2 # Copyright (C) 2010-2011 Laurentian University
3 # Author: Dan Scott <dscott@laurentian.ca>
5 # This program is free software; you can redistribute it and/or
6 # modify it under the terms of the GNU General Public License
7 # as published by the Free Software Foundation; either version 2
8 # of the License, or (at your option) any later version.
10 # This program is distributed in the hope that it will be useful,
11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 # GNU General Public License for more details.
14 # ---------------------------------------------------------------
21 use MARC::File::XML (BinaryEncoding => 'UTF-8');
24 use OpenILS::Utils::Fieldmapper;
25 use OpenSRF::Utils::SettingsClient;
26 use OpenSRF::EX qw/:try/;
28 use Unicode::Normalize;
29 use OpenILS::Application::AppUtils;
31 use Pod::Usage qw/ pod2usage /;
33 MARC::Charset->assume_unicode(1);
35 my ($start_id, $end_id);
36 my $bootstrap = '@sysconfdir@/opensrf_core.xml';
40 my $result = GetOptions(
42 'configuration=s' => \$bootstrap,
43 'record=i' => \@records,
45 'start_id=i' => \$start_id,
46 'end_id=i' => \$end_id,
49 if (!$result or $options{help}) {
53 OpenSRF::System->bootstrap_client(config_file => $bootstrap);
54 Fieldmapper->import(IDL => OpenSRF::Utils::SettingsClient->new->config_value("IDL"));
56 # must be loaded and initialized after the IDL is parsed
57 use OpenILS::Utils::CStoreEditor;
58 OpenILS::Utils::CStoreEditor::init();
60 my $e = OpenILS::Utils::CStoreEditor->new;
63 # get a list of all non-deleted records from Evergreen
64 # open-ils.cstore open-ils.cstore.direct.biblio.record_entry.id_list.atomic {"deleted":"f"}
65 $undeleted = $e->request(
66 'open-ils.cstore.direct.biblio.record_entry.id_list.atomic',
67 [{deleted => 'f'}, {id => { '>' => 0}}]
69 @records = @$undeleted;
72 if ($start_id and $end_id) {
73 @records = ($start_id .. $end_id);
75 # print Dumper($undeleted, \@records);
77 # Hash of controlled fields & subfields in bibliographic records, and their
78 # corresponding controlling fields & subfields in the authority record
80 # So, if the bib 650$a can be controlled by an auth 150$a, that maps to:
81 # 650 => { a => { 150 => 'a'}}
83 100 => { a => { 100 => 'a' },
100 110 => { a => { 110 => 'a' },
115 111 => { a => { 111 => 'a' },
131 130 => { a => { 130 => 'a' },
146 600 => { a => { 100 => 'a' },
171 610 => { a => { 110 => 'a' },
193 611 => { a => { 111 => 'a' },
213 630 => { a => { 130 => 'a' },
232 648 => { a => { 148 => 'a' },
238 650 => { a => { 150 => 'a' },
245 651 => { a => { 151 => 'a' },
251 655 => { a => { 155 => 'a' },
257 700 => { a => { 100 => 'a' },
274 710 => { a => { 110 => 'a' },
289 711 => { a => { 111 => 'a' },
305 730 => { a => { 130 => 'a' },
320 751 => { a => { 151 => 'a' },
326 830 => { a => { 830 => 'a' },
343 foreach my $rec_id (@records) {
346 # State variable; was the record changed?
350 my $record = $e->retrieve_biblio_record_entry($rec_id);
352 # print Dumper($record);
355 my $marc = MARC::Record->new_from_xml($record->marc());
357 # get the list of controlled fields
358 my @c_fields = keys %controllees;
360 foreach my $c_tag (@c_fields) {
361 my @c_subfields = keys %{$controllees{"$c_tag"}};
362 # print "Field: $field subfields: ";
363 # foreach (@subfields) { print "$_ "; }
365 # Get the MARCXML from the record and check for controlled fields/subfields
366 my @bib_fields = ($marc->field($c_tag));
367 foreach my $bib_field (@bib_fields) {
368 # print $_->as_formatted();
372 foreach my $c_subfield (@c_subfields) {
373 my $sf = $bib_field->subfield($c_subfield);
375 # Give me the first element of the list of authority controlling tags for this subfield
376 # XXX Will we need to support more than one controlling tag per subfield? Probably. That
377 # will suck. Oh well, leave that up to Ole to implement.
