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 open qw( :std :encoding(UTF-8) );
35 binmode( STDOUT, ":encoding(UTF-8)" );
36 my ( $input_marc_file, $number, $offset) = ('',0,0);
37 my ($version, $delete, $test_parameter, $skip_marc8_conversion, $char_encoding, $verbose, $commit, $fk_off,$format,$biblios,$authorities,$keepids,$match, $isbn_check, $logfile);
38 my ( $insert, $filters, $update, $all, $yamlfile, $authtypes, $append );
40 my ($sourcetag,$sourcesubfield,$idmapfl, $dedup_barcode);
47 'commit:f' => \$commit,
48 'file:s' => \$input_marc_file,
50 'o|offset:f' => \$offset,
53 't|test' => \$test_parameter,
54 's' => \$skip_marc8_conversion,
55 'c:s' => \$char_encoding,
61 'k|keepids:s' => \$keepids,
62 'b|biblios' => \$biblios,
63 'a|authorities' => \$authorities,
64 'authtypes:s' => \$authtypes,
65 'filter=s@' => \$filters,
69 'match=s@' => \$match,
70 'i|isbn' => \$isbn_check,
72 'y:s' => \$sourcesubfield,
73 'idmap:s' => \$idmapfl,
74 'cleanisbn!' => \$cleanisbn,
75 'yaml:s' => \$yamlfile,
76 'dedupbarcode' => \$dedup_barcode,
77 'framework=s' => \$framework,
78 'custom:s' => \$localcust,
80 $biblios ||= !$authorities;
82 my $writemode = ($append) ? "a" : "w";
89 if ($version || ($input_marc_file eq '')) {
90 pod2usage( -verbose => 2 );
94 if(defined $localcust) { #local customize module
96 $localcust= $localcust||'LocalChanges'; #default name
97 $localcust=~ s/^.*\/([^\/]+)$/$1/; #extract file name only
98 $localcust=~ s/\.pm$//; #remove extension
99 my $fqcust= $FindBin::Bin."/$localcust.pm"; #try migration_tools dir
104 print "WARNING: customize module $localcust.pm not found!\n";
108 require $localcust if $localcust;
109 $localcust=\&customize if $localcust;
112 my $dbh = C4::Context->dbh;
113 my $heading_fields=get_heading_fields();
115 if (defined $idmapfl) {
116 open(IDMAP,">$idmapfl") or die "cannot open $idmapfl \n";
119 if ((not defined $sourcesubfield) && (not defined $sourcetag)){
125 # Disable logging for the biblios and authorities import operation. It would unnecesarily
128 # Disable the syspref cache so we can change logging settings
129 C4::Context->disable_syspref_cache();
130 # Save current CataloguingLog and AuthoritiesLog sysprefs values
131 my $CataloguingLog = C4::Context->preference( 'CataloguingLog' );
132 my $AuthoritiesLog = C4::Context->preference( 'AuthoritiesLog' );
133 # Disable logging for both
134 C4::Context->set_preference( 'CataloguingLog', 0 );
135 C4::Context->set_preference( 'AuthoritiesLog', 0 );
138 $dbh->do("SET FOREIGN_KEY_CHECKS = 0");
144 print "deleting biblios\n";
145 $dbh->do("truncate biblio");
146 $dbh->do("truncate biblioitems");
147 $dbh->do("truncate items");
150 print "deleting authorities\n";
151 $dbh->do("truncate auth_header");
153 $dbh->do("truncate zebraqueue");
158 if ($test_parameter) {
159 print "TESTING MODE ONLY\n DOING NOTHING\n===============\n";
162 my $marcFlavour = C4::Context->preference('marcflavour') || 'MARC21';
164 print "Characteristic MARC flavour: $marcFlavour\n" if $verbose;
165 my $starttime = gettimeofday;
167 my $fh = IO::File->new($input_marc_file); # don't let MARC::Batch open the file, as it applies the ':utf8' IO layer
168 if (defined $format && $format =~ /XML/i) {
169 # ugly hack follows -- MARC::File::XML, when used by MARC::Batch,
170 # appears to try to convert incoming XML records from MARC-8
171 # to UTF-8. Setting the BinaryEncoding key turns that off
172 # TODO: see what happens to ISO-8859-1 XML files.
173 # TODO: determine if MARC::Batch can be fixed to handle
174 # XML records properly -- it probably should be
175 # be using a proper push or pull XML parser to
176 # extract the records, not using regexes to look
177 # for <record>.*</record>.
