[4726] | 1 | ###########################################################################
|
---|
| 2 | #
|
---|
| 3 | # OAIPlug.pm -- basic Open Archives Initiative (OAI) plugin
|
---|
| 4 | #
|
---|
| 5 | # A component of the Greenstone digital library software
|
---|
| 6 | # from the New Zealand Digital Library Project at the
|
---|
| 7 | # University of Waikato, New Zealand.
|
---|
| 8 | #
|
---|
| 9 | # Copyright (C) 1999 New Zealand Digital Library Project
|
---|
| 10 | #
|
---|
| 11 | # This program is free software; you can redistribute it and/or modify
|
---|
| 12 | # it under the terms of the GNU General Public License as published by
|
---|
| 13 | # the Free Software Foundation; either version 2 of the License, or
|
---|
| 14 | # (at your option) any later version.
|
---|
| 15 | #
|
---|
| 16 | # This program is distributed in the hope that it will be useful,
|
---|
| 17 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
| 18 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
| 19 | # GNU General Public License for more details.
|
---|
| 20 | #
|
---|
| 21 | # You should have received a copy of the GNU General Public License
|
---|
| 22 | # along with this program; if not, write to the Free Software
|
---|
| 23 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
| 24 | #
|
---|
| 25 | ###########################################################################
|
---|
| 26 |
|
---|
[15872] | 27 | package OAIPlugin;
|
---|
[4726] | 28 |
|
---|
[24951] | 29 | use extrametautil;
|
---|
[4726] | 30 | use unicode;
|
---|
| 31 | use util;
|
---|
| 32 |
|
---|
[10254] | 33 | use strict;
|
---|
| 34 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
---|
| 35 |
|
---|
[15872] | 36 | use ReadXMLFile;
|
---|
[17066] | 37 | use ReadTextFile; # needed for subroutine textcat_get_language_encoding
|
---|
[17216] | 38 | use metadatautil;
|
---|
[24547] | 39 | use MetadataRead;
|
---|
[24971] | 40 | use util;
|
---|
[9958] | 41 |
|
---|
[24547] | 42 | # methods with identical signatures take precedence in the order given in the ISA list.
|
---|
[4726] | 43 | sub BEGIN {
|
---|
[24547] | 44 | @OAIPlugin::ISA = ('MetadataRead', 'ReadXMLFile', 'ReadTextFile');
|
---|
[4726] | 45 | }
|
---|
| 46 |
|
---|
[19213] | 47 | my $set_list =
|
---|
| 48 | [ { 'name' => "auto",
|
---|
| 49 | 'desc' => "{OAIPlugin.metadata_set.auto}" },
|
---|
| 50 | { 'name' => "dc",
|
---|
| 51 | 'desc' => "{OAIPlugin.metadata_set.dc}" }
|
---|
| 52 | ];
|
---|
[9958] | 53 |
|
---|
[6408] | 54 | my $arguments =
|
---|
| 55 | [ { 'name' => "process_exp",
|
---|
[16013] | 56 | 'desc' => "{BasePlugin.process_exp}",
|
---|
[6408] | 57 | 'type' => "regexp",
|
---|
| 58 | 'reqd' => "no",
|
---|
[17290] | 59 | 'deft' => &get_default_process_exp() },
|
---|
[19213] | 60 | { 'name' => "metadata_set",
|
---|
| 61 | 'desc' => "{OAIPlugin.metadata_set}",
|
---|
| 62 | 'type' => "enumstring",
|
---|
| 63 | 'reqd' => "no",
|
---|
| 64 | 'list' => $set_list,
|
---|
| 65 | 'deft' => "dc" },
|
---|
[17319] | 66 | { 'name' => "document_field",
|
---|
| 67 | 'desc' => "{OAIPlugin.document_field}",
|
---|
[17290] | 68 | 'type' => "metadata",
|
---|
| 69 | 'reqd' => "no",
|
---|
| 70 | 'deft' => "gi.Sourcedoc" }
|
---|
[6408] | 71 | ];
|
---|
| 72 |
|
---|
[15872] | 73 | my $options = { 'name' => "OAIPlugin",
