source: gsdl/trunk/bin/script/explode_metadata_database.pl@ 20822

Last change on this file since 20822 was 20822, checked in by kjdon, 12 years ago

added collectdir and site options so can be used with a -collection in greenstone3

  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
File size: 19.0 KB
Line 
1#!/usr/bin/perl -w
2
3
4BEGIN {
5 die "GSDLHOME not set\n" unless defined $ENV{'GSDLHOME'};
6 unshift (@INC, "$ENV{'GSDLHOME'}/perllib");
7 unshift (@INC, "$ENV{'GSDLHOME'}/perllib/plugins");
8}
9
10use strict;
11no strict 'subs'; # allow barewords (eg STDERR) as function arguments
12no strict 'refs'; # allow filehandles to be variables and vice versa
13
14use encodings;
15use printusage;
16use parse2;
17use colcfg;
18
19use FileHandle;
20
21use File::Spec;
22use File::Basename;
23
24my $unicode_list =
25 [ { 'name' => "auto",
26 'desc' => "{ReadTextFile.input_encoding.auto}" },
27 { 'name' => "ascii",
28 'desc' => "{BasePlugin.encoding.ascii}" },
29 { 'name' => "utf8",
30 'desc' => "{BasePlugin.encoding.utf8}" },
31 { 'name' => "unicode",
32 'desc' => "{BasePlugin.encoding.unicode}" } ];
33
34my $e = $encodings::encodings;
35foreach my $enc (sort {$e->{$a}->{'name'} cmp $e->{$b}->{'name'}} keys (%$e))
36{
37 my $hashEncode =
38 {'name' => $enc,
39 'desc' => $e->{$enc}->{'name'}};
40
41 push(@{$unicode_list},$hashEncode);
42}
43
44my $arguments =
45 [
46 { 'name' => "language",
47 'desc' => "{scripts.language}",
48 'type' => "string",
49 'reqd' => "no",
50 'hiddengli' => "yes" },
51 { 'name' => "plugin",
52 'desc' => "{explode.plugin}",
53 'type' => "string",
54 'reqd' => "yes",
55 'hiddengli' => "yes"},
56 { 'name' => "input_encoding",
57 'desc' => "{explode.encoding}",
58 'type' => "enum",
59 'deft' => "auto",
60 'list' => $unicode_list,
61 'reqd' => "no" },
62 { 'name' => "metadata_set",
63 'desc' => "{explode.metadata_set}",
64 'type' => "string",
65 'reqd' => "no" },
66 { 'name' => "document_field",
67 'desc' => "{explode.document_field}",
68 'type' => "string",
69 'reqd' => "no"},
70 { 'name' => "document_prefix",
71 'desc' => "{explode.document_prefix}",
72 'type' => "string",
73 'reqd' => "no"},
74 { 'name' => "document_suffix",
75 'desc' => "{explode.document_suffix}",
76 'type' => "string",
77 'reqd' => "no"},
78 { 'name' => "records_per_folder",
79 'desc' => "{explode.records_per_folder}",
80 'type' => "int",
81 'range' => "0,",
82 'deft' => "100",
83 'reqd' => "no" },
84 { 'name' => "collectdir",
85 'desc' => "{import.collectdir}",
86 'type' => "string",
87 # parsearg left "" as default
88 #'deft' => &util::filename_cat ($ENV{'GSDLHOME'}, "collect"),
89 'deft' => "",
90 'reqd' => "no",
91 'hiddengli' => "yes" },
92 { 'name' => "site",
93 'desc' => "{import.site}",
94 'type' => "string",
95 'deft' => "",
96 'reqd' => "no",
97 'hiddengli' => "yes" },
98 { 'name' => "collection",
99 'desc' => "{explode.collection}",
100 'type' => "string",
101 'reqd' => "no",
102 'hiddengli' => "yes"},
103 { 'name' => "use_collection_plugin_options",
104 'desc' => "{explode.use_collection_plugin_options}",
105 'type' => "flag",
106 'reqd' => "no",
107 'hiddengli' => "yes"},
108 { 'name' => "plugin_options",
109 'desc' => "{explode.plugin_options}",
110 'type' => "string",
111 'reqd' => "no",
112 'hiddengli' => "yes"},
113 { 'name' => "verbosity",
114 'desc' => "{import.verbosity}",
115 'type' => "int",
116 'range' => "0,",
117 'deft' => "1",
118 'reqd' => "no",