178 $MARC::File::XML::_load_args{BinaryEncoding} = 'utf-8';
179 my $recordformat= ($marcFlavour eq "MARC21"?"USMARC":uc($marcFlavour));
180 #UNIMARC Authorities have a different way to manage encoding than UNIMARC biblios.
181 $recordformat=$recordformat."AUTH" if ($authorities and $marcFlavour ne "MARC21");
182 $MARC::File::XML::_load_args{RecordFormat} = $recordformat;
183 $batch = MARC::Batch->new( 'XML', $fh );
185 $batch = MARC::Batch->new( 'USMARC', $fh );
187 $batch->warnings_off();
188 $batch->strict_off();
190 my $commitnum = $commit ? $commit : 50;
195 print "Skipping file offset: $offset records\n";
196 $batch->next() while ($offset--);
199 my ($tagid,$subfieldid);
204 ( $tagid, $subfieldid ) =
205 GetMarcFromKohaField( "biblio.biblionumber", $framework );
209 # the SQL query to search on isbn
210 my $sth_isbn = $dbh->prepare("SELECT biblionumber,biblioitemnumber FROM biblioitems WHERE isbn=?");
212 $dbh->{AutoCommit} = 0;
215 $loghandle= IO::File->new($logfile, $writemode) ;
216 print $loghandle "id;operation;status\n";
221 eval { $record = $batch->next() };
223 print "Bad MARC record $i: $@ skipped\n";
224 # FIXME - because MARC::Batch->next() combines grabbing the next
225 # blob and parsing it into one operation, a correctable condition
226 # such as a MARC-8 record claiming that it's UTF-8 can't be recovered
227 # from because we don't have access to the original blob. Note
228 # that the staging import can deal with this condition (via
229 # C4::Charset::MarcToUTF8Record) because it doesn't use MARC::Batch.
232 # skip if we get an empty record (that is MARC valid, but will result in AddBiblio failure
233 last unless ( $record );
235 if( ($verbose//1)==1 ) { #no dot for verbose==2
236 print "." . ( $i % 100==0 ? "\n$i" : '' );
239 # transcode the record to UTF8 if needed & applicable.
240 if ($record->encoding() eq 'MARC-8' and not $skip_marc8_conversion) {
241 # FIXME update condition
242 my ($guessed_charset, $charset_errors);
243 ($record, $guessed_charset, $charset_errors) = MarcToUTF8Record($record, $marcFlavour.(($authorities and $marcFlavour ne "MARC21")?'AUTH':''));
244 if ($guessed_charset eq 'failed') {
245 warn "ERROR: failed to perform character conversion for record $i\n";
249 SetUTF8Flag($record);
250 &$localcust($record) if $localcust;
252 # remove trailing - in isbn (only for biblios, of course)
253 if ($biblios && $cleanisbn) {
254 my $tag = $marcFlavour eq 'UNIMARC' ? '010' : '020';
255 my $field = $record->field($tag);
256 my $isbn = $field && $field->subfield('a');
259 $field->update('a' => $isbn);
263 # search for duplicates (based on Local-number)
265 $originalid = GetRecordId( $record, $tagid, $subfieldid );
268 my $query = build_query( $match, $record );
269 my $server = ( $authorities ? 'authorityserver' : 'biblioserver' );
270 $debug && warn $query;
271 my ( $error, $results, $totalhits ) = C4::Search::SimpleSearch( $query, 0, 3, [$server] );
272 # changed to warn so able to continue with one broken record
273 if ( defined $error ) {
274 warn "unable to search the database for duplicates : $error";
277 $debug && warn "$query $server : $totalhits";
278 if ( $results && scalar(@$results) == 1 ) {
279 my $marcrecord = C4::Search::new_record_from_zebra( $server, $results->[0] );
280 SetUTF8Flag($marcrecord);
281 $id = GetRecordId( $marcrecord, $tagid, $subfieldid );
282 if ( $authorities && $marcFlavour ) {
283 #Skip if authority in database is the same as the on in database
284 if ( $marcrecord->field('005') && $record->field('005') &&
285 $marcrecord->field('005')->data && $record->field('005')->data &&
286 $marcrecord->field('005')->data >= $record->field('005')->data ) {
288 $yamlhash->{$originalid}->{'authid'} = $id;
290 # we recover all subfields of the heading authorities
292 foreach my $field ( $marcrecord->field("2..") ) {
293 push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
295 $yamlhash->{$originalid}->{'subfields'} = \@subfields;
300 } elsif ( $results && scalar(@$results) > 1 ) {
301 $debug && warn "more than one match for $query";
303 $debug && warn "nomatch for $query";
306 if ($keepids && $originalid) {
308 if ( length($keepids) == 3 ) {
309 $storeidfield = MARC::Field->new( $keepids, $originalid );
311 $storeidfield = MARC::Field->new( substr( $keepids, 0, 3 ), "", "", substr( $keepids, 3, 1 ), $originalid );
313 $record->insert_fields_ordered($storeidfield);
314 $record->delete_field( $record->field($tagid) );
