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";
87 pod2usage( -msg => "\nYou must specify either --biblios or --authorities, not both.\n", -exitval ) if $biblios && $authorities;
94 if ($version || ($input_marc_file eq '')) {
95 pod2usage( -verbose => 2 );
99 if(defined $localcust) { #local customize module
101 $localcust= $localcust||'LocalChanges'; #default name
102 $localcust=~ s/^.*\/([^\/]+)$/$1/; #extract file name only
103 $localcust=~ s/\.pm$//; #remove extension
104 my $fqcust= $FindBin::Bin."/$localcust.pm"; #try migration_tools dir
109 print "WARNING: customize module $localcust.pm not found!\n";
113 require $localcust if $localcust;
114 $localcust=\&customize if $localcust;
117 my $dbh = C4::Context->dbh;
118 my $heading_fields=get_heading_fields();
120 if (defined $idmapfl) {
121 open(IDMAP,">$idmapfl") or die "cannot open $idmapfl \n";
124 if ((not defined $sourcesubfield) && (not defined $sourcetag)){
130 # Disable logging for the biblios and authorities import operation. It would unnecessarily
133 # Disable the syspref cache so we can change logging settings
134 C4::Context->disable_syspref_cache();
135 # Save current CataloguingLog and AuthoritiesLog sysprefs values
136 my $CataloguingLog = C4::Context->preference( 'CataloguingLog' );
137 my $AuthoritiesLog = C4::Context->preference( 'AuthoritiesLog' );
138 # Disable logging for both
139 C4::Context->set_preference( 'CataloguingLog', 0 );
140 C4::Context->set_preference( 'AuthoritiesLog', 0 );
143 $dbh->do("SET FOREIGN_KEY_CHECKS = 0");
149 print "deleting biblios\n";
150 $dbh->do("truncate biblio");
151 $dbh->do("truncate biblioitems");
152 $dbh->do("truncate items");
155 print "deleting authorities\n";
156 $dbh->do("truncate auth_header");
158 $dbh->do("truncate zebraqueue");
163 if ($test_parameter) {
164 print "TESTING MODE ONLY\n DOING NOTHING\n===============\n";
167 my $marcFlavour = C4::Context->preference('marcflavour') || 'MARC21';
169 print "Characteristic MARC flavour: $marcFlavour\n" if $verbose;
170 my $starttime = gettimeofday;
172 my $fh = IO::File->new($input_marc_file); # don't let MARC::Batch open the file, as it applies the ':utf8' IO layer
173 if (defined $format && $format =~ /XML/i) {
174 # ugly hack follows -- MARC::File::XML, when used by MARC::Batch,
175 # appears to try to convert incoming XML records from MARC-8
176 # to UTF-8. Setting the BinaryEncoding key turns that off
177 # TODO: see what happens to ISO-8859-1 XML files.
178 # TODO: determine if MARC::Batch can be fixed to handle
179 # XML records properly -- it probably should be
180 # be using a proper push or pull XML parser to
181 # extract the records, not using regexes to look
182 # for <record>.*</record>.
183 $MARC::File::XML::_load_args{BinaryEncoding} = 'utf-8';
184 my $recordformat= ($marcFlavour eq "MARC21"?"USMARC":uc($marcFlavour));
185 #UNIMARC Authorities have a different way to manage encoding than UNIMARC biblios.
186 $recordformat=$recordformat."AUTH" if ($authorities and $marcFlavour ne "MARC21");
187 $MARC::File::XML::_load_args{RecordFormat} = $recordformat;
188 $batch = MARC::Batch->new( 'XML', $fh );
190 $batch = MARC::Batch->new( 'USMARC', $fh );
192 $batch->warnings_off();
193 $batch->strict_off();
195 my $commitnum = $commit ? $commit : 50;
200 print "Skipping file offset: $offset records\n";
201 $batch->next() while ($offset--);
204 my ($tagid,$subfieldid);
209 ( $tagid, $subfieldid ) =
210 GetMarcFromKohaField( "biblio.biblionumber", $framework );
214 # the SQL query to search on isbn
215 my $sth_isbn = $dbh->prepare("SELECT biblionumber,biblioitemnumber FROM biblioitems WHERE isbn=?");
217 $dbh->{AutoCommit} = 0;
220 $loghandle= IO::File->new($logfile, $writemode) ;
221 print $loghandle "id;operation;status\n";
224 my $searcher = Koha::SearchEngine::Search->new(
228 ? $Koha::SearchEngine::AUTHORITIES_INDEX
229 : $Koha::SearchEngine::BIBLIOS_INDEX
237 eval { $record = $batch->next() };
239 print "Bad MARC record $i: $@ skipped\n";
240 # FIXME - because MARC::Batch->next() combines grabbing the next
241 # blob and parsing it into one operation, a correctable condition
242 # such as a MARC-8 record claiming that it's UTF-8 can't be recovered
243 # from because we don't have access to the original blob. Note
244 # that the staging import can deal with this condition (via
245 # C4::Charset::MarcToUTF8Record) because it doesn't use MARC::Batch.
