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);
46 'commit:f' => \$commit,
47 'file:s' => \$input_marc_file,
49 'o|offset:f' => \$offset,
52 't|test' => \$test_parameter,
53 's' => \$skip_marc8_conversion,
54 'c:s' => \$char_encoding,
60 'k|keepids:s' => \$keepids,
61 'b|biblios' => \$biblios,
62 'a|authorities' => \$authorities,
63 'authtypes:s' => \$authtypes,
64 'filter=s@' => \$filters,
68 'match=s@' => \$match,
69 'i|isbn' => \$isbn_check,
71 'y:s' => \$sourcesubfield,
72 'idmap:s' => \$idmapfl,
73 'cleanisbn!' => \$cleanisbn,
74 'yaml:s' => \$yamlfile,
75 'dedupbarcode' => \$dedup_barcode,
76 'framework=s' => \$framework,
78 $biblios ||= !$authorities;
80 my $writemode = ($append) ? "a" : "w";
87 if ($version || ($input_marc_file eq '')) {
88 pod2usage( -verbose => 2 );
92 my $dbh = C4::Context->dbh;
93 my $heading_fields=get_heading_fields();
95 if (defined $idmapfl) {
96 open(IDMAP,">$idmapfl") or die "cannot open $idmapfl \n";
99 if ((not defined $sourcesubfield) && (not defined $sourcetag)){
104 # save the CataloguingLog property : we don't want to log a bulkmarcimport. It will slow the import &
105 # will create problems in the action_logs table, that can't handle more than 1 entry per second per user.
106 my $CataloguingLog = C4::Context->preference('CataloguingLog');
107 $dbh->do("UPDATE systempreferences SET value=0 WHERE variable='CataloguingLog'");
110 $dbh->do("SET FOREIGN_KEY_CHECKS = 0");
116 print "deleting biblios\n";
117 $dbh->do("truncate biblio");
118 $dbh->do("truncate biblioitems");
119 $dbh->do("truncate items");
122 print "deleting authorities\n";
123 $dbh->do("truncate auth_header");
125 $dbh->do("truncate zebraqueue");
130 if ($test_parameter) {
131 print "TESTING MODE ONLY\n DOING NOTHING\n===============\n";
134 my $marcFlavour = C4::Context->preference('marcflavour') || 'MARC21';
136 print "Characteristic MARC flavour: $marcFlavour\n" if $verbose;
137 my $starttime = gettimeofday;
139 my $fh = IO::File->new($input_marc_file); # don't let MARC::Batch open the file, as it applies the ':utf8' IO layer
140 if (defined $format && $format =~ /XML/i) {
141 # ugly hack follows -- MARC::File::XML, when used by MARC::Batch,
142 # appears to try to convert incoming XML records from MARC-8
143 # to UTF-8. Setting the BinaryEncoding key turns that off
144 # TODO: see what happens to ISO-8859-1 XML files.
145 # TODO: determine if MARC::Batch can be fixed to handle
146 # XML records properly -- it probably should be
147 # be using a proper push or pull XML parser to
148 # extract the records, not using regexes to look
149 # for <record>.*</record>.
150 $MARC::File::XML::_load_args{BinaryEncoding} = 'utf-8';
151 my $recordformat= ($marcFlavour eq "MARC21"?"USMARC":uc($marcFlavour));
152 #UNIMARC Authorities have a different way to manage encoding than UNIMARC biblios.
