| 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 |
|
|---|
| 31 | package import;
|
|---|
| 32 |
|
|---|
| 33 | BEGIN {
|
|---|
| 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/classify");
|
|---|
| 40 | }
|
|---|
| 41 |
|
|---|
| 42 | use arcinfo;
|
|---|
| 43 | use colcfg;
|
|---|
| 44 | use plugin;
|
|---|
| 45 | use docprint;
|
|---|
| 46 | use util;
|
|---|
| 47 | use parsargv;
|
|---|
| 48 | use FileHandle;
|
|---|
| 49 | use gsprintf;
|
|---|
| 50 | use printusage;
|
|---|
| 51 |
|
|---|
| 52 | my $oidtype_list =
|
|---|
| 53 | [ { 'name' => "hash",
|
|---|
| 54 | 'desc' => "{import.OIDtype.hash}" },
|
|---|
| 55 | { 'name' => "incremental",
|
|---|
| 56 | 'desc' => "{import.OIDtype.incremental}" } ];
|
|---|
| 57 |
|
|---|
| 58 | # Possible attributes for each argument
|
|---|
| 59 | # name: The name of the argument
|
|---|
| 60 | # desc: A description (or more likely a reference to a description) for this argument
|
|---|
| 61 | # type: The type of control used to represent the argument. Options include: string, int, flag, regexp, metadata, metadatum, language, enum etc
|
|---|
| 62 | # reqd: Is this argument required?
|
|---|
| 63 | # hiddengli: Is this argument hidden in GLI?
|
|---|
| 64 | # modegli: The lowest detail mode this argument is visible at in GLI
|
|---|
| 65 |
|
|---|
| 66 | my $arguments =
|
|---|
| 67 | [ { 'name' => "archivedir",
|
|---|
| 68 | 'desc' => "{import.archivedir}",
|
|---|
| 69 | 'type' => "string",
|
|---|
| 70 | 'reqd' => "no",
|
|---|
| 71 | 'hiddengli' => "yes" },
|
|---|
| 72 | { 'name' => "collectdir",
|
|---|
| 73 | 'desc' => "{import.collectdir}",
|
|---|
| 74 | 'type' => "string",
|
|---|
| 75 | 'deft' => &util::filename_cat ($ENV{'GSDLHOME'}, "collect"),
|
|---|
| 76 | 'reqd' => "no",
|
|---|
| 77 | 'hiddengli' => "yes" },
|
|---|
| 78 | { 'name' => "debug",
|
|---|
| 79 | 'desc' => "{import.debug}",
|
|---|
| 80 | 'type' => "flag",
|
|---|
| 81 | 'reqd' => "no",
|
|---|
| 82 | 'hiddengli' => "yes" },
|
|---|
| 83 | { 'name' => "faillog",
|
|---|
| 84 | 'desc' => "{import.faillog}",
|
|---|
| 85 | 'type' => "string",
|
|---|
| 86 | 'deft' => &util::filename_cat("<collectdir>", "colname", "etc", "fail.log"),
|
|---|
| 87 | 'reqd' => "no",
|
|---|
| 88 | 'modegli' => "4" },
|
|---|
| 89 | { 'name' => "groupsize",
|
|---|
| 90 | 'desc' => "{import.groupsize}",
|
|---|
| 91 | 'type' => "int",
|
|---|
| 92 | 'deft' => "1",
|
|---|
| 93 | 'reqd' => "no",
|
|---|
| 94 | 'modegli' => "3" },
|
|---|
| 95 | { 'name' => "gzip",
|
|---|
| 96 | 'desc' => "{import.gzip}",
|
|---|
| 97 | 'type' => "flag",
|
|---|
| 98 | 'reqd' => "no",
|
|---|
| 99 | 'modegli' => "4" },
|
|---|
| 100 | { 'name' => "importdir",
|
|---|
| 101 | 'desc' => "{import.importdir}",
|
|---|
| 102 | 'type' => "string",
|