119 'modegli' => "3" },
120 { 'name' => "xml",
121 'desc' => "",
122 'type' => "flag",
123 'reqd' => "no",
124 'hiddengli' => "yes" }
125 ];
126
127my $options = { 'name' => "explode_metadata_database.pl",
128 'desc' => "{explode.desc}",
129 'args' => $arguments };
130
131
132
133sub main
134{
135 my ($language, $input_encoding, $metadata_set, $plugin,
136 $document_field, $document_prefix, $document_suffix,
137 $records_per_folder, $plugin_options, $collectdir, $site, $collection,
138 $use_collection_plugin_options, $verbosity);
139
140 my $xml = 0;
141
142 my $hashParsingResult = {};
143 # parse the options
144 my $intArgLeftinAfterParsing = parse2::parse(\@ARGV,$arguments,$hashParsingResult,"allow_extra_options");
145
146 # If parse returns -1 then something has gone wrong
147 if ($intArgLeftinAfterParsing == -1)
148 {
149 &PrintUsage::print_txt_usage($options, "{explode.params}");
150 die "\n";
151 }
152
153 foreach my $strVariable (keys %$hashParsingResult)
154 {
155 eval "\$$strVariable = \$hashParsingResult->{\"\$strVariable\"}";
156 }
157
158 # If $language has been specified, load the appropriate resource bundle
159 # (Otherwise, the default resource bundle will be loaded automatically)
160 if ($language && $language =~ /\S/) {
161 &gsprintf::load_language_specific_resource_bundle($language);
162 }
163
164 if ($xml) {
165 &PrintUsage::print_xml_usage($options);
166 print "\n";
167 return;
168 }
169
170
171 # There should one arg left after parsing (the filename)
172 # Or the user may have specified -h, in which case we output the usage
173 if($intArgLeftinAfterParsing != 1 || (@ARGV && $ARGV[0] =~ /^\-+h/))
174 {
175 &PrintUsage::print_txt_usage($options, "{explode.params}");
176 die "\n";
177 }
178
179 # The metadata database filename is the first value that remains after the options have been parsed out
180 my $filename = $ARGV[0];
181 if (!defined $filename || $filename !~ /\w/) {
182 &PrintUsage::print_txt_usage($options, "{explode.params}");
183 print STDERR "You need to specify a filename";
184 die "\n";
185 }
186 # check that file exists
187 if (!-e $filename) {
188 print STDERR "File $filename doesn't exist...\n";
189 die "\n";
190 }
191 # check required options
192 if (!defined $plugin || $plugin !~ /\w/) {
193 &PrintUsage::print_txt_usage($options, "{explode.params}");
194 print STDERR "You need to specify a plugin";
195 die "\n";
196 }
197
198 # check metadata set
199 if (defined $metadata_set && $metadata_set =~ /\w/) {
200 $metadata_set .= ".";
201 } else {
202 $metadata_set = "";
203 }
204 if (defined $collection && $collection =~ /\w/) {
205 if (($collection = &colcfg::use_collection($site, $collection, $collectdir)) eq "") {
206 print STDERR "Collection $collection does not exist\n";
207 die "\n";
208 }
209 } else {
210 undef $collection;
211 }
212
213 if ($use_collection_plugin_options) {
214 if (defined $plugin_options && $plugin_options =~ /\w/) {
215 print STDERR "Error: you cannot have -use_collection_plugin_options and -plugin_options set at the same time\n";
216 die "\n";
217 }
218 if (not defined $collection) {
219 print STDERR "Error: you must specify a collection using -collection to use -use_collection_plugin_options\n";
220 die "\n";
221 }
222 }
223 my $plugobj;
224 require "$plugin.pm";
225
226 my $plugin_options_string = "";
227 if ($use_collection_plugin_options) {
228 # read in the collect.cfg file
229 # Read in the collection configuration file.