153 $recordformat=$recordformat."AUTH" if ($authorities and $marcFlavour ne "MARC21");
154 $MARC::File::XML::_load_args{RecordFormat} = $recordformat;
155 $batch = MARC::Batch->new( 'XML', $fh );
157 $batch = MARC::Batch->new( 'USMARC', $fh );
159 $batch->warnings_off();
160 $batch->strict_off();
162 my $commitnum = $commit ? $commit : 50;
167 print "Skipping file offset: $offset records\n";
168 $batch->next() while ($offset--);
171 my ($tagid,$subfieldid);
176 ( $tagid, $subfieldid ) =
177 GetMarcFromKohaField( "biblio.biblionumber", $framework );
181 # the SQL query to search on isbn
182 my $sth_isbn = $dbh->prepare("SELECT biblionumber,biblioitemnumber FROM biblioitems WHERE isbn=?");
184 $dbh->{AutoCommit} = 0;
187 $loghandle= IO::File->new($logfile, $writemode) ;
188 print $loghandle "id;operation;status\n";
193 eval { $record = $batch->next() };
195 print "Bad MARC record $i: $@ skipped\n";
196 # FIXME - because MARC::Batch->next() combines grabbing the next
197 # blob and parsing it into one operation, a correctable condition
198 # such as a MARC-8 record claiming that it's UTF-8 can't be recovered
199 # from because we don't have access to the original blob. Note
200 # that the staging import can deal with this condition (via
201 # C4::Charset::MarcToUTF8Record) because it doesn't use MARC::Batch.
204 # skip if we get an empty record (that is MARC valid, but will result in AddBiblio failure
205 last unless ( $record );
208 print "\n$i" unless $i % 100;
210 # transcode the record to UTF8 if needed & applicable.
211 if ($record->encoding() eq 'MARC-8' and not $skip_marc8_conversion) {
212 # FIXME update condition
213 my ($guessed_charset, $charset_errors);
214 ($record, $guessed_charset, $charset_errors) = MarcToUTF8Record($record, $marcFlavour.(($authorities and $marcFlavour ne "MARC21")?'AUTH':''));
215 if ($guessed_charset eq 'failed') {
216 warn "ERROR: failed to perform character conversion for record $i\n";
220 SetUTF8Flag($record);
222 # remove trailing - in isbn (only for biblios, of course)
223 if ($biblios && $cleanisbn) {
224 my $tag = $marcFlavour eq 'UNIMARC' ? '010' : '020';
225 my $field = $record->field($tag);
226 my $isbn = $field && $field->subfield('a');
229 $field->update('a' => $isbn);
233 # search for duplicates (based on Local-number)
235 $originalid = GetRecordId( $record, $tagid, $subfieldid );
238 my $query = build_query( $match, $record );
239 my $server = ( $authorities ? 'authorityserver' : 'biblioserver' );
240 $debug && warn $query;
241 my ( $error, $results, $totalhits ) = C4::Search::SimpleSearch( $query, 0, 3, [$server] );
242 die "unable to search the database for duplicates : $error" if ( defined $error );
243 $debug && warn "$query $server : $totalhits";
244 if ( $results && scalar(@$results) == 1 ) {
245 my $marcrecord = C4::Search::new_record_from_zebra( $server, $results->[0] );
246 SetUTF8Flag($marcrecord);
247 $id = GetRecordId( $marcrecord, $tagid, $subfieldid );
248 if ( $authorities && $marcFlavour ) {
249 #Skip if authority in database is the same as the on in database
250 if ( $marcrecord->field('005') && $record->field('005') &&
251 $marcrecord->field('005')->data && $record->field('005')->data &&
252 $marcrecord->field('005')->data >= $record->field('005')->data ) {
254 $yamlhash->{$originalid}->{'authid'} = $id;
256 # we recover all subfields of the heading authorities
258 foreach my $field ( $marcrecord->field("2..") ) {
259 push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
261 $yamlhash->{$originalid}->{'subfields'} = \@subfields;
266 } elsif ( $results && scalar(@$results) > 1 ) {
267 $debug && warn "more than one match for $query";
269 $debug && warn "nomatch for $query";
272 if ($keepids && $originalid) {
274 if ( length($keepids) == 3 ) {
275 $storeidfield = MARC::Field->new( $keepids, $originalid );
277 $storeidfield = MARC::Field->new( substr( $keepids, 0, 3 ), "", "", substr( $keepids, 3, 1 ), $originalid );
279 $record->insert_fields_ordered($storeidfield);
280 $record->delete_field( $record->field($tagid) );
282 foreach my $stringfilter (@$filters) {
283 if ( length($stringfilter) == 3 ) {
284 foreach my $field ( $record->field($stringfilter) ) {
285 $record->delete_field($field);