|---|
| 103 | 'reqd' => "no",
|
|---|
| 104 | 'hiddengli' => "yes" },
|
|---|
| 105 | { 'name' => "keepold",
|
|---|
| 106 | 'desc' => "{import.keepold}",
|
|---|
| 107 | 'type' => "flag",
|
|---|
| 108 | 'reqd' => "no",
|
|---|
| 109 | 'hiddengli' => "yes" },
|
|---|
| 110 | { 'name' => "language",
|
|---|
| 111 | 'desc' => "{scripts.language}",
|
|---|
| 112 | 'type' => "string",
|
|---|
| 113 | 'reqd' => "no",
|
|---|
| 114 | 'modegli' => "4" },
|
|---|
| 115 | { 'name' => "maxdocs",
|
|---|
| 116 | 'desc' => "{import.maxdocs}",
|
|---|
| 117 | 'type' => "int",
|
|---|
| 118 | 'reqd' => "no",
|
|---|
| 119 | 'range' => "1,",
|
|---|
| 120 | 'modegli' => "1" },
|
|---|
| 121 | { 'name' => "OIDtype",
|
|---|
| 122 | 'desc' => "{import.OIDtype}",
|
|---|
| 123 | 'type' => "enum",
|
|---|
| 124 | 'list' => $oidtype_list,
|
|---|
| 125 | 'deft' => "hash",
|
|---|
| 126 | 'reqd' => "no",
|
|---|
| 127 | 'modegli' => "3" },
|
|---|
| 128 | { 'name' => "out",
|
|---|
| 129 | 'desc' => "{import.out}",
|
|---|
| 130 | 'type' => "string",
|
|---|
| 131 | 'deft' => "STDERR",
|
|---|
| 132 | 'reqd' => "no",
|
|---|
| 133 | 'hiddengli' => "yes" },
|
|---|
| 134 | { 'name' => "removeold",
|
|---|
| 135 | 'desc' => "{import.removeold}",
|
|---|
| 136 | 'type' => "flag",
|
|---|
| 137 | 'reqd' => "no",
|
|---|
| 138 | 'modegli' => "3" },
|
|---|
| 139 | { 'name' => "sortmeta",
|
|---|
| 140 | 'desc' => "{import.sortmeta}",
|
|---|
| 141 | 'type' => "metadata",
|
|---|
| 142 | 'reqd' => "no",
|
|---|
| 143 | 'modegli' => "2" },
|
|---|
| 144 | { 'name' => "statsfile",
|
|---|
| 145 | 'desc' => "{import.statsfile}",
|
|---|
| 146 | 'type' => "string",
|
|---|
| 147 | 'deft' => "STDERR",
|
|---|
| 148 | 'reqd' => "no",
|
|---|
| 149 | 'hiddengli' => "yes" },
|
|---|
| 150 | { 'name' => "verbosity",
|
|---|
| 151 | 'desc' => "{import.verbosity}",
|
|---|
| 152 | 'type' => "int",
|
|---|
| 153 | 'range' => "0,3",
|
|---|
| 154 | 'deft' => "2",
|
|---|
| 155 | 'reqd' => "no",
|
|---|
| 156 | 'modegli' => "4" } ];
|
|---|
| 157 |
|
|---|
| 158 | my $options = { 'name' => "import.pl",
|
|---|
| 159 | 'desc' => "{import.desc}",
|
|---|
| 160 | 'args' => $arguments };
|
|---|
| 161 |
|
|---|
| 162 | sub gsprintf
|
|---|
| 163 | {
|
|---|
| 164 | return &gsprintf::gsprintf(@_);
|
|---|
| 165 | }
|
|---|
| 166 |
|
|---|
| 167 |
|
|---|
| 168 | # sub print_usage {
|
|---|
| 169 | # print STDOUT "\n";
|
|---|
| 170 | # print STDOUT "import.pl: Converts documents in collections -importdir directory into\n";
|
|---|
| 171 | # print STDOUT " xml documents which are written to the -archivedir directory.\n\n";
|
|---|
| 172 | # print STDOUT " usage: $0 [options] collection-name\n\n";
|
|---|
| 173 | # print STDOUT " options:\n";
|
|---|
| 174 | # print STDOUT " -verbosity number 0=none, 3=lots\n";
