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);
35 use Koha::SearchEngine;
36 use Koha::SearchEngine::Search;
38 use open qw( :std :encoding(UTF-8) );
39 binmode( STDOUT, ":encoding(UTF-8)" );
40 my ( $input_marc_file, $number, $offset) = ('',0,0);
41 my ($version, $delete, $test_parameter, $skip_marc8_conversion, $char_encoding, $verbose, $commit, $fk_off,$format,$biblios,$authorities,$keepids,$match, $isbn_check, $logfile);
42 my ( $insert, $filters, $update, $all, $yamlfile, $authtypes, $append );
44 my ($sourcetag,$sourcesubfield,$idmapfl, $dedup_barcode);
51 'commit:f' => \$commit,
52 'file:s' => \$input_marc_file,
54 'o|offset:f' => \$offset,
57 't|test' => \$test_parameter,
58 's' => \$skip_marc8_conversion,
59 'c:s' => \$char_encoding,
65 'k|keepids:s' => \$keepids,
66 'b|biblios' => \$biblios,
67 'a|authorities' => \$authorities,
68 'authtypes:s' => \$authtypes,
69 'filter=s@' => \$filters,
73 'match=s@' => \$match,
74 'i|isbn' => \$isbn_check,
76 'y:s' => \$sourcesubfield,
77 'idmap:s' => \$idmapfl,
78 'cleanisbn!' => \$cleanisbn,
79 'yaml:s' => \$yamlfile,
80 'dedupbarcode' => \$dedup_barcode,
81 'framework=s' => \$framework,
82 'custom:s' => \$localcust,
84 $biblios ||= !$authorities;
86 my $writemode = ($append) ? "a" : "w";
88 pod2usage( -msg => "\nYou must specify either --biblios or --authorities, not both.\n", -exitval ) if $biblios && $authorities;
95 if ($version || ($input_marc_file eq '')) {
96 pod2usage( -verbose => 2 );
100 if(defined $localcust) { #local customize module
102 $localcust= $localcust||'LocalChanges'; #default name
103 $localcust=~ s/^.*\/([^\/]+)$/$1/; #extract file name only
104 $localcust=~ s/\.pm$//; #remove extension
105 my $fqcust= $FindBin::Bin."/$localcust.pm"; #try migration_tools dir
110 print "WARNING: customize module $localcust.pm not found!\n";
114 require $localcust if $localcust;
115 $localcust=\&customize if $localcust;
118 my $dbh = C4::Context->dbh;
119 my $heading_fields=get_heading_fields();
121 if (defined $idmapfl) {
122 open(IDMAP,">$idmapfl") or die "cannot open $idmapfl \n";
125 if ((not defined $sourcesubfield) && (not defined $sourcetag)){
131 # Disable logging for the biblios and authorities import operation. It would unnecessarily
134 # Disable the syspref cache so we can change logging settings
135 C4::Context->disable_syspref_cache();
136 # Save current CataloguingLog and AuthoritiesLog sysprefs values
137 my $CataloguingLog = C4::Context->preference( 'CataloguingLog' );
138 my $AuthoritiesLog = C4::Context->preference( 'AuthoritiesLog' );
139 # Disable logging for both
140 C4::Context->set_preference( 'CataloguingLog', 0 );
141 C4::Context->set_preference( 'AuthoritiesLog', 0 );
144 $dbh->do("SET FOREIGN_KEY_CHECKS = 0");
150 print "deleting biblios\n";
151 $dbh->do("truncate biblio");
152 $dbh->do("truncate biblioitems");
153 $dbh->do("truncate items");
156 print "deleting authorities\n";
157 $dbh->do("truncate auth_header");
159 $dbh->do("truncate zebraqueue");
164 if ($test_parameter) {
165 print "TESTING MODE ONLY\n DOING NOTHING\n===============\n";
168 my $marcFlavour = C4::Context->preference('marcflavour') || 'MARC21';
170 print "Characteristic MARC flavour: $marcFlavour\n" if $verbose;
171 my $starttime = gettimeofday;
173 my $fh = IO::File->new($input_marc_file); # don't let MARC::Batch open the file, as it applies the ':utf8' IO layer
174 if (defined $format && $format =~ /XML/i) {
175 # ugly hack follows -- MARC::File::XML, when used by MARC::Batch,
176 # appears to try to convert incoming XML records from MARC-8
177 # to UTF-8. Setting the BinaryEncoding key turns that off
178 # TODO: see what happens to ISO-8859-1 XML files.