286 $debug && warn "removed : ", $field->as_string;
288 } elsif ($stringfilter =~ /([0-9]{3})([a-z0-9])(.*)/) {
290 my $removesubfield = $2;
291 my $removematch = $3;
292 if ( ( $removetag > "010" ) && $removesubfield ) {
293 foreach my $field ( $record->field($removetag) ) {
294 $field->delete_subfield( code => "$removesubfield", match => $removematch );
295 $debug && warn "Potentially removed : ", $field->subfield($removesubfield);
300 unless ($test_parameter) {
302 use C4::AuthoritiesMarc;
303 my $authtypecode=GuessAuthTypeCode($record, $heading_fields);
304 my $authid= ($id?$id:GuessAuthId($record));
305 if ($authid && GetAuthority($authid) && $update ){
306 ## Authority has an id and is in database : Replace
307 eval { ( $authid ) = ModAuthority($authid,$record, $authtypecode) };
309 warn "Problem with authority $authid Cannot Modify";
310 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ERROR"}) if ($logfile);
313 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ok"}) if ($logfile);
316 elsif (defined $authid) {
317 ## An authid is defined but no authority in database : add
318 eval { ( $authid ) = AddAuthority($record,$authid, $authtypecode) };
320 warn "Problem with authority $authid Cannot Add ".$@;
321 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
324 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
328 ## True insert in database
329 eval { ( $authid ) = AddAuthority($record,"", $authtypecode) };
331 warn "Problem with authority $authid Cannot Add".$@;
332 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
335 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
339 $yamlhash->{$originalid}->{'authid'} = $authid;
341 foreach my $field ( $record->field("2..") ) {
342 push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
344 $yamlhash->{$originalid}->{'subfields'} = \@subfields;
348 my ( $biblionumber, $biblioitemnumber, $itemnumbers_ref, $errors_ref );
350 # check for duplicate, based on ISBN (skip it if we already have found a duplicate with match parameter
351 if (!$biblionumber && $isbn_check && $isbn) {
352 # warn "search ISBN : $isbn";
353 $sth_isbn->execute($isbn);
354 ($biblionumber,$biblioitemnumber) = $sth_isbn->fetchrow;
356 if (defined $idmapfl) {
357 if ($sourcetag < "010"){
358 if ($record->field($sourcetag)){
359 my $source = $record->field($sourcetag)->data();
360 printf(IDMAP "%s|%s\n",$source,$biblionumber);
363 my $source=$record->subfield($sourcetag,$sourcesubfield);
364 printf(IDMAP "%s|%s\n",$source,$biblionumber);
367 # create biblio, unless we already have it ( either match or isbn )
369 eval{$biblioitemnumber=GetBiblioData($biblionumber)->{biblioitemnumber};};
371 eval { ( $biblionumber, $biblioitemnumber ) = ModBiblio( $record, $biblionumber, GetFrameworkcode($biblionumber) ) };
373 warn "ERROR: Edit biblio $biblionumber failed: $@\n";
374 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ERROR" } ) if ($logfile);
377 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ok" } ) if ($logfile);
380 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "warning : already in database" } ) if ($logfile);
384 eval { ( $biblionumber, $biblioitemnumber ) = AddBiblio( $record, '', { defer_marc_save => 1 } ) };
386 warn "ERROR: Adding biblio $biblionumber failed: $@\n";
387 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ERROR" } ) if ($logfile);
390 printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ok" } ) if ($logfile);
393 printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "warning : not in database" } ) if ($logfile);
396 eval { ( $itemnumbers_ref, $errors_ref ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
397 my $error_adding = $@;
398 # Work on a clone so that if there are real errors, we can maybe
400 my $clone_record = $record->clone();
401 C4::Biblio::_strip_item_fields($clone_record, '');
402 # This sets the marc fields if there was an error, and also calls
404 ModBiblioMarc( $clone_record, $biblionumber, $framework );
405 if ( $error_adding ) {
406 warn "ERROR: Adding items to bib $biblionumber failed: $error_adding";
407 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
408 # if we failed because of an exception, assume that
409 # the MARC columns in biblioitems were not set.