|
|---|
| 175 | # print STDOUT " -importdir directory Where the original material lives\n";
|
|---|
| 176 | # print STDOUT " -archivedir directory Where the converted material ends up\n";
|
|---|
| 177 | # print STDOUT " -keepold Will not destroy the current contents of the\n";
|
|---|
| 178 | # print STDOUT " archives directory (the default)\n";
|
|---|
| 179 | # print STDOUT " -removeold Will remove the old contents of the archives\n";
|
|---|
| 180 | # print STDOUT " directory -- use with care\n";
|
|---|
| 181 | # print STDOUT " -gzip Use gzip to compress resulting xml documents\n";
|
|---|
| 182 | # print STDOUT " (don't forget to include ZIPPlug in your plugin\n";
|
|---|
| 183 | # print STDOUT " list when building from compressed documents)\n";
|
|---|
| 184 | # print STDOUT " -maxdocs number Maximum number of documents to import\n";
|
|---|
| 185 | # print STDOUT " -groupsize number Number of import documents to group into one XML file\n";
|
|---|
| 186 | # print STDOUT " -OIDtype hash|incremental The method to use when generating unique\n";
|
|---|
| 187 | # print STDOUT " identifiers for each document. \"hash\" (the\n";
|
|---|
| 188 | # print STDOUT " default) hashes the contents of the file and so\n";
|
|---|
| 189 | # print STDOUT " will be the same every time the collection is\n";
|
|---|
| 190 | # print STDOUT " imported. \"incremental\" is a simple document\n";
|
|---|
| 191 | # print STDOUT " count and so will be significantly faster than\n";
|
|---|
| 192 | # print STDOUT " \"hash\". It is not guaranteed to always assign\n";
|
|---|
| 193 | # print STDOUT " the same identifier to a given document though\n";
|
|---|
| 194 | # print STDOUT " and does not allow further documents to be added\n";
|
|---|
| 195 | # print STDOUT " to existing xml archives\n";
|
|---|
| 196 | # print STDOUT " -sortmeta metadata Sort documents alphabetically by metadata for\n";
|
|---|
| 197 | # print STDOUT " building. This will be disabled if groupsize > 1\n";
|
|---|
| 198 | # print STDOUT " -debug Print imported text to STDOUT\n";
|
|---|
| 199 | # print STDOUT " -collectdir directory Collection directory (defaults to " .
|
|---|
| 200 | # &util::filename_cat ($ENV{'GSDLHOME'}, "collect") . ")\n";
|
|---|
| 201 | # print STDOUT " -out name Filename or handle to print output status to.\n";
|
|---|
| 202 | # print STDOUT " The default is STDERR\n";
|
|---|
| 203 | # print STDOUT " -statsfile name Filename or handle to print import statistics to.\n";
|
|---|
| 204 | # print STDOUT " The default is STDERR\n";
|
|---|
| 205 | # print STDOUT " -faillog name Fail log filename. This log receives the filenames\n";
|
|---|
| 206 | # print STDOUT " of any files which fail to be processed (defaults\n";
|
|---|
| 207 | # print STDOUT " to " .
