2 # Import an iso2709 file into Koha 3
8 # find Koha's Perl modules
9 # test carefully before changing this
11 eval { require "$FindBin::Bin/../kohalib.pl" };
15 use MARC::File::USMARC;
28 use Unicode::Normalize;
29 use Time::HiRes qw(gettimeofday);
34 use Koha::SearchEngine;
35 use Koha::SearchEngine::Search;
37 use open qw( :std :encoding(UTF-8) );
38 binmode( STDOUT, ":encoding(UTF-8)" );
39 my ( $input_marc_file, $number, $offset) = ('',0,0);
40 my ($version, $delete, $test_parameter, $skip_marc8_conversion, $char_encoding, $verbose, $commit, $fk_off,$format,$biblios,$authorities,$keepids,$match, $isbn_check, $logfile);
41 my ( $insert, $filters, $update, $all, $yamlfile, $authtypes, $append );
43 my ($sourcetag,$sourcesubfield,$idmapfl, $dedup_barcode);
50 'commit:f' => \$commit,
51 'file:s' => \$input_marc_file,
53 'o|offset:f' => \$offset,
56 't|test' => \$test_parameter,
57 's' => \$skip_marc8_conversion,
58 'c:s' => \$char_encoding,
64 'k|keepids:s' => \$keepids,
65 'b|biblios' => \$biblios,
66 'a|authorities' => \$authorities,
67 'authtypes:s' => \$authtypes,
68 'filter=s@' => \$filters,
72 'match=s@' => \$match,
73 'i|isbn' => \$isbn_check,
75 'y:s' => \$sourcesubfield,
76 'idmap:s' => \$idmapfl,
77 'cleanisbn!' => \$cleanisbn,
78 'yaml:s' => \$yamlfile,
79 'dedupbarcode' => \$dedup_barcode,
80 'framework=s' => \$framework,
81 'custom:s' => \$localcust,
83 $biblios ||= !$authorities;
85 my $writemode = ($append) ? "a" : "w";
92 if ($version || ($input_marc_file eq '')) {
93 pod2usage( -verbose => 2 );
97 if(defined $localcust) { #local customize module
99 $localcust= $localcust||'LocalChanges'; #default name
100 $localcust=~ s/^.*\/([^\/]+)$/$1/; #extract file name only
101 $localcust=~ s/\.pm$//; #remove extension
102 my $fqcust= $FindBin::Bin."/$localcust.pm"; #try migration_tools dir
107 print "WARNING: customize module $localcust.pm not found!\n";
111 require $localcust if $localcust;
112 $localcust=\&customize if $localcust;
115 my $dbh = C4::Context->dbh;
116 my $heading_fields=get_heading_fields();
118 if (defined $idmapfl) {
119 open(IDMAP,">$idmapfl") or die "cannot open $idmapfl \n";
122 if ((not defined $sourcesubfield) && (not defined $sourcetag)){
128 # Disable logging for the biblios and authorities import operation. It would unnecessarily
131 # Disable the syspref cache so we can change logging settings
132 C4::Context->disable_syspref_cache();
133 # Save current CataloguingLog and AuthoritiesLog sysprefs values
134 my $CataloguingLog = C4::Context->preference( 'CataloguingLog' );
135 my $AuthoritiesLog = C4::Context->preference( 'AuthoritiesLog' );
136 # Disable logging for both
137 C4::Context->set_preference( 'CataloguingLog', 0 );
138 C4::Context->set_preference( 'AuthoritiesLog', 0 );
141 $dbh->do("SET FOREIGN_KEY_CHECKS = 0");
147 print "deleting biblios\n";
148 $dbh->do("truncate biblio");
149 $dbh->do("truncate biblioitems");
150 $dbh->do("truncate items");
153 print "deleting authorities\n";
154 $dbh->do("truncate auth_header");
156 $dbh->do("truncate zebraqueue");
161 if ($test_parameter) {
162 print "TESTING MODE ONLY\n DOING NOTHING\n===============\n";
165 my $marcFlavour = C4::Context->preference('marcflavour') || 'MARC21';
167 print "Characteristic MARC flavour: $marcFlavour\n" if $verbose;
168 my $starttime = gettimeofday;
170 my $fh = IO::File->new($input_marc_file); # don't let MARC::Batch open the file, as it applies the ':utf8' IO layer
171 if (defined $format && $format =~ /XML/i) {
172 # ugly hack follows -- MARC::File::XML, when used by MARC::Batch,
173 # appears to try to convert incoming XML records from MARC-8
174 # to UTF-8. Setting the BinaryEncoding key turns that off
175 # TODO: see what happens to ISO-8859-1 XML files.