413 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
415 if ($dedup_barcode && grep { exists $_->{error_code} && $_->{error_code} eq 'duplicate_barcode' } @$errors_ref) {
416 # Find the record called 'barcode'
417 my ($tag, $sub) = C4::Biblio::GetMarcFromKohaField('items.barcode', $framework);
418 # Now remove any items that didn't have a duplicate_barcode error,
419 # erase the barcodes on items that did, and re-add those items.
421 foreach my $i (0 .. $#{$errors_ref}) {
422 my $ref = $errors_ref->[$i];
423 if ($ref && ($ref->{error_code} eq 'duplicate_barcode')) {
424 $dupes{$ref->{item_sequence}} = 1;
425 # Delete the error message because we're going to
427 delete $errors_ref->[$i];
431 foreach my $field ($record->field($tag)) {
434 # Here we remove the barcode
435 $field->delete_subfield(code => $sub);
437 # otherwise we delete the field because we don't want
439 $record->delete_fields($field);
442 # Now re-add the record as before, adding errors to the prev list
444 eval { ( $itemnumbers_ref, $more_errors ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
446 warn "ERROR: Adding items to bib $biblionumber failed: $@\n";
447 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
448 # if we failed because of an exception, assume that
449 # the MARC columns in biblioitems were not set.
450 ModBiblioMarc( $record, $biblionumber, $framework );
453 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
455 push @$errors_ref, @{ $more_errors };
457 if ($#{ $errors_ref } > -1) {
458 report_item_errors($biblionumber, $errors_ref);
460 $yamlhash->{$originalid} = $biblionumber if ($yamlfile);
462 $dbh->commit() if (0 == $i % $commitnum);
464 last if $i == $number;
467 $dbh->{AutoCommit} = 1;
471 $dbh->do("SET FOREIGN_KEY_CHECKS = 1");
474 # restore CataloguingLog
475 $dbh->do("UPDATE systempreferences SET value=$CataloguingLog WHERE variable='CataloguingLog'");
477 my $timeneeded = gettimeofday - $starttime;
478 print "\n$i MARC records done in $timeneeded seconds\n";
480 print $loghandle "file : $input_marc_file\n";
481 print $loghandle "$i MARC records done in $timeneeded seconds\n";
485 open my $yamlfileout, q{>}, "$yamlfile" or die "cannot open $yamlfile \n";
486 print $yamlfileout Dump($yamlhash);
491 my $marcrecord=shift;
496 return $marcrecord->field($tag)->data() if $marcrecord->field($tag);
499 if ($marcrecord->field($tag)){
500 return $marcrecord->subfield($tag,$subfield);
509 foreach my $matchingpoint (@$match){
510 my $string = build_simplequery($matchingpoint,$record);
511 push @searchstrings,$string if (length($string)>0);
514 $QParser = C4::Context->queryparser if (C4::Context->preference('UseQueryParser'));
521 return join(" $op ",@searchstrings);
523 sub build_simplequery {
527 my ($index,$recorddata)=split /,/,$element;
528 if ($recorddata=~/(\d{3})(.*)/) {
529 my ($tag,$subfields) =($1,$2);
530 foreach my $field ($record->field($tag)){
531 if (length($field->as_string("$subfields"))>0){
532 push @searchstrings,"$index:\"".$field->as_string("$subfields")."\"";
537 $QParser = C4::Context->queryparser if (C4::Context->preference('UseQueryParser'));
544 return join(" $op ",@searchstrings);
546 sub report_item_errors {
547 my $biblionumber = shift;
548 my $errors_ref = shift;
550 foreach my $error (@{ $errors_ref }) {
552 my $msg = "Item not added (bib $biblionumber, item tag #$error->{'item_sequence'}, barcode $error->{'item_barcode'}): ";
553 my $error_code = $error->{'error_code'};
554 $error_code =~ s/_/ /g;
555 $msg .= "$error_code $error->{'error_information'}";
560 my $logelements=shift;
561 print $loghandle join( ";", map { defined $_ ? $_ : "" } @$logelements{qw<id op status>} ), "\n";
563 sub get_heading_fields{
566 $headingfields=YAML::LoadFile($authtypes);
567 $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
568 $debug && warn YAML::Dump($headingfields);
570 unless ($headingfields){
571 $headingfields=$dbh->selectall_hashref("SELECT auth_tag_to_report, authtypecode from auth_types",'auth_tag_to_report',{Slice=>{}});
572 $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
574 return $headingfields;
579 bulkmarcimport.pl - Import bibliographic/authority records into Koha
583 $ export KOHA_CONF=/etc/koha.conf
584 $ perl misc/migration_tools/bulkmarcimport.pl -d -commit 1000 \\
585 -file /home/jmf/koha.mrc -n 3000
589 Don't use this script before you've entered and checked your MARC parameters
590 tables twice (or more!). Otherwise, the import won't work correctly and you
591 will get invalid data.
