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)){
124 #Disable the syspref cache or else the following UPDATE statements will have no effect
125 C4::Context->disable_syspref_cache();
127 # save the CataloguingLog and AuthoritiesLog properties : we don't want to log a bulkmarcimport. It will slow the import &
128 # will create problems in the action_logs table, that can't handle more than 1 entry per second per user.
129 my $CataloguingLog = C4::Context->preference('CataloguingLog');
130 $dbh->do("UPDATE systempreferences SET value=0 WHERE variable='CataloguingLog'");
132 my $AuthoritiesLog = C4::Context->preference('AuthoritiesLog');
133 $dbh->do("UPDATE systempreferences SET value=0 WHERE variable='AuthoritiesLog'");
136 $dbh->do("SET FOREIGN_KEY_CHECKS = 0");
142 print "deleting biblios\n";
143 $dbh->do("truncate biblio");
144 $dbh->do("truncate biblioitems");
145 $dbh->do("truncate items");
148 print "deleting authorities\n";
149 $dbh->do("truncate auth_header");
151 $dbh->do("truncate zebraqueue");
156 if ($test_parameter) {
157 print "TESTING MODE ONLY\n DOING NOTHING\n===============\n";
160 my $marcFlavour = C4::Context->preference('marcflavour') || 'MARC21';
162 print "Characteristic MARC flavour: $marcFlavour\n" if $verbose;
163 my $starttime = gettimeofday;
165 my $fh = IO::File->new($input_marc_file); # don't let MARC::Batch open the file, as it applies the ':utf8' IO layer
166 if (defined $format && $format =~ /XML/i) {
167 # ugly hack follows -- MARC::File::XML, when used by MARC::Batch,
168 # appears to try to convert incoming XML records from MARC-8
169 # to UTF-8. Setting the BinaryEncoding key turns that off
170 # TODO: see what happens to ISO-8859-1 XML files.
171 # TODO: determine if MARC::Batch can be fixed to handle
172 # XML records properly -- it probably should be
173 # be using a proper push or pull XML parser to
174 # extract the records, not using regexes to look
175 # for <record>.*</record>.
176 $MARC::File::XML::_load_args{BinaryEncoding} = 'utf-8';
177 my $recordformat= ($marcFlavour eq "MARC21"?"USMARC":uc($marcFlavour));
178 #UNIMARC Authorities have a different way to manage encoding than UNIMARC biblios.
179 $recordformat=$recordformat."AUTH" if ($authorities and $marcFlavour ne "MARC21");
180 $MARC::File::XML::_load_args{RecordFormat} = $recordformat;
181 $batch = MARC::Batch->new( 'XML', $fh );
183 $batch = MARC::Batch->new( 'USMARC', $fh );
185 $batch->warnings_off();
186 $batch->strict_off();
188 my $commitnum = $commit ? $commit : 50;
193 print "Skipping file offset: $offset records\n";
194 $batch->next() while ($offset--);
197 my ($tagid,$subfieldid);
202 ( $tagid, $subfieldid ) =
203 GetMarcFromKohaField( "biblio.biblionumber", $framework );
207 # the SQL query to search on isbn
208 my $sth_isbn = $dbh->prepare("SELECT biblionumber,biblioitemnumber FROM biblioitems WHERE isbn=?");
210 $dbh->{AutoCommit} = 0;
213 $loghandle= IO::File->new($logfile, $writemode) ;
214 print $loghandle "id;operation;status\n";
219 eval { $record = $batch->next() };
221 print "Bad MARC record $i: $@ skipped\n";
222 # FIXME - because MARC::Batch->next() combines grabbing the next
223 # blob and parsing it into one operation, a correctable condition
224 # such as a MARC-8 record claiming that it's UTF-8 can't be recovered
225 # from because we don't have access to the original blob. Note
226 # that the staging import can deal with this condition (via
227 # C4::Charset::MarcToUTF8Record) because it doesn't use MARC::Batch.
