source: gsdl/trunk/bin/script/import.pl@ 19500

Last change on this file since 19500 was 19500, checked in by davidb, 15 years ago

Code now tracks if a document needs to be re-processed on an incremental build when it's metadata has changed.
  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
File size: 21.6 KB
Line 
1#!/usr/bin/perl -w
2
3###########################################################################
4#
5# import.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 import a number of files into a particular collection
30
31package import;
32
33BEGIN {
34 die "GSDLHOME not set\n" unless defined $ENV{'GSDLHOME'};
35 die "GSDLOS not set\n" unless defined $ENV{'GSDLOS'};
36 unshift (@INC, "$ENV{'GSDLHOME'}/perllib");
37 unshift (@INC, "$ENV{'GSDLHOME'}/perllib/cpan");
38 unshift (@INC, "$ENV{'GSDLHOME'}/perllib/plugins");
39 unshift (@INC, "$ENV{'GSDLHOME'}/perllib/plugouts");
40 unshift (@INC, "$ENV{'GSDLHOME'}/perllib/classify");
41
42 if (defined $ENV{'GSDLEXTS'}) {
43 my @extensions = split(/:/,$ENV{'GSDLEXTS'});
44 foreach my $e (@extensions) {
45 my $ext_prefix = "$ENV{'GSDLHOME'}/ext/$e";
46
47 unshift (@INC, "$ext_prefix/perllib");
48 unshift (@INC, "$ext_prefix/perllib/cpan");
49 unshift (@INC, "$ext_prefix/perllib/plugins");
50 unshift (@INC, "$ext_prefix/perllib/plugouts");
51 unshift (@INC, "$ext_prefix/perllib/classify");
52 }
53 }
54}
55
56use strict;
57no strict 'refs'; # allow filehandles to be variables and vice versa
58no strict 'subs'; # allow barewords (eg STDERR) as function arguments
59
60use arcinfo;
61use colcfg;
62use plugin;
63use plugout;
64use manifest;
65use inexport;
66use util;
67use scriptutil;
68use FileHandle;
69use gsprintf 'gsprintf';
70use printusage;
71use parse2;
72
73
74
75my $oidtype_list =
76 [ { 'name' => "hash",
77 'desc' => "{import.OIDtype.hash}" },
78 { 'name' => "assigned",
79 'desc' => "{import.OIDtype.assigned}" },
80 { 'name' => "incremental",
81 'desc' => "{import.OIDtype.incremental}" },
82 { 'name' => "dirname",
83 'desc' => "{import.OIDtype.dirname}" } ];
84
85
86# used to control output file format
87my $saveas_list =
88 [ { 'name' => "GreenstoneXML",
89 'desc' => "{export.saveas.GreenstoneXML}"},
90 { 'name' => "GreenstoneMETS",
91 'desc' => "{export.saveas.GreenstoneMETS}"},
92 ];
93
94
95# Possible attributes for each argument
96# name: The name of the argument
97# desc: A description (or more likely a reference to a description) for this argument
98# type: The type of control used to represent the argument. Options include: string, int, flag, regexp, metadata, language, enum etc
99# reqd: Is this argument required?
100# hiddengli: Is this argument hidden in GLI?