179 # TODO: determine if MARC::Batch can be fixed to handle
180 # XML records properly -- it probably should be
181 # be using a proper push or pull XML parser to
182 # extract the records, not using regexes to look
183 # for <record>.*</record>.
184 $MARC::File::XML::_load_args{BinaryEncoding} = 'utf-8';
185 my $recordformat= ($marcFlavour eq "MARC21"?"USMARC":uc($marcFlavour));
186 #UNIMARC Authorities have a different way to manage encoding than UNIMARC biblios.
187 $recordformat=$recordformat."AUTH" if ($authorities and $marcFlavour ne "MARC21");
188 $MARC::File::XML::_load_args{RecordFormat} = $recordformat;
189 $batch = MARC::Batch->new( 'XML', $fh );
191 $batch = MARC::Batch->new( 'USMARC', $fh );
193 $batch->warnings_off();
194 $batch->strict_off();
196 my $commitnum = $commit ? $commit : 50;
201 print "Skipping file offset: $offset records\n";
202 $batch->next() while ($offset--);
205 my ($tagid,$subfieldid);
210 ( $tagid, $subfieldid ) =
211 GetMarcFromKohaField( "biblio.biblionumber", $framework );
215 # the SQL query to search on isbn
216 my $sth_isbn = $dbh->prepare("SELECT biblionumber,biblioitemnumber FROM biblioitems WHERE isbn=?");
218 $dbh->{AutoCommit} = 0;
221 $loghandle= IO::File->new($logfile, $writemode) ;
222 print $loghandle "id;operation;status\n";
225 my $searcher = Koha::SearchEngine::Search->new(
229 ? $Koha::SearchEngine::AUTHORITIES_INDEX
230 : $Koha::SearchEngine::BIBLIOS_INDEX
238 eval { $record = $batch->next() };
240 print "Bad MARC record $i: $@ skipped\n";
241 # FIXME - because MARC::Batch->next() combines grabbing the next
242 # blob and parsing it into one operation, a correctable condition
243 # such as a MARC-8 record claiming that it's UTF-8 can't be recovered
244 # from because we don't have access to the original blob. Note
245 # that the staging import can deal with this condition (via
246 # C4::Charset::MarcToUTF8Record) because it doesn't use MARC::Batch.
249 # skip if we get an empty record (that is MARC valid, but will result in AddBiblio failure
250 last unless ( $record );
252 if( ($verbose//1)==1 ) { #no dot for verbose==2
253 print "." . ( $i % 100==0 ? "\n$i" : '' );
256 # transcode the record to UTF8 if needed & applicable.