316 foreach my $stringfilter (@$filters) {
317 if ( length($stringfilter) == 3 ) {
318 foreach my $field ( $record->field($stringfilter) ) {
319 $record->delete_field($field);
320 $debug && warn "removed : ", $field->as_string;
322 } elsif ($stringfilter =~ /([0-9]{3})([a-z0-9])(.*)/) {
324 my $removesubfield = $2;
325 my $removematch = $3;
326 if ( ( $removetag > "010" ) && $removesubfield ) {
327 foreach my $field ( $record->field($removetag) ) {
328 $field->delete_subfield( code => "$removesubfield", match => $removematch );
329 $debug && warn "Potentially removed : ", $field->subfield($removesubfield);
334 unless ($test_parameter) {
336 use C4::AuthoritiesMarc;
337 my $authtypecode=GuessAuthTypeCode($record, $heading_fields);
338 my $authid= ($id?$id:GuessAuthId($record));
339 if ($authid && GetAuthority($authid) && $update ){
340 ## Authority has an id and is in database : Replace
341 eval { ( $authid ) = ModAuthority($authid,$record, $authtypecode) };
343 warn "Problem with authority $authid Cannot Modify";
344 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ERROR"}) if ($logfile);
347 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ok"}) if ($logfile);
350 elsif (defined $authid) {
351 ## An authid is defined but no authority in database : add
352 eval { ( $authid ) = AddAuthority($record,$authid, $authtypecode) };
354 warn "Problem with authority $authid Cannot Add ".$@;
355 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
358 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
362 ## True insert in database
363 eval { ( $authid ) = AddAuthority($record,"", $authtypecode) };
365 warn "Problem with authority $authid Cannot Add".$@;
366 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
369 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
373 $yamlhash->{$originalid}->{'authid'} = $authid;
375 foreach my $field ( $record->field("2..") ) {
376 push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
378 $yamlhash->{$originalid}->{'subfields'} = \@subfields;
382 my ( $biblionumber, $biblioitemnumber, $itemnumbers_ref, $errors_ref );
384 # check for duplicate, based on ISBN (skip it if we already have found a duplicate with match parameter
385 if (!$biblionumber && $isbn_check && $isbn) {
386 # warn "search ISBN : $isbn";
387 $sth_isbn->execute($isbn);
388 ($biblionumber,$biblioitemnumber) = $sth_isbn->fetchrow;
390 if (defined $idmapfl) {
391 if ($sourcetag < "010"){
392 if ($record->field($sourcetag)){
393 my $source = $record->field($sourcetag)->data();
394 printf(IDMAP "%s|%s\n",$source,$biblionumber);
397 my $source=$record->subfield($sourcetag,$sourcesubfield);
398 printf(IDMAP "%s|%s\n",$source,$biblionumber);
401 # create biblio, unless we already have it ( either match or isbn )
403 eval{$biblioitemnumber=GetBiblioData($biblionumber)->{biblioitemnumber};};
405 eval { ( $biblionumber, $biblioitemnumber ) = ModBiblio( $record, $biblionumber, GetFrameworkCode($biblionumber) ) };
407 warn "ERROR: Edit biblio $biblionumber failed: $@\n";
408 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ERROR" } ) if ($logfile);
411 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ok" } ) if ($logfile);
414 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "warning : already in database" } ) if ($logfile);
418 eval { ( $biblionumber, $biblioitemnumber ) = AddBiblio( $record, '', { defer_marc_save => 1 } ) };
420 warn "ERROR: Adding biblio $biblionumber failed: $@\n";
421 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ERROR" } ) if ($logfile);
424 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ok" } ) if ($logfile);
427 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "warning : not in database" } ) if ($logfile);
430 eval { ( $itemnumbers_ref, $errors_ref ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
431 my $error_adding = $@;
432 # Work on a clone so that if there are real errors, we can maybe
434 my $clone_record = $record->clone();
435 C4::Biblio::_strip_item_fields($clone_record, '');
436 # This sets the marc fields if there was an error, and also calls
438 ModBiblioMarc( $clone_record, $biblionumber, $framework );
439 if ( $error_adding ) {
440 warn "ERROR: Adding items to bib $biblionumber failed: $error_adding";
441 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
442 # if we failed because of an exception, assume that
443 # the MARC columns in biblioitems were not set.