230 my ($configfilename, $gs_mode) = &colcfg::get_collect_cfg_name(STDERR);
231 my $collectcfg = &colcfg::read_collect_cfg ($configfilename, $gs_mode);
232 $plugin_options_string = &get_plugin_options($collectcfg, $plugin);
233 }
234 elsif (defined $plugin_options && $plugin_options =~ /\w/) {
235 my @options = split(/\s/, $plugin_options);
236 map { $_ = "\"$_\"" unless $_ =~ /^\"/; } @options;
237 $plugin_options_string= join (",", @options);
238 }
239
240 if ($plugin_options_string eq "") {
241 eval ("\$plugobj = new $plugin()");
242 die "$@" if $@;
243 } else {
244 eval ("\$plugobj = new $plugin([], [$plugin_options_string])");
245 die "$@" if $@;
246 }
247
248 # ...and initialize it
249 $plugobj->init($verbosity, "STDERR", "STDERR");
250
251 if ($input_encoding eq "auto") {
252 ($language, $input_encoding) = $plugobj->textcat_get_language_encoding ($filename);
253 }
254
255 # Create a directory to store the document files...
256 my ($exploded_base_dir) = ($filename =~ /(.*)\.[^\.]+$/);
257
258 my $orig_base_dir = &File::Basename::dirname($filename);
259
260
261 my $split_exp = $plugobj->{'split_exp'};
262 if (defined $split_exp) {
263 # Read in file, and then split and process individual records
264
265 my $text = "";
266 # Use the plugin's read_file function to avoid duplicating code
267 $plugobj->read_file($filename, $input_encoding, undef, \$text);
268 # is there any text in the file??
269 die "\n" unless length($text);
270
271 # Split the text into records, using the plugin's split_exp
272
273 my @metadata_records = split(/$split_exp/, $text);
274 my $total_num_records = scalar(@metadata_records);
275 print STDERR "Number of records: $total_num_records\n";
276
277 # Write the metadata from each record to the metadata.xml file
278 my $record_number = 1;
279 my $documents_directory;
280 foreach my $record_text (@metadata_records) {
281
282 # Check if we need to start a new directory for these records
283 check_need_new_directory($exploded_base_dir,$record_number,
284 $records_per_folder,$total_num_records,
285 \$documents_directory);
286 # Use the plugin's process function to avoid duplicating code
287 my $doc_obj = new doc($filename, "nonindexed_doc", $plugobj->get_file_rename_method());
288 $plugobj->process(\$record_text, undef, undef, $filename, undef, $doc_obj, 0);
289
290
291 # Try to get a doc to attach the metadata to
292 # If no match found, create a dummy .nul file
293 attach_metadata_or_make_nul_doc($document_field, $doc_obj, $record_number,
294 $documents_directory, $orig_base_dir,
295 $document_prefix, $document_suffix, $metadata_set, $verbosity);
296
297
298 check_close_directory($record_number,$records_per_folder,$total_num_records);
299
300 $record_number = $record_number + 1;
301 }
302 }
303 else {
304 # Call metadata_read to set up associated metadata
305
306 my $pluginfo = undef;
307 my $block_hash = {};
308
309 my $processor = undef;
310 my $maxdocs = undef;
311 my $gli = undef;
312
313 my $extrametakeys = [];
314 my $extrametadata = {};
315 my $extrametafile = {};
316
317 $plugobj->metadata_read($pluginfo, "", $filename, $block_hash,
318 $extrametakeys, $extrametadata, $extrametafile,
319 $processor, $maxdocs, $gli);
320
321 my $total_num_records = scalar (@$extrametakeys);
322 print STDERR "Number of records: $total_num_records\n";
