[8879] | 1 | #!/usr/bin/perl -w
|
---|
| 2 |
|
---|
| 3 | ###########################################################################
|
---|
| 4 | #
|
---|
| 5 | # export.pl --
|
---|
| 6 | # A component of the Greenstone digital library software
|
---|
| 7 | # from the New Zealand Digital Library Project at the
|
---|
| 8 | # University of Waikato, New Zealand.
|
---|
| 9 | #
|
---|
| 10 | # Copyright (C) 1999 New Zealand Digital Library Project
|
---|
| 11 | #
|
---|
| 12 | # This program is free software; you can redistribute it and/or modify
|
---|
| 13 | # it under the terms of the GNU General Public License as published by
|
---|
| 14 | # the Free Software Foundation; either version 2 of the License, or
|
---|
| 15 | # (at your option) any later version.
|
---|
| 16 | #
|
---|
| 17 | # This program is distributed in the hope that it will be useful,
|
---|
| 18 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
| 19 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
| 20 | # GNU General Public License for more details.
|
---|
| 21 | #
|
---|
| 22 | # You should have received a copy of the GNU General Public License
|
---|
| 23 | # along with this program; if not, write to the Free Software
|
---|
| 24 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
| 25 | #
|
---|
| 26 | ###########################################################################
|
---|
| 27 |
|
---|
| 28 |
|
---|
| 29 | # This program will export a particular collection into a specific Format (e.g. METS or DSpace)
|
---|
| 30 | # Author: Chi-Yu Huang Date: 08-10-2004
|
---|
| 31 |
|
---|
| 32 | package export;
|
---|
| 33 |
|
---|
| 34 | BEGIN {
|
---|
| 35 | die "GSDLHOME not set\n" unless defined $ENV{'GSDLHOME'};
|
---|
| 36 | die "GSDLOS not set\n" unless defined $ENV{'GSDLOS'};
|
---|
| 37 | unshift (@INC, "$ENV{'GSDLHOME'}/perllib");
|
---|
| 38 | unshift (@INC, "$ENV{'GSDLHOME'}/perllib/cpan");
|
---|
| 39 | unshift (@INC, "$ENV{'GSDLHOME'}/perllib/plugins");
|
---|
[12333] | 40 | unshift (@INC, "$ENV{'GSDLHOME'}/perllib/plugouts");
|
---|
[8879] | 41 | unshift (@INC, "$ENV{'GSDLHOME'}/perllib/classify");
|
---|
[14942] | 42 |
|
---|
| 43 | if (defined $ENV{'GSDLEXTS'}) {
|
---|
| 44 | my @extensions = split(/:/,$ENV{'GSDLEXTS'});
|
---|
| 45 | foreach my $e (@extensions) {
|
---|
| 46 | my $ext_prefix = "$ENV{'GSDLHOME'}/ext/$e";
|
---|
| 47 |
|
---|
| 48 | unshift (@INC, "$ext_prefix/perllib");
|
---|
| 49 | unshift (@INC, "$ext_prefix/perllib/cpan");
|
---|
| 50 | }
|
---|
| 51 | }
|
---|
[8879] | 52 | }
|
---|
| 53 |
|
---|
[10340] | 54 | use strict;
|
---|
| 55 | no strict 'refs'; # allow filehandles to be variables and vice versa
|
---|
| 56 | no strict 'subs'; # allow barewords (eg STDERR) as function arguments
|
---|
| 57 |
|
---|
[10162] | 58 | use arcinfo;
|
---|
[8879] | 59 | use colcfg;
|
---|
| 60 | use plugin;
|
---|
[12333] | 61 | use plugout;
|
---|
[14935] | 62 | use manifest;
|
---|
[8879] | 63 | use util;
|
---|
[10417] | 64 | use scriptutil;
|
---|
[8879] | 65 | use FileHandle;
|
---|
[14935] | 66 | use gsprintf 'gsprintf';
|
---|
[8879] | 67 | use printusage;
|
---|
[10340] | 68 | use parse2;
|
---|
[8879] | 69 |
|
---|
[10340] | 70 |
|
---|
[8879] | 71 | my $oidtype_list =
|
---|
| 72 | [ { 'name' => "hash",
|
---|
[12691] | 73 | 'desc' => "{import.OIDtype.hash}" },
|
---|
[14935] | 74 | { 'name' => "assigned",
|
---|
| 75 | 'desc' => "{import.OIDtype.assigned}" },
|
---|
[8879] | 76 | { 'name' => "incremental",
|
---|
[12691] | 77 | 'desc' => "{import.OIDtype.incremental}" },
|
---|
| 78 | { 'name' => "dirname",
|
---|
| 79 | 'desc' => "{import.OIDtype.dirname}" } ];
|
---|
[8879] | 80 |
|
---|
| 81 | #** define to use the METS format or DSpace format
|
---|
| 82 | my $saveas_list =
|
---|
| 83 | [ { 'name' => "DSpace",
|
---|
| 84 | 'desc' => "{export.saveas.DSpace}" },
|
---|
[14935] | 85 | { 'name' => "GreenstoneMETS",
|
---|
| 86 | 'desc' => "{export.saveas.GreenstoneMETS}"},
|
---|
| 87 | { 'name' => "FedoraMETS",
|
---|
| 88 | 'desc' => "{export.saveas.FedoraMETS}"},
|
---|
[12333] | 89 | { 'name' => "GA",
|
---|
| 90 | 'desc' => "{export.saveas.GA}"},
|
---|
| 91 | { 'name' => "MARCXML",
|
---|
| 92 | 'desc' => "{export.saveas.MARCXML}"}
|
---|
| 93 | ];
|
---|
[8879] | 94 |
|
---|
| 95 |
|
---|
| 96 | # Possible attributes for each argument
|
---|
| 97 | # name: The name of the argument
|
---|
| 98 | # desc: A description (or more likely a reference to a description) for this argument
|
---|
[12361] | 99 | # type: The type of control used to represent the argument. Options include: string, int, flag, regexp, metadata, language, enum etc
|
---|
[8879] | 100 | # reqd: Is this argument required?
