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, $authtypes );
40 my ($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,
59 'k|keepids:s' => \
$keepids,
60 'b|biblios' => \
$biblios,
61 'a|authorities' => \
$authorities,
62 'authtypes:s' => \
$authtypes,
63 'filter=s@' => \
$filters,
67 'match=s@' => \
$match,
68 'i|isbn' => \
$isbn_check,
70 'y:s' => \
$sourcesubfield,
71 'idmap:s' => \
$idmapfl,
72 'cleanisbn!' => \
$cleanisbn,
73 'yaml:s' => \
$yamlfile,
74 'dedupbarcode' => \
$dedup_barcode,
75 'framework=s' => \
$framework,
77 $biblios ||= !$authorities;
85 if ($version || ($input_marc_file eq '')) {
86 pod2usage
( -verbose
=> 2 );
90 my $dbh = C4
::Context
->dbh;
91 my $heading_fields=get_heading_fields
();
93 if (defined $idmapfl) {
94 open(IDMAP
,">$idmapfl") or die "cannot open $idmapfl \n";
97 if ((not defined $sourcesubfield) && (not defined $sourcetag)){
102 # save the CataloguingLog property : we don't want to log a bulkmarcimport. It will slow the import &
103 # will create problems in the action_logs table, that can't handle more than 1 entry per second per user.
104 my $CataloguingLog = C4
::Context
->preference('CataloguingLog');
105 $dbh->do("UPDATE systempreferences SET value=0 WHERE variable='CataloguingLog'");
108 $dbh->do("SET FOREIGN_KEY_CHECKS = 0");
114 print "deleting biblios\n";
115 $dbh->do("truncate biblio");
116 $dbh->do("truncate biblioitems");
117 $dbh->do("truncate items");
120 print "deleting authorities\n";
121 $dbh->do("truncate auth_header");
123 $dbh->do("truncate zebraqueue");
128 if ($test_parameter) {
129 print "TESTING MODE ONLY\n DOING NOTHING\n===============\n";
132 my $marcFlavour = C4
::Context
->preference('marcflavour') || 'MARC21';
134 print "Characteristic MARC flavour: $marcFlavour\n" if $verbose;
135 my $starttime = gettimeofday
;
137 my $fh = IO
::File
->new($input_marc_file); # don't let MARC::Batch open the file, as it applies the ':utf8' IO layer
138 if (defined $format && $format =~ /XML/i) {
139 # ugly hack follows -- MARC::File::XML, when used by MARC::Batch,
140 # appears to try to convert incoming XML records from MARC-8
141 # to UTF-8. Setting the BinaryEncoding key turns that off
142 # TODO: see what happens to ISO-8859-1 XML files.
143 # TODO: determine if MARC::Batch can be fixed to handle
144 # XML records properly -- it probably should be
145 # be using a proper push or pull XML parser to
146 # extract the records, not using regexes to look
147 # for <record>.*</record>.
148 $MARC::File
::XML
::_load_args
{BinaryEncoding
} = 'utf-8';
149 my $recordformat= ($marcFlavour eq "MARC21"?
"USMARC":uc($marcFlavour));
150 #UNIMARC Authorities have a different way to manage encoding than UNIMARC biblios.
151 $recordformat=$recordformat."AUTH" if ($authorities and $marcFlavour ne "MARC21");
152 $MARC::File
::XML
::_load_args
{RecordFormat
} = $recordformat;
153 $batch = MARC
::Batch
->new( 'XML', $fh );
155 $batch = MARC
::Batch
->new( 'USMARC', $fh );
157 $batch->warnings_off();
158 $batch->strict_off();
160 my $commitnum = $commit ?
$commit : 50;
165 print "Skipping file offset: $offset records\n";
166 $batch->next() while ($offset--);
169 my ($tagid,$subfieldid);
174 ( $tagid, $subfieldid ) =
175 GetMarcFromKohaField
( "biblio.biblionumber", $framework );
179 # the SQL query to search on isbn
180 my $sth_isbn = $dbh->prepare("SELECT biblionumber,biblioitemnumber FROM biblioitems WHERE isbn=?");
182 $dbh->{AutoCommit
} = 0;
185 $loghandle= IO
::File
->new($logfile,"w") ;
186 print $loghandle "id;operation;status\n";
191 eval { $record = $batch->next() };
193 print "Bad MARC record $i: $@ skipped\n";
194 # FIXME - because MARC::Batch->next() combines grabbing the next
195 # blob and parsing it into one operation, a correctable condition
196 # such as a MARC-8 record claiming that it's UTF-8 can't be recovered
197 # from because we don't have access to the original blob. Note
198 # that the staging import can deal with this condition (via
199 # C4::Charset::MarcToUTF8Record) because it doesn't use MARC::Batch.
