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);
41 'commit:f' => \$commit,
42 'file:s' => \$input_marc_file,
44 'o|offset:f' => \$offset,
47 't' => \$test_parameter,
48 's' => \$skip_marc8_conversion,
49 'c:s' => \$char_encoding,
54 'k|keepids:s' => \$keepids,
55 'b|biblios' => \$biblios,
56 'a|authorities' => \$authorities,
57 'match=s@' => \$match,
58 'i|isbn' => \$isbn_check,
60 'y:s' => \$sourcesubfield,
61 'idmap:s' => \$idmapfl,
63 $biblios=!$authorities||$biblios;
65 if ($version || ($input_marc_file eq '')) {
66 pod2usage( -verbose => 2 );
70 if (defined $idmapfl) {
71 open(IDMAP,">$idmapfl") or die "cannot open $idmapfl \n";
74 if ((not defined $sourcesubfield) && (not defined $sourcetag)){
79 my $dbh = C4::Context->dbh;
81 # save the CataloguingLog property : we don't want to log a bulkmarcimport. It will slow the import &
82 # will create problems in the action_logs table, that can't handle more than 1 entry per second per user.
83 my $CataloguingLog = C4::Context->preference('CataloguingLog');
84 $dbh->do("UPDATE systempreferences SET value=0 WHERE variable='CataloguingLog'");
87 $dbh->do("SET FOREIGN_KEY_CHECKS = 0");
93 print "deleting biblios\n";
94 $dbh->do("truncate biblio");
95 $dbh->do("truncate biblioitems");
96 $dbh->do("truncate items");
99 print "deleting authorities\n";
100 $dbh->do("truncate auth_header");
102 $dbh->do("truncate zebraqueue");
107 if ($test_parameter) {
108 print "TESTING MODE ONLY\n DOING NOTHING\n===============\n";
111 my $marcFlavour = C4::Context->preference('marcflavour') || 'MARC21';
113 print "Characteristic MARC flavour: $marcFlavour\n" if $verbose;
114 my $starttime = gettimeofday;
116 my $fh = IO::File->new($input_marc_file); # don't let MARC::Batch open the file, as it applies the ':utf8' IO layer
117 if (defined $format && $format =~ /XML/i) {
118 # ugly hack follows -- MARC::File::XML, when used by MARC::Batch,
119 # appears to try to convert incoming XML records from MARC-8
120 # to UTF-8. Setting the BinaryEncoding key turns that off
121 # TODO: see what happens to ISO-8859-1 XML files.
122 # TODO: determine if MARC::Batch can be fixed to handle
123 # XML records properly -- it probably should be
124 # be using a proper push or pull XML parser to
125 # extract the records, not using regexes to look
126 # for <record>.*</record>.
127 $MARC::File::XML::_load_args{BinaryEncoding} = 'utf-8';
128 my $recordformat= ($marcFlavour eq "MARC21"?"USMARC":uc($marcFlavour));
129 #UNIMARC Authorities have a different way to manage encoding than UNIMARC biblios.
130 $recordformat=$recordformat."AUTH" if ($authorities and $marcFlavour ne "MARC21");
131 $MARC::File::XML::_load_args{RecordFormat} = $recordformat;
132 $batch = MARC::Batch->new( 'XML', $fh );
134 $batch = MARC::Batch->new( 'USMARC', $fh );
136 $batch->warnings_off();
137 $batch->strict_off();
139 my $commitnum = $commit ? $commit : 50;
144 print "Skipping file offset: $offset records\n";
145 $batch->next() while ($offset--);
148 my ($tagid,$subfieldid);
153 ( $tagid, $subfieldid ) =
154 GetMarcFromKohaField( "biblio.biblionumber", '' );
158 # the SQL query to search on isbn
159 my $sth_isbn = $dbh->prepare("SELECT biblionumber,biblioitemnumber FROM biblioitems WHERE isbn=?");
161 $dbh->{AutoCommit} = 0;
164 $loghandle= IO::File->new($logfile,"w") ;
165 print $loghandle "id;operation;status\n";
170 eval { $record = $batch->next() };
172 print "Bad MARC record: skipped\n";
173 # FIXME - because MARC::Batch->next() combines grabbing the next
174 # blob and parsing it into one operation, a correctable condition
175 # such as a MARC-8 record claiming that it's UTF-8 can't be recovered
176 # from because we don't have access to the original blob. Note
177 # that the staging import can deal with this condition (via
178 # C4::Charset::MarcToUTF8Record) because it doesn't use MARC::Batch.
181 # skip if we get an empty record (that is MARC valid, but will result in AddBiblio failure
182 last unless ( $record );
185 print "\r$i" unless $i % 100;
187 # transcode the record to UTF8 if needed & applicable.