101# modegli: The lowest detail mode this argument is visible at in GLI
102
103my $saveas_argument
104 = { 'name' => "saveas",
105 'desc' => "{import.saveas}",
106 'type' => "enum",
107 'list' => $saveas_list,
108 'deft' => "GreenstoneXML",
109 'reqd' => "no",
110 'modegli' => "3" };
111
112
113my $arguments =
114 [
115 $saveas_argument,
116 { 'name' => "archivedir",
117 'desc' => "{import.archivedir}",
118 'type' => "string",
119 'reqd' => "no",
120 'hiddengli' => "yes" },
121 { 'name' => "importdir",
122 'desc' => "{import.importdir}",
123 'type' => "string",
124 'reqd' => "no",
125 'hiddengli' => "yes" },
126 { 'name' => "collectdir",
127 'desc' => "{import.collectdir}",
128 'type' => "string",
129 # parsearg left "" as default
130 #'deft' => &util::filename_cat ($ENV{'GSDLHOME'}, "collect"),
131 'deft' => "",
132 'reqd' => "no",
133 'hiddengli' => "yes" },
134 { 'name' => "site",
135 'desc' => "{import.site}",
136 'type' => "string",
137 'deft' => "",
138 'reqd' => "no",
139 'hiddengli' => "yes" },
140 { 'name' => "manifest",
141 'desc' => "{import.manifest}",
142 'type' => "string",
143 'deft' => "",
144 'reqd' => "no",
145 'hiddengli' => "yes" },
146 { 'name' => "debug",
147 'desc' => "{import.debug}",
148 'type' => "flag",
149 'reqd' => "no",
150 'hiddengli' => "yes" },
151 { 'name' => "faillog",
152 'desc' => "{import.faillog}",
153 'type' => "string",
154 # parsearg left "" as default
155 #'deft' => &util::filename_cat("<collectdir>", "colname", "etc", "fail.log"),
156 'deft' => "",
157 'reqd' => "no",
158 'modegli' => "3" },
159 { 'name' => "incremental",
160 'desc' => "{import.incremental}",
161 'type' => "flag",
162 'hiddengli' => "yes" },
163 { 'name' => "keepold",
164 'desc' => "{import.keepold}",
165 'type' => "flag",
166 'reqd' => "no",
167 'hiddengli' => "yes" },
168 { 'name' => "removeold",
169 'desc' => "{import.removeold}",
170 'type' => "flag",
171 'reqd' => "no",
172 'hiddengli' => "yes" },
173 { 'name' => "language",
174 'desc' => "{scripts.language}",
175 'type' => "string",
176 'reqd' => "no",
177 'hiddengli' => "yes" },
178 { 'name' => "maxdocs",
179 'desc' => "{import.maxdocs}",
180 'type' => "int",
181 'reqd' => "no",
182 # parsearg left "" as default
183 #'deft' => "-1",
184 'range' => "1,",
185 'modegli' => "1" },
186 # don't set the default to hash - want to allow this to come from
187 # entry in collect.cfg but want to override it here
188 { 'name' => "OIDtype",
189 'desc' => "{import.OIDtype}",
190 'type' => "enum",
191 'list' => $oidtype_list,
192 # parsearg left "" as default
193 #'deft' => "hash",
194 'reqd' => "no",
195 'modegli' => "2" },
196 { 'name' => "OIDmetadata",
197 'desc' => "{import.OIDmetadata}",
198 'type' => "metadata",
199 # parsearg left "" as default
200 #'deft' => "dc.Identifier",
201 'reqd' => "no",
202 'modegli' => "2" },
203 { 'name' => "out",
204 'desc' => "{import.out}",
205 'type' => "string",
206 'deft' => "STDERR",
207 'reqd' => "no",
208 'hiddengli' => "yes" },
209 { 'name' => "sortmeta",
210 'desc' => "{import.sortmeta}",
211 'type' => "metadata",
212 'reqd' => "no",
213 'modegli' => "2" },
214 { 'name' => "reversesort",
215 'desc' => "{import.reversesort}",
216 'type' => "flag",
217 'reqd' => "no",
218 'modegli' => "2" },
219 { 'name' => "removeprefix",
220 'desc' => "{BasClas.removeprefix}",
221 'type' => "regexp",
222 'deft' => "",
223 'reqd' => "no",
224 'modegli' => "3" },
225 { 'name' => "removesuffix",
226 'desc' => "{BasClas.removesuffix}",
227 'type' => "regexp",
228 'deft' => "",
229 'reqd' => "no",
230 'modegli' => "3" },
231 { 'name' => "groupsize",
232 'desc' => "{import.groupsize}",
233 'type' => "int",
234 'deft' => "1",
235 'reqd' => "no",
236 'modegli' => "2" },
237 { 'name' => "gzip",
238 'desc' => "{import.gzip}",
239 'type' => "flag",
240 'reqd' => "no",