|
---|
| 101 | # hiddengli: Is this argument hidden in GLI?
|
---|
| 102 | # modegli: The lowest detail mode this argument is visible at in GLI
|
---|
| 103 |
|
---|
[14935] | 104 | my $saveas_argument =
|
---|
[9233] | 105 | { 'name' => "saveas",
|
---|
| 106 | 'desc' => "{export.saveas}",
|
---|
| 107 | 'type' => "enum",
|
---|
| 108 | 'list' => $saveas_list,
|
---|
| 109 | 'deft' => "METS",
|
---|
| 110 | 'reqd' => "no",
|
---|
| 111 | 'modegli' => "3" };
|
---|
| 112 |
|
---|
| 113 |
|
---|
[8879] | 114 | my $arguments =
|
---|
[10340] | 115 | [
|
---|
| 116 | $saveas_argument,
|
---|
| 117 | { 'name' => "exportdir",
|
---|
[8879] | 118 | 'desc' => "{export.exportdir}",
|
---|
| 119 | 'type' => "string",
|
---|
| 120 | 'reqd' => "no",
|
---|
| 121 | 'hiddengli' => "yes" },
|
---|
[10340] | 122 | { 'name' => "importdir",
|
---|
| 123 | 'desc' => "{import.importdir}",
|
---|
| 124 | 'type' => "string",
|
---|
| 125 | 'reqd' => "no",
|
---|
| 126 | 'hiddengli' => "yes" },
|
---|
[14925] | 127 | { 'name' => "site",
|
---|
| 128 | 'desc' => "{import.site}",
|
---|
| 129 | 'type' => "string",
|
---|
| 130 | 'deft' => "",
|
---|
| 131 | 'reqd' => "no",
|
---|
| 132 | 'hiddengli' => "yes" },
|
---|
[8879] | 133 | { 'name' => "collectdir",
|
---|
| 134 | 'desc' => "{export.collectdir}",
|
---|
| 135 | 'type' => "string",
|
---|
[14925] | 136 | # parsearg left "" as default
|
---|
| 137 | #'deft' => &util::filename_cat ($ENV{'GSDLHOME'}, "collect"),
|
---|
[14935] | 138 | 'deft' => "",
|
---|
| 139 | 'reqd' => "no",
|
---|
| 140 | 'hiddengli' => "yes" },
|
---|
| 141 | { 'name' => "manifest",
|
---|
| 142 | 'desc' => "{import.manifest}",
|
---|
| 143 | 'type' => "string",
|
---|
| 144 | 'deft' => "",
|
---|
| 145 | 'reqd' => "no",
|
---|
[8879] | 146 | 'hiddengli' => "yes" },
|
---|
[9233] | 147 | { 'name' => "listall",
|
---|
[12691] | 148 | 'desc' => "{export.listall}",
|
---|
[9233] | 149 | 'type' => "flag",
|
---|
| 150 | 'reqd' => "no" },
|
---|
[8879] | 151 | { 'name' => "debug",
|
---|
| 152 | 'desc' => "{export.debug}",
|
---|
| 153 | 'type' => "flag",
|
---|
| 154 | 'reqd' => "no",
|
---|
| 155 | 'hiddengli' => "yes" },
|
---|
| 156 | { 'name' => "faillog",
|
---|
| 157 | 'desc' => "{export.faillog}",
|
---|
| 158 | 'type' => "string",
|
---|
[10340] | 159 | 'deft' => "",
|
---|
[8879] | 160 | 'reqd' => "no",
|
---|
| 161 | 'modegli' => "4" },
|
---|
[14935] | 162 | { 'name' => "incremental",
|
---|
| 163 | 'desc' => "{import.incremental}",
|
---|
| 164 | 'type' => "flag",
|
---|
| 165 | 'hiddengli' => "yes" },
|
---|
[8879] | 166 | { 'name' => "keepold",
|
---|
| 167 | 'desc' => "{export.keepold}",
|
---|
| 168 | 'type' => "flag",
|
---|
| 169 | 'reqd' => "no",
|
---|
[14935] | 170 | 'hiddengli' => "yes" },
|
---|
[10340] | 171 | { 'name' => "removeold",
|
---|
| 172 | 'desc' => "{export.removeold}",
|
---|
| 173 | 'type' => "flag",
|
---|
| 174 | 'reqd' => "no",
|
---|
| 175 | 'modegli' => "3" },
|
---|
[8879] | 176 | { 'name' => "language",
|
---|
| 177 | 'desc' => "{scripts.language}",
|
---|
| 178 | 'type' => "string",
|
---|
| 179 | 'reqd' => "no",
|
---|
| 180 | 'modegli' => "4" },
|
---|
| 181 | { 'name' => "maxdocs",
|
---|
| 182 | 'desc' => "{export.maxdocs}",
|
---|
| 183 | 'type' => "int",
|
---|
| 184 | 'reqd' => "no",
|
---|
| 185 | 'range' => "1,",
|
---|
| 186 | 'modegli' => "1" },
|
---|
[12691] | 187 | { 'name' => "OIDtype",
