed on a page.
|
|---|
| 295 | my $min = 10;
|
|---|
| 296 | my $max = 30;
|
|---|
| 297 |
|
|---|
| 298 | foreach my $subsection (sort keys %$classhashref) {
|
|---|
| 299 | if ($subsection eq '0-9') {
|
|---|
| 300 | # leave this bin as-is... copy it straight across
|
|---|
| 301 | @{$compactedhash->{$subsection}} = @{$classhashref->{$subsection}};
|
|---|
| 302 | next;
|
|---|
| 303 | }
|
|---|
| 304 | $currentfirstletter = $subsection if $currentfirstletter eq "";
|
|---|
| 305 | if ((scalar (@currentOIDs) < $min) ||
|
|---|
| 306 | ((scalar (@currentOIDs) + scalar (@{$classhashref->{$subsection}})) <= $max)) {
|
|---|
| 307 | # add this letter to the bin and continue
|
|---|
| 308 | push (@currentOIDs, @{$classhashref->{$subsection}});
|
|---|
| 309 | $currentlastletter = $subsection;
|
|---|
| 310 | } else {
|
|---|
| 311 | # too many or too few for a separate bin
|
|---|
| 312 | if ($currentfirstletter eq $currentlastletter) {
|
|---|
| 313 | @{$compactedhash->{$currentfirstletter}} = @currentOIDs;
|
|---|
| 314 | $lastkey = $currentfirstletter;
|
|---|
| 315 | } else {
|
|---|
| 316 | @{$compactedhash->{"$currentfirstletter-$currentlastletter"}} = @currentOIDs;
|
|---|
| 317 | $lastkey = "$currentfirstletter-$currentlastletter";
|
|---|
| 318 | }
|
|---|
| 319 | if (scalar (@{$classhashref->{$subsection}}) >= $max) {
|
|---|
| 320 | # this key is now complete. Start a new one
|
|---|
| 321 | $compactedhash->{$subsection} = $classhashref->{$subsection};
|
|---|
| 322 | @currentOIDs = ();
|
|---|
| 323 | $currentfirstletter = "";
|
|---|
| 324 | $lastkey = $subsection;
|
|---|
| 325 | } else {
|
|---|
| 326 | @currentOIDs = @{$classhashref->{$subsection}};
|
|---|
| 327 | $currentfirstletter = $subsection;
|
|---|
| 328 | $currentlastletter = $subsection;
|
|---|
| 329 | }
|
|---|
| 330 | }
|
|---|
| 331 | }
|
|---|
| 332 |
|
|---|
| 333 | # add final OIDs to last sub-classification if there aren't many otherwise
|
|---|
| 334 | # add final sub-classification
|
|---|
| 335 | # BUG FIX: don't add anything if there are no currentOIDs (thanks to Don Gourley)
|
|---|
| 336 | if (! scalar(@currentOIDs)) {return $compactedhash;}
|
|---|
| 337 |
|
|---|
| 338 | if (scalar (@currentOIDs) < $min) {
|
|---|
| 339 | my ($newkey) = $lastkey =~ /^(.)/;
|
|---|
| 340 | @currentOIDs = (@{$compactedhash->{$lastkey}}, @currentOIDs);
|
|---|
| 341 | delete $compactedhash->{$lastkey};
|
|---|
| 342 | @{$compactedhash->{"$newkey-$currentlastletter"}} = @currentOIDs;
|
|---|
| 343 | } else {
|
|---|
| 344 | if ($currentfirstletter eq $currentlastletter) {
|
|---|
| 345 | @{$compactedhash->{$currentfirstletter}} = @currentOIDs;
|
|---|
| 346 | }
|
|---|
| 347 | else {
|
|---|
| 348 | @{$compactedhash->{"$currentfirstletter-$currentlastletter"}} =
|
|---|
| 349 | @currentOIDs;
|
|---|
| 350 | }
|
|---|
| 351 | }
|
|---|
| 352 |
|
|---|
| 353 | return $compactedhash;
|
|---|
| 354 | }
|
|---|
| 355 |
|
|---|
| 356 | 1;
|
|---|