202 # skip if we get an empty record (that is MARC valid, but will result in AddBiblio failure
203 last unless ( $record );
206 print "\n$i" unless $i % 100;
208 # transcode the record to UTF8 if needed & applicable.
209 if ($record->encoding() eq 'MARC-8' and not $skip_marc8_conversion) {
210 # FIXME update condition
211 my ($guessed_charset, $charset_errors);
212 ($record, $guessed_charset, $charset_errors) = MarcToUTF8Record
($record, $marcFlavour.(($authorities and $marcFlavour ne "MARC21")?
'AUTH':''));
213 if ($guessed_charset eq 'failed') {
214 warn "ERROR: failed to perform character conversion for record $i\n";
218 SetUTF8Flag
($record);
220 # remove trailing - in isbn (only for biblios, of course)
221 if ($biblios && $cleanisbn) {
222 my $tag = $marcFlavour eq 'UNIMARC' ?
'010' : '020';
223 my $field = $record->field($tag);
224 my $isbn = $field && $field->subfield('a');
227 $field->update('a' => $isbn);
231 # search for duplicates (based on Local-number)
233 $originalid = GetRecordId
( $record, $tagid, $subfieldid );
236 my $query = build_query
( $match, $record );
237 my $server = ( $authorities ?
'authorityserver' : 'biblioserver' );
238 $debug && warn $query;
239 my ( $error, $results, $totalhits ) = C4
::Search
::SimpleSearch
( $query, 0, 3, [$server] );
240 die "unable to search the database for duplicates : $error" if ( defined $error );
241 $debug && warn "$query $server : $totalhits";
242 if ( $results && scalar(@
$results) == 1 ) {
243 my $marcrecord = MARC
::File
::USMARC
::decode
( $results->[0] );
244 SetUTF8Flag
($marcrecord);
245 $id = GetRecordId
( $marcrecord, $tagid, $subfieldid );
246 if ( $authorities && $marcFlavour ) {
247 #Skip if authority in database is the same as the on in database
248 if ( $marcrecord->field('005')->data >= $record->field('005')->data ) {
250 $yamlhash->{$originalid}->{'authid'} = $id;
252 # we recover all subfields of the heading authorities
254 foreach my $field ( $marcrecord->field("2..") ) {
255 push @subfields, map { ( $_->[0] =~ /[a-z]/ ?
$_->[1] : () ) } $field->subfields();
257 $yamlhash->{$originalid}->{'subfields'} = \
@subfields;
262 } elsif ( $results && scalar(@
$results) > 1 ) {
263 $debug && warn "more than one match for $query";
265 $debug && warn "nomatch for $query";
268 if ($keepids && $originalid) {
270 if ( length($keepids) == 3 ) {
271 $storeidfield = MARC
::Field
->new( $keepids, $originalid );
273 $storeidfield = MARC
::Field
->new( substr( $keepids, 0, 3 ), "", "", substr( $keepids, 3, 1 ), $originalid );
275 $record->insert_fields_ordered($storeidfield);
276 $record->delete_field( $record->field($tagid) );
278 foreach my $stringfilter (@
$filters) {
279 if ( length($stringfilter) == 3 ) {
280 foreach my $field ( $record->field($stringfilter) ) {
281 $record->delete_field($field);
282 $debug && warn "removed : ", $field->as_string;
284 } elsif ($stringfilter =~ /([0-9]{3})([a-z0-9])(.*)/) {
286 my $removesubfield = $2;
287 my $removematch = $3;
288 if ( ( $removetag > "010" ) && $removesubfield ) {
289 foreach my $field ( $record->field($removetag) ) {
290 $field->delete_subfield( code
=> "$removesubfield", match
=> $removematch );
291 $debug && warn "Potentially removed : ", $field->subfield($removesubfield);
296 unless ($test_parameter) {
298 use C4
::AuthoritiesMarc
;
299 my $authtypecode=GuessAuthTypeCode
($record, $heading_fields);
300 my $authid= ($id?