|
---|
| 188 | 'desc' => "{import.OIDtype}",
|
---|
| 189 | 'type' => "enum",
|
---|
| 190 | 'list' => $oidtype_list,
|
---|
| 191 | # parsearg left "" as default
|
---|
| 192 | #'deft' => "hash",
|
---|
| 193 | 'reqd' => "no",
|
---|
| 194 | 'modegli' => "3" },
|
---|
| 195 | { 'name' => "OIDmetadata",
|
---|
| 196 | 'desc' => "{import.OIDmetadata}",
|
---|
| 197 | 'type' => "metadata",
|
---|
| 198 | 'deft' => "dc.Identifier",
|
---|
| 199 | 'reqd' => "no",
|
---|
| 200 | 'modegli' => "3" },
|
---|
[8879] | 201 | { 'name' => "out",
|
---|
| 202 | 'desc' => "{export.out}",
|
---|
| 203 | 'type' => "string",
|
---|
| 204 | 'deft' => "STDERR",
|
---|
| 205 | 'reqd' => "no",
|
---|
| 206 | 'hiddengli' => "yes" },
|
---|
[9951] | 207 | { 'name' => "statsfile",
|
---|
| 208 | 'desc' => "{export.statsfile}",
|
---|
| 209 | 'type' => "string",
|
---|
| 210 | 'deft' => "STDERR",
|
---|
| 211 | 'reqd' => "no",
|
---|
| 212 | 'hiddengli' => "yes" },
|
---|
[12333] | 213 | { 'name' => "xsltfile",
|
---|
| 214 | 'desc' => "{BasPlugout.xslt_file}",
|
---|
| 215 | 'type' => "string",
|
---|
| 216 | 'reqd' => "no",
|
---|
| 217 | 'hiddengli' => "yes" },
|
---|
| 218 | { 'name' => "xslt_txt",
|
---|
| 219 | 'desc' => "{METSPlugout.xslt_txt}",
|
---|
| 220 | 'type' => "string",
|
---|
| 221 | 'reqd' => "no",
|
---|
| 222 | 'hiddengli' => "no" },
|
---|
| 223 | { 'name' => "xslt_mets",
|
---|
| 224 | 'desc' => "{METSPlugout.xslt_mets}",
|
---|
| 225 | 'type' => "string",
|
---|
| 226 | 'reqd' => "no",
|
---|
| 227 | 'hiddengli' => "no" },
|
---|
[14935] | 228 | { 'name' => "fedora_namespace",
|
---|
| 229 | 'desc' => "{FedoraMETSPlugout.fedora_namespace}",
|
---|
| 230 | 'type' => "string",
|
---|
| 231 | 'deft' => "greenstone",
|
---|
| 232 | 'reqd' => "no",
|
---|
| 233 | 'hiddengli' => "no" },
|
---|
[12598] | 234 | { 'name' => "mapping_file",
|
---|
| 235 | 'desc' => "{MARCXMLPlugout.mapping_file}",
|
---|
| 236 | 'type' => "string",
|
---|
| 237 | 'reqd' => "no",
|
---|
| 238 | 'hiddengli' => "no" },
|
---|
[12333] | 239 | { 'name' => "group_marc",
|
---|
| 240 | 'desc' => "{MARCXMLPlugout.group}",
|
---|
| 241 | 'type' => "flag",
|
---|
| 242 | 'reqd' => "no",
|
---|
| 243 | 'hiddengli' => "no" },
|
---|
[8879] | 244 | { 'name' => "verbosity",
|
---|
| 245 | 'desc' => "{export.verbosity}",
|
---|
| 246 | 'type' => "int",
|
---|
| 247 | 'range' => "0,3",
|
---|
| 248 | 'deft' => "2",
|
---|
| 249 | 'reqd' => "no",
|
---|
[10340] | 250 | 'modegli' => "4" },
|
---|
| 251 | { 'name' => "gli",
|
---|
| 252 | 'desc' => "",
|
---|
| 253 | 'type' => "flag",
|
---|
| 254 | 'reqd' => "no",
|
---|
| 255 | 'hiddengli' => "yes" },
|
---|
| 256 | { 'name' => "xml",
|
---|
| 257 | 'desc' => "{scripts.xml}",
|
---|
| 258 | 'type' => "flag",
|
---|
| 259 | 'reqd' => "no",
|
---|
| 260 | 'hiddengli' => "yes" }
|
---|
| 261 | ];
|
---|
[8879] | 262 |
|
---|
| 263 | my $options = { 'name' => "export.pl",
|
---|
| 264 | 'desc' => "{export.desc}",
|
---|
| 265 | 'args' => $arguments };
|
---|
| 266 |
|
---|
[9233] | 267 | my $listall_options = { 'name' => "export.pl",
|
---|
| 268 | 'desc' => "{export.desc}",
|
---|
| 269 | 'args' => [ $saveas_argument ] };
|
---|
| 270 |
|
---|
[14935] | 271 | #sub gsprintf
|
---|
| 272 | #{
|
---|
| 273 | # return &gsprintf::gsprintf(@_);
|
---|
| 274 | #}
|
---|
| 275 |
|
---|
| 276 |
|
---|
| 277 | sub store_collectionmeta