|
|---|
| 208 | # &util::filename_cat("<collectdir>", "colname", "etc", "fail.log") . ")\n";
|
|---|
| 209 | # print STDOUT " [Type \"perl -S import.pl | more\" if this help text scrolled off your screen]";
|
|---|
| 210 | # print STDOUT "\n" unless $ENV{'GSDLOS'} =~ /^windows$/i;
|
|---|
| 211 | # }
|
|---|
| 212 |
|
|---|
| 213 | &main();
|
|---|
| 214 |
|
|---|
| 215 | sub main {
|
|---|
| 216 | my ($verbosity, $importdir, $archivedir, $keepold,
|
|---|
| 217 | $removeold, $gzip, $groupsize, $OIDtype, $debug,
|
|---|
| 218 | $maxdocs, $collection, $configfilename, $collectcfg,
|
|---|
| 219 | $pluginfo, $sortmeta, $archive_info_filename, $statsfile,
|
|---|
| 220 | $archive_info, $processor, $out, $faillog, $collectdir, $gli);
|
|---|
| 221 |
|
|---|
| 222 | # ***** 11-04-03 - John Thompson *****
|
|---|
| 223 | my $xml = 0;
|
|---|
| 224 | # ************************************
|
|---|
| 225 |
|
|---|
| 226 | # note that no defaults are passed for most options as they're set
|
|---|
| 227 | # later (after we check the collect.cfg file)
|
|---|
| 228 | if (!parsargv::parse(\@ARGV,
|
|---|
| 229 | 'language/.*/', \$language,
|
|---|
| 230 | 'verbosity/\d+/', \$verbosity,
|
|---|
| 231 | 'importdir/.*/', \$importdir,
|
|---|
| 232 | 'archivedir/.*/', \$archivedir,
|
|---|
| 233 | 'keepold', \$keepold,
|
|---|
| 234 | 'removeold', \$removeold,
|
|---|
| 235 | 'gzip', \$gzip,
|
|---|
| 236 | 'groupsize/\d+/1', \$groupsize,
|
|---|
| 237 | 'OIDtype/^(hash|incremental)$/', \$OIDtype,
|
|---|
| 238 | 'sortmeta/.*/', \$sortmeta,
|
|---|
| 239 | 'debug', \$debug,
|
|---|
| 240 | 'maxdocs/^\-?\d+/', \$maxdocs,
|
|---|
| 241 | 'collectdir/.*/', \$collectdir,
|
|---|
| 242 | 'out/.*/STDERR', \$out,
|
|---|
| 243 | 'statsfile/.*/STDERR', \$statsfile,
|
|---|
| 244 | 'faillog/.*/', \$faillog,
|
|---|
| 245 | 'gli', \$gli,
|
|---|
| 246 | q^xml^, \$xml)) {
|
|---|
| 247 | &PrintUsage::print_txt_usage($options, "{import.params}");
|
|---|
| 248 | die "\n";
|
|---|
| 249 | }
|
|---|
| 250 |
|
|---|
| 251 | # If $language has been specified, load the appropriate resource bundle
|
|---|
| 252 | # (Otherwise, the default resource bundle will be loaded automatically)
|
|---|
| 253 | if ($language) {
|
|---|
| 254 | &gsprintf::load_language_specific_resource_bundle($language);
|
|---|
| 255 | }
|
|---|
| 256 |
|
|---|
| 257 | if ($xml) {
|
|---|
| 258 | &PrintUsage::print_xml_usage($options);
|
|---|
| 259 | die "\n";
|
|---|
| 260 | }
|
|---|
| 261 |
|
|---|
| 262 | my $close_out = 0;
|
|---|
| 263 | if ($out !~ /^(STDERR|STDOUT)$/i) {
|
|---|
| 264 | open (OUT, ">$out") ||
|
|---|
| 265 | (&gsprintf(STDERR, "{common.cannot_open_output_file}\n", $out) && die);
|
|---|
| 266 | $out = 'import::OUT';
|
|---|
| 267 | $close_out = 1;
|
|---|
| 268 | }
|
|---|
| 269 | $out->autoflush(1);
|
|---|
| 270 |
|
|---|
| 271 | # set removeold to false if it has been defined
|
|---|
| 272 | $removeold = 0 if ($keepold);
|
|---|
| 273 |
|
|---|
| 274 | # get and check the collection name
|
|---|
| 275 | if (($collection = &util::use_collection(@ARGV, $collectdir)) eq "") {
|
|---|
| 276 | &PrintUsage::print_txt_usage($options, "{import.params}");
|
|---|
| 277 | die "\n";
|
|---|
| 278 | }
|
|---|
| 279 |
|
|---|
| 280 | if ($faillog eq "") {
|
|---|
| 281 | $faillog = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "etc", "fail.log");
|
|---|
| 282 | }
|
|---|
| 283 | open (FAILLOG, ">$faillog") ||
|
|---|
| 284 | (&gsprintf(STDERR, "{import.cannot_open_fail_log}\n", $faillog) && die);
|
|---|
| 285 | my $faillogname = $faillog;
|
|---|
| 286 | $faillog = 'import::FAILLOG';
|
|---|
| 287 | $faillog->autoflush(1);
|
|---|
| 288 |
|
|---|
| 289 | # check sortmeta
|
|---|
| 290 | $sortmeta = undef unless defined $sortmeta && $sortmeta =~ /\S/;
|
|---|
| 291 | if (defined $sortmeta && $groupsize > 1) {
|
|---|
| 292 | &gsprintf($out, "{import.cannot_sort}\n\n");
|
|---|
| 293 | $sortmeta = undef;
|
|---|
| 294 | }
|
|---|
| 295 |
|
|---|
| 296 | # dynamically load 'docsave' module so it can pick up on a collection
|
|---|
| 297 | # specific docsave.pm is specified.