|
---|
| 74 | 'desc' => "{OAIPlugin.desc}",
|
---|
[6408] | 75 | 'abstract' => "no",
|
---|
| 76 | 'inherits' => "yes",
|
---|
[17103] | 77 | 'explodes' => "yes",
|
---|
[6408] | 78 | 'args' => $arguments };
|
---|
[4747] | 79 |
|
---|
[10254] | 80 |
|
---|
[4726] | 81 | sub new {
|
---|
[10218] | 82 | my ($class) = shift (@_);
|
---|
| 83 | my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
|
---|
| 84 | push(@$pluginlist, $class);
|
---|
[4873] | 85 |
|
---|
[15872] | 86 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
---|
| 87 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
---|
[4726] | 88 |
|
---|
[17126] | 89 | new ReadTextFile($pluginlist, $inputargs, $hashArgOptLists,1);
|
---|
[15872] | 90 | my $self = new ReadXMLFile($pluginlist, $inputargs, $hashArgOptLists);
|
---|
[4726] | 91 |
|
---|
[21763] | 92 | if ($self->{'info_only'}) {
|
---|
| 93 | # don't worry about modifying options
|
---|
| 94 | return bless $self, $class;
|
---|
| 95 | }
|
---|
[24404] | 96 | # trim any ex. from document field iff it's the only metadata namespace prefix
|
---|
| 97 | $self->{'document_field'} =~ s/^ex\.([^.]+)$/$1/;
|
---|
[4726] | 98 | return bless $self, $class;
|
---|
| 99 | }
|
---|
| 100 |
|
---|
| 101 | sub get_default_process_exp {
|
---|
| 102 | my $self = shift (@_);
|
---|
| 103 |
|
---|
| 104 | return q^(?i)(\.oai)$^;
|
---|
| 105 | }
|
---|
| 106 |
|
---|
[13222] | 107 | sub get_doctype {
|
---|
| 108 | my $self = shift(@_);
|
---|
| 109 |
|
---|
| 110 | return "OAI-PMH";
|
---|
| 111 | }
|
---|
| 112 |
|
---|
[9958] | 113 | sub xml_start_document {
|
---|
[10254] | 114 | my $self = shift (@_);
|
---|
[9958] | 115 | $self->{'in_metadata_node'} = 0;
|
---|
| 116 | $self->{'rawxml'} = "";
|
---|
[17290] | 117 | $self->{'saved_metadata'} = {};
|
---|
[9958] | 118 | }
|
---|
[4726] | 119 |
|
---|
[9958] | 120 | sub xml_end_document {
|
---|
| 121 | }
|
---|
[4726] | 122 |
|
---|
[9958] | 123 | sub xml_doctype {
|
---|
| 124 | my $self = shift(@_);
|
---|
| 125 |
|
---|
| 126 | my ($expat, $name, $sysid, $pubid, $internal) = @_;
|
---|
| 127 |
|
---|
[13886] | 128 | ##die "" if ($name !~ /^OAI-PMH$/);
|
---|
[9958] | 129 |
|
---|
[4726] | 130 | my $outhandle = $self->{'outhandle'};
|
---|
[15872] | 131 | print $outhandle "OAIPlugin: processing $self->{'file'}\n" if $self->{'verbosity'} > 1;
|
---|
| 132 | print STDERR "<Processing n='$self->{'file'}' p='OAIPlugin'>\n" if $self->{'gli'};
|
---|
[4726] | 133 |
|
---|
[9958] | 134 | }
|
---|
[4726] | 135 |
|
---|
[9958] | 136 |
|
---|
| 137 | sub xml_start_tag {
|
---|
| 138 | my $self = shift(@_);
|
---|
| 139 | my ($expat,$element) = @_;
|
---|
| 140 |
|
---|
| 141 | my %attr_hash = %_;
|
---|
| 142 |
|
---|
| 143 | my $attr = "";
|
---|
| 144 | map { $attr .= " $_=$attr_hash{$_}"; } keys %attr_hash;
|
---|
| 145 |
|
---|
| 146 | $self->{'rawxml'} .= "<$element$attr>";
|
---|
| 147 |
|
---|
| 148 | if ($element eq "metadata") {
|
---|
| 149 | $self->{'in_metadata_node'} = 1;
|
---|
| 150 | $self->{'metadata_xml'} = "";
|
---|
[4726] | 151 | }
|
---|
[9958] | 152 |
|
---|
| 153 | if ($self->{'in_metadata_node'}) {
|
---|
| 154 | $self->{'metadata_xml'} .= "<$element$attr>";
|
---|
[4726] | 155 | }
|
---|
[9958] | 156 | }
|
---|
[4726] | 157 |
|
---|
[9958] | 158 | sub xml_end_tag {