248 # skip if we get an empty record (that is MARC valid, but will result in AddBiblio failure
249 last unless ( $record );
251 if( ($verbose//1)==1 ) { #no dot for verbose==2
252 print "." . ( $i % 100==0 ? "\n$i" : '' );
255 # transcode the record to UTF8 if needed & applicable.
256 if ($record->encoding() eq 'MARC-8' and not $skip_marc8_conversion) {
257 # FIXME update condition
258 my ($guessed_charset, $charset_errors);
259 ($record, $guessed_charset, $charset_errors) = MarcToUTF8Record($record, $marcFlavour.(($authorities and $marcFlavour ne "MARC21")?'AUTH':''));
260 if ($guessed_charset eq 'failed') {
261 warn "ERROR: failed to perform character conversion for record $i\n";
265 SetUTF8Flag($record);
266 &$localcust($record) if $localcust;
268 # remove trailing - in isbn (only for biblios, of course)
269 if ($biblios && $cleanisbn) {
270 my $tag = $marcFlavour eq 'UNIMARC' ? '010' : '020';
271 my $field = $record->field($tag);
272 my $isbn = $field && $field->subfield('a');
275 $field->update('a' => $isbn);
279 # search for duplicates (based on Local-number)
281 $originalid = GetRecordId( $record, $tagid, $subfieldid );
284 my $query = build_query( $match, $record );
285 my $server = ( $authorities ? 'authorityserver' : 'biblioserver' );
286 $debug && warn $query;
287 my ( $error, $results, $totalhits ) = $searcher->simple_search_compat( $query, 0, 3, [$server] );
288 # changed to warn so able to continue with one broken record
289 if ( defined $error ) {
290 warn "unable to search the database for duplicates : $error";
291 printlog( { id => $id || $originalid || $match, op => "match", status => "ERROR" } ) if ($logfile);
294 $debug && warn "$query $server : $totalhits";
295 if ( $results && scalar(@$results) == 1 ) {
296 my $marcrecord = C4::Search::new_record_from_zebra( $server, $results->[0] );
297 SetUTF8Flag($marcrecord);
298 $id = GetRecordId( $marcrecord, $tagid, $subfieldid );
299 if ( $authorities && $marcFlavour ) {
300 #Skip if authority in database is the same as the on in database
301 if ( $marcrecord->field('005') && $record->field('005') &&
302 $marcrecord->field('005')->data && $record->field('005')->data &&
303 $marcrecord->field('005')->data >= $record->field('005')->data ) {
305 $yamlhash->{$originalid}->{'authid'} = $id;
307 # we recover all subfields of the heading authorities
309 foreach my $field ( $marcrecord->field("2..") ) {
310 push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
312 $yamlhash->{$originalid}->{'subfields'} = \@subfields;
317 } elsif ( $results && scalar(@$results) > 1 ) {
318 $debug && warn "more than one match for $query";
320 $debug && warn "nomatch for $query";
323 if ($keepids && $originalid) {
325 if ( length($keepids) == 3 ) {
326 $storeidfield = MARC::Field->new( $keepids, $originalid );
328 $storeidfield = MARC::Field->new( substr( $keepids, 0, 3 ), "", "", substr( $keepids, 3, 1 ), $originalid );
330 $record->insert_fields_ordered($storeidfield);
331 $record->delete_field( $record->field($tagid) );
333 foreach my $stringfilter (@$filters) {
334 if ( length($stringfilter) == 3 ) {
335 foreach my $field ( $record->field($stringfilter) ) {
336 $record->delete_field($field);
337 $debug && warn "removed : ", $field->as_string;
339 } elsif ($stringfilter =~ /([0-9]{3})([a-z0-9])(.*)/) {
341 my $removesubfield = $2;
342 my $removematch = $3;
343 if ( ( $removetag > "010" ) && $removesubfield ) {
344 foreach my $field ( $record->field($removetag) ) {
345 $field->delete_subfield( code => "$removesubfield", match => $removematch );
346 $debug && warn "Potentially removed : ", $field->subfield($removesubfield);
351 unless ($test_parameter) {
353 use C4::AuthoritiesMarc;