599 This version/help screen
601 =item B<-b, -biblios>
603 Type of import: bibliographic records
605 =item B<-a, -authorities>
607 Type of import: authority records
609 =item B<-file>=I<FILE>
611 The I<FILE> to import
615 Verbose mode. 1 means "some infos", 2 means "MARC dumping"
619 Turn off foreign key checks during import.
621 =item B<-n>=I<NUMBER>
623 The I<NUMBER> of records to import. If missing, all the file is imported
625 =item B<-o, -offset>=I<NUMBER>
627 File offset before importing, ie I<NUMBER> of records to skip.
629 =item B<-commit>=I<NUMBER>
631 The I<NUMBER> of records to wait before performing a 'commit' operation
635 File logs actions done for each record and their status into file
639 If specified, data will be appended to the logfile. If not, the logfile will be erased for each execution.
643 Test mode: parses the file, saying what he would do, but doing nothing.
647 Skip automatic conversion of MARC-8 to UTF-8. This option is provided for
650 =item B<-c>=I<CHARACTERISTIC>
652 The I<CHARACTERISTIC> MARC flavour. At the moment, only I<MARC21> and
653 I<UNIMARC> are supported. MARC21 by default.
657 Delete EVERYTHING related to biblio in koha-DB before import. Tables: biblio,
660 =item B<-m>=I<FORMAT>
662 Input file I<FORMAT>: I<MARCXML> or I<ISO2709> (defaults to ISO2709)
666 file yamlfile with authoritiesTypes and distinguishable record field in order
667 to store the correct authtype
671 yaml file format a yaml file with ids
675 list of fields that will not be imported. Can be any from 000 to 999 or field,
676 subfield and subfield's matching value such as 200avalue
680 if set, only insert when possible
684 if set, only updates (any biblio should have a matching record)
688 if set, do whatever is required
690 =item B<-k, -keepids>=<FIELD>
692 Field store ids in I<FIELD> (usefull for authorities, where 001 contains the
693 authid for Koha, that can contain a very valuable info for authorities coming
694 from LOC or BNF. useless for biblios probably)
696 =item B<-match>=<FIELD>
698 I<FIELD> matchindex,fieldtomatch matchpoint to use to deduplicate fieldtomatch
699 can be either 001 to 999 or field and list of subfields as such 100abcde
703 If set, a search will be done on isbn, and, if the same isbn is found, the
704 biblio is not added. It's another method to deduplicate. B<-match> & B<-isbn>
709 Clean ISBN fields from entering biblio records, ie removes hyphens. By default,
710 ISBN are cleaned. --nocleanisbn will keep ISBN unchanged.
714 Source bib I<TAG> for reporting the source bib number
716 =item B<-y>=I<SUBFIELD>
718 Source I<SUBFIELD> for reporting the source bib number
720 =item B<-idmap>=I<FILE>
722 I<FILE> for the koha bib and source id
726 Store ids in 009 (usefull for authorities, where 001 contains the authid for
727 Koha, that can contain a very valuable info for authorities coming from LOC or
728 BNF. useless for biblios probably)
730 =item B<-dedupbarcode>
732 If set, whenever a duplicate barcode is detected, it is removed and the attempt
733 to add the record is retried, thereby giving the record a blank barcode. This
734 is useful when something has set barcodes to be a biblio ID, or similar
735 (usually other software.)
739 This is the code for the framework that the requested records will have attached
740 to them when they are created. If not specified, then the default framework