[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 |
|
---|
| 27 | package OAIPlug;
|
---|
| 28 |
|
---|
| 29 | use BasPlug;
|
---|
| 30 | use unicode;
|
---|
| 31 | use util;
|
---|
| 32 |
|
---|
[10254] | 33 | use strict;
|
---|
| 34 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
---|
| 35 |
|
---|
[9958] | 36 | use XMLPlug;
|
---|
| 37 |
|
---|
[4726] | 38 | sub BEGIN {
|
---|
[9958] | 39 | @OAIPlug::ISA = ('XMLPlug');
|
---|
[4726] | 40 | }
|
---|
| 41 |
|
---|
[9958] | 42 |
|
---|
[6408] | 43 | my $arguments =
|
---|
| 44 | [ { 'name' => "process_exp",
|
---|
| 45 | 'desc' => "{BasPlug.process_exp}",
|
---|
| 46 | 'type' => "regexp",
|
---|
| 47 | 'reqd' => "no",
|
---|
| 48 | 'deft' => &get_default_process_exp() },
|
---|
| 49 | ];
|
---|
| 50 |
|
---|
[4747] | 51 | my $options = { 'name' => "OAIPlug",
|
---|
[5680] | 52 | 'desc' => "{OAIPlug.desc}",
|
---|
[6408] | 53 | 'abstract' => "no",
|
---|
| 54 | 'inherits' => "yes",
|
---|
| 55 | 'args' => $arguments };
|
---|
[4747] | 56 |
|
---|
[10254] | 57 |
|
---|
[4726] | 58 | sub new {
|
---|
[10218] | 59 | my ($class) = shift (@_);
|
---|
| 60 | my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
|
---|
| 61 | push(@$pluginlist, $class);
|
---|
[4873] | 62 |
|
---|
[10218] | 63 | if(defined $arguments){ push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});}
|
---|
| 64 | if(defined $options) { push(@{$hashArgOptLists->{"OptList"}},$options)};
|
---|
[4726] | 65 |
|
---|
[10218] | 66 | my $self = (defined $hashArgOptLists)? new XMLPlug($pluginlist,$inputargs,$hashArgOptLists): new XMLPlug($pluginlist,$inputargs);
|
---|
[4726] | 67 |
|
---|
| 68 | return bless $self, $class;
|
---|
| 69 | }
|
---|
| 70 |
|
---|
| 71 | sub get_default_process_exp {
|
---|
| 72 | my $self = shift (@_);
|
---|
| 73 |
|
---|
| 74 | return q^(?i)(\.oai)$^;
|
---|
| 75 | }
|
---|
| 76 |
|
---|
[9958] | 77 | sub xml_start_document {
|
---|
[10254] | 78 | my $self = shift (@_);
|
---|
[9958] | 79 | $self->{'in_metadata_node'} = 0;
|
---|
| 80 | $self->{'rawxml'} = "";
|
---|
| 81 | }
|
---|
[4726] | 82 |
|
---|
[9958] | 83 | sub xml_end_document {
|
---|
| 84 | }
|
---|
[4726] | 85 |
|
---|
[9958] | 86 | sub xml_doctype {
|
---|
| 87 | my $self = shift(@_);
|
---|
| 88 |
|
---|
| 89 | my ($expat, $name, $sysid, $pubid, $internal) = @_;
|
---|
| 90 |
|
---|
| 91 | # allow the short-lived and badly named "GreenstoneArchive" files to be processed
|
---|
| 92 | # as well as the "Archive" files which should now be created by import.pl
|
---|
| 93 | die "" if ($name !~ /^OAI-PMH$/);
|
---|
| 94 |
|
---|
[4726] | 95 | my $outhandle = $self->{'outhandle'};
|
---|
[9958] | 96 | print $outhandle "OAIPlug: processing $self->{'file'}\n" if $self->{'verbosity'} > 1;
