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' => $arguments };
|
---|
161 |
|
---|
162 | sub gsprintf
|
---|
163 | {
|
---|
164 | return &gsprintf::gsprintf(@_);
|
---|
165 | }
|
---|
166 |
|
---|
167 |
|
---|
168 | &main();
|
---|
169 |
|
---|
170 | sub main {
|
---|
171 | my ($verbosity, $importdir, $archivedir, $keepold,
|
---|
172 | $removeold, $gzip, $groupsize, $OIDtype, $debug,
|
---|
173 | $maxdocs, $collection, $configfilename, $collectcfg,
|
---|
174 | $pluginfo, $sortmeta, $archive_info_filename, $statsfile,
|
---|
175 | $archive_info, $processor, $out, $faillog, $collectdir, $gli);
|
---|
176 |
|
---|
177 | # ***** 11-04-03 - John Thompson *****
|
---|
178 | my $xml = 0;
|
---|
179 | # ************************************
|
---|
180 |
|
---|
181 | # note that no defaults are passed for most options as they're set
|
---|
182 | # later (after we check the collect.cfg file)
|
---|
183 | if (!parsargv::parse(\@ARGV,
|
---|
184 | 'language/.*/', \$language,
|
---|
185 | 'verbosity/\d+/', \$verbosity,
|
---|
186 | 'importdir/.*/', \$importdir,
|
---|
187 | 'archivedir/.*/', \$archivedir,
|
---|
188 | 'keepold', \$keepold,
|
---|
189 | 'removeold', \$removeold,
|
---|
190 | 'gzip', \$gzip,
|
---|
191 | 'groupsize/\d+/1', \$groupsize,
|
---|
192 | 'OIDtype/^(hash|incremental)$/', \$OIDtype,
|
---|
193 | 'sortmeta/.*/', \$sortmeta,
|
---|
194 | 'debug', \$debug,
|
---|
195 | 'maxdocs/^\-?\d+/', \$maxdocs,
|
---|
196 | 'collectdir/.*/', \$collectdir,
|
---|
197 | 'out/.*/STDERR', \$out,
|
---|
198 | 'statsfile/.*/STDERR', \$statsfile,
|
---|
199 | 'faillog/.*/', \$faillog,
|
---|
200 | 'gli', \$gli,
|
---|
201 | q^xml^, \$xml)) {
|
---|
202 | &PrintUsage::print_txt_usage($options, "{import.params}");
|
---|
203 | die "\n";
|
---|
204 | }
|
---|
205 |
|
---|
206 | # If $language has been specified, load the appropriate resource bundle
|
---|
207 | # (Otherwise, the default resource bundle will be loaded automatically)
|
---|
208 | if ($language) {
|
---|
209 | &gsprintf::load_language_specific_resource_bundle($language);
|
---|
210 | }
|
---|
211 |
|
---|
212 | if ($xml) {
|
---|
213 | &PrintUsage::print_xml_usage($options);
|
---|
214 | die "\n";
|
---|
215 | }
|
---|
216 |
|
---|
217 | if ($gli) { # the gli wants strings to be in UTF-8
|
---|
218 | &gsprintf::output_strings_in_UTF8;
|
---|
219 | }
|
---|
220 | my $close_out = 0;
|
---|
221 | if ($out !~ /^(STDERR|STDOUT)$/i) {
|
---|
222 | open (OUT, ">$out") ||
|
---|
223 | (&gsprintf(STDERR, "{common.cannot_open_output_file}\n", $out) && die);
|
---|
224 | $out = 'import::OUT';
|
---|
225 | $close_out = 1;
|
---|
226 | }
|
---|
227 | $out->autoflush(1);
|
---|
228 |
|
---|
229 | # set removeold to false if it has been defined
|
---|
230 | $removeold = 0 if ($keepold);
|
---|
231 |
|
---|
232 | # get and check the collection name
|
---|
233 | if (($collection = &util::use_collection(@ARGV, $collectdir)) eq "") {
|
---|
234 | &PrintUsage::print_txt_usage($options, "{import.params}");
|
---|
235 | die "\n";
|
---|
236 | }
|
---|
237 |
|
---|
238 | if ($faillog eq "") {
|
---|
239 | $faillog = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "etc", "fail.log");
|
---|
240 | }
|
---|
241 | open (FAILLOG, ">$faillog") ||
|
---|
242 | (&gsprintf(STDERR, "{import.cannot_open_fail_log}\n", $faillog) && die);
|
---|
243 | my $faillogname = $faillog;
|
---|
244 | $faillog = 'import::FAILLOG';
|
---|
245 | $faillog->autoflush(1);
|
---|
246 |
|
---|
247 | # check sortmeta
|
---|
248 | $sortmeta = undef unless defined $sortmeta && $sortmeta =~ /\S/;
|
---|
249 | if (defined $sortmeta && $groupsize > 1) {
|
---|
250 | &gsprintf($out, "{import.cannot_sort}\n\n");
|
---|
251 | $sortmeta = undef;
|
---|
252 | }
|
---|
253 |
|
---|
254 | # dynamically load 'docsave' module so it can pick up on a collection
|
---|
255 | # specific docsave.pm is specified.