241 'modegli' => "3" },
242 { 'name' => "statsfile",
243 'desc' => "{import.statsfile}",
244 'type' => "string",
245 'deft' => "STDERR",
246 'reqd' => "no",
247 'hiddengli' => "yes" },
248 { 'name' => "verbosity",
249 'desc' => "{import.verbosity}",
250 'type' => "int",
251 'range' => "0,",
252 # parsearg left "" as default
253 #'deft' => "2",
254 'reqd' => "no",
255 'modegli' => "3" },
256 { 'name' => "gli",
257 'desc' => "{scripts.gli}",
258 'type' => "flag",
259 'reqd' => "no",
260 'hiddengli' => "yes" },
261 { 'name' => "xml",
262 'desc' => "{scripts.xml}",
263 'type' => "flag",
264 'reqd' => "no",
265 'hiddengli' => "yes" }];
266
267my $options = { 'name' => "import.pl",
268 'desc' => "{import.desc}",
269 'args' => $arguments };
270
271
272&main();
273
274sub main {
275 # params
276 my ($language, $verbosity, $debug,
277 $collectdir, $importdir, $archivedir, $site, $manifest,
278 $incremental, $keepold, $removeold,
279 $saveas,
280 $OIDtype, $OIDmetadata,
281 $maxdocs, $statsfile,
282 $out, $faillog, $gli,
283 $gzip, $groupsize,
284 $sortmeta, $reversesort, $removeprefix, $removesuffix
285 );
286
287 my $xml = 0;
288
289 # other vars
290 my ($configfilename, $collection, $collectcfg,
291 $arcinfo_doc_filename, $arcinfo_src_filename, $archive_info,
292 $gs_mode,
293 $processor, $pluginfo);
294
295 my $service = "import";
296
297 my $hashParsingResult = {};
298 # general options available to all plugins
299 my $intArgLeftinAfterParsing = parse2::parse(\@ARGV,$arguments,$hashParsingResult,"allow_extra_options");
300 # Parse returns -1 if something has gone wrong
301 if ($intArgLeftinAfterParsing == -1)
302 {
303 &PrintUsage::print_txt_usage($options, "{import.params}");
304 die "\n";
305 }
306
307 foreach my $strVariable (keys %$hashParsingResult)
308 {
309 eval "\$$strVariable = \$hashParsingResult->{\"\$strVariable\"}";
310 }
311
312 # If $language has been specified, load the appropriate resource bundle
313 # (Otherwise, the default resource bundle will be loaded automatically)
314 if ($language && $language =~ /\S/) {
315 &gsprintf::load_language_specific_resource_bundle($language);
316 }
317
318 if ($xml) {
319 &PrintUsage::print_xml_usage($options);
320 print "\n";
321 return;
322 }
323
324 if ($gli) { # the gli wants strings to be in UTF-8
325 &gsprintf::output_strings_in_UTF8;
326 }
327
328 # now check that we had exactly one leftover arg, which should be
329 # the collection name. We don't want to do this earlier, cos
330 # -xml arg doesn't need a collection name
331 # Or if the user specified -h, then we output the usage also
332 if ($intArgLeftinAfterParsing != 1 || (@ARGV && $ARGV[0] =~ /^\-+h/))
333 {
334 &PrintUsage::print_txt_usage($options, "{import.params}");
335 die "\n";
336 }
337
338 my $close_out = 0;
339 if ($out !~ /^(STDERR|STDOUT)$/i) {
340 open (OUT, ">$out") ||
341 (&gsprintf(STDERR, "{common.cannot_open_output_file}: $!\n", $out) && die);
342 $out = 'import::OUT';
343 $close_out = 1;
344 }
345 $out->autoflush(1);
346
347 # get and check the collection name
348 if (($collection = &colcfg::use_collection($site, @ARGV, $collectdir)) eq "") {
349 &PrintUsage::print_txt_usage($options, "{import.params}");
350 die "\n";
351 }
352
353 # add collection's perllib dir into include path in
354 # case we have collection specific modules
355 unshift (@INC, "$ENV{'GSDLCOLLECTDIR'}/perllib");
356
357 # check that we can open the faillog
358 if ($faillog eq "") {
359 $faillog = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "etc", "fail.log");
360 }
361 open (FAILLOG, ">$faillog") ||
362 (&gsprintf(STDERR, "{import.cannot_open_fail_log}\n", $faillog) && die);
363
364
365 my $faillogname = $faillog;
366 $faillog = 'import::FAILLOG';
367 $faillog->autoflush(1);
368
369 # Read in the collection configuration file.