|
---|
[8879] | 278 | {
|
---|
[14935] | 279 | my ($collectionmeta,$field,$doc_obj) = @_;
|
---|
| 280 |
|
---|
| 281 | my $section = $doc_obj->get_top_section();
|
---|
| 282 |
|
---|
| 283 | my $field_hash = $collectionmeta->{$field};
|
---|
| 284 |
|
---|
| 285 | foreach my $k (keys %$field_hash)
|
---|
| 286 | {
|
---|
| 287 | my $val = $field_hash->{$k};
|
---|
| 288 |
|
---|
| 289 |
|
---|
| 290 | ### print STDERR "*** $k = $field_hash->{$k}\n";
|
---|
| 291 |
|
---|
| 292 | my $md_label = "ex.$field";
|
---|
| 293 |
|
---|
| 294 |
|
---|
| 295 | if ($k =~ m/^\[l=(.*?)\]$/)
|
---|
| 296 | {
|
---|
| 297 |
|
---|
| 298 | my $md_suffix = $1;
|
---|
| 299 | $md_label .= "^$md_suffix";
|
---|
| 300 | }
|
---|
| 301 |
|
---|
| 302 |
|
---|
| 303 | $doc_obj->add_utf8_metadata($section,$md_label, $val);
|
---|
| 304 |
|
---|
| 305 | if (($md_label eq "ex.collectionname^en") || ($md_label eq "ex.collectionname"))
|
---|
| 306 | {
|
---|
| 307 | $doc_obj->add_utf8_metadata($section,"dc.Title", $val);
|
---|
| 308 | }
|
---|
| 309 |
|
---|
| 310 | }
|
---|
[8879] | 311 | }
|
---|
| 312 |
|
---|
| 313 |
|
---|
| 314 | &main();
|
---|
| 315 |
|
---|
| 316 | sub main {
|
---|
[10340] | 317 | # params
|
---|
[14935] | 318 | my ($language, $verbosity, $debug,
|
---|
| 319 | $importdir, $manifest, $incremental, $keepold,
|
---|
| 320 | $exportdir, $site, $listall,
|
---|
| 321 | $removeold, $saveas,
|
---|
| 322 | $OIDtype, $OIDmetadata,
|
---|
| 323 | $out, $faillog, $collectdir, $gli,
|
---|
| 324 | $gs_mode, $collectcfg,
|
---|
| 325 | $maxdocs, $statsfile,
|
---|
| 326 | $mapping_file,
|
---|
| 327 | $xsltfile,
|
---|
| 328 | $xslt_mets, $xslt_txt, $fedora_namespace, $group_marc);
|
---|
| 329 |
|
---|
[8879] | 330 | my $xml = 0;
|
---|
[10340] | 331 |
|
---|
| 332 | # other vars
|
---|
[14935] | 333 | my ($configfilename, $collection,
|
---|
| 334 | $export_info_filename, $export_info,
|
---|
| 335 | $processor, $pluginfo);
|
---|
| 336 |
|
---|
[8879] | 337 | my $service = "export";
|
---|
| 338 |
|
---|
[10340] | 339 | my $hashParsingResult = {};
|
---|
| 340 | # general options available to all plugins
|
---|
| 341 | my $intArgLeftinAfterParsing = parse2::parse(\@ARGV,$arguments,$hashParsingResult,"allow_extra_options");
|
---|
[12545] | 342 |
|
---|
| 343 | # If parse returns -1 then something has gone wrong
|
---|
| 344 | if ($intArgLeftinAfterParsing == -1)
|
---|
| 345 | {
|
---|
| 346 | &PrintUsage::print_txt_usage($options, "{export.params}");
|
---|
| 347 | die "\n";
|
---|
| 348 | }
|
---|
| 349 |
|
---|
[10340] | 350 | foreach my $strVariable (keys %$hashParsingResult)
|
---|
| 351 | {
|
---|
| 352 | eval "\$$strVariable = \$hashParsingResult->{\"\$strVariable\"}";
|
---|
[8879] | 353 | }
|
---|
| 354 |
|
---|
[10340] | 355 |
|
---|
| 356 | # these are options used by other things - we just set default values
|
---|
| 357 | # undef means will be set from config file if there
|
---|
| 358 | my $gzip = undef;
|
---|
| 359 | my $groupsize = 1;
|
---|
[12691] | 360 | #my $OIDtype = undef;
|
---|
[10340] | 361 | my $sortmeta = undef;
|
---|
[9233] | 362 |
|
---|
[9235] | 363 | my $explicit_exportdir = (defined $exportdir) ? 1 : 0;
|
---|
| 364 |
|
---|
[10417] | 365 | # save these command line settings. don't want config file settings in one
|
---|
| 366 | # coll used for other colls
|
---|
| 367 | # does this apply to other vars???