|
---|
256 |
|
---|
257 | unshift (@INC, "$ENV{'GSDLCOLLECTDIR'}/perllib");
|
---|
258 | require docsave;
|
---|
259 |
|
---|
260 |
|
---|
261 | # get the list of plugins for this collection and set any options that
|
---|
262 | # were specified in the collect.cfg (all import.pl options except
|
---|
263 | # -collectdir, -out and -faillog may be specified in the collect.cfg (these
|
---|
264 | # options must be known before we read the collect.cfg))
|
---|
265 | my $plugins = [];
|
---|
266 | my @global_opts = ();
|
---|
267 |
|
---|
268 | $configfilename = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "etc", "collect.cfg");
|
---|
269 | if (-e $configfilename) {
|
---|
270 | $collectcfg = &colcfg::read_collect_cfg ($configfilename);
|
---|
271 | if (defined $collectcfg->{'plugin'}) {
|
---|
272 | $plugins = $collectcfg->{'plugin'};
|
---|
273 | }
|
---|
274 |
|
---|
275 | if ($verbosity !~ /\d+/) {
|
---|
276 | if (defined $collectcfg->{'verbosity'} && $collectcfg->{'verbosity'} =~ /\d+/) {
|
---|
277 | $verbosity = $collectcfg->{'verbosity'};
|
---|
278 | } else {
|
---|
279 | $verbosity = 2; # the default
|
---|
280 | }
|
---|
281 | }
|
---|
282 | if (defined $collectcfg->{'importdir'} && $importdir eq "") {
|
---|
283 | $importdir = $collectcfg->{'importdir'};
|
---|
284 | }
|
---|
285 | if (defined $collectcfg->{'archivedir'} && $archivedir eq "") {
|
---|
286 | $archivedir = $collectcfg->{'archivedir'};
|
---|
287 | }
|
---|
288 | if (defined $collectcfg->{'removeold'}) {
|
---|
289 | if ($collectcfg->{'removeold'} =~ /^true$/i && !$keepold) {
|
---|
290 | $removeold = 1;
|
---|
291 | }
|
---|
292 | if ($collectcfg->{'removeold'} =~ /^false$/i && !$removeold) {
|
---|
293 | $removeold = 0;
|
---|
294 | }
|
---|
295 | }
|
---|
296 | if (defined $collectcfg->{'keepold'}) {
|
---|
297 | if ($collectcfg->{'keepold'} =~ /^false$/i && !$keepold) {
|
---|
298 | $removeold = 1;
|
---|
299 | }
|
---|
300 | }
|
---|
301 | if (defined $collectcfg->{'gzip'} && !$gzip) {
|
---|
302 | if ($collectcfg->{'gzip'} =~ /^true$/i) {
|
---|
303 | $gzip = 1;
|
---|
304 | }
|
---|
305 | }
|
---|
306 | if ($maxdocs !~ /\-?\d+/) {
|
---|
307 | if (defined $collectcfg->{'maxdocs'} && $collectcfg->{'maxdocs'} =~ /\-?\d+/) {
|
---|
308 | $maxdocs = $collectcfg->{'maxdocs'};
|
---|
309 | } else {
|
---|
310 | $maxdocs = -1; # the default
|
---|
311 | }
|
---|
312 | }
|
---|
313 | if ($groupsize == 1) {
|
---|
314 | if (defined $collectcfg->{'groupsize'} && $collectcfg->{'groupsize'} =~ /\d+/) {
|
---|
315 | $groupsize = $collectcfg->{'groupsize'};
|
---|
316 | }
|
---|
317 | }
|
---|
318 | if ($OIDtype !~ /^(hash|incremental)$/) {
|
---|
319 | if (defined $collectcfg->{'OIDtype'} && $collectcfg->{'OIDtype'} =~ /^(hash|incremental)$/) {
|
---|
320 | $OIDtype = $collectcfg->{'OIDtype'};
|
---|
321 | } else {
|
---|
322 | $OIDtype = "hash"; # the default
|
---|
323 | }
|
---|
324 | }
|
---|
325 | if (defined $collectcfg->{'sortmeta'} && $sortmeta eq "") {
|
---|
326 | $sortmeta = $collectcfg->{'sortmeta'};
|
---|
327 | }
|
---|
328 | if (defined $collectcfg->{'debug'} && $collectcfg->{'debug'} =~ /^true$/i) {
|
---|
329 | $debug = 1;
|
---|
330 | }
|
---|
331 | if (defined $collectcfg->{'gli'} && $collectcfg->{'gli'} =~ /^true$/i) {
|
---|
332 | $gli = 1;
|
---|
333 | }
|
---|
334 |
|
---|
335 | # global plugin stuff
|
---|
336 | if (defined $collectcfg->{'separate_cjk'}&& $collectcfg->{'separate_cjk'} =~ /^true$/i) {
|
---|
337 | push @global_opts, "-separate_cjk";
|
---|
338 | }
|
---|
339 |
|
---|
340 |
|
---|
341 | } else {
|
---|