176 # TODO: determine if MARC::Batch can be fixed to handle
177 # XML records properly -- it probably should be
178 # be using a proper push or pull XML parser to
179 # extract the records, not using regexes to look
180 # for <record>.*</record>.
181 $MARC::File::XML::_load_args{BinaryEncoding} = 'utf-8';
182 my $recordformat= ($marcFlavour eq "MARC21"?"USMARC":uc($marcFlavour));
183 #UNIMARC Authorities have a different way to manage encoding than UNIMARC biblios.
184 $recordformat=$recordformat."AUTH" if ($authorities and $marcFlavour ne "MARC21");
185 $MARC::File::XML::_load_args{RecordFormat} = $recordformat;
186 $batch = MARC::Batch->new( 'XML', $fh );
188 $batch = MARC::Batch->new( 'USMARC', $fh );
190 $batch->warnings_off();
191 $batch->strict_off();
193 my $commitnum = $commit ? $commit : 50;
198 print "Skipping file offset: $offset records\n";
199 $batch->next() while ($offset--);
202 my ($tagid,$subfieldid);
207 ( $tagid, $subfieldid ) =
208 GetMarcFromKohaField( "biblio.biblionumber", $framework );
212 # the SQL query to search on isbn
213 my $sth_isbn = $dbh->prepare("SELECT biblionumber,biblioitemnumber FROM biblioitems WHERE isbn=?");
215 $dbh->{AutoCommit} = 0;
218 $loghandle= IO::File->new($logfile, $writemode) ;
219 print $loghandle "id;operation;status\n";
222 my $searcher = Koha::SearchEngine::Search->new(
226 ? $Koha::SearchEngine::AUTHORITIES_INDEX
227 : $Koha::SearchEngine::BIBLIOS_INDEX
235 eval { $record = $batch->next() };
237 print "Bad MARC record $i: $@ skipped\n";
238 # FIXME - because MARC::Batch->next() combines grabbing the next
239 # blob and parsing it into one operation, a correctable condition
240 # such as a MARC-8 record claiming that it's UTF-8 can't be recovered
241 # from because we don't have access to the original blob. Note
242 # that the staging import can deal with this condition (via
243 # C4::Charset::MarcToUTF8Record) because it doesn't use MARC::Batch.
246 # skip if we get an empty record (that is MARC valid, but will result in AddBiblio failure
247 last unless ( $record );
249 if( ($verbose//1)==1 ) { #no dot for verbose==2
250 print "." . ( $i % 100==0 ? "\n$i" : '' );
253 # transcode the record to UTF8 if needed & applicable.