|
---|
| 97 | print STDERR "<Processing n='$self->{'file'}' p='OAIPlug'>\n" if $self->{'gli'};
|
---|
[4726] | 98 |
|
---|
[9958] | 99 | }
|
---|
[4726] | 100 |
|
---|
[9958] | 101 |
|
---|
| 102 | sub xml_start_tag {
|
---|
| 103 | my $self = shift(@_);
|
---|
| 104 | my ($expat,$element) = @_;
|
---|
| 105 |
|
---|
| 106 | my %attr_hash = %_;
|
---|
| 107 |
|
---|
| 108 | my $attr = "";
|
---|
| 109 | map { $attr .= " $_=$attr_hash{$_}"; } keys %attr_hash;
|
---|
| 110 |
|
---|
| 111 | $self->{'rawxml'} .= "<$element$attr>";
|
---|
| 112 |
|
---|
| 113 | if ($element eq "metadata") {
|
---|
| 114 | $self->{'in_metadata_node'} = 1;
|
---|
| 115 | $self->{'metadata_xml'} = "";
|
---|
[4726] | 116 | }
|
---|
[9958] | 117 |
|
---|
| 118 | if ($self->{'in_metadata_node'}) {
|
---|
| 119 | $self->{'metadata_xml'} .= "<$element$attr>";
|
---|
[4726] | 120 | }
|
---|
[9958] | 121 | }
|
---|
[4726] | 122 |
|
---|
[9958] | 123 | sub xml_end_tag {
|
---|
| 124 | my $self = shift(@_);
|
---|
| 125 | my ($expat, $element) = @_;
|
---|
[4726] | 126 |
|
---|
[9958] | 127 | $self->{'rawxml'} .= "</$element>";
|
---|
[4726] | 128 |
|
---|
[9958] | 129 | if ($self->{'in_metadata_node'}) {
|
---|
| 130 | $self->{'metadata_xml'} .= "</$element>";
|
---|
[4726] | 131 | }
|
---|
| 132 |
|
---|
[9958] | 133 | if ($element eq "metadata") {
|
---|
| 134 | my $textref = \$self->{'metadata_xml'};
|
---|
| 135 | my $metadata = $self->{'metadata'};
|
---|
| 136 | $self->extract_oai_metadata($textref,$metadata);
|
---|
[4726] | 137 |
|
---|
[9958] | 138 | $self->{'in_metadata_node'} = 0;
|
---|
| 139 | }
|
---|
[4726] | 140 |
|
---|
| 141 |
|
---|
[9958] | 142 | }
|
---|
[4726] | 143 |
|
---|
[9958] | 144 | sub xml_text {
|
---|
| 145 | my $self = shift(@_);
|
---|
| 146 | my ($expat) = @_;
|
---|
[8684] | 147 |
|
---|
[9958] | 148 | $self->{'rawxml'} .= $_;
|
---|
[4726] | 149 |
|
---|
[9958] | 150 | if ($self->{'in_metadata_node'}) {
|
---|
| 151 | $self->{'metadata_xml'} .= $_;
|
---|
[4726] | 152 | }
|
---|
[9958] | 153 | }
|
---|
[4726] | 154 |
|
---|
[8121] | 155 |
|
---|
[4726] | 156 |
|
---|
[5919] | 157 |
|
---|
[9958] | 158 | sub read {
|
---|
| 159 | my $self = shift (@_);
|
---|
| 160 |
|
---|
| 161 | my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
|
---|
[4726] | 162 |
|
---|
[9958] | 163 | my $outhandle = $self->{'outhandle'};
|
---|
[4726] | 164 |
|
---|
[9958] | 165 | my $filename = $file;
|
---|
| 166 | $filename = &util::filename_cat ($base_dir, $file) if $base_dir =~ /\w/;
|
---|
[4726] | 167 |
|
---|
[9958] | 168 | return 0 if ((-d $filename) && ($filename =~ m/srcdocs$/));
|
---|
[4726] | 169 |
|
---|