230 # skip if we get an empty record (that is MARC valid, but will result in AddBiblio failure
231 last unless ( $record );
233 if( ($verbose//1)==1 ) { #no dot for verbose==2
234 print "." . ( $i % 100==0 ? "\n$i" : '' );
237 # transcode the record to UTF8 if needed & applicable.
238 if ($record->encoding() eq 'MARC-8' and not $skip_marc8_conversion) {
239 # FIXME update condition
240 my ($guessed_charset, $charset_errors);
241 ($record, $guessed_charset, $charset_errors) = MarcToUTF8Record($record, $marcFlavour.(($authorities and $marcFlavour ne "MARC21")?'AUTH':''));
242 if ($guessed_charset eq 'failed') {
243 warn "ERROR: failed to perform character conversion for record $i\n";
247 SetUTF8Flag($record);
248 &$localcust($record) if $localcust;
250 # remove trailing - in isbn (only for biblios, of course)
251 if ($biblios && $cleanisbn) {
252 my $tag = $marcFlavour eq 'UNIMARC' ? '010' : '020';
253 my $field = $record->field($tag);
254 my $isbn = $field && $field->subfield('a');
257 $field->update('a' => $isbn);
261 # search for duplicates (based on Local-number)
263 $originalid = GetRecordId( $record, $tagid, $subfieldid );
266 my $query = build_query( $match, $record );
267 my $server = ( $authorities ? 'authorityserver' : 'biblioserver' );
268 $debug && warn $query;
269 my ( $error, $results, $totalhits ) = C4::Search::SimpleSearch( $query, 0, 3, [$server] );
270 die "unable to search the database for duplicates : $error" if ( defined $error );
271 $debug && warn "$query $server : $totalhits";
272 if ( $results && scalar(@$results) == 1 ) {
273 my $marcrecord = C4::Search::new_record_from_zebra( $server, $results->[0] );
274 SetUTF8Flag($marcrecord);
275 $id = GetRecordId( $marcrecord, $tagid, $subfieldid );
276 if ( $authorities && $marcFlavour ) {
277 #Skip if authority in database is the same as the on in database
278 if ( $marcrecord->field('005') && $record->field('005') &&
279 $marcrecord->field('005')->data && $record->field('005')->data &&
280 $marcrecord->field('005')->data >= $record->field('005')->data ) {
282 $yamlhash->{$originalid}->{'authid'} = $id;
284 # we recover all subfields of the heading authorities
286 foreach my $field ( $marcrecord->field("2..") ) {
287 push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
289 $yamlhash->{$originalid}->{'subfields'} = \@subfields;
294 } elsif ( $results && scalar(@$results) > 1 ) {
295 $debug && warn "more than one match for $query";
297 $debug && warn "nomatch for $query";
300 if ($keepids && $originalid) {
302 if ( length($keepids) == 3 ) {
303 $storeidfield = MARC::Field->new( $keepids, $originalid );
305 $storeidfield = MARC::Field->new( substr( $keepids, 0, 3 ), "", "", substr( $keepids, 3, 1 ), $originalid );
307 $record->insert_fields_ordered($storeidfield);
308 $record->delete_field( $record->field($tagid) );
310 foreach my $stringfilter (@$filters) {
311 if ( length($stringfilter) == 3 ) {
312 foreach my $field ( $record->field($stringfilter) ) {
313 $record->delete_field($field);
314 $debug && warn "removed : ", $field->as_string;
316 } elsif ($stringfilter =~ /([0-9]{3})([a-z0-9])(.*)/) {
318 my $removesubfield = $2;
319 my $removematch = $3;
320 if ( ( $removetag > "010" ) && $removesubfield ) {
321 foreach my $field ( $record->field($removetag) ) {
322 $field->delete_subfield( code => "$removesubfield", match => $removematch );
323 $debug && warn "Potentially removed : ", $field->subfield($removesubfield);
328 unless ($test_parameter) {
330 use C4::AuthoritiesMarc;
331 my $authtypecode=GuessAuthTypeCode($record, $heading_fields);
332 my $authid= ($id?$id:GuessAuthId($record));
333 if ($authid && GetAuthority($authid) && $update ){
334 ## Authority has an id and is in database : Replace
335 eval { ( $authid ) = ModAuthority($authid,$record, $authtypecode) };
337 warn "Problem with authority $authid Cannot Modify";