|
---|
| 159 | my $self = shift(@_);
|
---|
| 160 | my ($expat, $element) = @_;
|
---|
[4726] | 161 |
|
---|
[9958] | 162 | $self->{'rawxml'} .= "</$element>";
|
---|
[4726] | 163 |
|
---|
[9958] | 164 | if ($self->{'in_metadata_node'}) {
|
---|
| 165 | $self->{'metadata_xml'} .= "</$element>";
|
---|
[4726] | 166 | }
|
---|
| 167 |
|
---|
[9958] | 168 | if ($element eq "metadata") {
|
---|
| 169 | my $textref = \$self->{'metadata_xml'};
|
---|
[17290] | 170 | #my $metadata = $self->{'metadata'};
|
---|
| 171 | my $metadata = $self->{'saved_metadata'};
|
---|
[9958] | 172 | $self->extract_oai_metadata($textref,$metadata);
|
---|
[4726] | 173 |
|
---|
[9958] | 174 | $self->{'in_metadata_node'} = 0;
|
---|
| 175 | }
|
---|
[4726] | 176 |
|
---|
| 177 |
|
---|
[9958] | 178 | }
|
---|
[4726] | 179 |
|
---|
[9958] | 180 | sub xml_text {
|
---|
| 181 | my $self = shift(@_);
|
---|
| 182 | my ($expat) = @_;
|
---|
[8684] | 183 |
|
---|
[9958] | 184 | $self->{'rawxml'} .= $_;
|
---|
[4726] | 185 |
|
---|
[9958] | 186 | if ($self->{'in_metadata_node'}) {
|
---|
| 187 | $self->{'metadata_xml'} .= $_;
|
---|
[4726] | 188 | }
|
---|
[9958] | 189 | }
|
---|
[4726] | 190 |
|
---|
[8121] | 191 |
|
---|
[17216] | 192 | sub metadata_read {
|
---|
| 193 | my $self = shift (@_);
|
---|
[4726] | 194 |
|
---|
[19493] | 195 | my ($pluginfo, $base_dir, $file, $block_hash,
|
---|
| 196 | $extrametakeys, $extrametadata, $extrametafile,
|
---|
[23212] | 197 | $processor, $gli, $aux) = @_;
|
---|
[5919] | 198 |
|
---|
[17216] | 199 | # can we process this file??
|
---|
| 200 | my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
|
---|
[24403] | 201 | return undef unless $self->can_process_this_file_for_metadata($filename_full_path);
|
---|
[17216] | 202 |
|
---|
[17290] | 203 | if (!$self->parse_file($filename_full_path, $file, $gli)) {
|
---|
| 204 | $self->{'saved_metadata'} = undef;
|
---|
| 205 | return undef;
|
---|
| 206 | }
|
---|
| 207 |
|
---|
| 208 | my $new_metadata = $self->{'saved_metadata'};
|
---|
| 209 | $self->{'saved_metadata'} = undef;
|
---|
[17319] | 210 |
|
---|
[17290] | 211 | # add the pretty metadata table as metadata
|
---|
| 212 | my $ppmd_table = $self->{'ppmd_table'};
|
---|
| 213 | $new_metadata->{'prettymd'} = $ppmd_table;
|
---|
| 214 | $self->{'ppmd_table'} = undef;
|
---|
[17319] | 215 |
|
---|
| 216 | my $document_metadata_field = $self->{'document_field'};
|
---|
| 217 | my $url_array = $new_metadata->{$document_metadata_field};
|
---|
[22316] | 218 | if (!defined $url_array) {
|
---|
| 219 | # try ex.
|
---|
| 220 | $url_array = $new_metadata->{"ex.$document_metadata_field"};
|
---|
| 221 | }
|
---|
[17290] | 222 | my $num_urls = (defined $url_array) ? scalar(@$url_array) : 0;
|
---|
[17319] | 223 | ##print STDERR "$num_urls urls for $file\n";
|
---|
[17290] | 224 | my $srcdoc_exists = 0;
|
---|
| 225 | my $srcdoc_pos = 0;
|
---|
| 226 | my $filename_dir = &util::filename_head($filename_full_path);
|
---|
[17590] | 227 |
|
---|
[17591] | 228 | # filenames in extrametadata must be relative to current dir, as
|
---|
| 229 | # DirectoryPlugin adds path info on itself
|
---|
[17590] | 230 | my ($filename_for_metadata) = $file =~ /([^\\\/]+)$/; # this assumes there will only be one record per oai file - is this always the case??