447 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
449 if ($dedup_barcode && grep { exists $_->{error_code} && $_->{error_code} eq 'duplicate_barcode' } @$errors_ref) {
450 # Find the record called 'barcode'
451 my ($tag, $sub) = C4::Biblio::GetMarcFromKohaField('items.barcode', $framework);
452 # Now remove any items that didn't have a duplicate_barcode error,
453 # erase the barcodes on items that did, and re-add those items.
455 foreach my $i (0 .. $#{$errors_ref}) {
456 my $ref = $errors_ref->[$i];
457 if ($ref && ($ref->{error_code} eq 'duplicate_barcode')) {
458 $dupes{$ref->{item_sequence}} = 1;
459 # Delete the error message because we're going to
461 delete $errors_ref->[$i];
465 foreach my $field ($record->field($tag)) {
468 # Here we remove the barcode
469 $field->delete_subfield(code => $sub);
471 # otherwise we delete the field because we don't want
473 $record->delete_fields($field);
476 # Now re-add the record as before, adding errors to the prev list
478 eval { ( $itemnumbers_ref, $more_errors ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
480 warn "ERROR: Adding items to bib $biblionumber failed: $@\n";
481 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
482 # if we failed because of an exception, assume that
483 # the MARC columns in biblioitems were not set.
484 ModBiblioMarc( $record, $biblionumber, $framework );
487 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
489 push @$errors_ref, @{ $more_errors };
491 if ($#{ $errors_ref } > -1) {
492 report_item_errors($biblionumber, $errors_ref);
494 $yamlhash->{$originalid} = $biblionumber if ($yamlfile);
496 $dbh->commit() if (0 == $i % $commitnum);
498 print $record->as_formatted()."\n" if ($verbose//0)==2;
499 last if $i == $number;
502 $dbh->{AutoCommit} = 1;
506 $dbh->do("SET FOREIGN_KEY_CHECKS = 1");
509 # Restore CataloguingLog
510 C4::Context->set_preference( 'CataloguingLog', $CataloguingLog );
511 # Restore AuthoritiesLog
512 C4::Context->set_preference( 'AuthoritiesLog', $AuthoritiesLog );
514 my $timeneeded = gettimeofday - $starttime;
515 print "\n$i MARC records done in $timeneeded seconds\n";
517 print $loghandle "file : $input_marc_file\n";
518 print $loghandle "$i MARC records done in $timeneeded seconds\n";
522 open my $yamlfileout, q{>}, "$yamlfile" or die "cannot open $yamlfile \n";
523 print $yamlfileout Dump($yamlhash);
528 my $marcrecord=shift;
533 return $marcrecord->field($tag)->data() if $marcrecord->field($tag);
536 if ($marcrecord->field($tag)){
537 return $marcrecord->subfield($tag,$subfield);
546 foreach my $matchingpoint (@$match){
547 my $string = build_simplequery($matchingpoint,$record);
548 push @searchstrings,$string if (length($string)>0);
551 $QParser = C4::Context->queryparser if (C4::Context->preference('UseQueryParser'));
558 return join(" $op ",@searchstrings);
560 sub build_simplequery {
564 my ($index,$recorddata)=split /,/,$element;
565 if ($recorddata=~/(\d{3})(.*)/) {
566 my ($tag,$subfields) =($1,$2);
567 foreach my $field ($record->field($tag)){
568 if (length($field->as_string("$subfields"))>0){
569 push @searchstrings,"$index:\"".$field->as_string("$subfields")."\"";
574 $QParser = C4::Context->queryparser if (C4::Context->preference('UseQueryParser'));
581 return join(" $op ",@searchstrings);
583 sub report_item_errors {
584 my $biblionumber = shift;
585 my $errors_ref = shift;
587 foreach my $error (@{ $errors_ref }) {
589 my $msg = "Item not added (bib $biblionumber, item tag #$error->{'item_sequence'}, barcode $error->{'item_barcode'}): ";
590 my $error_code = $error->{'error_code'};
591 $error_code =~ s/_/ /g;
592 $msg .= "$error_code $error->{'error_information'}";
597 my $logelements=shift;
598 print $loghandle join( ";", map { defined $_ ? $_ : "" } @$logelements{qw<id op status>} ), "\n";
600 sub get_heading_fields{
603 $headingfields=YAML::LoadFile($authtypes);
604 $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
605 $debug && warn YAML::Dump($headingfields);
607 unless ($headingfields){
608 $headingfields=$dbh->selectall_hashref("SELECT auth_tag_to_report, authtypecode from auth_types",'auth_tag_to_report',{Slice=>{}});
609 $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
611 return $headingfields;
616 bulkmarcimport.pl - Import bibliographic/authority records into Koha
620 $ export KOHA_CONF=/etc/koha.conf
621 $ perl misc/migration_tools/bulkmarcimport.pl -d -commit 1000 \\
622 -file /home/jmf/koha.mrc -n 3000
626 Don't use this script before you've entered and checked your MARC parameters
627 tables twice (or more!). Otherwise, the import won't work correctly and you
628 will get invalid data.