254 if ($record->encoding() eq 'MARC-8' and not $skip_marc8_conversion) {
255 # FIXME update condition
256 my ($guessed_charset, $charset_errors);
257 ($record, $guessed_charset, $charset_errors) = MarcToUTF8Record($record, $marcFlavour.(($authorities and $marcFlavour ne "MARC21")?'AUTH':''));
258 if ($guessed_charset eq 'failed') {
259 warn "ERROR: failed to perform character conversion for record $i\n";
263 SetUTF8Flag($record);
264 &$localcust($record) if $localcust;
266 # remove trailing - in isbn (only for biblios, of course)
267 if ($biblios && $cleanisbn) {
268 my $tag = $marcFlavour eq 'UNIMARC' ? '010' : '020';
269 my $field = $record->field($tag);
270 my $isbn = $field && $field->subfield('a');
273 $field->update('a' => $isbn);
277 # search for duplicates (based on Local-number)
279 $originalid = GetRecordId( $record, $tagid, $subfieldid );
282 my $query = build_query( $match, $record );
283 my $server = ( $authorities ? 'authorityserver' : 'biblioserver' );
284 $debug && warn $query;
285 my ( $error, $results, $totalhits ) = $searcher->simple_search_compat( $query, 0, 3, [$server] );
286 # changed to warn so able to continue with one broken record
287 if ( defined $error ) {
288 warn "unable to search the database for duplicates : $error";
289 printlog( { id => $id || $originalid || $match, op => "match", status => "ERROR" } ) if ($logfile);
292 $debug && warn "$query $server : $totalhits";
293 if ( $results && scalar(@$results) == 1 ) {
294 my $marcrecord = C4::Search::new_record_from_zebra( $server, $results->[0] );
295 SetUTF8Flag($marcrecord);
296 $id = GetRecordId( $marcrecord, $tagid, $subfieldid );
297 if ( $authorities && $marcFlavour ) {
298 #Skip if authority in database is the same as the on in database
299 if ( $marcrecord->field('005') && $record->field('005') &&
300 $marcrecord->field('005')->data && $record->field('005')->data &&
301 $marcrecord->field('005')->data >= $record->field('005')->data ) {
303 $yamlhash->{$originalid}->{'authid'} = $id;
305 # we recover all subfields of the heading authorities
307 foreach my $field ( $marcrecord->field("2..") ) {
308 push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
310 $yamlhash->{$originalid}->{'subfields'} = \@subfields;
315 } elsif ( $results && scalar(@$results) > 1 ) {
316 $debug && warn "more than one match for $query";
318 $debug && warn "nomatch for $query";
321 if ($keepids && $originalid) {
323 if ( length($keepids) == 3 ) {
324 $storeidfield = MARC::Field->new( $keepids, $originalid );
326 $storeidfield = MARC::Field->new( substr( $keepids, 0, 3 ), "", "", substr( $keepids, 3, 1 ), $originalid );
328 $record->insert_fields_ordered($storeidfield);
329 $record->delete_field( $record->field($tagid) );
331 foreach my $stringfilter (@$filters) {
332 if ( length($stringfilter) == 3 ) {
333 foreach my $field ( $record->field($stringfilter) ) {
334 $record->delete_field($field);
335 $debug && warn "removed : ", $field->as_string;
337 } elsif ($stringfilter =~ /([0-9]{3})([a-z0-9])(.*)/) {
339 my $removesubfield = $2;
340 my $removematch = $3;
341 if ( ( $removetag > "010" ) && $removesubfield ) {
342 foreach my $field ( $record->field($removetag) ) {
343 $field->delete_subfield( code => "$removesubfield", match => $removematch );
344 $debug && warn "Potentially removed : ", $field->subfield($removesubfield);
349 unless ($test_parameter) {
351 use C4::AuthoritiesMarc;
352 my $authtypecode=GuessAuthTypeCode($record, $heading_fields);
353 my $authid= ($id?$id:GuessAuthId($record));
354 if ($authid && GetAuthority($authid) && $update ){
355 ## Authority has an id and is in database : Replace
356 eval { ( $authid ) = ModAuthority($authid,$record, $authtypecode) };
358 warn "Problem with authority $authid Cannot Modify";
359 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ERROR"}) if ($logfile);
362 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ok"}) if ($logfile);
365 elsif (defined $authid) {
366 ## An authid is defined but no authority in database : add
367 eval { ( $authid ) = AddAuthority($record,$authid, $authtypecode) };
369 warn "Problem with authority $authid Cannot Add ".$@;
370 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
373 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