354 my $authtypecode=GuessAuthTypeCode($record, $heading_fields);
355 my $authid= ($id?$id:GuessAuthId($record));
356 if ($authid && GetAuthority($authid) && $update ){
357 ## Authority has an id and is in database : Replace
358 eval { ( $authid ) = ModAuthority($authid,$record, $authtypecode) };
360 warn "Problem with authority $authid Cannot Modify";
361 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ERROR"}) if ($logfile);
364 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ok"}) if ($logfile);
367 elsif (defined $authid) {
368 ## An authid is defined but no authority in database : add
369 eval { ( $authid ) = AddAuthority($record,$authid, $authtypecode) };
371 warn "Problem with authority $authid Cannot Add ".$@;
372 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
375 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
379 ## True insert in database
380 eval { ( $authid ) = AddAuthority($record,"", $authtypecode) };
382 warn "Problem with authority $authid Cannot Add".$@;
383 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
386 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
390 $yamlhash->{$originalid}->{'authid'} = $authid;
392 foreach my $field ( $record->field("2..") ) {
393 push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
395 $yamlhash->{$originalid}->{'subfields'} = \@subfields;
399 my ( $biblionumber, $biblioitemnumber, $itemnumbers_ref, $errors_ref );
401 # check for duplicate, based on ISBN (skip it if we already have found a duplicate with match parameter
402 if (!$biblionumber && $isbn_check && $isbn) {
403 # warn "search ISBN : $isbn";
404 $sth_isbn->execute($isbn);
405 ($biblionumber,$biblioitemnumber) = $sth_isbn->fetchrow;
407 if (defined $idmapfl) {
408 if ($sourcetag < "010"){
409 if ($record->field($sourcetag)){
410 my $source = $record->field($sourcetag)->data();
411 printf(IDMAP "%s|%s\n",$source,$biblionumber);
414 my $source=$record->subfield($sourcetag,$sourcesubfield);
415 printf(IDMAP "%s|%s\n",$source,$biblionumber);
418 # create biblio, unless we already have it ( either match or isbn )
420 eval{$biblioitemnumber=GetBiblioData($biblionumber)->{biblioitemnumber};};
422 eval { ( $biblionumber, $biblioitemnumber ) = ModBiblio( $record, $biblionumber, GetFrameworkCode($biblionumber) ) };
424 warn "ERROR: Edit biblio $biblionumber failed: $@\n";
425 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ERROR" } ) if ($logfile);
428 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ok" } ) if ($logfile);
431 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "warning : already in database" } ) if ($logfile);
435 eval { ( $biblionumber, $biblioitemnumber ) = AddBiblio( $record, '', { defer_marc_save => 1 } ) };
437 warn "ERROR: Adding biblio $biblionumber failed: $@\n";
438 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ERROR" } ) if ($logfile);
441 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ok" } ) if ($logfile);
444 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "warning : not in database" } ) if ($logfile);
447 eval { ( $itemnumbers_ref, $errors_ref ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
448 my $error_adding = $@;
449 # Work on a clone so that if there are real errors, we can maybe
451 my $clone_record = $record->clone();
452 C4::Biblio::_strip_item_fields($clone_record, '');
453 # This sets the marc fields if there was an error, and also calls
455 ModBiblioMarc( $clone_record, $biblionumber, $framework );
456 if ( $error_adding ) {
457 warn "ERROR: Adding items to bib $biblionumber failed: $error_adding";
458 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
459 # if we failed because of an exception, assume that
460 # the MARC columns in biblioitems were not set.