$id:GuessAuthId
($record));
301 if ($authid && GetAuthority
($authid) && $update ){
302 ## Authority has an id and is in database : Replace
303 eval { ( $authid ) = ModAuthority
($authid,$record, $authtypecode) };
305 warn "Problem with authority $authid Cannot Modify";
306 printlog
({id
=>$originalid||$id||$authid, op
=>"edit",status
=>"ERROR"}) if ($logfile);
309 printlog
({id
=>$originalid||$id||$authid, op
=>"edit",status
=>"ok"}) if ($logfile);
312 elsif (defined $authid) {
313 ## An authid is defined but no authority in database : add
314 eval { ( $authid ) = AddAuthority
($record,$authid, $authtypecode) };
316 warn "Problem with authority $authid Cannot Add ".$@
;
317 printlog
({id
=>$originalid||$id||$authid, op
=>"insert",status
=>"ERROR"}) if ($logfile);
320 printlog
({id
=>$originalid||$id||$authid, op
=>"insert",status
=>"ok"}) if ($logfile);
324 ## True insert in database
325 eval { ( $authid ) = AddAuthority
($record,"", $authtypecode) };
327 warn "Problem with authority $authid Cannot Add".$@
;
328 printlog
({id
=>$originalid||$id||$authid, op
=>"insert",status
=>"ERROR"}) if ($logfile);
331 printlog
({id
=>$originalid||$id||$authid, op
=>"insert",status
=>"ok"}) if ($logfile);
335 $yamlhash->{$originalid}->{'authid'} = $authid;
337 foreach my $field ( $record->field("2..") ) {
338 push @subfields, map { ( $_->[0] =~ /[a-z]/ ?
$_->[1] : () ) } $field->subfields();
340 $yamlhash->{$originalid}->{'subfields'} = \
@subfields;
344 my ( $biblionumber, $biblioitemnumber, $itemnumbers_ref, $errors_ref );
346 # check for duplicate, based on ISBN (skip it if we already have found a duplicate with match parameter
347 if (!$biblionumber && $isbn_check && $isbn) {
348 # warn "search ISBN : $isbn";
349 $sth_isbn->execute($isbn);
350 ($biblionumber,$biblioitemnumber) = $sth_isbn->fetchrow;
352 if (defined $idmapfl) {
353 if ($sourcetag < "010"){
354 if ($record->field($sourcetag)){
355 my $source = $record->field($sourcetag)->data();
356 printf(IDMAP
"%s|%s\n",$source,$biblionumber);
359 my $source=$record->subfield($sourcetag,$sourcesubfield);
360 printf(IDMAP
"%s|%s\n",$source,$biblionumber);
363 # create biblio, unless we already have it ( either match or isbn )
365 eval{$biblioitemnumber=GetBiblioData
($biblionumber)->{biblioitemnumber
};};
367 eval { ( $biblionumber, $biblioitemnumber ) = ModBiblio
( $record, $biblionumber, GetFrameworkcode
($biblionumber) ) };
369 warn "ERROR: Edit biblio $biblionumber failed: $@\n";
370 printlog
( { id
=> $id || $originalid || $biblionumber, op
=> "update", status
=> "ERROR" } ) if ($logfile);
373 printlog
( { id
=> $id || $originalid || $biblionumber, op
=> "update", status
=> "ok" } ) if ($logfile);
376 printlog
( { id
=> $id || $originalid || $biblionumber, op
=> "insert", status
=> "warning : already in database" } ) if ($logfile);
380 eval { ( $biblionumber, $biblioitemnumber ) = AddBiblio
( $record, '', { defer_marc_save
=> 1 } ) };
382 warn "ERROR: Adding biblio $biblionumber failed: $@\n";
383 printlog
( { id
=> $id || $originalid || $biblionumber, op
=> "insert", status
=> "ERROR" } ) if ($logfile);
386 printlog
( { id
=> $id || $originalid || $biblionumber, op
=> "insert", status
=> "ok" } ) if ($logfile);
389 printlog
( { id
=> $id || $originalid || $biblionumber, op
=> "update", status
=> "warning : not in database" } ) if ($logfile);
392 eval { ( $itemnumbers_ref, $errors_ref ) = AddItemBatchFromMarc
( $record, $biblionumber, $biblioitemnumber, '' ); };
393 my $error_adding = $@
;
394 # Work on a clone so that if there are real errors, we can maybe
396 my $clone_record = $record->clone();
397 C4
::Biblio
::_strip_item_fields
($clone_record, '');
398 # This sets the marc fields if there was an error, and also calls
400 ModBiblioMarc
( $clone_record, $biblionumber, $framework );
401 if ( $error_adding ) {
402 warn "ERROR: Adding items to bib $biblionumber failed: $error_adding";
403 printlog
({id
=>$id||$originalid||$biblionumber, op
=>"insertitem",status
=>"ERROR"}) if ($logfile);
404 # if we failed because of an exception, assume that
405 # the MARC columns in biblioitems were not set.