377 ## True insert in database
378 eval { ( $authid ) = AddAuthority($record,"", $authtypecode) };
380 warn "Problem with authority $authid Cannot Add".$@;
381 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
384 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
388 $yamlhash->{$originalid}->{'authid'} = $authid;
390 foreach my $field ( $record->field("2..") ) {
391 push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
393 $yamlhash->{$originalid}->{'subfields'} = \@subfields;
397 my ( $biblionumber, $biblioitemnumber, $itemnumbers_ref, $errors_ref );
399 # check for duplicate, based on ISBN (skip it if we already have found a duplicate with match parameter
400 if (!$biblionumber && $isbn_check && $isbn) {
401 # warn "search ISBN : $isbn";
402 $sth_isbn->execute($isbn);
403 ($biblionumber,$biblioitemnumber) = $sth_isbn->fetchrow;
405 if (defined $idmapfl) {
406 if ($sourcetag < "010"){
407 if ($record->field($sourcetag)){
408 my $source = $record->field($sourcetag)->data();
409 printf(IDMAP "%s|%s\n",$source,$biblionumber);
412 my $source=$record->subfield($sourcetag,$sourcesubfield);
413 printf(IDMAP "%s|%s\n",$source,$biblionumber);
416 # create biblio, unless we already have it ( either match or isbn )
418 eval{$biblioitemnumber=GetBiblioData($biblionumber)->{biblioitemnumber};};
420 eval { ( $biblionumber, $biblioitemnumber ) = ModBiblio( $record, $biblionumber, GetFrameworkCode($biblionumber) ) };
422 warn "ERROR: Edit biblio $biblionumber failed: $@\n";
423 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ERROR" } ) if ($logfile);
426 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ok" } ) if ($logfile);
429 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "warning : already in database" } ) if ($logfile);
433 eval { ( $biblionumber, $biblioitemnumber ) = AddBiblio( $record, '', { defer_marc_save => 1 } ) };
435 warn "ERROR: Adding biblio $biblionumber failed: $@\n";
436 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ERROR" } ) if ($logfile);
439 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ok" } ) if ($logfile);
442 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "warning : not in database" } ) if ($logfile);
445 eval { ( $itemnumbers_ref, $errors_ref ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
446 my $error_adding = $@;
447 # Work on a clone so that if there are real errors, we can maybe
449 my $clone_record = $record->clone();
450 C4::Biblio::_strip_item_fields($clone_record, '');
451 # This sets the marc fields if there was an error, and also calls
453 ModBiblioMarc( $clone_record, $biblionumber, $framework );
454 if ( $error_adding ) {
455 warn "ERROR: Adding items to bib $biblionumber failed: $error_adding";
456 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
457 # if we failed because of an exception, assume that
458 # the MARC columns in biblioitems were not set.
462 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
464 if ($dedup_barcode && grep { exists $_->{error_code} && $_->{error_code} eq 'duplicate_barcode' } @$errors_ref) {
465 # Find the record called 'barcode'
466 my ($tag, $sub) = C4::Biblio::GetMarcFromKohaField('items.barcode', $framework);
467 # Now remove any items that didn't have a duplicate_barcode error,
468 # erase the barcodes on items that did, and re-add those items.
470 foreach my $i (0 .. $#{$errors_ref}) {
471 my $ref = $errors_ref->[$i];
472 if ($ref && ($ref->{error_code} eq 'duplicate_barcode')) {
473 $dupes{$ref->{item_sequence}} = 1;
474 # Delete the error message because we're going to
476 delete $errors_ref->[$i];
480 foreach my $field ($record->field($tag)) {
483 # Here we remove the barcode
484 $field->delete_subfield(code => $sub);
486 # otherwise we delete the field because we don't want
488 $record->delete_fields($field);
491 # Now re-add the record as before, adding errors to the prev list
493 eval { ( $itemnumbers_ref, $more_errors ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
495 warn "ERROR: Adding items to bib $biblionumber failed: $@\n";
496 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
497 # if we failed because of an exception, assume that
498 # the MARC columns in biblioitems were not set.