|
---|
| 368 | my $global_removeold = $removeold;
|
---|
| 369 | my $global_keepold = $keepold;
|
---|
[8879] | 370 | # If $language has been specified, load the appropriate resource bundle
|
---|
| 371 | # (Otherwise, the default resource bundle will be loaded automatically)
|
---|
| 372 | if ($language) {
|
---|
| 373 | &gsprintf::load_language_specific_resource_bundle($language);
|
---|
| 374 | }
|
---|
| 375 |
|
---|
[9233] | 376 | if ($listall) {
|
---|
| 377 | if ($xml) {
|
---|
| 378 | &PrintUsage::print_xml_usage($listall_options);
|
---|
| 379 | }
|
---|
| 380 | else
|
---|
| 381 | {
|
---|
| 382 | &PrintUsage::print_txt_usage($listall_options,"{export.params}");
|
---|
| 383 | }
|
---|
| 384 | die "\n";
|
---|
| 385 | }
|
---|
| 386 | elsif ($xml) {
|
---|
[8879] | 387 | &PrintUsage::print_xml_usage($options);
|
---|
| 388 | die "\n";
|
---|
| 389 | }
|
---|
| 390 |
|
---|
[12545] | 391 | # can have more than one collection name,
|
---|
| 392 | # if the first extra option is -h, then output the help
|
---|
| 393 | if (scalar(@ARGV) == 0 || (@ARGV && $ARGV[0] =~ /^\-+h/)) {
|
---|
[10340] | 394 | &PrintUsage::print_txt_usage($options, "{export.params}");
|
---|
| 395 | die "\n";
|
---|
| 396 | }
|
---|
| 397 |
|
---|
[8879] | 398 | if ($gli) { # the gli wants strings to be in UTF-8
|
---|
| 399 | &gsprintf::output_strings_in_UTF8;
|
---|
| 400 | }
|
---|
| 401 | my $close_out = 0;
|
---|
| 402 | if ($out !~ /^(STDERR|STDOUT)$/i) {
|
---|
| 403 | open (OUT, ">$out") ||
|
---|
| 404 | (&gsprintf(STDERR, "{common.cannot_open_output_file}\n", $out) && die);
|
---|
| 405 | $out = 'export::OUT';
|
---|
| 406 | $close_out = 1;
|
---|
| 407 | }
|
---|
| 408 | $out->autoflush(1);
|
---|
| 409 |
|
---|
[9233] | 410 | while (scalar(@ARGV)>0) {
|
---|
| 411 | my $collect_name = shift @ARGV;
|
---|
| 412 | $ENV{'GSDLCOLLECTION'} = $collect_name;
|
---|
[8879] | 413 |
|
---|
[9233] | 414 | eval {
|
---|
| 415 | # get and check the collection name
|
---|
[14925] | 416 | if (($collection = &colcfg::use_collection($site, $collect_name, $collectdir)) eq "") {
|
---|
[9233] | 417 | &PrintUsage::print_txt_usage($options, "{export.params}");
|
---|
| 418 | die "\n";
|
---|
| 419 | }
|
---|
[12358] | 420 | # add collection's perllib dir into include path in
|
---|
| 421 | # case we have collection specific modules
|
---|
| 422 | unshift (@INC, "$ENV{'GSDLCOLLECTDIR'}/perllib");
|
---|
| 423 |
|
---|
[9233] | 424 | if ($faillog eq "") {
|
---|
| 425 | $faillog = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "etc", "fail.log");
|
---|
| 426 | }
|
---|
| 427 | open (FAILLOG, ">$faillog") ||
|
---|
| 428 | (&gsprintf(STDERR, "{export.cannot_open_fail_log}\n", $faillog) && die);
|
---|
| 429 | my $faillogname = $faillog;
|
---|
| 430 | $faillog = 'export::FAILLOG';
|
---|
| 431 | $faillog->autoflush(1);
|
---|
| 432 |
|
---|
| 433 | # check sortmeta
|
---|
| 434 | $sortmeta = undef unless defined $sortmeta && $sortmeta =~ /\S/;
|
---|
| 435 | if (defined $sortmeta && $groupsize > 1) {
|
---|
| 436 | &gsprintf($out, "{export.cannot_sort}\n\n");
|
---|
| 437 | $sortmeta = undef;
|
---|
| 438 | }
|
---|
| 439 |
|
---|
| 440 | # get the list of plugins for this collection and set any options that
|
---|
| 441 | # were specified in the collect.cfg (all export.pl options except
|
---|
| 442 | # -collectdir, -out and -faillog may be specified in the collect.cfg (these
|
---|
| 443 | # options must be known before we read the collect.cfg))
|
---|
| 444 | my $plugins = [];
|
---|
| 445 | my @global_opts = ();
|
---|
[14935] | 446 |
|
---|
[14925] | 447 | # Read in the collection configuration file.