|
---|
[17290] | 231 | for (my $i=0; $i<$num_urls; $i++) {
|
---|
[17216] | 232 |
|
---|
[17290] | 233 | if ($url_array->[$i] !~ m/^(https?|ftp):/) {
|
---|
[17216] | 234 |
|
---|
[17290] | 235 | my $src_filename = &util::filename_cat($filename_dir, $url_array->[$i]);
|
---|
| 236 | if (-e $src_filename) {
|
---|
| 237 | $srcdoc_pos = $i;
|
---|
| 238 | $srcdoc_exists = 1;
|
---|
[19622] | 239 | # get the slashes the right way, use filename_cat
|
---|
| 240 | $filename_for_metadata = &util::filename_cat($url_array->[$i]);
|
---|
[17290] | 241 | last;
|
---|
[17216] | 242 | }
|
---|
| 243 | }
|
---|
| 244 | }
|
---|
[17290] | 245 |
|
---|
[17319] | 246 | if ($srcdoc_exists) {
|
---|
[17290] | 247 | $self->{'oai-files'}->{$file}->{'srcdoc_exists'} = 1;
|
---|
| 248 | }
|
---|
[17216] | 249 | else {
|
---|
[17290] | 250 | # save the rawxml for the source document
|
---|
| 251 | $self->{'oai-files'}->{$file}->{'srcdoc_exists'} = 0;
|
---|
| 252 | $self->{'oai-files'}->{$file}->{'rawxml'} = $self->{'rawxml'};
|
---|
| 253 | $self->{'rawxml'} = "";
|
---|
[17216] | 254 | }
|
---|
[17290] | 255 |
|
---|
| 256 | # return all the metadata we have extracted to the caller.
|
---|
| 257 | # Directory plug will pass it back in at read time, so we don't need to extract it again.
|
---|
[24971] | 258 |
|
---|
| 259 | # Extrametadata keys should be regular expressions
|
---|
| 260 | # Indexing into the extrameta data structures requires the filename's style of slashes to be in URL format
|
---|
| 261 | # Then need to convert the filename to a regex, no longer to protect windows directory chars \, but for
|
---|
| 262 | # protecting special characters like brackets in the filepath such as "C:\Program Files (x86)\Greenstone".
|
---|
| 263 | $filename_for_metadata = &util::filepath_to_url_format($filename_for_metadata);
|
---|
[17513] | 264 | $filename_for_metadata = &util::filename_to_regex($filename_for_metadata);
|
---|
[24971] | 265 |
|
---|
[24951] | 266 | &extrametautil::setmetadata($extrametadata, $filename_for_metadata, $new_metadata);
|
---|
| 267 | &extrametautil::addmetakey($extrametakeys, $filename_for_metadata);
|
---|
| 268 | if ($srcdoc_exists) {
|
---|
| 269 | if (!defined &extrametautil::getmetafile($extrametafile, $filename_for_metadata)) {
|
---|
| 270 | &extrametautil::setmetafile($extrametafile, $filename_for_metadata, {});
|
---|
[20792] | 271 | }
|
---|
| 272 | #maps the file to full path
|
---|
[24951] | 273 | &extrametautil::setmetafile_for_named_file($extrametafile, $filename_for_metadata, $file, $filename_full_path);
|
---|
| 274 |
|
---|
[20792] | 275 | }
|
---|
[17290] | 276 | return 1;
|
---|
| 277 |
|
---|
[17216] | 278 | }
|
---|
| 279 |
|
---|
| 280 |
|
---|
[9958] | 281 | sub read {
|
---|
| 282 | my $self = shift (@_);
|
---|
[17290] | 283 |
|
---|
[16392] | 284 | my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
|
---|
[4726] | 285 |
|
---|
[17290] | 286 | if (!defined $self->{'oai-files'}->{$file}) {
|
---|
| 287 | return undef;
|
---|
| 288 | }
|
---|
[17319] | 289 |
|
---|
[17290] | 290 | my $srcdoc_exists = $self->{'oai-files'}->{$file}->{'srcdoc_exists'};
|
---|