499 ModBiblioMarc( $record, $biblionumber, $framework );
502 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
504 push @$errors_ref, @{ $more_errors };
506 if ($#{ $errors_ref } > -1) {
507 report_item_errors($biblionumber, $errors_ref);
509 $yamlhash->{$originalid} = $biblionumber if ($yamlfile);
511 $dbh->commit() if (0 == $i % $commitnum);
513 print $record->as_formatted()."\n" if ($verbose//0)==2;
514 last if $i == $number;
517 $dbh->{AutoCommit} = 1;
521 $dbh->do("SET FOREIGN_KEY_CHECKS = 1");
524 # Restore CataloguingLog
525 C4::Context->set_preference( 'CataloguingLog', $CataloguingLog );
526 # Restore AuthoritiesLog
527 C4::Context->set_preference( 'AuthoritiesLog', $AuthoritiesLog );
529 my $timeneeded = gettimeofday - $starttime;
530 print "\n$i MARC records done in $timeneeded seconds\n";
532 print $loghandle "file : $input_marc_file\n";
533 print $loghandle "$i MARC records done in $timeneeded seconds\n";
537 open my $yamlfileout, q{>}, "$yamlfile" or die "cannot open $yamlfile \n";
538 print $yamlfileout Dump($yamlhash);
543 my $marcrecord=shift;
548 return $marcrecord->field($tag)->data() if $marcrecord->field($tag);
551 if ($marcrecord->field($tag)){
552 return $marcrecord->subfield($tag,$subfield);
561 foreach my $matchingpoint (@$match){
562 my $string = build_simplequery($matchingpoint,$record);
563 push @searchstrings,$string if (length($string)>0);
566 $QParser = C4::Context->queryparser if (C4::Context->preference('UseQueryParser'));
573 return join(" $op ",@searchstrings);
575 sub build_simplequery {
579 my ($index,$recorddata)=split /,/,$element;
580 if ($recorddata=~/(\d{3})(.*)/) {
581 my ($tag,$subfields) =($1,$2);
582 foreach my $field ($record->field($tag)){
583 if (length($field->as_string("$subfields"))>0){
584 push @searchstrings,"$index:\"".$field->as_string("$subfields")."\"";
589 $QParser = C4::Context->queryparser if (C4::Context->preference('UseQueryParser'));
596 return join(" $op ",@searchstrings);
598 sub report_item_errors {
599 my $biblionumber = shift;
600 my $errors_ref = shift;
602 foreach my $error (@{ $errors_ref }) {
604 my $msg = "Item not added (bib $biblionumber, item tag #$error->{'item_sequence'}, barcode $error->{'item_barcode'}): ";
605 my $error_code = $error->{'error_code'};
606 $error_code =~ s/_/ /g;
607 $msg .= "$error_code $error->{'error_information'}";
612 my $logelements=shift;
613 print $loghandle join( ";", map { defined $_ ? $_ : "" } @$logelements{qw<id op status>} ), "\n";
615 sub get_heading_fields{
618 $headingfields=YAML::LoadFile($authtypes);
619 $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
620 $debug && warn YAML::Dump($headingfields);
622 unless ($headingfields){
623 $headingfields=$dbh->selectall_hashref("SELECT auth_tag_to_report, authtypecode from auth_types",'auth_tag_to_report',{Slice=>{}});
624 $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
626 return $headingfields;
631 bulkmarcimport.pl - Import bibliographic/authority records into Koha
635 $ export KOHA_CONF=/etc/koha.conf
636 $ perl misc/migration_tools/bulkmarcimport.pl -d -commit 1000 \\
637 -file /home/jmf/koha.mrc -n 3000
641 Don't use this script before you've entered and checked your MARC parameters
642 tables twice (or more!). Otherwise, the import won't work correctly and you
643 will get invalid data.