342 | (&gsprintf($out, "{common.cannot_find_cfg_file}\n", $configfilename) && die);
|
---|
343 | }
|
---|
344 |
|
---|
345 | $gli = 0 unless defined $gli;
|
---|
346 |
|
---|
347 | print STDERR "<Import>\n" if $gli;
|
---|
348 |
|
---|
349 | # fill in the default import and archives directories if none
|
---|
350 | # were supplied, turn all \ into / and remove trailing /
|
---|
351 | $importdir = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "import") if $importdir eq "";
|
---|
352 | $importdir =~ s/[\\\/]+/\//g;
|
---|
353 | $importdir =~ s/\/$//;
|
---|
354 | $archivedir = &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, "archives") if $archivedir eq "";
|
---|
355 | $archivedir =~ s/[\\\/]+/\//g;
|
---|
356 | $archivedir =~ s/\/$//;
|
---|
357 |
|
---|
358 | # load all the plugins
|
---|
359 | $pluginfo = &plugin::load_plugins ($plugins, $verbosity, $out, $faillog, \@global_opts);
|
---|
360 | if (scalar(@$pluginfo) == 0) {
|
---|
361 | &gsprintf($out, "{import.no_plugins_loaded}\n");
|
---|
362 | die "\n";
|
---|
363 | }
|
---|
364 |
|
---|
365 | # remove the old contents of the archives directory if needed
|
---|
366 | if ($removeold && -e $archivedir) {
|
---|
367 | &gsprintf($out, "{import.removing_archives}\n");
|
---|
368 | sleep(3); # just in case...
|
---|
369 | &util::rm_r ($archivedir);
|
---|
370 | }
|
---|
371 |
|
---|
372 | # read the archive information file
|
---|
373 | if (!$debug) {
|
---|
374 | $archive_info_filename = &util::filename_cat ($archivedir, "archives.inf");
|
---|
375 | $archive_info = new arcinfo ();
|
---|
376 | $archive_info->load_info ($archive_info_filename);
|
---|
377 |
|
---|
378 | # create a docsave object to process the documents
|
---|
379 | $processor = new docsave ($collection, $archive_info, $verbosity, $gzip, $groupsize, $out);
|
---|
380 | $processor->setarchivedir ($archivedir);
|
---|
381 | $processor->set_sortmeta ($sortmeta) if defined $sortmeta;
|
---|
382 | $processor->set_OIDtype ($OIDtype);
|
---|
383 | } else {
|
---|
384 | $processor = new docprint ();
|
---|
385 | }
|
---|
386 |
|
---|
387 | &plugin::begin($pluginfo, $importdir, $processor, $maxdocs);
|
---|
388 |
|
---|
389 | # process the import directory
|
---|
390 | &plugin::read ($pluginfo, $importdir, "", {}, $processor, $maxdocs, $gli);
|
---|
391 |
|
---|
392 | &plugin::end($pluginfo, $processor);
|
---|
393 |
|
---|
394 | # write out the archive information file
|
---|
395 | if (!$debug) {
|
---|
396 | $processor->close_file_output() if $groupsize > 1;
|
---|
397 | $archive_info->save_info($archive_info_filename);
|
---|
398 | }
|
---|
399 |
|
---|
400 | # write out import stats
|
---|
401 | my $close_stats = 0;
|
---|
402 | if ($statsfile !~ /^(STDERR|STDOUT)$/i) {
|
---|
403 | if (open (STATS, ">$statsfile")) {
|
---|
404 | $statsfile = 'import::STATS';
|
---|
405 | $close_stats = 1;
|
---|
406 | } else {
|
---|
407 | &gsprintf($out, "{import.cannot_open_stats_file}", $statsfile);
|
---|
408 | &gsprintf($out, "{import.stats_backup}\n");
|
---|
409 | $statsfile = 'STDERR';
|
---|
410 | }
|
---|
411 | }
|
---|
412 |
|
---|
413 | &gsprintf($out, "\n");
|
---|
414 | &gsprintf($out, "*********************************************\n");
|
---|
415 | &gsprintf($out, "{import.complete}\n");
|
---|
416 | &gsprintf($out, "*********************************************\n");
|
---|
417 |
|
---|
418 | &plugin::write_stats($pluginfo, $statsfile, $faillogname, $gli);
|
---|
419 | if ($close_stats) {
|
---|
420 | close STATS;
|
---|
421 | }
|
---|
422 |
|
---|
423 | close OUT if $close_out;
|
---|
424 | close FAILLOG;
|
---|
425 | }
|
---|