[9958] | 170 | if ($self->SUPER::read(@_)) {
|
---|
| 171 |
|
---|
| 172 | # Do encoding stuff
|
---|
| 173 | my ($language, $encoding) = $self->textcat_get_language_encoding ($filename);
|
---|
| 174 |
|
---|
| 175 | my $url_array = $metadata->{'URL'};
|
---|
[10254] | 176 | my $num_urls = (defined $url_array) ? scalar(@$url_array) : 0;
|
---|
[9958] | 177 |
|
---|
| 178 | my $srcdoc_exists = 0;
|
---|
| 179 | my $srcdoc_pos = 0;
|
---|
| 180 | my $filename_dir = &util::filename_head($filename);
|
---|
| 181 |
|
---|
| 182 | for (my $i=0; $i<$num_urls; $i++) {
|
---|
| 183 |
|
---|
| 184 | if ($url_array->[$i] !~ m/^(http|ftp):/) {
|
---|
| 185 |
|
---|
| 186 | my $src_filename = &util::filename_cat($filename_dir, $url_array->[$i]);
|
---|
| 187 |
|
---|
| 188 | if (-e $src_filename) {
|
---|
| 189 | $srcdoc_pos = $i;
|
---|
| 190 | $srcdoc_exists = 1;
|
---|
| 191 | }
|
---|
| 192 | }
|
---|
| 193 | }
|
---|
| 194 |
|
---|
| 195 | if ($srcdoc_exists)
|
---|
| 196 | {
|
---|
| 197 | print $outhandle "OAIPlug: passing metadata on to $url_array->[0]\n"
|
---|
| 198 | if ($self->{'verbosity'}>1);
|
---|
| 199 |
|
---|
| 200 |
|
---|
| 201 | # Make pretty print metadata table stick with src filename
|
---|
| 202 | my $ppmd_table = $self->{'ppmd_table'};
|
---|
| 203 | $metadata->{'prettymd'} = [ $ppmd_table ];
|
---|
| 204 | $self->{'ppmd_table'} = undef;
|
---|
| 205 |
|
---|
| 206 | return &plugin::read ($pluginfo, $filename_dir, $url_array->[0],
|
---|
| 207 | $metadata, $processor, $maxdocs, $total_count, $gli);
|
---|
| 208 | }
|
---|
| 209 | else
|
---|
| 210 | {
|
---|
| 211 | # create a new document
|
---|
| 212 | my $doc_obj = new doc ($filename, "indexed_doc");
|
---|
| 213 | my $top_section = $doc_obj->get_top_section;
|
---|
| 214 | my $plugin_type = $self->{'plugin_type'};
|
---|
| 215 |
|
---|
| 216 | $doc_obj->add_utf8_metadata($top_section, "Language", $language);
|
---|
| 217 | $doc_obj->add_utf8_metadata($top_section, "Encoding", $encoding);
|
---|
| 218 | $doc_obj->add_utf8_metadata($top_section, "Plugin", $plugin_type);
|
---|
| 219 | $doc_obj->add_metadata($top_section, "FileFormat", "OAI");
|
---|
| 220 | $doc_obj->add_metadata($top_section, "FileSize", (-s $filename));
|
---|
| 221 |
|
---|
| 222 | # include any metadata passed in from previous plugins
|
---|
| 223 | # note that this metadata is associated with the top level section
|
---|
| 224 | $self->extra_metadata ($doc_obj, $doc_obj->get_top_section(), $metadata);
|
---|
| 225 |
|
---|
| 226 | # do plugin specific processing of doc_obj
|
---|
| 227 | my $textref = \$self->{'rawxml'};
|
---|
| 228 | unless (defined ($self->process($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj))) {
|
