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;
27 use Unicode::Normalize;
28 use Time::HiRes qw(gettimeofday);
33 binmode(STDOUT, ":utf8");
34 my ( $input_marc_file, $number, $offset) = ('',0,0);
35 my ($version, $delete, $test_parameter, $skip_marc8_conversion, $char_encoding, $verbose, $commit, $fk_off,$format,$biblios,$authorities,$keepids,$match, $isbn_check, $logfile);
36 my ($sourcetag,$sourcesubfield,$idmapfl);
42 'commit:f' => \$commit,
43 'file:s' => \$input_marc_file,
45 'o|offset:f' => \$offset,
48 't' => \$test_parameter,
49 's' => \$skip_marc8_conversion,
50 'c:s' => \$char_encoding,
55 'k|keepids:s' => \$keepids,
56 'b|biblios' => \$biblios,
57 'a|authorities' => \$authorities,
58 'match=s@' => \$match,
59 'i|isbn' => \$isbn_check,
61 'y:s' => \$sourcesubfield,
62 'idmap:s' => \$idmapfl,
63 'cleanisbn!' => \$cleanisbn,
65 $biblios=!$authorities||$biblios;
67 if ($version || ($input_marc_file eq '')) {
68 pod2usage( -verbose => 2 );
72 if (defined $idmapfl) {
73 open(IDMAP,">$idmapfl") or die "cannot open $idmapfl \n";
76 if ((not defined $sourcesubfield) && (not defined $sourcetag)){
81 my $dbh = C4::Context->dbh;
83 # save the CataloguingLog property : we don't want to log a bulkmarcimport. It will slow the import &
84 # will create problems in the action_logs table, that can't handle more than 1 entry per second per user.
85 my $CataloguingLog = C4::Context->preference('CataloguingLog');
86 $dbh->do("UPDATE systempreferences SET value=0 WHERE variable='CataloguingLog'");
89 $dbh->do("SET FOREIGN_KEY_CHECKS = 0");
95 print "deleting biblios\n";
96 $dbh->do("truncate biblio");
97 $dbh->do("truncate biblioitems");
98 $dbh->do("truncate items");
101 print "deleting authorities\n";
102 $dbh->do("truncate auth_header");
104 $dbh->do("truncate zebraqueue");
109 if ($test_parameter) {
110 print "TESTING MODE ONLY\n DOING NOTHING\n===============\n";
113 my $marcFlavour = C4::Context->preference('marcflavour') || 'MARC21';
115 print "Characteristic MARC flavour: $marcFlavour\n" if $verbose;
116 my $starttime = gettimeofday;
118 my $fh = IO::File->new($input_marc_file); # don't let MARC::Batch open the file, as it applies the ':utf8' IO layer
119 if (defined $format && $format =~ /XML/i) {
120 # ugly hack follows -- MARC::File::XML, when used by MARC::Batch,
121 # appears to try to convert incoming XML records from MARC-8
122 # to UTF-8. Setting the BinaryEncoding key turns that off
123 # TODO: see what happens to ISO-8859-1 XML files.
124 # TODO: determine if MARC::Batch can be fixed to handle
125 # XML records properly -- it probably should be
126 # be using a proper push or pull XML parser to
127 # extract the records, not using regexes to look
128 # for <record>.*</record>.
129 $MARC::File::XML::_load_args{BinaryEncoding} = 'utf-8';
130 my $recordformat= ($marcFlavour eq "MARC21"?"USMARC":uc($marcFlavour));
131 #UNIMARC Authorities have a different way to manage encoding than UNIMARC biblios.
132 $recordformat=$recordformat."AUTH" if ($authorities and $marcFlavour ne "MARC21");
133 $MARC::File::XML::_load_args{RecordFormat} = $recordformat;
134 $batch = MARC::Batch->new( 'XML', $fh );
136 $batch = MARC::Batch->new( 'USMARC', $fh );
138 $batch->warnings_off();
139 $batch->strict_off();
141 my $commitnum = $commit ? $commit : 50;
146 print "Skipping file offset: $offset records\n";
147 $batch->next() while ($offset--);
150 my ($tagid,$subfieldid);
155 ( $tagid, $subfieldid ) =
156 GetMarcFromKohaField( "biblio.biblionumber", '' );
160 # the SQL query to search on isbn
161 my $sth_isbn = $dbh->prepare("SELECT biblionumber,biblioitemnumber FROM biblioitems WHERE isbn=?");
163 $dbh->{AutoCommit} = 0;
166 $loghandle= IO::File->new($logfile,"w") ;
167 print $loghandle "id;operation;status\n";
172 eval { $record = $batch->next() };
174 print "Bad MARC record $i: $@ skipped\n";
175 # FIXME - because MARC::Batch->next() combines grabbing the next
176 # blob and parsing it into one operation, a correctable condition
177 # such as a MARC-8 record claiming that it's UTF-8 can't be recovered
178 # from because we don't have access to the original blob. Note
179 # that the staging import can deal with this condition (via
180 # C4::Charset::MarcToUTF8Record) because it doesn't use MARC::Batch.