464 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
466 if ($dedup_barcode && grep { exists $_->{error_code} && $_->{error_code} eq 'duplicate_barcode' } @$errors_ref) {
467 # Find the record called 'barcode'
468 my ($tag, $sub) = C4::Biblio::GetMarcFromKohaField('items.barcode', $framework);
469 # Now remove any items that didn't have a duplicate_barcode error,
470 # erase the barcodes on items that did, and re-add those items.
472 foreach my $i (0 .. $#{$errors_ref}) {
473 my $ref = $errors_ref->[$i];
474 if ($ref && ($ref->{error_code} eq 'duplicate_barcode')) {
475 $dupes{$ref->{item_sequence}} = 1;
476 # Delete the error message because we're going to
478 delete $errors_ref->[$i];
482 foreach my $field ($record->field($tag)) {
485 # Here we remove the barcode
486 $field->delete_subfield(code => $sub);
488 # otherwise we delete the field because we don't want
490 $record->delete_fields($field);
493 # Now re-add the record as before, adding errors to the prev list
495 eval { ( $itemnumbers_ref, $more_errors ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
497 warn "ERROR: Adding items to bib $biblionumber failed: $@\n";
498 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
499 # if we failed because of an exception, assume that
500 # the MARC columns in biblioitems were not set.
501 ModBiblioMarc( $record, $biblionumber, $framework );
504 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
506 push @$errors_ref, @{ $more_errors };
508 if ($#{ $errors_ref } > -1) {
509 report_item_errors($biblionumber, $errors_ref);
511 $yamlhash->{$originalid} = $biblionumber if ($yamlfile);
513 $dbh->commit() if (0 == $i % $commitnum);
515 print $record->as_formatted()."\n" if ($verbose//0)==2;
516 last if $i == $number;
519 $dbh->{AutoCommit} = 1;
523 $dbh->do("SET FOREIGN_KEY_CHECKS = 1");
526 # Restore CataloguingLog
527 C4::Context->set_preference( 'CataloguingLog', $CataloguingLog );
528 # Restore AuthoritiesLog
529 C4::Context->set_preference( 'AuthoritiesLog', $AuthoritiesLog );
531 my $timeneeded = gettimeofday - $starttime;
532 print "\n$i MARC records done in $timeneeded seconds\n";
534 print $loghandle "file : $input_marc_file\n";
535 print $loghandle "$i MARC records done in $timeneeded seconds\n";
539 open my $yamlfileout, q{>}, "$yamlfile" or die "cannot open $yamlfile \n";
540 print $yamlfileout Dump($yamlhash);
545 my $marcrecord=shift;
550 return $marcrecord->field($tag)->data() if $marcrecord->field($tag);
553 if ($marcrecord->field($tag)){
554 return $marcrecord->subfield($tag,$subfield);
563 foreach my $matchingpoint (@$match){
564 my $string = build_simplequery($matchingpoint,$record);
565 push @searchstrings,$string if (length($string)>0);
568 $QParser = C4::Context->queryparser if (C4::Context->preference('UseQueryParser'));
575 return join(" $op ",@searchstrings);
577 sub build_simplequery {
581 my ($index,$recorddata)=split /,/,$element;
582 if ($recorddata=~/(\d{3})(.*)/) {
583 my ($tag,$subfields) =($1,$2);
584 foreach my $field ($record->field($tag)){
585 if (length($field->as_string("$subfields"))>0){
586 push @searchstrings,"$index:\"".$field->as_string("$subfields")."\"";
591 $QParser = C4::Context->queryparser if (C4::Context->preference('UseQueryParser'));
598 return join(" $op ",@searchstrings);
600 sub report_item_errors {
601 my $biblionumber = shift;
602 my $errors_ref = shift;
604 foreach my $error (@{ $errors_ref }) {
606 my $msg = "Item not added (bib $biblionumber, item tag #$error->{'item_sequence'}, barcode $error->{'item_barcode'}): ";
607 my $error_code = $error->{'error_code'};
608 $error_code =~ s/_/ /g;
609 $msg .= "$error_code $error->{'error_information'}";
614 my $logelements=shift;
615 print $loghandle join( ";", map { defined $_ ? $_ : "" } @$logelements{qw<id op status>} ), "\n";
617 sub get_heading_fields{
620 $headingfields=YAML::LoadFile($authtypes);
621 $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
622 $debug && warn YAML::Dump($headingfields);
624 unless ($headingfields){
625 $headingfields=$dbh->selectall_hashref("SELECT auth_tag_to_report, authtypecode from auth_types",'auth_tag_to_report',{Slice=>{}});
626 $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
628 return $headingfields;
633 bulkmarcimport.pl - Import bibliographic/authority records into Koha
637 $ export KOHA_CONF=/etc/koha.conf
638 $ perl misc/migration_tools/bulkmarcimport.pl -d -commit 1000 \\
639 -file /home/jmf/koha.mrc -n 3000
643 Don't use this script before you've entered and checked your MARC parameters
644 tables twice (or more!). Otherwise, the import won't work correctly and you
645 will get invalid data.