---|
| 229 | print STDERR "<ProcessingError n='$file'>\n" if ($gli);
|
---|
| 230 | return -1;
|
---|
| 231 | }
|
---|
| 232 |
|
---|
| 233 | # do any automatic metadata extraction
|
---|
| 234 | $self->auto_extract_metadata ($doc_obj);
|
---|
| 235 |
|
---|
| 236 | # add an OID
|
---|
| 237 | $doc_obj->set_OID();
|
---|
| 238 |
|
---|
| 239 | my $ppmd_table = $self->{'ppmd_table'};
|
---|
| 240 | $doc_obj->add_utf8_metadata($top_section,"prettymd",$ppmd_table);
|
---|
| 241 | $self->{'ppmd_table'} = undef;
|
---|
| 242 |
|
---|
| 243 | # process the document
|
---|
| 244 | $processor->process($doc_obj);
|
---|
| 245 |
|
---|
| 246 | $self->{'num_processed'} ++;
|
---|
| 247 |
|
---|
| 248 | return 1; # processed the file
|
---|
| 249 | }
|
---|
[4726] | 250 | }
|
---|
[9958] | 251 | else {
|
---|
| 252 | return undef;
|
---|
| 253 | }
|
---|
[4726] | 254 | }
|
---|
| 255 |
|
---|
| 256 |
|
---|
| 257 | # do plugin specific processing of doc_obj
|
---|
| 258 | sub process {
|
---|
| 259 | my $self = shift (@_);
|
---|
[6332] | 260 | my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
|
---|
[4726] | 261 | my $outhandle = $self->{'outhandle'};
|
---|
| 262 |
|
---|
[6332] | 263 | print STDERR "<Processing n='$file' p='OAIPlug'>\n" if ($gli);
|
---|
[4726] | 264 | print $outhandle "OAIPlug: processing $file\n"
|
---|
| 265 | if $self->{'verbosity'} > 1;
|
---|
| 266 |
|
---|
| 267 | my $cursection = $doc_obj->get_top_section();
|
---|
| 268 |
|
---|
| 269 | ## $self->extract_metadata ($textref, $metadata, $doc_obj, $cursection);
|
---|
| 270 |
|
---|
| 271 | # add text to document object
|
---|
| 272 |
|
---|
| 273 | # $$textref =~ s/<(.*?)>/$1 /g;
|
---|
| 274 | $$textref =~ s/</</g;
|
---|
| 275 | $$textref =~ s/>/>/g;
|
---|
| 276 |
|
---|
| 277 | ## print STDERR "*** adding text: $$textref\n";
|
---|
| 278 |
|
---|
| 279 | $doc_obj->add_utf8_text($cursection, $$textref);
|
---|
| 280 |
|
---|
| 281 | return 1;
|
---|
| 282 | }
|
---|
| 283 |
|
---|
| 284 |
|
---|
[9958] | 285 | # Improvement is to merge this with newer version in MetadataPass
|
---|
[4726] | 286 |
|
---|
[9958] | 287 | sub open_prettyprint_metadata_table
|
---|
| 288 | {
|
---|
| 289 | my $self = shift(@_);
|
---|
| 290 |
|
---|
| 291 | my $att = "width=100% cellspacing=2";
|
---|
| 292 | my $style = "style=\'border-bottom: 4px solid #000080\'";
|
---|
| 293 |
|
---|
| 294 | $self->{'ppmd_table'} = "\n<table $att $style>";
|
---|
| 295 | }
|
---|
| 296 |
|
---|
| 297 | sub add_prettyprint_metadata_line
|
---|
| 298 | {
|
---|
| 299 | my $self = shift(@_);
|
---|
| 300 | my ($metaname, $metavalue_utf8) = @_;
|
---|
| 301 |
|
---|
| 302 | $metavalue_utf8 =~ s/hdl\.handle\.net/mcgonagall.cs.waikato.ac.nz:8080\/dspace\/handle/;