183 # skip if we get an empty record (that is MARC valid, but will result in AddBiblio failure
184 last unless ( $record );
187 print "\r$i" unless $i % 100;
189 # transcode the record to UTF8 if needed & applicable.
190 if ($record->encoding() eq 'MARC-8' and not $skip_marc8_conversion) {
191 # FIXME update condition
192 my ($guessed_charset, $charset_errors);
193 ($record, $guessed_charset, $charset_errors) = MarcToUTF8Record($record, $marcFlavour.(($authorities and $marcFlavour ne "MARC21")?'AUTH':''));
194 if ($guessed_charset eq 'failed') {
195 warn "ERROR: failed to perform character conversion for record $i\n";
200 # remove trailing - in isbn (only for biblios, of course)
201 if ($biblios && $cleanisbn) {
202 my $tag = $marcFlavour eq 'UNIMARC' ? '010' : '020';
203 my $field = $record->field($tag);
204 my $isbn = $field && $field->subfield('a');
207 $field->update('a' => $isbn);
211 # search for duplicates (based on Local-number)
214 my $query=build_query($match,$record);
215 my $server=($authorities?'authorityserver':'biblioserver');
216 my ($error, $results,$totalhits)=C4::Search::SimpleSearch( $query, 0, 3, [$server] );
217 die "unable to search the database for duplicates : $error" if (defined $error);
218 #warn "$query $server : $totalhits";
219 if ( @{$results} == 1 ){
220 my $marcrecord = MARC::File::USMARC::decode($results->[0]);
221 $id=GetRecordId($marcrecord,$tagid,$subfieldid);
223 elsif ( @{$results} > 1){
224 $debug && warn "more than one match for $query";
227 $debug && warn "nomatch for $query";
232 $originalid=GetRecordId($record,$tagid,$subfieldid);
235 if (length($keepids)==3){
236 $storeidfield=MARC::Field->new($keepids,$originalid);
239 $storeidfield=MARC::Field->new(substr($keepids,0,3),"","",substr($keepids,3,1),$originalid);
241 $record->insert_fields_ordered($storeidfield);
242 $record->delete_field($record->field($tagid));
245 unless ($test_parameter) {
247 use C4::AuthoritiesMarc;
248 my $authtypecode=GuessAuthTypeCode($record);
249 my $authid= ($id?$id:GuessAuthId($record));
250 if ($authid && GetAuthority($authid)){
251 ## Authority has an id and is in database : Replace
252 eval { ( $authid ) = ModAuthority($authid,$record, $authtypecode) };
254 warn "Problem with authority $authid Cannot Modify";
255 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ERROR"}) if ($logfile);
258 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ok"}) if ($logfile);
261 elsif (defined $authid) {
262 ## An authid is defined but no authority in database : add
263 eval { ( $authid ) = AddAuthority($record,$authid, $authtypecode) };
265 warn "Problem with authority $authid Cannot Add ".$@;
266 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
269 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
273 ## True insert in database
274 eval { ( $authid ) = AddAuthority($record,"", $authtypecode) };
276 warn "Problem with authority $authid Cannot Add".$@;
277 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
280 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
285 my ( $biblionumber, $biblioitemnumber, $itemnumbers_ref, $errors_ref );
287 # check for duplicate, based on ISBN (skip it if we already have found a duplicate with match parameter
288 if (!$biblionumber && $isbn_check && $isbn) {
289 # warn "search ISBN : $isbn";
290 $sth_isbn->execute($isbn);
291 ($biblionumber,$biblioitemnumber) = $sth_isbn->fetchrow;
293 if (defined $idmapfl) {
294 if ($sourcetag < "010"){
295 if ($record->field($sourcetag)){
296 my $source = $record->field($sourcetag)->data();
297 printf(IDMAP "%s|%s\n",$source,$biblionumber);
300 my $source=$record->subfield($sourcetag,$sourcesubfield);
301 printf(IDMAP "%s|%s\n",$source,$biblionumber);
304 # create biblio, unless we already have it ( either match or isbn )
306 eval{$biblioitemnumber=GetBiblioData($biblionumber)->{biblioitemnumber};}
310 eval { ( $biblionumber, $biblioitemnumber ) = AddBiblio($record, '', { defer_marc_save => 1 }) };
313 warn "ERROR: Adding biblio $biblionumber failed: $@\n";
314 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ERROR"}) if ($logfile);
318 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
320 eval { ( $itemnumbers_ref, $errors_ref ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
322 warn "ERROR: Adding items to bib $biblionumber failed: $@\n";
323 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
324 # if we failed because of an exception, assume that
325 # the MARC columns in biblioitems were not set.