653 This version/help screen
655 =item B<-b, -biblios>
657 Type of import: bibliographic records
659 =item B<-a, -authorities>
661 Type of import: authority records
663 =item B<-file>=I<FILE>
665 The I<FILE> to import
669 Verbose mode. 1 means "some infos", 2 means "MARC dumping"
673 Turn off foreign key checks during import.
675 =item B<-n>=I<NUMBER>
677 The I<NUMBER> of records to import. If missing, all the file is imported
679 =item B<-o, -offset>=I<NUMBER>
681 File offset before importing, ie I<NUMBER> of records to skip.
683 =item B<-commit>=I<NUMBER>
685 The I<NUMBER> of records to wait before performing a 'commit' operation
689 File logs actions done for each record and their status into file
693 If specified, data will be appended to the logfile. If not, the logfile will be erased for each execution.
697 Test mode: parses the file, saying what he would do, but doing nothing.
701 Skip automatic conversion of MARC-8 to UTF-8. This option is provided for
704 =item B<-c>=I<CHARACTERISTIC>
706 The I<CHARACTERISTIC> MARC flavour. At the moment, only I<MARC21> and
707 I<UNIMARC> are supported. MARC21 by default.
711 Delete EVERYTHING related to biblio in koha-DB before import. Tables: biblio,
714 =item B<-m>=I<FORMAT>
716 Input file I<FORMAT>: I<MARCXML> or I<ISO2709> (defaults to ISO2709)
720 file yamlfile with authoritiesTypes and distinguishable record field in order
721 to store the correct authtype
725 yaml file format a yaml file with ids
729 list of fields that will not be imported. Can be any from 000 to 999 or field,
730 subfield and subfield's matching value such as 200avalue
734 if set, only insert when possible
738 if set, only updates (any biblio should have a matching record)
742 if set, do whatever is required
744 =item B<-k, -keepids>=<FIELD>
746 Field store ids in I<FIELD> (useful for authorities, where 001 contains the
747 authid for Koha, that can contain a very valuable info for authorities coming
748 from LOC or BNF. useless for biblios probably)
750 =item B<-match>=<FIELD>
752 I<FIELD> matchindex,fieldtomatch matchpoint to use to deduplicate fieldtomatch
753 can be either 001 to 999 or field and list of subfields as such 100abcde
757 If set, a search will be done on isbn, and, if the same isbn is found, the
758 biblio is not added. It's another method to deduplicate. B<-match> & B<-isbn>
763 Clean ISBN fields from entering biblio records, ie removes hyphens. By default,
764 ISBN are cleaned. --nocleanisbn will keep ISBN unchanged.
768 Source bib I<TAG> for reporting the source bib number
770 =item B<-y>=I<SUBFIELD>
772 Source I<SUBFIELD> for reporting the source bib number
774 =item B<-idmap>=I<FILE>
776 I<FILE> for the koha bib and source id
780 Store ids in 009 (useful for authorities, where 001 contains the authid for
781 Koha, that can contain a very valuable info for authorities coming from LOC or
782 BNF. useless for biblios probably)
784 =item B<-dedupbarcode>
786 If set, whenever a duplicate barcode is detected, it is removed and the attempt
787 to add the record is retried, thereby giving the record a blank barcode. This
788 is useful when something has set barcodes to be a biblio ID, or similar
789 (usually other software.)
793 This is the code for the framework that the requested records will have attached
794 to them when they are created. If not specified, then the default framework
797 =item B<-custom>=I<MODULE>
799 This parameter allows you to use a local module with a customize subroutine
800 that is called for each MARC record.
801 If no filename is passed, LocalChanges.pm is assumed to be in the
802 migration_tools subdirectory. You may pass an absolute file name or a file name
803 from the migration_tools directory.