651 This version/help screen
653 =item B<-b, -biblios>
655 Type of import: bibliographic records
657 =item B<-a, -authorities>
659 Type of import: authority records
661 =item B<-file>=I<FILE>
663 The I<FILE> to import
667 Verbose mode. 1 means "some infos", 2 means "MARC dumping"
671 Turn off foreign key checks during import.
673 =item B<-n>=I<NUMBER>
675 The I<NUMBER> of records to import. If missing, all the file is imported
677 =item B<-o, -offset>=I<NUMBER>
679 File offset before importing, ie I<NUMBER> of records to skip.
681 =item B<-commit>=I<NUMBER>
683 The I<NUMBER> of records to wait before performing a 'commit' operation
687 File logs actions done for each record and their status into file
691 If specified, data will be appended to the logfile. If not, the logfile will be erased for each execution.
695 Test mode: parses the file, saying what he would do, but doing nothing.
699 Skip automatic conversion of MARC-8 to UTF-8. This option is provided for
702 =item B<-c>=I<CHARACTERISTIC>
704 The I<CHARACTERISTIC> MARC flavour. At the moment, only I<MARC21> and
705 I<UNIMARC> are supported. MARC21 by default.
709 Delete EVERYTHING related to biblio in koha-DB before import. Tables: biblio,
712 =item B<-m>=I<FORMAT>
714 Input file I<FORMAT>: I<MARCXML> or I<ISO2709> (defaults to ISO2709)
718 file yamlfile with authoritiesTypes and distinguishable record field in order
719 to store the correct authtype
723 yaml file format a yaml file with ids
727 list of fields that will not be imported. Can be any from 000 to 999 or field,
728 subfield and subfield's matching value such as 200avalue
732 if set, only insert when possible
736 if set, only updates (any biblio should have a matching record)
740 if set, do whatever is required
742 =item B<-k, -keepids>=<FIELD>
744 Field store ids in I<FIELD> (useful for authorities, where 001 contains the
745 authid for Koha, that can contain a very valuable info for authorities coming
746 from LOC or BNF. useless for biblios probably)
748 =item B<-match>=<FIELD>
750 I<FIELD> matchindex,fieldtomatch matchpoint to use to deduplicate fieldtomatch
751 can be either 001 to 999 or field and list of subfields as such 100abcde
755 If set, a search will be done on isbn, and, if the same isbn is found, the
756 biblio is not added. It's another method to deduplicate. B<-match> & B<-isbn>
761 Clean ISBN fields from entering biblio records, ie removes hyphens. By default,
762 ISBN are cleaned. --nocleanisbn will keep ISBN unchanged.
766 Source bib I<TAG> for reporting the source bib number
768 =item B<-y>=I<SUBFIELD>
770 Source I<SUBFIELD> for reporting the source bib number
772 =item B<-idmap>=I<FILE>
774 I<FILE> for the koha bib and source id
778 Store ids in 009 (useful for authorities, where 001 contains the authid for
779 Koha, that can contain a very valuable info for authorities coming from LOC or
780 BNF. useless for biblios probably)
782 =item B<-dedupbarcode>
784 If set, whenever a duplicate barcode is detected, it is removed and the attempt
785 to add the record is retried, thereby giving the record a blank barcode. This
786 is useful when something has set barcodes to be a biblio ID, or similar
787 (usually other software.)
791 This is the code for the framework that the requested records will have attached
792 to them when they are created. If not specified, then the default framework
795 =item B<-custom>=I<MODULE>
797 This parameter allows you to use a local module with a customize subroutine
798 that is called for each MARC record.
799 If no filename is passed, LocalChanges.pm is assumed to be in the
800 migration_tools subdirectory. You may pass an absolute file name or a file name
801 from the migration_tools directory.