323 my $record_number = 1;
324 my $documents_directory;
325 foreach my $record (@$extrametakeys) {
326 &check_need_new_directory($exploded_base_dir, $record_number, $records_per_folder, $total_num_records, \$documents_directory);
327
328 # Attach metadata to object
329 # => use the plugin's extra_metadata function to avoid duplicating code
330 my $doc_obj = new doc($filename, "nonindexed_doc", $plugobj->get_file_rename_method());
331 # all the metadata has been extracted into extrametadata
332 $plugobj->extra_metadata ($doc_obj, $doc_obj->get_top_section(), $extrametadata->{$record});
333
334 # Try to get a doc to attach the metadata to
335 # If no match found, create a dummy .nul file
336 attach_metadata_or_make_nul_doc($document_field, $doc_obj, $record_number, $documents_directory, $orig_base_dir, $document_prefix, $document_suffix, $metadata_set, $verbosity);
337
338 &check_close_directory($record_number,$records_per_folder,$total_num_records);
339
340 $record_number = $record_number + 1;
341
342 }
343 }
344
345 # Explode means just that: the original file is deleted
346 &util::rm($filename);
347 $plugobj->clean_up_after_exploding();
348
349}
350
351
352sub need_new_directory
353{
354 my ($exploded_base_dir) = @_;
355
356 my $documents_directory = $exploded_base_dir;
357
358 if (-d $documents_directory) {
359 die "Error: document directory $documents_directory already exists (bailing).\n";
360 }
361 &util::mk_dir($documents_directory);
362
363 my $documents_metadata_xml_file = &util::filename_cat($documents_directory, "metadata.xml");
364 if (-e $documents_metadata_xml_file) {
365 die "Error: documents metadata.xml file $documents_metadata_xml_file already exists (bailing).\n";
366 }
367
368 # Start the metadata.xml file
369 open(METADATA_XML_FILE, ">$documents_metadata_xml_file");
370 print METADATA_XML_FILE
371 "<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"no\"?>\n" .
372 "<!DOCTYPE DirectoryMetadata SYSTEM \"http://greenstone.org/dtd/DirectoryMetadata/1.0/DirectoryMetadata.dtd\">\n" .
373 "<DirectoryMetadata>\n";
374
375 return $documents_directory;
376}
377
378sub check_need_new_directory
379{
380 my ($exploded_base_dir,$record_number, $records_per_folder,
381 $total_num_records, $documents_dir_ref) = @_;
382
383
384 # Check if we need to start a new directory for these records
385 if (($record_number % $records_per_folder) == 1) {
386 my $documents_directory = $exploded_base_dir;
387
388 if ($total_num_records > $records_per_folder) {
389 $documents_directory .= "." . sprintf("%8.8d", $record_number);
390 }
391
392 $$documents_dir_ref = need_new_directory($documents_directory);
393 }
394}
395
396
397
398
399
400sub attach_metadata_or_make_nul_doc
401{
402 my ($document_field, $doc_obj, $record_number,
403 $documents_directory, $orig_base_dir,
404 $document_prefix, $document_suffix, $metadata_set, $verbosity) = @_;
405
406 my $record_metadata = $doc_obj->get_all_metadata($doc_obj->get_top_section());
407 my $document_file;
408
409 # try to get a doc to attach the metadata to
410 if (defined $document_field) {
411 foreach my $pair (@$record_metadata) {
412 my ($field, $value) = (@$pair);
413
414 $value =~ s/\\\\/\\/g;
415 my $document_file_full;
416
417 # Does this metadata element specify a document to obtain?