| 291 | if ($srcdoc_exists) {
|
---|
[17319] | 292 | # do nothing more - all the metadata has been extracted and associated with the srcdoc
|
---|
[17216] | 293 | # no more need to access details of this $file => tidy up as you go
|
---|
| 294 | delete $self->{'oai-files'}->{$file};
|
---|
[17290] | 295 | return 0; # not processed here, but don't pass on to rest of plugins
|
---|
| 296 | }
|
---|
[17216] | 297 |
|
---|
[17290] | 298 | my $filename = $file;
|
---|
| 299 | $filename = &util::filename_cat ($base_dir, $file) if $base_dir =~ /\w/;
|
---|
[17216] | 300 |
|
---|
[17290] | 301 | # Do encoding stuff on metadata
|
---|
| 302 | my ($language, $encoding) = $self->textcat_get_language_encoding ($filename);
|
---|
[17216] | 303 |
|
---|
[17290] | 304 | # create a new document
|
---|
[18327] | 305 | my $doc_obj = new doc ($filename, "indexed_doc", $self->{'file_rename_method'});
|
---|
[17290] | 306 | my $top_section = $doc_obj->get_top_section;
|
---|
| 307 | my $plugin_type = $self->{'plugin_type'};
|
---|
| 308 |
|
---|
[17588] | 309 | my ($filemeta) = $file =~ /([^\\\/]+)$/;
|
---|
[23352] | 310 | my $plugin_filename_encoding = $self->{'filename_encoding'};
|
---|
[23349] | 311 | my $filename_encoding = $self->deduce_filename_encoding($file,$metadata,$plugin_filename_encoding);
|
---|
[23352] | 312 | $self->set_Source_metadata($doc_obj, $filename, $filename_encoding);
|
---|
[23349] | 313 |
|
---|
[17290] | 314 | $doc_obj->add_utf8_metadata($top_section, "Language", $language);
|
---|
| 315 | $doc_obj->add_utf8_metadata($top_section, "Encoding", $encoding);
|
---|
| 316 | $doc_obj->add_utf8_metadata($top_section, "Plugin", $plugin_type);
|
---|
| 317 | $doc_obj->add_metadata($top_section, "FileFormat", "OAI");
|
---|
| 318 | $doc_obj->add_metadata($top_section, "FileSize", (-s $filename));
|
---|
| 319 |
|
---|
| 320 | # include any metadata passed in from previous plugins
|
---|
| 321 | # note that this metadata is associated with the top level section
|
---|
[17319] | 322 | # this will include all the metadata from the oai file that we extracted
|
---|
| 323 | # during metadata_read
|
---|
[17290] | 324 | $self->extra_metadata ($doc_obj, $doc_obj->get_top_section(), $metadata);
|
---|
| 325 |
|
---|
| 326 | # do plugin specific processing of doc_obj
|
---|
| 327 | my $text = $self->{'oai-files'}->{$file}->{'rawxml'};
|
---|
| 328 | delete $self->{'oai-files'}->{$file};
|
---|
[17216] | 329 |
|
---|
[17290] | 330 | unless (defined ($self->process(\$text, $pluginfo, $base_dir, $file, $metadata, $doc_obj))) {
|
---|
| 331 | print STDERR "<ProcessingError n='$file'>\n" if ($gli);
|
---|
| 332 | return -1;
|
---|
[17216] | 333 | }
|
---|
[17290] | 334 |
|
---|
| 335 | # do any automatic metadata extraction
|
---|
| 336 | $self->auto_extract_metadata ($doc_obj);
|
---|
| 337 |
|
---|
| 338 | # add an OID
|
---|
| 339 | $self->add_OID($doc_obj);
|
---|
| 340 |
|
---|
| 341 | # process the document
|
---|
| 342 | $processor->process($doc_obj);
|
---|
| 343 |
|
---|
| 344 | $self->{'num_processed'} ++;
|
---|
| 345 |
|
---|
| 346 | return 1; # processed the file
|
---|
[17216] | 347 | }
|
---|
| 348 |
|
---|