|
---|
| 448 | ($configfilename, $gs_mode) = &colcfg::get_collect_cfg_name($out);
|
---|
[14935] | 449 |
|
---|
[14925] | 450 | if ($gs_mode eq "gs2") {
|
---|
| 451 | $collectcfg = &colcfg::read_collect_cfg ($configfilename);
|
---|
| 452 | } elsif ($gs_mode eq "gs3") {
|
---|
| 453 | $collectcfg = &colcfg::read_collection_cfg_xml ($configfilename);
|
---|
[10417] | 454 | }
|
---|
[14925] | 455 |
|
---|
[14935] | 456 | if (defined $collectcfg->{'importdir'} && $importdir eq "") {
|
---|
| 457 | $importdir = $collectcfg->{'importdir'};
|
---|
| 458 | }
|
---|
| 459 | if (defined $collectcfg->{'exportdir'} && $exportdir eq "") {
|
---|
| 460 | $exportdir = $collectcfg->{'exportdir'};
|
---|
| 461 | }
|
---|
| 462 |
|
---|
| 463 | # fill in the default import and export directories if none
|
---|
| 464 | # were supplied, turn all \ into / and remove trailing /
|
---|
| 465 | $importdir = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "import") if $importdir eq "";
|
---|
| 466 | $importdir =~ s/[\\\/]+/\//g;
|
---|
| 467 | $importdir =~ s/\/$//;
|
---|
| 468 | if (!-e $importdir) {
|
---|
| 469 | &gsprintf($out, "{import.no_import_dir}\n\n", $importdir);
|
---|
| 470 | die "\n";
|
---|
| 471 | }
|
---|
| 472 |
|
---|
| 473 | $exportdir = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "export") if $exportdir eq "";
|
---|
| 474 | $exportdir =~ s/[\\\/]+/\//g;
|
---|
| 475 | $exportdir =~ s/\/$//;
|
---|
| 476 |
|
---|
[10417] | 477 | if (defined $collectcfg->{'plugin'}) {
|
---|
| 478 | $plugins = $collectcfg->{'plugin'};
|
---|
| 479 | }
|
---|
| 480 |
|
---|
| 481 | if ($verbosity !~ /\d+/) {
|
---|
| 482 | if (defined $collectcfg->{'verbosity'} && $collectcfg->{'verbosity'} =~ /\d+/) {
|
---|
| 483 | $verbosity = $collectcfg->{'verbosity'};
|
---|
| 484 | } else {
|
---|
| 485 | $verbosity = 2; # the default
|
---|
[9233] | 486 | }
|
---|
[10417] | 487 | }
|
---|
| 488 |
|
---|
| 489 | if (defined $collectcfg->{'gzip'} && !$gzip) {
|
---|
| 490 | if ($collectcfg->{'gzip'} =~ /^true$/i) {
|
---|
| 491 | $gzip = 1;
|
---|
[9233] | 492 | }
|
---|
[10417] | 493 | }
|
---|
| 494 | if ($maxdocs !~ /\-?\d+/) {
|
---|
| 495 | if (defined $collectcfg->{'maxdocs'} && $collectcfg->{'maxdocs'} =~ /\-?\d+/) {
|
---|
| 496 | $maxdocs = $collectcfg->{'maxdocs'};
|
---|
| 497 | } else {
|
---|
| 498 | $maxdocs = -1; # the default
|
---|
[9233] | 499 | }
|
---|
[10417] | 500 | }
|
---|
| 501 | if ($groupsize == 1) {
|
---|
| 502 | if (defined $collectcfg->{'groupsize'} && $collectcfg->{'groupsize'} =~ /\d+/) {
|
---|
| 503 | $groupsize = $collectcfg->{'groupsize'};
|
---|
[9233] | 504 | }
|
---|
[10417] | 505 | }
|
---|
[12691] | 506 | if (!defined $OIDtype || ($OIDtype !~ /^(hash|incremental|assigned|dirname)$/)) {
|
---|
[10417] | 507 | if (defined $collectcfg->{'OIDtype'} && $collectcfg->{'OIDtype'} =~ /^(hash|incremental)$/) {
|
---|
| 508 | $OIDtype = $collectcfg->{'OIDtype'};
|
---|
| 509 | } else {
|
---|
| 510 | $OIDtype = "hash"; # the default
|
---|
[9233] | 511 | }
|
---|
[8879] | 512 | }
|
---|
[10417] | 513 | if (defined $collectcfg->{'sortmeta'} && $sortmeta eq "") {
|
---|
| 514 | $sortmeta = $collectcfg->{'sortmeta'};
|
---|
| 515 | }
|
---|
| 516 | if (defined $collectcfg->{'debug'} && $collectcfg->{'debug'} =~ /^true$/i) {
|
---|
| 517 | $debug = 1;
|
---|
| 518 | }
|
---|
| 519 | if (defined $collectcfg->{'gli'} && $collectcfg->{'gli'} =~ /^true$/i) {
|
---|
| 520 | $gli = 1;
|
---|
| 521 | }
|
---|
[9233] | 522 |
|
---|
[10417] | 523 | # global plugin stuff
|
---|
| 524 | if (defined $collectcfg->{'separate_cjk'}&& $collectcfg->{'separate_cjk'} =~ /^true$/i) {
|
---|
| 525 | push @global_opts, "-separate_cjk";
|
---|
| 526 | }
|
---|
| 527 |
|
---|