188 if ($record->encoding() eq 'MARC-8' and not $skip_marc8_conversion) {
189 # FIXME update condition
190 my ($guessed_charset, $charset_errors);
191 ($record, $guessed_charset, $charset_errors) = MarcToUTF8Record($record, $marcFlavour.(($authorities and $marcFlavour ne "MARC21")?'AUTH':''));
192 if ($guessed_charset eq 'failed') {
193 warn "ERROR: failed to perform character conversion for record $i\n";
198 # remove trailing - in isbn (only for biblios, of course)
200 if ($marcFlavour eq 'UNIMARC') {
201 if (my $f010 = $record->field('010')) {
202 $isbn = $f010->subfield('a');
204 $f010->update('a' => $isbn);
207 if (my $f020 = $record->field('020')) {
208 $isbn = $f020->subfield('a');
210 $f020->update('a' => $isbn);
215 # search for duplicates (based on Local-number)
218 my $query=build_query($match,$record);
219 my $server=($authorities?'authorityserver':'biblioserver');
220 my ($error, $results,$totalhits)=C4::Search::SimpleSearch( $query, 0, 3, [$server] );
221 die "unable to search the database for duplicates : $error" if (defined $error);
222 #warn "$query $server : $totalhits";
223 if ($results && scalar(@$results)==1){
224 my $marcrecord = MARC::File::USMARC::decode($results->[0]);
225 $id=GetRecordId($marcrecord,$tagid,$subfieldid);
227 elsif ($results && scalar(@$results)>1){
228 $debug && warn "more than one match for $query";
231 $debug && warn "nomatch for $query";
236 $originalid=GetRecordId($record,$tagid,$subfieldid);
239 if (length($keepids)==3){
240 $storeidfield=MARC::Field->new($keepids,$originalid);
243 $storeidfield=MARC::Field->new(substr($keepids,0,3),"","",substr($keepids,3,1),$originalid);
245 $record->insert_fields_ordered($storeidfield);
246 $record->delete_field($record->field($tagid));
249 unless ($test_parameter) {
251 use C4::AuthoritiesMarc;
252 my $authtypecode=GuessAuthTypeCode($record);
253 my $authid= ($id?$id:GuessAuthId($record));
254 if ($authid && GetAuthority($authid)){
255 ## Authority has an id and is in database : Replace
256 eval { ( $authid ) = ModAuthority($authid,$record, $authtypecode) };
258 warn "Problem with authority $authid Cannot Modify";
259 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ERROR"}) if ($logfile);
262 printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ok"}) if ($logfile);
265 elsif (defined $authid) {
266 ## An authid is defined but no authority in database : add
267 eval { ( $authid ) = AddAuthority($record,$authid, $authtypecode) };
269 warn "Problem with authority $authid Cannot Add ".$@;
270 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
273 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
277 ## True insert in database
278 eval { ( $authid ) = AddAuthority($record,"", $authtypecode) };
280 warn "Problem with authority $authid Cannot Add".$@;
281 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
284 printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
289 my ( $biblionumber, $biblioitemnumber, $itemnumbers_ref, $errors_ref );
291 # check for duplicate, based on ISBN (skip it if we already have found a duplicate with match parameter
292 if (!$biblionumber && $isbn_check && $isbn) {
293 # warn "search ISBN : $isbn";
294 $sth_isbn->execute($isbn);
295 ($biblionumber,$biblioitemnumber) = $sth_isbn->fetchrow;
297 if (defined $idmapfl) {
298 if ($sourcetag < "010"){
299 if ($record->field($sourcetag)){
300 my $source = $record->field($sourcetag)->data();
301 printf(IDMAP "%s|%s\n",$source,$biblionumber);
304 my $source=$record->subfield($sourcetag,$sourcesubfield);
305 printf(IDMAP "%s|%s\n",$source,$biblionumber);
308 # create biblio, unless we already have it ( either match or isbn )
310 eval{$biblioitemnumber=GetBiblioData($biblionumber)->{biblioitemnumber};}
314 eval { ( $biblionumber, $biblioitemnumber ) = AddBiblio($record, '', { defer_marc_save => 1 }) };
317 warn "ERROR: Adding biblio $biblionumber failed: $@\n";
318 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ERROR"}) if ($logfile);
322 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
324 eval { ( $itemnumbers_ref, $errors_ref ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
326 warn "ERROR: Adding items to bib $biblionumber failed: $@\n";
327 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
328 # if we failed because of an exception, assume that
329 # the MARC columns in biblioitems were not set.