| 349 |
|
---|
[4726] | 350 | # do plugin specific processing of doc_obj
|
---|
| 351 | sub process {
|
---|
| 352 | my $self = shift (@_);
|
---|
[6332] | 353 | my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
|
---|
[4726] | 354 | my $outhandle = $self->{'outhandle'};
|
---|
| 355 |
|
---|
[15872] | 356 | print STDERR "<Processing n='$file' p='OAIPlugin'>\n" if ($gli);
|
---|
| 357 | print $outhandle "OAIPlugin: processing $file\n"
|
---|
[4726] | 358 | if $self->{'verbosity'} > 1;
|
---|
| 359 |
|
---|
| 360 | my $cursection = $doc_obj->get_top_section();
|
---|
| 361 |
|
---|
| 362 | ## $self->extract_metadata ($textref, $metadata, $doc_obj, $cursection);
|
---|
| 363 |
|
---|
| 364 | # add text to document object
|
---|
| 365 |
|
---|
| 366 | # $$textref =~ s/<(.*?)>/$1 /g;
|
---|
| 367 | $$textref =~ s/</</g;
|
---|
| 368 | $$textref =~ s/>/>/g;
|
---|
[14963] | 369 | $$textref =~ s/\[/[/g;
|
---|
| 370 | $$textref =~ s/\]/]/g;
|
---|
[4726] | 371 |
|
---|
| 372 | $doc_obj->add_utf8_text($cursection, $$textref);
|
---|
| 373 |
|
---|
| 374 | return 1;
|
---|
| 375 | }
|
---|
| 376 |
|
---|
| 377 |
|
---|
[9958] | 378 | # Improvement is to merge this with newer version in MetadataPass
|
---|
[4726] | 379 |
|
---|
[9958] | 380 | sub open_prettyprint_metadata_table
|
---|
| 381 | {
|
---|
| 382 | my $self = shift(@_);
|
---|
| 383 |
|
---|
| 384 | my $att = "width=100% cellspacing=2";
|
---|
| 385 | my $style = "style=\'border-bottom: 4px solid #000080\'";
|
---|
| 386 |
|
---|
[17290] | 387 | $self->{'ppmd_table'} = "\n<table $att $style>";
|
---|
[9958] | 388 | }
|
---|
| 389 |
|
---|
| 390 | sub add_prettyprint_metadata_line
|
---|
| 391 | {
|
---|
| 392 | my $self = shift(@_);
|
---|
| 393 | my ($metaname, $metavalue_utf8) = @_;
|
---|
| 394 |
|
---|
| 395 | $metavalue_utf8 = &util::hyperlink_text($metavalue_utf8);
|
---|
| 396 |
|
---|
| 397 | $self->{'ppmd_table'} .= " <tr bgcolor=#b5d3cd>\n";
|
---|
| 398 | $self->{'ppmd_table'} .= " <td width=30%>\n";
|
---|
| 399 | $self->{'ppmd_table'} .= " $metaname\n";
|
---|
| 400 | $self->{'ppmd_table'} .= " </td>\n";
|
---|
| 401 | $self->{'ppmd_table'} .= " <td>\n";
|
---|
| 402 | $self->{'ppmd_table'} .= " $metavalue_utf8\n";
|
---|
| 403 | $self->{'ppmd_table'} .= " </td>\n";
|
---|
| 404 | $self->{'ppmd_table'} .= " </tr>\n";
|
---|
| 405 |
|
---|
| 406 | }
|
---|
| 407 |
|
---|
| 408 | sub close_prettyprint_metadata_table
|
---|
| 409 | {
|
---|
| 410 | my $self = shift(@_);
|
---|
| 411 |
|
---|
| 412 | $self->{'ppmd_table'} .= "</table>\n";
|
---|
| 413 | }
|
---|
| 414 |
|
---|
[19213] | 415 | my $qualified_dc_mapping = {
|
---|
| 416 | "alternative" => "dc.title",
|
---|
| 417 | "tableOfContents" => "dc.description",
|
---|
| 418 | "abstract" => "dc.description",
|
---|
| 419 | "created" => "dc.date",
|
---|
| 420 | "valid" => "dc.date",
|
---|
| 421 | "available" => "dc.date",
|
---|
| 422 | "issued" => "dc.date",
|
---|
| 423 | "modified" => "dc.date",
|
---|
| 424 | "dateAccepted" => "dc.date",
|
---|
| 425 | "dateCopyrighted" => "dc.date",
|
---|
| 426 | "dateSubmitted" => "dc.date",
|
---|
| 427 | "extent" => "dc.format",
|