326 ModBiblioMarc( $record, $biblionumber, '' );
330 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
332 if ($#{ $errors_ref } > -1) {
333 report_item_errors($biblionumber, $errors_ref);
336 $dbh->commit() if (0 == $i % $commitnum);
338 last if $i == $number;
345 $dbh->do("SET FOREIGN_KEY_CHECKS = 1");
348 # restore CataloguingLog
349 $dbh->do("UPDATE systempreferences SET value=$CataloguingLog WHERE variable='CataloguingLog'");
351 my $timeneeded = gettimeofday - $starttime;
352 print "\n$i MARC records done in $timeneeded seconds\n";
354 print $loghandle "file : $input_marc_file\n";
355 print $loghandle "$i MARC records done in $timeneeded seconds\n";
361 my $marcrecord=shift;
366 return $marcrecord->field($tag)->data() if $marcrecord->field($tag);
369 if ($marcrecord->field($tag)){
370 return $marcrecord->subfield($tag,$subfield);
379 foreach my $matchingpoint (@$match){
380 my $string = build_simplequery($matchingpoint,$record);
381 push @searchstrings,$string if (length($string)>0);
383 return join(" and ",@searchstrings);
385 sub build_simplequery {
388 my ($index,$recorddata)=split /,/,$element;
389 my ($tag,$subfields) =($1,$2) if ($recorddata=~/(\d{3})(.*)/);
391 foreach my $field ($record->field($tag)){
392 if (length($field->as_string("$subfields"))>0){
393 push @searchstrings,"$index,wrdl=\"".$field->as_string("$subfields")."\"";
396 return join(" and ",@searchstrings);
398 sub report_item_errors {
399 my $biblionumber = shift;
400 my $errors_ref = shift;
402 foreach my $error (@{ $errors_ref }) {
403 my $msg = "Item not added (bib $biblionumber, item tag #$error->{'item_sequence'}, barcode $error->{'item_barcode'}): ";
404 my $error_code = $error->{'error_code'};
405 $error_code =~ s/_/ /g;
406 $msg .= "$error_code $error->{'error_information'}";
411 my $logelements=shift;
412 print $loghandle join (";",@$logelements{qw<id op status>}),"\n";
418 bulkmarcimport.pl - Import bibliographic/authority records into Koha
422 $ export KOHA_CONF=/etc/koha.conf
423 $ perl misc/migration_tools/bulkmarcimport.pl -d -commit 1000 \\
424 -file /home/jmf/koha.mrc -n 3000
428 Don't use this script before you've entered and checked your MARC parameters
429 tables twice (or more!). Otherwise, the import won't work correctly and you
430 will get invalid data.
438 This version/help screen
440 =item B<-b, -biblios>
442 Type of import: bibliographic records
444 =item B<-a, -authorities>
446 Type of import: authority records
448 =item B<-file>=I<FILE>
450 The I<FILE> to import
454 Verbose mode. 1 means "some infos", 2 means "MARC dumping"
458 Turn off foreign key checks during import.
460 =item B<-n>=I<NUMBER>
462 The I<NUMBER> of records to import. If missing, all the file is imported
464 =item B<-o, -offset>=I<NUMBER>
466 File offset before importing, ie I<NUMBER> of records to skip.
468 =item B<-commit>=I<NUMBER>
470 The I<NUMBER> of records to wait before performing a 'commit' operation
474 File logs actions done for each record and their status into file
478 Test mode: parses the file, saying what he would do, but doing nothing.
482 Skip automatic conversion of MARC-8 to UTF-8. This option is provided for
485 =item B<-c>=I<CHARACTERISTIC>
487 The I<CHARACTERISTIC> MARC flavour. At the moment, only I<MARC21> and
488 I<UNIMARC> are supported. MARC21 by default.
492 Delete EVERYTHING related to biblio in koha-DB before import. Tables: biblio,
495 =item B<-m>=I<FORMAT>
497 Input file I<FORMAT>: I<MARCXML> or I<ISO2709> (defaults to ISO2709)
499 =item B<-k, -keepids>=<FIELD>
501 Field store ids in I<FIELD> (usefull for authorities, where 001 contains the
502 authid for Koha, that can contain a very valuable info for authorities coming
503 from LOC or BNF. useless for biblios probably)
505 =item B<-match>=<FIELD>
507 I<FIELD> matchindex,fieldtomatch matchpoint to use to deduplicate fieldtomatch
508 can be either 001 to 999 or field and list of subfields as such 100abcde
512 If set, a search will be done on isbn, and, if the same isbn is found, the
513 biblio is not added. It's another method to deduplicate. B<-match> & B<-isbn>
518 Clean ISBN fields from entering biblio records, ie removes hyphens. By default,
519 ISBN are cleaned. --nocleanisbn will keep ISBN unchanged.
523 Source bib I<TAG> for reporting the source bib number
525 =item B<-y>=I<SUBFIELD>
527 Source I<SUBFIELD> for reporting the source bib number
529 =item B<-idmap>=I<FILE>
531 I<FILE> for the koha bib and source id
535 Store ids in 009 (usefull for authorities, where 001 contains the authid for
536 Koha, that can contain a very valuable info for authorities coming from LOC or
537 BNF. useless for biblios probably)