418 if ($field eq $document_field) {
419 if(-d $document_prefix && $document_prefix !~ m@^(http|ftp|https)://@ ) {
420 # if the document-prefix refers to a directory but not URL, ensure it has a file-separator at the end
421 # by first of all stripping any trailing slash and then always ensuring one is used through filename_cat
422 $document_prefix =~ s/(\/|\\)$//;
423 $document_file_full = &util::filename_cat($document_prefix, "$value$document_suffix");
424 } else { # the doc prefix may also contain the prefix of the actual *filename* following the directory
425 $document_file_full = $document_prefix . $value . $document_suffix;
426 }
427
428 # this either downloads/copies the document, or creates a nul file for it.
429 $document_file = &obtain_document($document_file_full, $documents_directory, $orig_base_dir, $verbosity);
430 &write_metadata_xml_file_entry(METADATA_XML_FILE, $document_file, $record_metadata, $metadata_set);
431 }
432 }
433 }
434
435 # Create a dummy .nul file if we haven't obtained a document (or null file) for this record
436 if (not defined $document_file) {
437
438 if (defined ($record_number)) {
439 $document_file = sprintf("%8.8d", $record_number) . ".nul";
440 }
441 else {
442 $document_file = "doc.nul";
443 }
444 open(DUMMY_FILE, ">$documents_directory/$document_file");
445 close(DUMMY_FILE);
446 &write_metadata_xml_file_entry(METADATA_XML_FILE, $document_file, $record_metadata, $metadata_set);
447 }
448
449}
450
451sub close_directory
452{
453 # Finish and close the metadata.xml file
454 print METADATA_XML_FILE "\n</DirectoryMetadata>\n";
455 close(METADATA_XML_FILE);
456
457}
458
459
460sub check_close_directory
461{
462 my ($record_number,$records_per_folder,$total_num_records) = @_;
463
464 if (($record_number % $records_per_folder) == 0 || $record_number == $total_num_records) {
465 # Finish and close the metadata.xml file
466 close_directory();
467 }
468}
469
470
471
472sub write_metadata_xml_file_entry
473{
474 my $metadata_xml_file = shift(@_);
475 my $file_name = shift(@_);
476 my $record_metadata = shift(@_);
477 my $meta_prefix = shift(@_);
478
479 # Make $file_name XML-safe
480 $file_name =~ s/&/&amp;/g;
481 $file_name =~ s/</&lt;/g;
482 $file_name =~ s/>/&gt;/g;
483
484 # Convert $file_name into a regular expression that matches it
485 $file_name =~ s/\./\\\./g;
486 $file_name =~ s/\(/\\\(/g;
487 $file_name =~ s/\)/\\\)/g;
488 $file_name =~ s/\{/\\\{/g;
489 $file_name =~ s/\}/\\\}/g;
490 $file_name =~ s/\[/\\\[/g;
491 $file_name =~ s/\]/\\\]/g;
492
493 print $metadata_xml_file
494 "\n" .
495 " <FileSet>\n" .
496 " <FileName>$file_name</FileName>\n" .
497 " <Description>\n";
498
499 foreach my $pair (@$record_metadata) {
500 my ($field, $value) = (@$pair);
501
502 # We're only interested in metadata from the database
503 next if ($field eq "lastmodified");
504 next if ($field eq "gsdlsourcefilename");
505 next if ($field eq "gsdldoctype");
506 next if ($field eq "FileFormat");
507
508 # Ignore the ^all metadata, since it will be invalid if the source metadata is changed
509 next if ($field =~ /\^all$/); # ISISPlug specific!
510
511 # Square brackets in metadata values need to be escaped so they don't confuse Greenstone/GLI
512 $value =~ s/\[/&\#091;/g;
513 $value =~ s/\]/&\#093;/g;
514
515 # Make $value XML-safe
516 $value =~ s/&/&amp;/g; # May mess up existing entities!
517 $value =~ s/</&lt;/g;
518 $value =~ s/>/&gt;/g;
519
520 # we are not allowed & in xml except in entities.
521 # if there are undefined entities then parsing will also crap out.
522 # should we be checking for them too?