---|
| 428 | "medium" => "dc.format",
|
---|
| 429 | "isVersionOf" => "dc.relation",
|
---|
| 430 | "hasVersion" => "dc.relation",
|
---|
| 431 | "isReplacedBy" => "dc.relation",
|
---|
| 432 | "replaces" => "dc.relation",
|
---|
| 433 | "isRequiredBy" => "dc.relation",
|
---|
| 434 | "requires" => "dc.relation",
|
---|
| 435 | "isPartOf" => "dc.relation",
|
---|
| 436 | "hasPart" => "dc.relation",
|
---|
| 437 | "isReferencedBy" => "dc.relation",
|
---|
| 438 | "references" => "dc.relation",
|
---|
| 439 | "isFormatOf" => "dc.relation",
|
---|
| 440 | "hasFormat" => "dc.relation",
|
---|
| 441 | "conformsTo" => "dc.relation",
|
---|
| 442 | "spatial" => "dc.coverage",
|
---|
| 443 | "temporal" => "dc.coverage",
|
---|
| 444 | # these are now top level elements in our qualified dc metadata set
|
---|
[18901] | 445 | # "audience" => "dc.any",
|
---|
| 446 | # "accrualMethod" => "dc.any",
|
---|
| 447 | # "accrualPeriodicity" => "dc.any",
|
---|
| 448 | # "accrualPolicy" => "dc.any",
|
---|
| 449 | # "instructionalMethod" => "dc.any",
|
---|
| 450 | # "provenance" => "dc.any",
|
---|
| 451 | # "rightsHolder" => "dc.any",
|
---|
[19213] | 452 | "mediator" => "dc.audience",
|
---|
| 453 | "educationLevel" => "dc.audience",
|
---|
| 454 | "accessRights" => "dc.rights",
|
---|
| 455 | "license" => "dc.rights",
|
---|
| 456 | "bibliographicCitation" => "dc.identifier"
|
---|
| 457 | };
|
---|
[14940] | 458 |
|
---|
[19213] | 459 | sub remap_dc_metadata
|
---|
| 460 | {
|
---|
| 461 | my $self = shift(@_);
|
---|
| 462 |
|
---|
| 463 | my ($metaname) = @_;
|
---|
| 464 |
|
---|
[14940] | 465 | my ($prefix,$name) = ($metaname =~ m/^(.*?)\.(.*?)$/);
|
---|
| 466 |
|
---|
[19213] | 467 | if (defined $qualified_dc_mapping->{$name}) {
|
---|
| 468 |
|
---|
| 469 | return $qualified_dc_mapping->{$name}."^".$name;
|
---|
[14940] | 470 | }
|
---|
[19213] | 471 |
|
---|
| 472 |
|
---|
[14940] | 473 | return $metaname; # didn't get a match, return param passed in unchanged
|
---|
| 474 | }
|
---|
| 475 |
|
---|
| 476 |
|
---|
[4726] | 477 | sub extract_oai_metadata {
|
---|
| 478 | my $self = shift (@_);
|
---|
| 479 | my ($textref, $metadata) = @_;
|
---|
| 480 | my $outhandle = $self->{'outhandle'};
|
---|
| 481 |
|
---|
[9958] | 482 | $self->open_prettyprint_metadata_table();
|
---|
| 483 |
|
---|
| 484 | if ($$textref =~ m/<metadata\s*>(.*?)<\/metadata\s*>/s)
|
---|
[4726] | 485 | {
|
---|
[10254] | 486 | my $metadata_text = $1;
|
---|
[4726] | 487 |
|
---|
[14940] | 488 | # locate and remove outermost tag (ignoring any attribute information in top-level tag)
|
---|
[19213] | 489 | my ($outer_tagname,$inner_metadata_text) = ($metadata_text =~ m/<([^ >]+).*?>(.*?)<\/\1>/s);
|
---|
[14940] | 490 | # split tag into namespace and tag name
|
---|
[19213] | 491 | my($namespace,$top_level_prefix) = ($outer_tagname =~ m/^(.*?):(.*?)$/);
|
---|
[17066] | 492 | # sometimes, the dc namespace is not specified as the prefix in each element (like <dc:title>)
|
---|
| 493 | # but is rather defined in the wrapper element containing the various dc meta elements,
|
---|
| 494 | # like <dc><title></title><creator></creator></dc>.