|
|---|
| 298 |
|
|---|
| 299 | unshift (@INC, "$ENV{'GSDLCOLLECTDIR'}/perllib");
|
|---|
| 300 | require docsave;
|
|---|
| 301 |
|
|---|
| 302 |
|
|---|
| 303 | # get the list of plugins for this collection and set any options that
|
|---|
| 304 | # were specified in the collect.cfg (all import.pl options except
|
|---|
| 305 | # -collectdir, -out and -faillog may be specified in the collect.cfg (these
|
|---|
| 306 | # options must be known before we read the collect.cfg))
|
|---|
| 307 | my $plugins = [];
|
|---|
| 308 | my @global_opts = ();
|
|---|
| 309 |
|
|---|
| 310 | $configfilename = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "etc", "collect.cfg");
|
|---|
| 311 | if (-e $configfilename) {
|
|---|
| 312 | $collectcfg = &colcfg::read_collect_cfg ($configfilename);
|
|---|
| 313 | if (defined $collectcfg->{'plugin'}) {
|
|---|
| 314 | $plugins = $collectcfg->{'plugin'};
|
|---|
| 315 | }
|
|---|
| 316 |
|
|---|
| 317 | if ($verbosity !~ /\d+/) {
|
|---|
| 318 | if (defined $collectcfg->{'verbosity'} && $collectcfg->{'verbosity'} =~ /\d+/) {
|
|---|
| 319 | $verbosity = $collectcfg->{'verbosity'};
|
|---|
| 320 | } else {
|
|---|
| 321 | $verbosity = 2; # the default
|
|---|
| 322 | }
|
|---|
| 323 | }
|
|---|
| 324 | if (defined $collectcfg->{'importdir'} && $importdir eq "") {
|
|---|
| 325 | $importdir = $collectcfg->{'importdir'};
|
|---|
| 326 | }
|
|---|
| 327 | if (defined $collectcfg->{'archivedir'} && $archivedir eq "") {
|
|---|
| 328 | $archivedir = $collectcfg->{'archivedir'};
|
|---|
| 329 | }
|
|---|
| 330 | if (defined $collectcfg->{'removeold'}) {
|
|---|
| 331 | if ($collectcfg->{'removeold'} =~ /^true$/i && !$keepold) {
|
|---|
| 332 | $removeold = 1;
|
|---|
| 333 | }
|
|---|
| 334 | if ($collectcfg->{'removeold'} =~ /^false$/i && !$removeold) {
|
|---|
| 335 | $removeold = 0;
|
|---|
| 336 | }
|
|---|
| 337 | }
|
|---|
| 338 | if (defined $collectcfg->{'keepold'}) {
|
|---|
| 339 | if ($collectcfg->{'keepold'} =~ /^false$/i && !$keepold) {
|
|---|
| 340 | $removeold = 1;
|
|---|
| 341 | }
|
|---|
| 342 | }
|
|---|
| 343 | if (defined $collectcfg->{'gzip'} && !$gzip) {
|
|---|
| 344 | if ($collectcfg->{'gzip'} =~ /^true$/i) {
|
|---|
| 345 | $gzip = 1;
|
|---|
| 346 | }
|
|---|
| 347 | }
|
|---|
| 348 | if ($maxdocs !~ /\-?\d+/) {
|
|---|
| 349 | if (defined $collectcfg->{'maxdocs'} && $collectcfg->{'maxdocs'} =~ /\-?\d+/) {
|
|---|
| 350 | $maxdocs = $collectcfg->{'maxdocs'};
|
|---|
| 351 | } else {
|
|---|
| 352 | $maxdocs = -1; # the default
|
|---|
| 353 | }
|
|---|
| 354 | }
|
|---|
| 355 | if ($groupsize == 1) {
|