[12965] | 528 | ($removeold, $keepold) = &scriptutil::check_removeold_and_keepold($global_removeold, $global_keepold, 0, "export", $collectcfg);
|
---|
[10417] | 529 |
|
---|
[9233] | 530 | $gli = 0 unless defined $gli;
|
---|
| 531 |
|
---|
| 532 | print STDERR "<export>\n" if $gli;
|
---|
| 533 |
|
---|
| 534 | # fill in the default import and export directories if none
|
---|
| 535 | # were supplied, turn all \ into / and remove trailing /
|
---|
| 536 | $importdir = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "import") if $importdir eq "";
|
---|
| 537 | $importdir =~ s/[\\\/]+/\//g;
|
---|
| 538 | $importdir =~ s/\/$//;
|
---|
| 539 | $exportdir = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "export") if $exportdir eq "";
|
---|
| 540 | $exportdir =~ s/[\\\/]+/\//g;
|
---|
| 541 | $exportdir =~ s/\/$//;
|
---|
| 542 |
|
---|
| 543 | # load all the plugins
|
---|
| 544 | $pluginfo = &plugin::load_plugins ($plugins, $verbosity, $out, $faillog, \@global_opts);
|
---|
| 545 |
|
---|
| 546 | if (scalar(@$pluginfo) == 0) {
|
---|
| 547 | &gsprintf($out, "{import.no_plugins_loaded}\n");
|
---|
| 548 | die "\n";
|
---|
[8879] | 549 | }
|
---|
[9233] | 550 |
|
---|
| 551 | # remove the old contents of the export directory if needed
|
---|
| 552 | if ($removeold && -e $exportdir) {
|
---|
| 553 | &gsprintf($out, "{export.removing_export}\n");
|
---|
| 554 | &util::rm_r ($exportdir);
|
---|
[8879] | 555 | }
|
---|
[9233] | 556 |
|
---|
| 557 | # read the export information file
|
---|
[13169] | 558 |
|
---|
| 559 | # Export to DSpace Archive format or METs format
|
---|
| 560 | # If saveas=DSpace, a "contents" file will be created, otherwise "export.inf"
|
---|
[9233] | 561 |
|
---|
[13169] | 562 | # the plugouts should be doing this!!
|
---|
| 563 | if ($saveas eq "DSpace"){
|
---|
| 564 | $export_info_filename = &util::filename_cat ($exportdir, "contents");
|
---|
[14935] | 565 | } elsif ($saveas =~ m/^.*METS$/ || $saveas eq "GA" || $saveas eq "MARC" ) {
|
---|
[13169] | 566 | $export_info_filename = &util::filename_cat ($exportdir, "export.inf");
|
---|
| 567 | }
|
---|
| 568 |
|
---|
| 569 | $export_info = new arcinfo();
|
---|
| 570 | $export_info -> load_info ($export_info_filename);
|
---|
| 571 |
|
---|
[14733] | 572 | my ($plugout);
|
---|
| 573 | if (defined $collectcfg->{'plugout'}) {
|
---|
| 574 | $plugout = $collectcfg->{'plugout'};
|
---|
| 575 | }
|
---|
| 576 | else{
|
---|
[14935] | 577 | if ($saveas !~ /^(GA|.*METS|DSpace|MARCXML)$/) {
|
---|
[14733] | 578 | push @$plugout,"GAPlugout";
|
---|
[9233] | 579 | }
|
---|
[14733] | 580 | else{
|
---|
| 581 | push @$plugout,$saveas."Plugout";
|
---|
| 582 | }
|
---|
[8879] | 583 | }
|
---|
[14733] | 584 |
|
---|
| 585 | my $plugout_name = $plugout->[0];
|
---|
[9233] | 586 |
|
---|
[14733] | 587 | push @$plugout,("-output_info",$export_info) if (defined $export_info);
|
---|
| 588 | push @$plugout,("-verbosity",$verbosity) if (defined $verbosity);
|
---|
| 589 | push @$plugout,("-debug") if ($debug);
|
---|
| 590 | push @$plugout,("-gzip_output",$gzip) if (defined $gzip);
|
---|
| 591 | push @$plugout,("-group_size",$groupsize) if (defined $groupsize);
|
---|
| 592 | push @$plugout,("-output_handle",$out) if (defined $out);
|
---|
| 593 | push @$plugout,("-xslt_file",$xsltfile) if (defined $xsltfile);
|
---|
[14935] | 594 | push @$plugout,("-group") if ($group_marc && $plugout_name =~ m/^MARCXMLPlugout$/);
|
---|
| 595 | push @$plugout,("-mapping_file",$mapping_file) if (defined $mapping_file && $plugout_name =~ m/^MARCXMLPlugout$/);
|
---|
| 596 | push @$plugout,("-xslt_mets",$xslt_mets) if (defined $xslt_mets && $plugout_name =~ m/^.*METSPlugout$/);
|
---|
| 597 | push @$plugout,("-xslt_txt",$xslt_txt) if (defined $xslt_txt && $plugout_name =~ m/^.*METSPlugout$/);