409 printlog
({id
=>$id||$originalid||$biblionumber, op
=>"insert",status
=>"ok"}) if ($logfile);
411 if ($dedup_barcode && grep { exists $_->{error_code
} && $_->{error_code
} eq 'duplicate_barcode' } @
$errors_ref) {
412 # Find the record called 'barcode'
413 my ($tag, $sub) = C4
::Biblio
::GetMarcFromKohaField
('items.barcode', $framework);
414 # Now remove any items that didn't have a duplicate_barcode error,
415 # erase the barcodes on items that did, and re-add those items.
417 foreach my $i (0 .. $#{$errors_ref}) {
418 my $ref = $errors_ref->[$i];
419 if ($ref && ($ref->{error_code
} eq 'duplicate_barcode')) {
420 $dupes{$ref->{item_sequence
}} = 1;
421 # Delete the error message because we're going to
423 delete $errors_ref->[$i];
427 foreach my $field ($record->field($tag)) {
430 # Here we remove the barcode
431 $field->delete_subfield(code
=> $sub);
433 # otherwise we delete the field because we don't want
435 $record->delete_fields($field);
438 # Now re-add the record as before, adding errors to the prev list
440 eval { ( $itemnumbers_ref, $more_errors ) = AddItemBatchFromMarc
( $record, $biblionumber, $biblioitemnumber, '' ); };
442 warn "ERROR: Adding items to bib $biblionumber failed: $@\n";
443 printlog
({id
=>$id||$originalid||$biblionumber, op
=>"insertitem",status
=>"ERROR"}) if ($logfile);
444 # if we failed because of an exception, assume that
445 # the MARC columns in biblioitems were not set.
446 ModBiblioMarc
( $record, $biblionumber, $framework );
449 printlog
({id
=>$id||$originalid||$biblionumber, op
=>"insert",status
=>"ok"}) if ($logfile);
451 push @
$errors_ref, @
{ $more_errors };
453 if ($#{ $errors_ref } > -1) {
454 report_item_errors
($biblionumber, $errors_ref);
456 $yamlhash->{$originalid} = $biblionumber if ($yamlfile);
458 $dbh->commit() if (0 == $i % $commitnum);
460 last if $i == $number;
463 $dbh->{AutoCommit
} = 1;
467 $dbh->do("SET FOREIGN_KEY_CHECKS = 1");
470 # restore CataloguingLog
471 $dbh->do("UPDATE systempreferences SET value=$CataloguingLog WHERE variable='CataloguingLog'");
473 my $timeneeded = gettimeofday
- $starttime;
474 print "\n$i MARC records done in $timeneeded seconds\n";
476 print $loghandle "file : $input_marc_file\n";
477 print $loghandle "$i MARC records done in $timeneeded seconds\n";
481 open my $yamlfileout, q{>}, "$yamlfile" or die "cannot open $yamlfile \n";
482 print $yamlfileout Dump
($yamlhash);
487 my $marcrecord=shift;
492 return $marcrecord->field($tag)->data() if $marcrecord->field($tag);
495 if ($marcrecord->field($tag)){
496 return $marcrecord->subfield($tag,$subfield);
505 foreach my $matchingpoint (@
$match){
506 my $string = build_simplequery
($matchingpoint,$record);
507 push @searchstrings,$string if (length($string)>0);
510 $QParser = C4
::Context
->queryparser if (C4
::Context
->preference('UseQueryParser'));
517 return join(" $op ",@searchstrings);
519 sub build_simplequery
{
523 my ($index,$recorddata)=split /,/,$element;
524 if ($recorddata=~/(\d{3})(.*)/) {
525 my ($tag,$subfields) =($1,$2);
526 foreach my $field ($record->field($tag)){
527 if (length($field->as_string("$subfields"))>0){
528 push @searchstrings,"$index:\"".$field->as_string("$subfields")."\"";
533 $QParser = C4
::Context
->queryparser if (C4
::Context
->preference('UseQueryParser'));
540 return join(" $op ",@searchstrings);
542 sub report_item_errors
{
543 my $biblionumber = shift;
544 my $errors_ref = shift;
546 foreach my $error (@
{ $errors_ref }) {
548 my $msg = "Item not added (bib $biblionumber, item tag #$error->{'item_sequence'}, barcode $error->{'item_barcode'}): ";
549 my $error_code = $error->{'error_code'};
550 $error_code =~ s/_/ /g;
551 $msg .= "$error_code $error->{'error_information'}";
556 my $logelements=shift;
557 print $loghandle join( ";", map { defined $_ ?