|---|
| 356 | if (defined $collectcfg->{'groupsize'} && $collectcfg->{'groupsize'} =~ /\d+/) {
|
|---|
| 357 | $groupsize = $collectcfg->{'groupsize'};
|
|---|
| 358 | }
|
|---|
| 359 | }
|
|---|
| 360 | if ($OIDtype !~ /^(hash|incremental)$/) {
|
|---|
| 361 | if (defined $collectcfg->{'OIDtype'} && $collectcfg->{'OIDtype'} =~ /^(hash|incremental)$/) {
|
|---|
| 362 | $OIDtype = $collectcfg->{'OIDtype'};
|
|---|
| 363 | } else {
|
|---|
| 364 | $OIDtype = "hash"; # the default
|
|---|
| 365 | }
|
|---|
| 366 | }
|
|---|
| 367 | if (defined $collectcfg->{'sortmeta'} && $sortmeta eq "") {
|
|---|
| 368 | $sortmeta = $collectcfg->{'sortmeta'};
|
|---|
| 369 | }
|
|---|
| 370 | if (defined $collectcfg->{'debug'} && $collectcfg->{'debug'} =~ /^true$/i) {
|
|---|
| 371 | $debug = 1;
|
|---|
| 372 | }
|
|---|
| 373 | if (defined $collectcfg->{'gli'} && $collectcfg->{'gli'} =~ /^true$/i) {
|
|---|
| 374 | $gli = 1;
|
|---|
| 375 | }
|
|---|
| 376 |
|
|---|
| 377 | # global plugin stuff
|
|---|
| 378 | if (defined $collectcfg->{'separate_cjk'}&& $collectcfg->{'separate_cjk'} =~ /^true$/i) {
|
|---|
| 379 | push @global_opts, "-separate_cjk";
|
|---|
| 380 | }
|
|---|
| 381 |
|
|---|
| 382 |
|
|---|
| 383 | } else {
|
|---|
| 384 | (&gsprintf($out, "{common.cannot_find_cfg_file}\n", $configfilename) && die);
|
|---|
| 385 | }
|
|---|
| 386 |
|
|---|
| 387 | $gli = 0 unless defined $gli;
|
|---|
| 388 |
|
|---|
| 389 | print STDERR "<Import>\n" if $gli;
|
|---|
| 390 |
|
|---|
| 391 | # fill in the default import and archives directories if none
|
|---|
| 392 | # were supplied, turn all \ into / and remove trailing /
|
|---|
| 393 | $importdir = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "import") if $importdir eq "";
|
|---|
| 394 | $importdir =~ s/[\\\/]+/\//g;
|
|---|
| 395 | $importdir =~ s/\/$//;
|
|---|
| 396 | $archivedir = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "archives") if $archivedir eq "";
|
|---|
| 397 | $archivedir =~ s/[\\\/]+/\//g;
|
|---|
| 398 | $archivedir =~ s/\/$//;
|
|---|
| 399 |
|
|---|
| 400 | # load all the plugins
|
|---|
| 401 | $pluginfo = &plugin::load_plugins ($plugins, $verbosity, $out, $faillog, \@global_opts);
|
|---|
| 402 | if (scalar(@$pluginfo) == 0) {
|
|---|
| 403 | &gsprintf($out, "{import.no_plugins_loaded}\n");
|
|---|
| 404 | die "\n";
|
|---|
| 405 | }
|
|---|
| 406 |
|
|---|
| 407 | # remove the old contents of the archives directory if needed
|
|---|
| 408 | if ($removeold && -e $archivedir) {
|
|---|
| 409 | &gsprintf($out, "{import.removing_archives}\n");
|
|---|
| 410 | sleep(3); # just in case...