378 $match_subfields{$c_subfield} = (keys %{$controllees{$c_tag}{$c_subfield}})[0];
379 $match_tag = $match_subfields{$c_subfield};
380 push @searches, {term => $sf, subfield => $c_subfield};
383 # print Dumper(\%match_subfields);
386 my @tags = ($match_tag);
388 # print "Controlling tag: $c_tag and match tag $match_tag\n";
389 # print Dumper(\@tags, \@searches);
391 # Now we've built up a complete set of matching controlled
392 # subfields for this particular field; let's check to see if
393 # we have a matching authority record
394 my $session = OpenSRF::AppSession->create("open-ils.search");
395 my $validates = $session->request("open-ils.search.authority.validate.tag.id_list",
396 "tags", \@tags, "searches", \@searches
398 $session->disconnect();
400 # print Dumper($validates);
402 # Protect against failed (error condition) search request
404 print STDERR "Search for matching authority failed; record # $rec_id\n";
408 if (scalar(@$validates) == 0) {
412 # Iterate through the returned authority record IDs to delete any
413 # matching $0 subfields already in the bib record
414 foreach my $auth_zero (@$validates) {
415 $bib_field->delete_subfield(code => '0', match => qr/\)$auth_zero$/);
418 # Okay, we have a matching authority control; time to
419 # add the magical subfield 0. Use the first returned auth
421 my $auth_id = @$validates[0];
422 my $auth_rec = $e->retrieve_authority_record_entry($auth_id);
423 my $auth_marc = MARC::Record->new_from_xml($auth_rec->marc());
424 my $cni = $auth_marc->field('003')->data();
426 $bib_field->add_subfields('0' => "($cni)$auth_id");
431 my $editor = OpenILS::Utils::CStoreEditor->new(xact=>1);
432 # print $marc->as_formatted();
433 my $xml = $marc->as_xml_record();
435 $xml =~ s/^<\?xml.+\?\s*>//go;
436 $xml =~ s/>\s+</></go;
437 $xml =~ s/\p{Cc}//go;
438 $xml = OpenILS::Application::AppUtils->entityize($xml);
441 $editor->update_biblio_record_entry($record);
446 print STDERR "\nRecord # $rec_id : $err\n";
447 import MARC::File::XML; # reset SAX parser so that one bad record doesn't kill the entire process
455 authority_control_fields.pl - Controls fields in bibliographic records with authorities in Evergreen
459 C<authority_control_fields.pl> [B<--configuration>=I<opensrf_core.conf>]
460 [[B<--record>=I<record>[ B<--record>=I<record>]]] | [B<--all>] | [B<--start_id>=I<start-ID> B<--end_id>=I<end-ID>]
464 For a given set of records:
468 =item * Iterate through the list of fields that are controlled fields
470 =item * Iterate through the list of subfields that are controlled for
473 =item * Search for a matching authority record for that combination of
478 =item * If we find a match, then add a $0 subfield to that field identifying
479 the controlling authority record
481 =item * If we do not find a match, then insert a row into an "uncontrolled"
482 table identifying the record ID, field, and subfield(s) that were not controlled
486 =item * Iterate through the list of floating subdivisions
490 =item * If we find a match, then add a $0 subfield to that field identifying
491 the controlling authority record
493 =item * If we do not find a match, then insert a row into an "uncontrolled"
494 table identifying the record ID, field, and subfield(s) that were not controlled
498 =item * If we changed the record, update it in the database
506 =item * B<-c> I<config-file>, B<--configuration>=I<config-file>
508 Specifies the OpenSRF configuration file used to connect to the OpenSRF router.
509 Defaults to F<@sysconfdir@/opensrf_core.xml>
511 =item * B<-r> I<record-ID>, B<--record>=I<record-ID>
513 Specifies the bibliographic record ID (found in the C<biblio.record_entry.id>
514 column) of the record to process. This option may be specified more than once
515 to process multiple records in a single run.
517 =item * B<-a>, B<--all>
519 Specifies that all bibliographic records should be processed. For large
520 databases, this may take an extraordinarily long amount of time.
522 =item * B<-s> I<start-ID>, B<--start_id>=I<start-ID>
524 Specifies the starting ID of the range of bibliographic records to process.
525 This option is ignored unless it is accompanied by the B<-e> or B<--end_id>
528 =item * B<-e> I<end-ID>, B<--end_id>=I<end-ID>
530 Specifies the ending ID of the range of bibliographic records to process.
531 This option is ignored unless it is accompanied by the B<-s> or B<--start>
538 authority_control_fields.pl --start_id 1 --end_id 50000
540 Processes the bibliographic records with IDs between 1 and 50,000 using the
541 default OpenSRF configuration file for connection information.
545 Dan Scott <dscott@laurentian.ca>
547 =head1 COPYRIGHT AND LICENSE
549 Copyright 2010-2011 by Dan Scott
551 This program is free software; you can redistribute it and/or
552 modify it under the terms of the GNU General Public License
553 as published by the Free Software Foundation; either version 2
554 of the License, or (at your option) any later version.
556 This program is distributed in the hope that it will be useful,
557 but WITHOUT ANY WARRANTY; without even the implied warranty of
558 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
559 GNU General Public License for more details.
561 You should have received a copy of the GNU General Public License
562 along with this program; if not, write to the Free Software
563 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.