257 if ($record->encoding() eq 'MARC-8' and not $skip_marc8_conversion) {
258 # FIXME update condition
259 my ($guessed_charset, $charset_errors);
260 ($record, $guessed_charset, $charset_errors) = MarcToUTF8Record($record, $marcFlavour.(($authorities and $marcFlavour ne "MARC21")?'AUTH':''));
261 if ($guessed_charset eq 'failed') {
262 warn "ERROR: failed to perform character conversion for record $i\n";
266 SetUTF8Flag($record);
267 &$localcust($record) if $localcust;
269 # remove trailing - in isbn (only for biblios, of course)
270 if ($biblios && $cleanisbn) {
271 my $tag = $marcFlavour eq 'UNIMARC' ? '010' : '020';
272 my $field = $record->field($tag);
273 my $isbn = $field && $field->subfield('a');
276 $field->update('a' => $isbn);
280 # search for duplicates (based on Local-number)
282 $originalid = GetRecordId( $record, $tagid, $subfieldid );
285 my $query = build_query( $match, $record );
286 my $server = ( $authorities ? 'authorityserver' : 'biblioserver' );
287 $debug && warn $query;
288 my ( $error, $results, $totalhits ) = $searcher->simple_search_compat( $query, 0, 3, [$server] );
289 # changed to warn so able to continue with one broken record
290 if ( defined $error ) {
291 warn "unable to search the database for duplicates : $error";
292 printlog( { id => $id || $originalid || $match, op => "match", status => "ERROR" } ) if ($logfile);
295 $debug && warn "$query $server : $totalhits";
296 if ( $results && scalar(@$results) == 1 ) {
297 my $marcrecord = C4::Search::new_record_from_zebra( $server, $results->[0] );
298 SetUTF8Flag($marcrecord);
299 $id = GetRecordId( $marcrecord, $tagid, $subfieldid );
300 if ( $authorities && $marcFlavour ) {
301 #Skip if authority in database is the same as the on in database
302 if ( $marcrecord->field('005') && $record->field('005') &&
303 $marcrecord->field('005')->data && $record->field('005')->data &&
304 $marcrecord->field('005')->data >= $record->field('005')->data ) {
306 $yamlhash->{$originalid}->{'authid'} = $id;
308 # we recover all subfields of the heading authorities
310 foreach my $field ( $marcrecord->field("2..") ) {
311 push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
313 $yamlhash->{$originalid}->{'subfields'} = \@subfields;
318 } elsif ( $results && scalar(@$results) > 1 ) {
319 $debug && warn "more than one match for $query";
321 $debug && warn "nomatch for $query";
324 if ($keepids && $originalid) {
326 if ( length($keepids) == 3 ) {
327 $storeidfield = MARC::Field->new( $keepids, $originalid );
329 $storeidfield = MARC::Field->new( substr( $keepids, 0, 3 ), "", "", substr( $keepids, 3, 1 ), $originalid );
331 $record->insert_fields_ordered($storeidfield);
332 $record->delete_field( $record->field($tagid) );
334 foreach my $stringfilter (@$filters) {
335 if ( length($stringfilter) == 3 ) {
336 foreach my $field ( $record->field($stringfilter) ) {
337 $record->delete_field($field);
338 $debug && warn "removed : ", $field->as_string;
340 } elsif ($stringfilter =~ /([0-9]{3})([a-z0-9])(.*)/) {
342 my $removesubfield = $2;
343 my $removematch = $3;
344 if ( ( $removetag > "010" ) && $removesubfield ) {
345 foreach my $field ( $record->field($removetag) ) {
346 $field->delete_subfield( code => "$removesubfield", match => $removematch );
347 $debug && warn "Potentially removed : ", $field->subfield($removesubfield);
352 unless ($test_parameter) {
354 use C4::AuthoritiesMarc;
355 my $authtypecode=GuessAuthTypeCode($record, $heading_fields);
356 my $authid= ($id?$id:GuessAuthId($record));
357 if ($authid && GetAuthority($authid) && $update ){
358 ## Authority has an id and is in database : Replace
359 eval { ( $authid ) = ModAuthority($authid,$record, $authtypecode) };
361 warn "Problem with authority $authid Cannot Modify";
362 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ERROR"}) if ($logfile);
365 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ok"}) if ($logfile);
368 elsif (defined $authid) {
369 ## An authid is defined but no authority in database : add
370 eval { ( $authid ) = AddAuthority($record,$authid, $authtypecode) };
372 warn "Problem with authority $authid Cannot Add ".$@;
373 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
376 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