523 # this may not get all possibilities
524 # $value =~ s/&([^;\s]*(\s|$))/&amp;$1/g;
525
526 # do we already have a namespace specified?
527 my $full_field = $field;
528 if ($meta_prefix ne "") {
529 $full_field =~ s/^\w+\.//;
530 $full_field = $meta_prefix.$full_field;
531 }
532
533 print $metadata_xml_file " <Metadata mode=\"accumulate\" name=\"$full_field\">$value</Metadata>\n";
534 }
535
536 print $metadata_xml_file
537 " </Description>\n" .
538 " </FileSet>\n";
539}
540
541sub obtain_document
542{
543 my ($document_file_full,$documents_directory,$orig_base_dir,$verbosity) = @_;
544
545 print STDERR "Obtaining document file $document_file_full...\n" if ($verbosity > 1);
546
547 my $document_file_name;
548 my $local_document_file;
549
550 # Document specified is on the web
551 if ($document_file_full =~ /^https?:/ || $document_file_full =~ /^ftp:/) {
552 $document_file_full =~ /([^\/]+)$/;
553 $document_file_name = $1;
554 $local_document_file = &util::filename_cat($documents_directory, $document_file_name);
555
556 my $wget_options = "--quiet";
557 $wget_options = "--verbose" if ($verbosity > 2);
558 $wget_options .= " --timestamping"; # Only re-download files if they're newer
559 my $wget_command = "wget $wget_options \"$document_file_full\" --output-document \"$local_document_file\"";
560 `$wget_command`;
561
562 # Check the document was obtained successfully
563 if (!-e $local_document_file) {
564 print STDERR "WARNING: Could not obtain document file $document_file_full\n";
565 }
566 }
567 # Document specified is on the disk
568 else {
569 # convert the dirseps in filepath to correct dir sep for OS
570 $document_file_full = &util::filename_cat($document_file_full);
571 my $dir_sep = &util::get_os_dirsep();
572
573 $document_file_full =~ m/(.+$dir_sep)?(.*)$/;
574 $document_file_name = $2;
575
576
577 my $is_absolute = File::Spec->file_name_is_absolute($document_file_full);
578 print STDERR "doc file full = $document_file_full\n";
579
580 if (!$is_absolute) {
581 $document_file_full
582 = &util::filename_cat($orig_base_dir,$document_file_full);
583 }
584
585 $local_document_file = &util::filename_cat($documents_directory, $document_file_name);
586
587 if (-e $document_file_full) {
588 &util::cp($document_file_full, $documents_directory);
589 }
590
591 # Check the document was obtained successfully
592 if (!-e $local_document_file) {
593 print STDERR "WARNING: Could not obtain document file $document_file_full\n";
594 }
595 else {
596 $orig_base_dir =~ s/\\/\\\\/g; # escape windows style slashes for the regex below
597 if ($document_file_full =~ m/^$orig_base_dir.*/) {
598 # file local to metadata record
599 # => copy has been made successfully, so remove original
600 &util::rm($document_file_full);
601 }
602 }
603 }
604
605 # If the document wasn't obtained successfully, create a .nul file for it
606 if (!-e $local_document_file) {
607 $document_file_name .= ".nul";
608 open(NULL_FILE, ">$local_document_file.nul");
609 close(NULL_FILE);
610 print STDERR "Creating a nul document $document_file_name\n";
611 }
612
613 return $document_file_name;
614}
615
616sub get_plugin_options {
617 my ($collectcfg, $plugin) = @_;
618
619 my $plugin_list = $collectcfg ->{'plugin'};
620
621 foreach my $pluginoptions (@$plugin_list) {
622 my $pluginname = shift @$pluginoptions;
623 next unless $pluginname eq $plugin;
624 map { $_ = "\"$_\""; } @$pluginoptions;
625 my $options = join (",", @$pluginoptions);
626 return $options;
627 }
628 return "";
629}
630
631&main(@ARGV);
632
Note: See TracBrowser for help on using the repository browser.