338 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ERROR"}) if ($logfile);
341 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ok"}) if ($logfile);
344 elsif (defined $authid) {
345 ## An authid is defined but no authority in database : add
346 eval { ( $authid ) = AddAuthority($record,$authid, $authtypecode) };
348 warn "Problem with authority $authid Cannot Add ".$@;
349 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
352 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
356 ## True insert in database
357 eval { ( $authid ) = AddAuthority($record,"", $authtypecode) };
359 warn "Problem with authority $authid Cannot Add".$@;
360 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
363 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
367 $yamlhash->{$originalid}->{'authid'} = $authid;
369 foreach my $field ( $record->field("2..") ) {
370 push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
372 $yamlhash->{$originalid}->{'subfields'} = \@subfields;
376 my ( $biblionumber, $biblioitemnumber, $itemnumbers_ref, $errors_ref );
378 # check for duplicate, based on ISBN (skip it if we already have found a duplicate with match parameter
379 if (!$biblionumber && $isbn_check && $isbn) {
380 # warn "search ISBN : $isbn";
381 $sth_isbn->execute($isbn);
382 ($biblionumber,$biblioitemnumber) = $sth_isbn->fetchrow;
384 if (defined $idmapfl) {
385 if ($sourcetag < "010"){
386 if ($record->field($sourcetag)){
387 my $source = $record->field($sourcetag)->data();
388 printf(IDMAP "%s|%s\n",$source,$biblionumber);
391 my $source=$record->subfield($sourcetag,$sourcesubfield);
392 printf(IDMAP "%s|%s\n",$source,$biblionumber);
395 # create biblio, unless we already have it ( either match or isbn )
397 eval{$biblioitemnumber=GetBiblioData($biblionumber)->{biblioitemnumber};};
399 eval { ( $biblionumber, $biblioitemnumber ) = ModBiblio( $record, $biblionumber, GetFrameworkcode($biblionumber) ) };
401 warn "ERROR: Edit biblio $biblionumber failed: $@\n";
402 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ERROR" } ) if ($logfile);
405 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ok" } ) if ($logfile);
408 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "warning : already in database" } ) if ($logfile);
412 eval { ( $biblionumber, $biblioitemnumber ) = AddBiblio( $record, '', { defer_marc_save => 1 } ) };
414 warn "ERROR: Adding biblio $biblionumber failed: $@\n";
415 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ERROR" } ) if ($logfile);
418 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ok" } ) if ($logfile);
421 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "warning : not in database" } ) if ($logfile);
424 eval { ( $itemnumbers_ref, $errors_ref ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
425 my $error_adding = $@;
426 # Work on a clone so that if there are real errors, we can maybe
428 my $clone_record = $record->clone();
429 C4::Biblio::_strip_item_fields($clone_record, '');
430 # This sets the marc fields if there was an error, and also calls
432 ModBiblioMarc( $clone_record, $biblionumber, $framework );
433 if ( $error_adding ) {
434 warn "ERROR: Adding items to bib $biblionumber failed: $error_adding";
435 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
436 # if we failed because of an exception, assume that
437 # the MARC columns in biblioitems were not set.
441 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
443 if ($dedup_barcode && grep { exists $_->{error_code} && $_->{error_code} eq 'duplicate_barcode' } @$errors_ref) {
444 # Find the record called 'barcode'
445 my ($tag, $sub) = C4::Biblio::GetMarcFromKohaField('items.barcode', $framework);
446 # Now remove any items that didn't have a duplicate_barcode error,
447 # erase the barcodes on items that did, and re-add those items.