$_ : "" } @
$logelements{qw
<id op status
>} ), "\n";
559 sub get_heading_fields
{
562 $headingfields=YAML
::LoadFile
($authtypes);
563 $headingfields={C4
::Context
->preference('marcflavour')=>$headingfields};
564 $debug && warn YAML
::Dump
($headingfields);
566 unless ($headingfields){
567 $headingfields=$dbh->selectall_hashref("SELECT auth_tag_to_report, authtypecode from auth_types",'auth_tag_to_report',{Slice
=>{}});
568 $headingfields={C4
::Context
->preference('marcflavour')=>$headingfields};
570 return $headingfields;
575 bulkmarcimport.pl - Import bibliographic/authority records into Koha
579 $ export KOHA_CONF=/etc/koha.conf
580 $ perl misc/migration_tools/bulkmarcimport.pl -d -commit 1000 \\
581 -file /home/jmf/koha.mrc -n 3000
585 Don't use this script before you've entered and checked your MARC parameters
586 tables twice (or more!). Otherwise, the import won't work correctly and you
587 will get invalid data.
595 This version/help screen
597 =item B<-b, -biblios>
599 Type of import: bibliographic records
601 =item B<-a, -authorities>
603 Type of import: authority records
605 =item B<-file>=I<FILE>
607 The I<FILE> to import
611 Verbose mode. 1 means "some infos", 2 means "MARC dumping"
615 Turn off foreign key checks during import.
617 =item B<-n>=I<NUMBER>
619 The I<NUMBER> of records to import. If missing, all the file is imported
621 =item B<-o, -offset>=I<NUMBER>
623 File offset before importing, ie I<NUMBER> of records to skip.
625 =item B<-commit>=I<NUMBER>
627 The I<NUMBER> of records to wait before performing a 'commit' operation
631 File logs actions done for each record and their status into file
635 Test mode: parses the file, saying what he would do, but doing nothing.
639 Skip automatic conversion of MARC-8 to UTF-8. This option is provided for
642 =item B<-c>=I<CHARACTERISTIC>
644 The I<CHARACTERISTIC> MARC flavour. At the moment, only I<MARC21> and
645 I<UNIMARC> are supported. MARC21 by default.
649 Delete EVERYTHING related to biblio in koha-DB before import. Tables: biblio,
652 =item B<-m>=I<FORMAT>
654 Input file I<FORMAT>: I<MARCXML> or I<ISO2709> (defaults to ISO2709)
658 file yamlfile with authoritiesTypes and distinguishable record field in order
659 to store the correct authtype
663 yaml file format a yaml file with ids
667 list of fields that will not be imported. Can be any from 000 to 999 or field,
668 subfield and subfield's matching value such as 200avalue
672 if set, only insert when possible
676 if set, only updates (any biblio should have a matching record)
680 if set, do whatever is required
682 =item B<-k, -keepids>=<FIELD>
684 Field store ids in I<FIELD> (usefull for authorities, where 001 contains the
685 authid for Koha, that can contain a very valuable info for authorities coming
686 from LOC or BNF. useless for biblios probably)
688 =item B<-match>=<FIELD>
690 I<FIELD> matchindex,fieldtomatch matchpoint to use to deduplicate fieldtomatch
691 can be either 001 to 999 or field and list of subfields as such 100abcde
695 If set, a search will be done on isbn, and, if the same isbn is found, the
696 biblio is not added. It's another method to deduplicate. B<-match> & B<-isbn>
701 Clean ISBN fields from entering biblio records, ie removes hyphens. By default,
702 ISBN are cleaned. --nocleanisbn will keep ISBN unchanged.
706 Source bib I<TAG> for reporting the source bib number
708 =item B<-y>=I<SUBFIELD>
710 Source I<SUBFIELD> for reporting the source bib number
712 =item B<-idmap>=I<FILE>
714 I<FILE> for the koha bib and source id
718 Store ids in 009 (usefull for authorities, where 001 contains the authid for
719 Koha, that can contain a very valuable info for authorities coming from LOC or
720 BNF. useless for biblios probably)
722 =item B<-dedupbarcode>
724 If set, whenever a duplicate barcode is detected, it is removed and the attempt
725 to add the record is retried, thereby giving the record a blank barcode. This
726 is useful when something has set barcodes to be a biblio ID, or similar
727 (usually other software.)
731 This is the code for the framework that the requested records will have attached
732 to them when they are created. If not specified, then the default framework