636 This version/help screen
638 =item B<-b, -biblios>
640 Type of import: bibliographic records
642 =item B<-a, -authorities>
644 Type of import: authority records
646 =item B<-file>=I<FILE>
648 The I<FILE> to import
652 Verbose mode. 1 means "some infos", 2 means "MARC dumping"
656 Turn off foreign key checks during import.
658 =item B<-n>=I<NUMBER>
660 The I<NUMBER> of records to import. If missing, all the file is imported
662 =item B<-o, -offset>=I<NUMBER>
664 File offset before importing, ie I<NUMBER> of records to skip.
666 =item B<-commit>=I<NUMBER>
668 The I<NUMBER> of records to wait before performing a 'commit' operation
672 File logs actions done for each record and their status into file
676 If specified, data will be appended to the logfile. If not, the logfile will be erased for each execution.
680 Test mode: parses the file, saying what he would do, but doing nothing.
684 Skip automatic conversion of MARC-8 to UTF-8. This option is provided for
687 =item B<-c>=I<CHARACTERISTIC>
689 The I<CHARACTERISTIC> MARC flavour. At the moment, only I<MARC21> and
690 I<UNIMARC> are supported. MARC21 by default.
694 Delete EVERYTHING related to biblio in koha-DB before import. Tables: biblio,
697 =item B<-m>=I<FORMAT>
699 Input file I<FORMAT>: I<MARCXML> or I<ISO2709> (defaults to ISO2709)
703 file yamlfile with authoritiesTypes and distinguishable record field in order
704 to store the correct authtype
708 yaml file format a yaml file with ids
712 list of fields that will not be imported. Can be any from 000 to 999 or field,
713 subfield and subfield's matching value such as 200avalue
717 if set, only insert when possible
721 if set, only updates (any biblio should have a matching record)
725 if set, do whatever is required
727 =item B<-k, -keepids>=<FIELD>
729 Field store ids in I<FIELD> (usefull for authorities, where 001 contains the
730 authid for Koha, that can contain a very valuable info for authorities coming
731 from LOC or BNF. useless for biblios probably)
733 =item B<-match>=<FIELD>
735 I<FIELD> matchindex,fieldtomatch matchpoint to use to deduplicate fieldtomatch
736 can be either 001 to 999 or field and list of subfields as such 100abcde
740 If set, a search will be done on isbn, and, if the same isbn is found, the
741 biblio is not added. It's another method to deduplicate. B<-match> & B<-isbn>
746 Clean ISBN fields from entering biblio records, ie removes hyphens. By default,
747 ISBN are cleaned. --nocleanisbn will keep ISBN unchanged.
751 Source bib I<TAG> for reporting the source bib number
753 =item B<-y>=I<SUBFIELD>
755 Source I<SUBFIELD> for reporting the source bib number
757 =item B<-idmap>=I<FILE>
759 I<FILE> for the koha bib and source id
763 Store ids in 009 (usefull for authorities, where 001 contains the authid for
764 Koha, that can contain a very valuable info for authorities coming from LOC or
765 BNF. useless for biblios probably)
767 =item B<-dedupbarcode>
769 If set, whenever a duplicate barcode is detected, it is removed and the attempt
770 to add the record is retried, thereby giving the record a blank barcode. This
771 is useful when something has set barcodes to be a biblio ID, or similar
772 (usually other software.)
776 This is the code for the framework that the requested records will have attached
777 to them when they are created. If not specified, then the default framework
780 =item B<-custom>=I<MODULE>
782 This parameter allows you to use a local module with a customize subroutine
783 that is called for each MARC record.
784 If no filename is passed, LocalChanges.pm is assumed to be in the
785 migration_tools subdirectory. You may pass an absolute file name or a file name
786 from the migration_tools directory.