449 foreach my $i (0 .. $#{$errors_ref}) {
450 my $ref = $errors_ref->[$i];
451 if ($ref && ($ref->{error_code} eq 'duplicate_barcode')) {
452 $dupes{$ref->{item_sequence}} = 1;
453 # Delete the error message because we're going to
455 delete $errors_ref->[$i];
459 foreach my $field ($record->field($tag)) {
462 # Here we remove the barcode
463 $field->delete_subfield(code => $sub);
465 # otherwise we delete the field because we don't want
467 $record->delete_fields($field);
470 # Now re-add the record as before, adding errors to the prev list
472 eval { ( $itemnumbers_ref, $more_errors ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
474 warn "ERROR: Adding items to bib $biblionumber failed: $@\n";
475 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
476 # if we failed because of an exception, assume that
477 # the MARC columns in biblioitems were not set.
478 ModBiblioMarc( $record, $biblionumber, $framework );
481 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
483 push @$errors_ref, @{ $more_errors };
485 if ($#{ $errors_ref } > -1) {
486 report_item_errors($biblionumber, $errors_ref);
488 $yamlhash->{$originalid} = $biblionumber if ($yamlfile);
490 $dbh->commit() if (0 == $i % $commitnum);
492 print $record->as_formatted()."\n" if ($verbose//0)==2;
493 last if $i == $number;
496 $dbh->{AutoCommit} = 1;
500 $dbh->do("SET FOREIGN_KEY_CHECKS = 1");
503 # restore CataloguingLog
504 $dbh->do("UPDATE systempreferences SET value=$CataloguingLog WHERE variable='CataloguingLog'");
505 # restore AuthoritiesLog
506 $dbh->do("UPDATE systempreferences SET value=$AuthoritiesLog WHERE variable='AuthoritiesLog'");
508 my $timeneeded = gettimeofday - $starttime;
509 print "\n$i MARC records done in $timeneeded seconds\n";
511 print $loghandle "file : $input_marc_file\n";
512 print $loghandle "$i MARC records done in $timeneeded seconds\n";
516 open my $yamlfileout, q{>}, "$yamlfile" or die "cannot open $yamlfile \n";
517 print $yamlfileout Dump($yamlhash);
522 my $marcrecord=shift;
527 return $marcrecord->field($tag)->data() if $marcrecord->field($tag);
530 if ($marcrecord->field($tag)){
531 return $marcrecord->subfield($tag,$subfield);
540 foreach my $matchingpoint (@$match){
541 my $string = build_simplequery($matchingpoint,$record);
542 push @searchstrings,$string if (length($string)>0);
545 $QParser = C4::Context->queryparser if (C4::Context->preference('UseQueryParser'));
552 return join(" $op ",@searchstrings);
554 sub build_simplequery {
558 my ($index,$recorddata)=split /,/,$element;
559 if ($recorddata=~/(\d{3})(.*)/) {
560 my ($tag,$subfields) =($1,$2);
561 foreach my $field ($record->field($tag)){
562 if (length($field->as_string("$subfields"))>0){
563 push @searchstrings,"$index:\"".$field->as_string("$subfields")."\"";
568 $QParser = C4::Context->queryparser if (C4::Context->preference('UseQueryParser'));
575 return join(" $op ",@searchstrings);
577 sub report_item_errors {
578 my $biblionumber = shift;
579 my $errors_ref = shift;
581 foreach my $error (@{ $errors_ref }) {
583 my $msg = "Item not added (bib $biblionumber, item tag #$error->{'item_sequence'}, barcode $error->{'item_barcode'}): ";
584 my $error_code = $error->{'error_code'};
585 $error_code =~ s/_/ /g;
586 $msg .= "$error_code $error->{'error_information'}";
591 my $logelements=shift;
592 print $loghandle join( ";", map { defined $_ ? $_ : "" } @$logelements{qw<id op status>} ), "\n";
594 sub get_heading_fields{
597 $headingfields=YAML::LoadFile($authtypes);
598 $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
599 $debug && warn YAML::Dump($headingfields);
601 unless ($headingfields){
602 $headingfields=$dbh->selectall_hashref("SELECT auth_tag_to_report, authtypecode from auth_types",'auth_tag_to_report',{Slice=>{}});
603 $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
605 return $headingfields;
610 bulkmarcimport.pl - Import bibliographic/authority records into Koha
614 $ export KOHA_CONF=/etc/koha.conf
615 $ perl misc/migration_tools/bulkmarcimport.pl -d -commit 1000 \\
616 -file /home/jmf/koha.mrc -n 3000
620 Don't use this script before you've entered and checked your MARC parameters
621 tables twice (or more!). Otherwise, the import won't work correctly and you
622 will get invalid data.