370 ($configfilename, $gs_mode) = &colcfg::get_collect_cfg_name($out);
371
372 if ($gs_mode eq "gs2") {
373 $collectcfg = &colcfg::read_collect_cfg ($configfilename);
374 } elsif ($gs_mode eq "gs3") {
375 $collectcfg = &colcfg::read_collection_cfg_xml ($configfilename);
376 }
377
378 if (defined $collectcfg->{'importdir'} && $importdir eq "") {
379 $importdir = $collectcfg->{'importdir'};
380 }
381 if (defined $collectcfg->{'archivedir'} && $archivedir eq "") {
382 $archivedir = $collectcfg->{'archivedir'};
383 }
384 # fill in the default import and archives directories if none
385 # were supplied, turn all \ into / and remove trailing /
386 $importdir = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "import") if $importdir eq "";
387 $importdir =~ s/[\\\/]+/\//g;
388 $importdir =~ s/\/$//;
389 if (!-e $importdir) {
390 &gsprintf($out, "{import.no_import_dir}\n\n", $importdir);
391 die "\n";
392 }
393
394 $archivedir = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "archives") if $archivedir eq "";
395 $archivedir =~ s/[\\\/]+/\//g;
396 $archivedir =~ s/\/$//;
397
398 my $plugins = [];
399 if (defined $collectcfg->{'plugin'}) {
400 $plugins = $collectcfg->{'plugin'};
401 }
402 #some global options for the plugins
403 my @global_opts = ();
404
405 if ($verbosity !~ /\d+/) {
406 if (defined $collectcfg->{'verbosity'} && $collectcfg->{'verbosity'} =~ /\d+/) {
407 $verbosity = $collectcfg->{'verbosity'};
408 } else {
409 $verbosity = 2; # the default
410 }
411 }
412 if (defined $collectcfg->{'manifest'} && $manifest eq "") {
413 $manifest = $collectcfg->{'manifest'};
414 }
415
416 if (defined $collectcfg->{'gzip'} && !$gzip) {
417 if ($collectcfg->{'gzip'} =~ /^true$/i) {
418 $gzip = 1;
419 }
420 }
421
422 if ($maxdocs !~ /\-?\d+/) {
423 if (defined $collectcfg->{'maxdocs'} && $collectcfg->{'maxdocs'} =~ /\-?\d+/) {
424 $maxdocs = $collectcfg->{'maxdocs'};
425 } else {
426 $maxdocs = -1; # the default
427 }
428 }
429 if ($groupsize == 1) {
430 if (defined $collectcfg->{'groupsize'} && $collectcfg->{'groupsize'} =~ /\d+/) {
431 $groupsize = $collectcfg->{'groupsize'};
432 }
433 }
434
435 if (!defined $OIDtype || ($OIDtype !~ /^(hash|incremental|assigned|dirname)$/ )) {
436 if (defined $collectcfg->{'OIDtype'} && $collectcfg->{'OIDtype'} =~ /^(hash|incremental|assigned|dirname)$/) {
437 $OIDtype = $collectcfg->{'OIDtype'};
438 } else {
439 $OIDtype = "hash"; # the default
440 }
441 }
442
443 if ((!defined $OIDmetadata) || ($OIDmetadata eq "")) {
444 if (defined $collectcfg->{'OIDmetadata'}) {
445 $OIDmetadata = $collectcfg->{'OIDmetadata'};
446 } else {
447 $OIDmetadata = "dc.Identifier"; # the default
448 }
449 }
450
451 if (defined $collectcfg->{'sortmeta'} && (!defined $sortmeta || $sortmeta eq "")) {
452 $sortmeta = $collectcfg->{'sortmeta'};
453 }
454 # sortmeta cannot be used with group size
455 $sortmeta = undef unless defined $sortmeta && $sortmeta =~ /\S/;
456 if (defined $sortmeta && $groupsize > 1) {
457 &gsprintf($out, "{import.cannot_sort}\n\n");
458 $sortmeta = undef;
459 }
460
461 if (defined $sortmeta) {
462 if (defined $collectcfg->{'reversesort'} && $collectcfg->{'reversesort'} =~ /^true$/i) {
463 $reversesort = 1;
464 }
465 } else {
466 # reversesort only valid with sortmeta
467 $reversesort = 0;
468 }
469 if (defined $collectcfg->{'removeprefix'} && $removeprefix eq "") {
470 $removeprefix = $collectcfg->{'removeprefix'};
471 }
472
473 if (defined $collectcfg->{'removesuffix'} && $removesuffix eq "") {
474 $removesuffix = $collectcfg->{'removesuffix'};
475 }
476 if (defined $collectcfg->{'debug'} && $collectcfg->{'debug'} =~ /^true$/i) {
477 $debug = 1;
478 }
479 if (defined $collectcfg->{'gli'} && $collectcfg->{'gli'} =~ /^true$/i) {
480 $gli = 1;
481 }
482 $gli = 0 unless defined $gli;