|
---|
| 495 | # In such a case, we use this wrapper element as the top_level_prefix
|
---|
[19213] | 496 |
|
---|
| 497 | # if there was no prefix, then the tag itself becomes the top_level_prefix
|
---|
| 498 | if(!defined $top_level_prefix && defined $outer_tagname) {
|
---|
| 499 | $top_level_prefix = $outer_tagname;
|
---|
[17066] | 500 | }
|
---|
| 501 |
|
---|
[19213] | 502 | #process each element one by one
|
---|
[14949] | 503 | while ($inner_metadata_text =~ m/<([^ >]+).*?>(.*?)<\/\1>(.*)/s)
|
---|
[4726] | 504 | {
|
---|
[9958] | 505 |
|
---|
[4726] | 506 | my $metaname = $1;
|
---|
| 507 | my $metavalue = $2;
|
---|
[14949] | 508 | $inner_metadata_text = $3;
|
---|
| 509 |
|
---|
[19213] | 510 | # greenstone uses . for namespace, while oai uses :
|
---|
[14940] | 511 | $metaname =~ s/:/\./;
|
---|
[19213] | 512 | # if there is no namespace, then we use the outer tag name or
|
---|
| 513 | # namespace for this element
|
---|
[14940] | 514 | if ($metaname !~ m/\./)
|
---|
[4726] | 515 | {
|
---|
[14940] | 516 | $metaname = "$top_level_prefix.$metaname";
|
---|
[4726] | 517 | }
|
---|
[19213] | 518 |
|
---|
| 519 | # if metadata set is auto, leave as is, otherwise convert to
|
---|
| 520 | # specified namespace
|
---|
| 521 | if ($self->{'metadata_set'} ne "auto") {
|
---|
[20787] | 522 | if ($metaname !~ /^gi\./) { # hack to not overwrite gi metadata
|
---|
| 523 | $metaname =~ s/^([^\.]*)\./$self->{'metadata_set'}\./;
|
---|
| 524 | if ($self->{'metadata_set'} eq "dc") {
|
---|
| 525 | # convert qualified dc terms to gs version, e.g.
|
---|
| 526 | # spatial becomes coverage^spatial
|
---|
| 527 | $metaname = $self->remap_dc_metadata($metaname);
|
---|
| 528 | }
|
---|
[19213] | 529 | }
|
---|
| 530 | }
|
---|
[4726] | 531 |
|
---|
[18901] | 532 | # uppercase the first char of the name
|
---|
| 533 | $metaname =~ s/\.(.)/\.\u$1/;
|
---|
[14963] | 534 | $metavalue =~ s/\[/[/g;
|
---|
| 535 | $metavalue =~ s/\]/]/g;
|
---|
| 536 |
|
---|
[22316] | 537 | # so that GLI can see this metadata, store here as ex.dc.Title etc
|
---|
[24404] | 538 | my $ex_metaname = $metaname;
|
---|
| 539 | $ex_metaname =~ s/^ex\.//; # remove any pre-existing ex. prefix
|
---|
| 540 | $ex_metaname = "ex.$ex_metaname"; # at last can prefix ex.
|
---|
[22316] | 541 |
|
---|
| 542 | if (defined $metadata->{$ex_metaname})
|
---|
[4726] | 543 | {
|
---|
[22316] | 544 | push(@{$metadata->{$ex_metaname}},$metavalue);
|
---|
[8121] | 545 |
|
---|
[4726] | 546 | }
|
---|
| 547 | else
|
---|
| 548 | {
|
---|
[22316] | 549 | $metadata->{$ex_metaname} = [ $metavalue ];
|
---|
[4726] | 550 | }
|
---|
| 551 |
|
---|
[22316] | 552 | # but don't add ex to the pretty print line
|
---|
[9958] | 553 | $self->add_prettyprint_metadata_line($metaname, $metavalue);
|
---|
| 554 |
|
---|
[4726] | 555 | }
|
---|
| 556 | }
|
---|
[9958] | 557 |
|
---|
| 558 | $self->close_prettyprint_metadata_table();
|
---|
[4726] | 559 | }
|
---|
| 560 |
|
---|
[13886] | 561 | ## we know from the file extension, so doesn't need to check the doctype
|
---|
| 562 | sub check_doctype {
|
---|
| 563 |
|
---|
| 564 | return 1;
|
---|
| 565 | }
|
---|
| 566 |
|
---|
[4726] | 567 | 1;
|
---|