|
---|
| 598 | push @$plugout,("-fedora_namespace",$fedora_namespace) if (defined $fedora_namespace && $plugout_name eq "FedoraMETSPlugout");
|
---|
| 599 |
|
---|
[14733] | 600 | $processor = &plugout::load_plugout($plugout);
|
---|
[13169] | 601 |
|
---|
| 602 | $processor->setoutputdir ($exportdir);
|
---|
| 603 |
|
---|
| 604 | $processor->set_sortmeta ($sortmeta) if defined $sortmeta;
|
---|
| 605 | $processor->set_OIDtype ($OIDtype, $OIDmetadata);
|
---|
| 606 |
|
---|
[9233] | 607 | &plugin::begin($pluginfo, $importdir, $processor, $maxdocs);
|
---|
| 608 |
|
---|
| 609 | # process the import directory
|
---|
[9854] | 610 | &plugin::read ($pluginfo, $importdir, "", {}, $processor, $maxdocs,0, $gli);
|
---|
[14935] | 611 |
|
---|
| 612 | if ($saveas eq "FedoraMETS")
|
---|
| 613 | {
|
---|
| 614 | # create collection "doc obj" for Fedora that contains
|
---|
| 615 | # collection-level metadata
|
---|
| 616 |
|
---|
| 617 | my $doc_obj = new doc($configfilename,"nonindexed_doc");
|
---|
| 618 | $doc_obj->set_OID("collection");
|
---|
| 619 |
|
---|
| 620 | my $col_name = undef;
|
---|
| 621 | my $col_meta = $collectcfg->{'collectionmeta'};
|
---|
| 622 | if (defined $col_meta)
|
---|
| 623 | {
|
---|
| 624 | store_collectionmeta($col_meta,"collectionname",$doc_obj);
|
---|
| 625 | store_collectionmeta($col_meta,"collectioextra",$doc_obj);
|
---|
| 626 |
|
---|
| 627 | }
|
---|
| 628 |
|
---|
| 629 | $processor->process($doc_obj);
|
---|
| 630 | }
|
---|
[9233] | 631 |
|
---|
| 632 | &plugin::end($pluginfo, $processor);
|
---|
[10162] | 633 |
|
---|
| 634 | &plugin::deinit($pluginfo, $processor);
|
---|
[9233] | 635 |
|
---|
| 636 | # write out the export information file
|
---|
[13169] | 637 | $processor->close_file_output() if $groupsize > 1;
|
---|
| 638 | $processor->close_group_output() if $processor->is_group();
|
---|
[14935] | 639 | if ($saveas =~ m/^.*METS$/) {
|
---|
[13169] | 640 | $export_info->save_info($export_info_filename);
|
---|
[8879] | 641 | }
|
---|
[9233] | 642 |
|
---|
[9951] | 643 | # write out export stats
|
---|
| 644 | my $close_stats = 0;
|
---|
| 645 | if ($statsfile !~ /^(STDERR|STDOUT)$/i) {
|
---|
| 646 | if (open (STATS, ">$statsfile")) {
|
---|
| 647 | $statsfile = 'import::STATS';
|
---|
| 648 | $close_stats = 1;
|
---|
| 649 | } else {
|
---|
| 650 | &gsprintf($out, "{import.cannot_open_stats_file}", $statsfile);
|
---|
| 651 | &gsprintf($out, "{import.stats_backup}\n");
|
---|
| 652 | $statsfile = 'STDERR';
|
---|
| 653 | }
|
---|
| 654 | }
|
---|
| 655 |
|
---|
| 656 | &gsprintf($out, "\n");
|
---|
| 657 | &gsprintf($out, "*********************************************\n");
|
---|
| 658 |
|
---|
| 659 | &plugin::write_stats($pluginfo, $statsfile, $faillogname, $gli);
|
---|
| 660 | if ($close_stats) {
|
---|
| 661 | close STATS;
|
---|
| 662 | }
|
---|
| 663 |
|
---|
| 664 | &gsprintf($out, "*********************************************\n");
|
---|
| 665 |
|
---|
| 666 | close OUT if $close_out;
|
---|
| 667 |
|
---|
[9233] | 668 | close FAILLOG;
|
---|
| 669 | };
|
---|
[8879] | 670 |
|
---|
[9951] | 671 | if ($@) {
|
---|
| 672 | print STDERR $@;
|
---|
| 673 | }
|
---|
| 674 |
|
---|
[9233] | 675 | ## $ENV{'GSDLCOLLECTION'} = undef;
|
---|
| 676 | $importdir = "";
|
---|
[9235] | 677 | $removeold = 0 if ($explicit_exportdir);
|
---|
[9233] | 678 |
|
---|
[10340] | 679 | } # while processing ARGV
|
---|
[9233] | 680 |
|
---|
[8879] | 681 | &gsprintf($out, "\n");
|
---|
| 682 | &gsprintf($out, "*********************************************\n");
|
---|
[9951] | 683 | &gsprintf($out, "* {export.complete}\n");
|
---|
| 684 | &gsprintf($out, "*********************************************\n");
|
---|
[8879] | 685 |
|
---|
| 686 | }
|
---|