483
484 # check keepold and removeold
485 ($removeold, $keepold, $incremental) = &scriptutil::check_removeold_and_keepold($removeold, $keepold, $incremental, "archives", $collectcfg);
486
487
488 print STDERR "<Import>\n" if $gli;
489
490 my $manifest_lookup = new manifest();
491 if ($manifest ne "") {
492 my $manifest_filename = $manifest;
493
494 if ($manifest_filename !~ m/^[\\\/]/) {
495 $manifest_filename = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, $manifest_filename);
496 }
497
498 $manifest =~ s/[\\\/]+/\//g;
499 $manifest =~ s/\/$//;
500
501 $manifest_lookup->parse($manifest_filename);
502 }
503
504
505 # load all the plugins
506 $pluginfo = &plugin::load_plugins ($plugins, $verbosity, $out, $faillog, \@global_opts, $incremental);
507 if (scalar(@$pluginfo) == 0) {
508 &gsprintf($out, "{import.no_plugins_loaded}\n");
509 die "\n";
510 }
511
512 # remove the old contents of the archives directory (and tmp directory) if needed
513 if ($removeold) {
514 if (-e $archivedir) {
515 &gsprintf($out, "{import.removing_archives}\n");
516 &util::rm_r ($archivedir);
517 }
518 my $tmpdir = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "tmp");
519 $tmpdir =~ s/[\\\/]+/\//g;
520 $tmpdir =~ s/\/$//;
521 if (-e $tmpdir) {
522 &gsprintf($out, "{import.removing_tmpdir}\n");
523 &util::rm_r ($tmpdir);
524 }
525 }
526 # create the archives dir if needed
527 &util::mk_all_dir($archivedir);
528
529 # read the archive information file
530## $arcinfo_doc_filename = &util::filename_cat ($archivedir, "archives.inf");
531
532 $arcinfo_doc_filename = &util::filename_cat ($archivedir, "archiveinf-doc");
533 &util::rename_gdbm_file($arcinfo_doc_filename); # ensures gdb
534 $arcinfo_doc_filename .= ".gdb";
535
536 $arcinfo_src_filename = &util::filename_cat ($archivedir, "archiveinf-src");
537 &util::rename_gdbm_file($arcinfo_src_filename); # ensures gdb
538 $arcinfo_src_filename .= ".gdb";
539
540
541 $archive_info = new arcinfo ();
542 $archive_info->load_info ($arcinfo_doc_filename);
543 if ($reversesort) {
544 $archive_info->reverse_sort();
545 }
546
547 if ($manifest eq "") {
548 # Load in list of files in import folder from last import (if present)
549 $archive_info->load_prev_import_filelist ($arcinfo_src_filename);
550 }
551
552 ####Use Plugout####
553 my ($plugout);
554 if (defined $collectcfg->{'plugout'} && $collectcfg->{'plugout'} =~ /^(GreenstoneXML|GreenstoneMETS)Plugout/) {
555 $plugout = $collectcfg->{'plugout'};
556 }
557 else{
558 if ($saveas !~ /^(GreenstoneXML|GreenstoneMETS)$/) {
559 push @$plugout,"GreenstoneXMLPlugout";
560 }
561 else{
562 push @$plugout,$saveas."Plugout";
563 }
564 }
565
566 push @$plugout,("-output_info",$archive_info) if (defined $archive_info);
567 push @$plugout,("-verbosity",$verbosity) if (defined $verbosity);
568 push @$plugout,("-gzip_output") if ($gzip);
569 push @$plugout,("-group_size",$groupsize) if (defined $groupsize);
570 push @$plugout,("-output_handle",$out) if (defined $out);
571 push @$plugout,("-debug") if ($debug);
572
573 $processor = &plugout::load_plugout($plugout);
574 $processor->setoutputdir ($archivedir);
575 $processor->set_sortmeta ($sortmeta, $removeprefix, $removesuffix) if defined $sortmeta;
576 $processor->set_OIDtype ($OIDtype, $OIDmetadata);
577
578 &plugin::begin($pluginfo, $importdir, $processor, $maxdocs, $gli);
579
580 if ($manifest eq "") {
581 # process the import directory
582 my $block_hash = {};
583 my $metadata = {};
584 # gobal blocking pass may set up some metadata
585 &plugin::file_block_read($pluginfo, $importdir, "", $block_hash, $metadata, $gli);
586
587
588 if (!$removeold) {
589
590 &inexport::prime_doc_oid_count($archivedir);
591
592
593 # Can now work out which files were new, already existed, and have
594 # been deleted