|
---|
| 303 | $metavalue_utf8 = &util::hyperlink_text($metavalue_utf8);
|
---|
| 304 |
|
---|
| 305 | $self->{'ppmd_table'} .= " <tr bgcolor=#b5d3cd>\n";
|
---|
| 306 | $self->{'ppmd_table'} .= " <td width=30%>\n";
|
---|
| 307 | $self->{'ppmd_table'} .= " $metaname\n";
|
---|
| 308 | $self->{'ppmd_table'} .= " </td>\n";
|
---|
| 309 | $self->{'ppmd_table'} .= " <td>\n";
|
---|
| 310 | $self->{'ppmd_table'} .= " $metavalue_utf8\n";
|
---|
| 311 | $self->{'ppmd_table'} .= " </td>\n";
|
---|
| 312 | $self->{'ppmd_table'} .= " </tr>\n";
|
---|
| 313 |
|
---|
| 314 | }
|
---|
| 315 |
|
---|
| 316 | sub close_prettyprint_metadata_table
|
---|
| 317 | {
|
---|
| 318 | my $self = shift(@_);
|
---|
| 319 |
|
---|
| 320 | $self->{'ppmd_table'} .= "</table>\n";
|
---|
| 321 | }
|
---|
| 322 |
|
---|
| 323 |
|
---|
| 324 |
|
---|
| 325 |
|
---|
[4726] | 326 | sub extract_oai_metadata {
|
---|
| 327 | my $self = shift (@_);
|
---|
| 328 | my ($textref, $metadata) = @_;
|
---|
| 329 | my $outhandle = $self->{'outhandle'};
|
---|
| 330 |
|
---|
[9958] | 331 | # Only handles DC metadata
|
---|
| 332 |
|
---|
| 333 | $self->open_prettyprint_metadata_table();
|
---|
| 334 |
|
---|
| 335 | if ($$textref =~ m/<metadata\s*>(.*?)<\/metadata\s*>/s)
|
---|
[4726] | 336 | {
|
---|
[10254] | 337 | my $metadata_text = $1;
|
---|
[4726] | 338 | $metadata_text =~ s/^.*?<(oai_dc:)?dc.*?>(.*?)<\/(oai_dc:)?dc>.*?/$2/s;
|
---|
| 339 |
|
---|
| 340 | while ($metadata_text =~ m/<(.*?)>(.*?)<\/(.*?)>(.*)/s)
|
---|
| 341 | {
|
---|
| 342 | # if URL given for document as identifier metadata, store it ...
|
---|
| 343 | # $doc_obj->add_utf8_metadata($cursection, "URL", $web_url);
|
---|
[9958] | 344 |
|
---|
[4726] | 345 | my $metaname = $1;
|
---|
| 346 | my $metavalue = $2;
|
---|
| 347 | $metadata_text = $4;
|
---|
| 348 |
|
---|
| 349 | $metaname =~ s/^(dc:)?(.)/\u$2/;
|
---|
| 350 |
|
---|
| 351 | if ($metaname eq "Identifier")
|
---|
| 352 | {
|
---|
| 353 | # name clashes with GSDL reserved metadata name for hash id
|
---|
| 354 | $metaname = "URL";
|
---|
| 355 | }
|
---|
| 356 |
|
---|
| 357 | if (defined $metadata->{$metaname})
|
---|
| 358 | {
|
---|
| 359 | push(@{$metadata->{$metaname}},$metavalue);
|
---|
[8121] | 360 |
|
---|
[4726] | 361 | }
|
---|
| 362 | else
|
---|
| 363 | {
|
---|
| 364 | $metadata->{$metaname} = [ $metavalue ];
|
---|
| 365 | }
|
---|
| 366 |
|
---|
[9958] | 367 | $self->add_prettyprint_metadata_line($metaname, $metavalue);
|
---|
| 368 |
|
---|
[4726] | 369 | }
|
---|
| 370 | }
|
---|
[9958] | 371 |
|
---|
| 372 | $self->close_prettyprint_metadata_table();
|
---|
[4726] | 373 | }
|
---|
| 374 |
|
---|
| 375 | 1;
|
---|