380 ## True insert in database
381 eval { ( $authid ) = AddAuthority($record,"", $authtypecode) };
383 warn "Problem with authority $authid Cannot Add".$@;
384 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
387 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
391 $yamlhash->{$originalid}->{'authid'} = $authid;
393 foreach my $field ( $record->field("2..") ) {
394 push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
396 $yamlhash->{$originalid}->{'subfields'} = \@subfields;
400 my ( $biblionumber, $biblioitemnumber, $itemnumbers_ref, $errors_ref );
402 # check for duplicate, based on ISBN (skip it if we already have found a duplicate with match parameter
403 if (!$biblionumber && $isbn_check && $isbn) {
404 # warn "search ISBN : $isbn";
405 $sth_isbn->execute($isbn);
406 ($biblionumber,$biblioitemnumber) = $sth_isbn->fetchrow;
408 if (defined $idmapfl) {
409 if ($sourcetag < "010"){
410 if ($record->field($sourcetag)){
411 my $source = $record->field($sourcetag)->data();
412 printf(IDMAP "%s|%s\n",$source,$biblionumber);
415 my $source=$record->subfield($sourcetag,$sourcesubfield);
416 printf(IDMAP "%s|%s\n",$source,$biblionumber);
419 # create biblio, unless we already have it ( either match or isbn )
422 $biblioitemnumber = Koha::Biblios->find( $biblionumber )->biblioitem->biblioitemnumber;
425 eval { ( $biblionumber, $biblioitemnumber ) = ModBiblio( $record, $biblionumber, GetFrameworkCode($biblionumber) ) };
427 warn "ERROR: Edit biblio $biblionumber failed: $@\n";
428 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ERROR" } ) if ($logfile);
431 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ok" } ) if ($logfile);
434 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "warning : already in database" } ) if ($logfile);
438 eval { ( $biblionumber, $biblioitemnumber ) = AddBiblio( $record, '', { defer_marc_save => 1 } ) };
440 warn "ERROR: Adding biblio $biblionumber failed: $@\n";
441 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ERROR" } ) if ($logfile);
444 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ok" } ) if ($logfile);
447 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "warning : not in database" } ) if ($logfile);
450 eval { ( $itemnumbers_ref, $errors_ref ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
451 my $error_adding = $@;
452 # Work on a clone so that if there are real errors, we can maybe
454 my $clone_record = $record->clone();
455 C4::Biblio::_strip_item_fields($clone_record, '');
456 # This sets the marc fields if there was an error, and also calls
458 ModBiblioMarc( $clone_record, $biblionumber, $framework );
459 if ( $error_adding ) {
460 warn "ERROR: Adding items to bib $biblionumber failed: $error_adding";
461 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
462 # if we failed because of an exception, assume that
463 # the MARC columns in biblioitems were not set.
467 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
469 if ($dedup_barcode && grep { exists $_->{error_code} && $_->{error_code} eq 'duplicate_barcode' } @$errors_ref) {
470 # Find the record called 'barcode'
471 my ($tag, $sub) = C4::Biblio::GetMarcFromKohaField('items.barcode', $framework);
472 # Now remove any items that didn't have a duplicate_barcode error,
473 # erase the barcodes on items that did, and re-add those items.
475 foreach my $i (0 .. $#{$errors_ref}) {
476 my $ref = $errors_ref->[$i];
477 if ($ref && ($ref->{error_code} eq 'duplicate_barcode')) {
478 $dupes{$ref->{item_sequence}} = 1;
479 # Delete the error message because we're going to
481 delete $errors_ref->[$i];
485 foreach my $field ($record->field($tag)) {
488 # Here we remove the barcode
489 $field->delete_subfield(code => $sub);
491 # otherwise we delete the field because we don't want
493 $record->delete_fields($field);
496 # Now re-add the record as before, adding errors to the prev list
498 eval { ( $itemnumbers_ref, $more_errors ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
500 warn "ERROR: Adding items to bib $biblionumber failed: $@\n";
501 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
502 # if we failed because of an exception, assume that
503 # the MARC columns in biblioitems were not set.