|
|---|
| 411 | &util::rm_r ($archivedir);
|
|---|
| 412 | }
|
|---|
| 413 |
|
|---|
| 414 | # read the archive information file
|
|---|
| 415 | if (!$debug) {
|
|---|
| 416 | $archive_info_filename = &util::filename_cat ($archivedir, "archives.inf");
|
|---|
| 417 | $archive_info = new arcinfo ();
|
|---|
| 418 | $archive_info->load_info ($archive_info_filename);
|
|---|
| 419 |
|
|---|
| 420 | # create a docsave object to process the documents
|
|---|
| 421 | $processor = new docsave ($collection, $archive_info, $verbosity, $gzip, $groupsize, $out);
|
|---|
| 422 | $processor->setarchivedir ($archivedir);
|
|---|
| 423 | $processor->set_sortmeta ($sortmeta) if defined $sortmeta;
|
|---|
| 424 | $processor->set_OIDtype ($OIDtype);
|
|---|
| 425 | } else {
|
|---|
| 426 | $processor = new docprint ();
|
|---|
| 427 | }
|
|---|
| 428 |
|
|---|
| 429 | &plugin::begin($pluginfo, $importdir, $processor, $maxdocs);
|
|---|
| 430 |
|
|---|
| 431 | # process the import directory
|
|---|
| 432 | &plugin::read ($pluginfo, $importdir, "", {}, $processor, $maxdocs, $gli);
|
|---|
| 433 |
|
|---|
| 434 | &plugin::end($pluginfo, $processor);
|
|---|
| 435 |
|
|---|
| 436 | # write out the archive information file
|
|---|
| 437 | if (!$debug) {
|
|---|
| 438 | $processor->close_file_output() if $groupsize > 1;
|
|---|
| 439 | $archive_info->save_info($archive_info_filename);
|
|---|
| 440 | }
|
|---|
| 441 |
|
|---|
| 442 | # write out import stats
|
|---|
| 443 | my $close_stats = 0;
|
|---|
| 444 | if ($statsfile !~ /^(STDERR|STDOUT)$/i) {
|
|---|
| 445 | if (open (STATS, ">$statsfile")) {
|
|---|
| 446 | $statsfile = 'import::STATS';
|
|---|
| 447 | $close_stats = 1;
|
|---|
| 448 | } else {
|
|---|
| 449 | &gsprintf($out, "{import.cannot_open_stats_file}", $statsfile);
|
|---|
| 450 | &gsprintf($out, "{import.stats_backup}\n");
|
|---|
| 451 | $statsfile = 'STDERR';
|
|---|
| 452 | }
|
|---|
| 453 | }
|
|---|
| 454 |
|
|---|
| 455 | &gsprintf($out, "\n");
|
|---|
| 456 | &gsprintf($out, "*********************************************\n");
|
|---|
| 457 | &gsprintf($out, "{import.complete}\n");
|
|---|
| 458 | &gsprintf($out, "*********************************************\n");
|
|---|
| 459 |
|
|---|
| 460 | &plugin::write_stats($pluginfo, $statsfile, $faillogname, $gli);
|
|---|
| 461 | if ($close_stats) {
|
|---|
| 462 | close STATS;
|
|---|
| 463 | }
|
|---|
| 464 |
|
|---|
| 465 | close OUT if $close_out;
|
|---|
| 466 | close FAILLOG;
|
|---|
| 467 | }
|
|---|