330 ModBiblioMarc( $record, $biblionumber, '' );
334 printlog({id=>$id||$originalid||$biblionumber, op=>"insert",status=>"ok"}) if ($logfile);
336 if ($#{ $errors_ref } > -1) {
337 report_item_errors($biblionumber, $errors_ref);
340 $dbh->commit() if (0 == $i % $commitnum);
342 last if $i == $number;
349 $dbh->do("SET FOREIGN_KEY_CHECKS = 1");
352 # restore CataloguingLog
353 $dbh->do("UPDATE systempreferences SET value=$CataloguingLog WHERE variable='CataloguingLog'");
355 my $timeneeded = gettimeofday - $starttime;
356 print "\n$i MARC records done in $timeneeded seconds\n";
358 print $loghandle "file : $input_marc_file\n";
359 print $loghandle "$i MARC records done in $timeneeded seconds\n";
365 my $marcrecord=shift;
370 return $marcrecord->field($tag)->data() if $marcrecord->field($tag);
373 if ($marcrecord->field($tag)){
374 return $marcrecord->subfield($tag,$subfield);
383 foreach my $matchingpoint (@$match){
384 my $string = build_simplequery($matchingpoint,$record);
385 push @searchstrings,$string if (length($string)>0);
387 return join(" and ",@searchstrings);
389 sub build_simplequery {
392 my ($index,$recorddata)=split /,/,$element;
393 my ($tag,$subfields) =($1,$2) if ($recorddata=~/(\d{3})(.*)/);
395 foreach my $field ($record->field($tag)){
396 if (length($field->as_string("$subfields"))>0){
397 push @searchstrings,"$index,wrdl=\"".$field->as_string("$subfields")."\"";
400 return join(" and ",@searchstrings);
402 sub report_item_errors {
403 my $biblionumber = shift;
404 my $errors_ref = shift;
406 foreach my $error (@{ $errors_ref }) {
407 my $msg = "Item not added (bib $biblionumber, item tag #$error->{'item_sequence'}, barcode $error->{'item_barcode'}): ";
408 my $error_code = $error->{'error_code'};
409 $error_code =~ s/_/ /g;
410 $msg .= "$error_code $error->{'error_information'}";
415 my $logelements=shift;
416 print $loghandle join (";",@$logelements{qw<id op status>}),"\n";
422 bulkmarcimport.pl - Import bibliographic/authority records into Koha
426 $ export KOHA_CONF=/etc/koha.conf
427 $ perl misc/migration_tools/bulkmarcimport.pl -d -commit 1000 \\
428 -file /home/jmf/koha.mrc -n 3000
432 Don't use this script before you've entered and checked your MARC parameters
433 tables twice (or more!). Otherwise, the import won't work correctly and you
434 will get invalid data.
442 This version/help screen
444 =item B<-b, -biblios>
446 Type of import: bibliographic records
448 =item B<-a, -authorities>
450 Type of import: authority records
452 =item B<-file>=I<FILE>
454 The I<FILE> to import
458 Verbose mode. 1 means "some infos", 2 means "MARC dumping"
462 Turn off foreign key checks during import.
464 =item B<-n>=I<NUMBER>
466 The I<NUMBER> of records to import. If missing, all the file is imported
468 =item B<-o, -offset>=I<NUMBER>
470 File offset before importing, ie I<NUMBER> of records to skip.
472 =item B<-commit>=I<NUMBER>
474 The I<NUMBER> of records to wait before performing a 'commit' operation
478 File logs actions done for each record and their status into file
482 Test mode: parses the file, saying what he would do, but doing nothing.
486 Skip automatic conversion of MARC-8 to UTF-8. This option is provided for
489 =item B<-c>=I<CHARACTERISTIC>
491 The I<CHARACTERISTIC> MARC flavour. At the moment, only I<MARC21> and
492 I<UNIMARC> are supported. MARC21 by default.
496 Delete EVERYTHING related to biblio in koha-DB before import. Tables: biblio,
499 =item B<-m>=I<FORMAT>
501 Input file I<FORMAT>: I<MARCXML> or I<ISO2709> (defaults to ISO2709)
503 =item B<-k, -keepids>=<FIELD>
505 Field store ids in I<FIELD> (usefull for authorities, where 001 contains the
506 authid for Koha, that can contain a very valuable info for authorities coming
507 from LOC or BNF. useless for biblios probably)
509 =item B<-match>=<FIELD>
511 I<FIELD> matchindex,fieldtomatch matchpoint to use to deduplicate fieldtomatch
512 can be either 001 to 999 or field and list of subfields as such 100abcde
516 If set, a search will be done on isbn, and, if the same isbn is found, the
517 biblio is not added. It's another method to deduplicate. B<-match> & B<-isbn>
522 Source bib I<TAG> for reporting the source bib number
524 =item B<-y>=I<SUBFIELD>
526 Source I<SUBFIELD> for reporting the source bib number
528 =item B<-idmap>=I<FILE>
530 I<FILE> for the koha bib and source id
534 Store ids in 009 (usefull for authorities, where 001 contains the authid for
535 Koha, that can contain a very valuable info for authorities coming from LOC or
536 BNF. useless for biblios probably)