504 ModBiblioMarc( $record, $biblionumber, $framework );
507 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
509 push @$errors_ref, @{ $more_errors };
511 if ($#{ $errors_ref } > -1) {
512 report_item_errors($biblionumber, $errors_ref);
514 $yamlhash->{$originalid} = $biblionumber if ($yamlfile);
516 $dbh->commit() if (0 == $i % $commitnum);
518 print $record->as_formatted()."\n" if ($verbose//0)==2;
519 last if $i == $number;
522 $dbh->{AutoCommit} = 1;
526 $dbh->do("SET FOREIGN_KEY_CHECKS = 1");
529 # Restore CataloguingLog
530 C4::Context->set_preference( 'CataloguingLog', $CataloguingLog );
531 # Restore AuthoritiesLog
532 C4::Context->set_preference( 'AuthoritiesLog', $AuthoritiesLog );
534 my $timeneeded = gettimeofday - $starttime;
535 print "\n$i MARC records done in $timeneeded seconds\n";
537 print $loghandle "file : $input_marc_file\n";
538 print $loghandle "$i MARC records done in $timeneeded seconds\n";
542 open my $yamlfileout, q{>}, "$yamlfile" or die "cannot open $yamlfile \n";
543 print $yamlfileout Dump($yamlhash);
548 my $marcrecord=shift;
553 return $marcrecord->field($tag)->data() if $marcrecord->field($tag);
556 if ($marcrecord->field($tag)){
557 return $marcrecord->subfield($tag,$subfield);
566 foreach my $matchingpoint (@$match){
567 my $string = build_simplequery($matchingpoint,$record);
568 push @searchstrings,$string if (length($string)>0);
571 $QParser = C4::Context->queryparser if (C4::Context->preference('UseQueryParser'));
578 return join(" $op ",@searchstrings);
580 sub build_simplequery {
584 my ($index,$recorddata)=split /,/,$element;
585 if ($recorddata=~/(\d{3})(.*)/) {
586 my ($tag,$subfields) =($1,$2);
587 foreach my $field ($record->field($tag)){
588 if (length($field->as_string("$subfields"))>0){
589 push @searchstrings,"$index:\"".$field->as_string("$subfields")."\"";
594 $QParser = C4::Context->queryparser if (C4::Context->preference('UseQueryParser'));
601 return join(" $op ",@searchstrings);
603 sub report_item_errors {
604 my $biblionumber = shift;
605 my $errors_ref = shift;
607 foreach my $error (@{ $errors_ref }) {
609 my $msg = "Item not added (bib $biblionumber, item tag #$error->{'item_sequence'}, barcode $error->{'item_barcode'}): ";
610 my $error_code = $error->{'error_code'};
611 $error_code =~ s/_/ /g;
612 $msg .= "$error_code $error->{'error_information'}";
617 my $logelements=shift;
618 print $loghandle join( ";", map { defined $_ ? $_ : "" } @$logelements{qw<id op status>} ), "\n";
620 sub get_heading_fields{
623 $headingfields=YAML::LoadFile($authtypes);
624 $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
625 $debug && warn YAML::Dump($headingfields);
627 unless ($headingfields){
628 $headingfields=$dbh->selectall_hashref("SELECT auth_tag_to_report, authtypecode from auth_types",'auth_tag_to_report',{Slice=>{}});
629 $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
631 return $headingfields;
636 bulkmarcimport.pl - Import bibliographic/authority records into Koha
640 $ export KOHA_CONF=/etc/koha.conf
641 $ perl misc/migration_tools/bulkmarcimport.pl -d -commit 1000 \\
642 -file /home/jmf/koha.mrc -n 3000
646 Don't use this script before you've entered and checked your MARC parameters
647 tables twice (or more!). Otherwise, the import won't work correctly and you
648 will get invalid data.