595
596 &inexport::new_vs_old_import_diff($archive_info,$block_hash,$importdir);
597
598 my @deleted_files = sort keys %{$block_hash->{'deleted_files'}};
599 if (scalar(@deleted_files>0)) {
600 print STDERR "Files deleted since last import:\n ";
601 print STDERR join("\n ",@deleted_files), "\n";
602 }
603
604 my @new_files = sort keys %{$block_hash->{'new_files'}};
605 if (scalar(@new_files>0)) {
606 print STDERR "New files since last import:\n ";
607 print STDERR join("\n ",@new_files), "\n";
608 }
609
610 &inexport::mark_docs_for_deletion($archive_info,$block_hash,\@deleted_files,
611 $archivedir,$verbosity);
612
613 &inexport::mark_docs_for_reindex($archive_info,$block_hash,
614 $archivedir,$verbosity);
615
616 my @reindex_files = sort keys %{$block_hash->{'reindex_files'}};
617
618 if (scalar(@reindex_files>0)) {
619 print STDERR "Files to reindex since last import:\n ";
620 print STDERR join("\n ",@reindex_files), "\n";
621 }
622
623
624 # not sure if the following will work -- wil the metadata datastructure be correctly initialized
625 # in the right order?
626# foreach my $file (@new_files, @reindex_files) {
627# &plugin::read ($pluginfo, $importdir, $file, $block_hash, $metadata, $processor, $maxdocs, 0, $gli);
628# }
629
630
631 # Play it safe, and run through the entire folder, only processing new or edited files
632 &plugin::read ($pluginfo, $importdir, "", $block_hash, $metadata, $processor, $maxdocs, 0, $gli);
633
634 }
635 else {
636 &plugin::read ($pluginfo, $importdir, "", $block_hash, $metadata, $processor, $maxdocs, 0, $gli);
637 }
638
639 }
640 else
641 {
642 # process any files marked for importing
643 foreach my $file (keys %{$manifest_lookup->{'import'}}) {
644 &plugin::read ($pluginfo, $importdir, $file, {}, {}, $processor, $maxdocs, 0, $gli);
645 }
646
647 my @deleted_files = keys %{$manifest_lookup->{'delete'}};
648
649 &inexport::mark_docs_for_deletion($archive_info,{},\@deleted_files,$archivedir);
650 }
651
652 &plugin::end($pluginfo, $processor);
653
654 &plugin::deinit($pluginfo, $processor);
655
656 # Store the value of OIDCount (used in doc.pm) so it can be
657 # restored correctly to this value on an incremental build
658 &inexport::store_doc_oid_count($archivedir);
659
660 # write out the archive information file
661 $processor->close_file_output() if $groupsize > 1;
662 $processor->close_group_output() if $processor->is_group();
663
664# The following 'if' statement is in the export.pl version of the script,
665# The reason for the 'if' statement is now given in export.pl
666# Unclear at this point if the same should be done here
667## if (($saveas =~ m/^.*METS$/) || ($saveas eq "MARC")) {
668 # Not all export types need this (e.g. DSpace)
669
670 # should we still do this in debug mode??
671
672 # for backwards compatability with archvies.inf file
673 if ($arcinfo_doc_filename =~ m/\.inf$/) {
674 $archive_info->save_info($arcinfo_doc_filename);
675 }
676
677## }
678
679 # write out import stats
680 my $close_stats = 0;
681 if ($statsfile !~ /^(STDERR|STDOUT)$/i) {
682 if (open (STATS, ">$statsfile")) {
683 $statsfile = 'import::STATS';
684 $close_stats = 1;
685 } else {
686 &gsprintf($out, "{import.cannot_open_stats_file}", $statsfile);
687 &gsprintf($out, "{import.stats_backup}\n");
688 $statsfile = 'STDERR';
689 }
690 }
691
692 &gsprintf($out, "\n");
693 &gsprintf($out, "*********************************************\n");
694 &gsprintf($out, "{import.complete}\n");
695 &gsprintf($out, "*********************************************\n");
696
697 &plugin::write_stats($pluginfo, $statsfile, $faillogname, $gli);
698 if ($close_stats) {
699 close STATS;
700 }
701
702 close OUT if $close_out;
703 close FAILLOG;
704}
Note: See TracBrowser for help on using the repository browser.