630 This version/help screen
632 =item B<-b, -biblios>
634 Type of import: bibliographic records
636 =item B<-a, -authorities>
638 Type of import: authority records
640 =item B<-file>=I<FILE>
642 The I<FILE> to import
646 Verbose mode. 1 means "some infos", 2 means "MARC dumping"
650 Turn off foreign key checks during import.
652 =item B<-n>=I<NUMBER>
654 The I<NUMBER> of records to import. If missing, all the file is imported
656 =item B<-o, -offset>=I<NUMBER>
658 File offset before importing, ie I<NUMBER> of records to skip.
660 =item B<-commit>=I<NUMBER>
662 The I<NUMBER> of records to wait before performing a 'commit' operation
666 File logs actions done for each record and their status into file
670 If specified, data will be appended to the logfile. If not, the logfile will be erased for each execution.
674 Test mode: parses the file, saying what he would do, but doing nothing.
678 Skip automatic conversion of MARC-8 to UTF-8. This option is provided for
681 =item B<-c>=I<CHARACTERISTIC>
683 The I<CHARACTERISTIC> MARC flavour. At the moment, only I<MARC21> and
684 I<UNIMARC> are supported. MARC21 by default.
688 Delete EVERYTHING related to biblio in koha-DB before import. Tables: biblio,
691 =item B<-m>=I<FORMAT>
693 Input file I<FORMAT>: I<MARCXML> or I<ISO2709> (defaults to ISO2709)
697 file yamlfile with authoritiesTypes and distinguishable record field in order
698 to store the correct authtype
702 yaml file format a yaml file with ids
706 list of fields that will not be imported. Can be any from 000 to 999 or field,
707 subfield and subfield's matching value such as 200avalue
711 if set, only insert when possible
715 if set, only updates (any biblio should have a matching record)
719 if set, do whatever is required
721 =item B<-k, -keepids>=<FIELD>
723 Field store ids in I<FIELD> (usefull for authorities, where 001 contains the
724 authid for Koha, that can contain a very valuable info for authorities coming
725 from LOC or BNF. useless for biblios probably)
727 =item B<-match>=<FIELD>
729 I<FIELD> matchindex,fieldtomatch matchpoint to use to deduplicate fieldtomatch
730 can be either 001 to 999 or field and list of subfields as such 100abcde
734 If set, a search will be done on isbn, and, if the same isbn is found, the
735 biblio is not added. It's another method to deduplicate. B<-match> & B<-isbn>
740 Clean ISBN fields from entering biblio records, ie removes hyphens. By default,
741 ISBN are cleaned. --nocleanisbn will keep ISBN unchanged.
745 Source bib I<TAG> for reporting the source bib number
747 =item B<-y>=I<SUBFIELD>
749 Source I<SUBFIELD> for reporting the source bib number
751 =item B<-idmap>=I<FILE>
753 I<FILE> for the koha bib and source id
757 Store ids in 009 (usefull for authorities, where 001 contains the authid for
758 Koha, that can contain a very valuable info for authorities coming from LOC or
759 BNF. useless for biblios probably)
761 =item B<-dedupbarcode>
763 If set, whenever a duplicate barcode is detected, it is removed and the attempt
764 to add the record is retried, thereby giving the record a blank barcode. This
765 is useful when something has set barcodes to be a biblio ID, or similar
766 (usually other software.)
770 This is the code for the framework that the requested records will have attached
771 to them when they are created. If not specified, then the default framework
774 =item B<-custom>=I<MODULE>
776 This parameter allows you to use a local module with a customize subroutine
777 that is called for each MARC record.
778 If no filename is passed, LocalChanges.pm is assumed to be in the
779 migration_tools subdirectory. You may pass an absolute file name or a file name
780 from the migration_tools directory.