656 This version/help screen
658 =item B<-b, -biblios>
660 Type of import: bibliographic records
662 =item B<-a, -authorities>
664 Type of import: authority records
666 =item B<-file>=I<FILE>
668 The I<FILE> to import
672 Verbose mode. 1 means "some infos", 2 means "MARC dumping"
676 Turn off foreign key checks during import.
678 =item B<-n>=I<NUMBER>
680 The I<NUMBER> of records to import. If missing, all the file is imported
682 =item B<-o, -offset>=I<NUMBER>
684 File offset before importing, ie I<NUMBER> of records to skip.
686 =item B<-commit>=I<NUMBER>
688 The I<NUMBER> of records to wait before performing a 'commit' operation
692 File logs actions done for each record and their status into file
696 If specified, data will be appended to the logfile. If not, the logfile will be erased for each execution.
700 Test mode: parses the file, saying what it would do, but doing nothing.
704 Skip automatic conversion of MARC-8 to UTF-8. This option is provided for
707 =item B<-c>=I<CHARACTERISTIC>
709 The I<CHARACTERISTIC> MARC flavour. At the moment, only I<MARC21> and
710 I<UNIMARC> are supported. MARC21 by default.
714 Delete EVERYTHING related to biblio in koha-DB before import. Tables: biblio,
717 =item B<-m>=I<FORMAT>
719 Input file I<FORMAT>: I<MARCXML> or I<ISO2709> (defaults to ISO2709)
723 file yamlfile with authoritiesTypes and distinguishable record field in order
724 to store the correct authtype
728 yaml file format a yaml file with ids
732 list of fields that will not be imported. Can be any from 000 to 999 or field,
733 subfield and subfield's matching value such as 200avalue
737 if set, only insert when possible
741 if set, only updates (any biblio should have a matching record)
745 if set, do whatever is required
747 =item B<-k, -keepids>=<FIELD>
749 Field store ids in I<FIELD> (useful for authorities, where 001 contains the
750 authid for Koha, that can contain a very valuable info for authorities coming
751 from LOC or BNF. useless for biblios probably)
753 =item B<-match>=<FIELD>
755 I<FIELD> matchindex,fieldtomatch matchpoint to use to deduplicate fieldtomatch
756 can be either 001 to 999 or field and list of subfields as such 100abcde
760 If set, a search will be done on isbn, and, if the same isbn is found, the
761 biblio is not added. It's another method to deduplicate. B<-match> & B<-isbn>
766 Clean ISBN fields from entering biblio records, ie removes hyphens. By default,
767 ISBN are cleaned. --nocleanisbn will keep ISBN unchanged.
771 Source bib I<TAG> for reporting the source bib number
773 =item B<-y>=I<SUBFIELD>
775 Source I<SUBFIELD> for reporting the source bib number
777 =item B<-idmap>=I<FILE>
779 I<FILE> for the koha bib and source id
783 Store ids in 009 (useful for authorities, where 001 contains the authid for
784 Koha, that can contain a very valuable info for authorities coming from LOC or
785 BNF. useless for biblios probably)
787 =item B<-dedupbarcode>
789 If set, whenever a duplicate barcode is detected, it is removed and the attempt
790 to add the record is retried, thereby giving the record a blank barcode. This
791 is useful when something has set barcodes to be a biblio ID, or similar
792 (usually other software.)
796 This is the code for the framework that the requested records will have attached
797 to them when they are created. If not specified, then the default framework
800 =item B<-custom>=I<MODULE>
802 This parameter allows you to use a local module with a customize subroutine
803 that is called for each MARC record.
804 If no filename is passed, LocalChanges.pm is assumed to be in the
805 migration_tools subdirectory. You may pass an absolute file name or a file name
806 from the migration_tools directory.