misc/migration_tools/bulkmarcimport.pl

   1 #!/usr/bin/perl
   2 # Import an iso2709 file into Koha 3
   3
   4 use Modern::Perl;
   5 #use diagnostics;
   6 BEGIN {
   7     # find Koha's Perl modules
   8     # test carefully before changing this
   9     use FindBin;
  10     eval { require "$FindBin::Bin/../kohalib.pl" };
  11 }
  12
  13 # Koha modules used
  14 use MARC::File::USMARC;
  15 use MARC::File::XML;
  16 use MARC::Record;
  17 use MARC::Batch;
  18 use MARC::Charset;
  19
  20 use Koha::Script;
  21 use C4::Context;
  22 use C4::Biblio;
  23 use C4::Koha;
  24 use C4::Debug;
  25 use C4::Charset;
  26 use C4::Items;
  27 use C4::MarcModificationTemplates;
  28
  29 use YAML;
  30 use Unicode::Normalize;
  31 use Time::HiRes qw(gettimeofday);
  32 use Getopt::Long;
  33 use IO::File;
  34 use Pod::Usage;
  35
  36 use Koha::Biblios;
  37 use Koha::SearchEngine;
  38 use Koha::SearchEngine::Search;
  39
  40 use open qw( :std :encoding(UTF-8) );
  41 binmode( STDOUT, ":encoding(UTF-8)" );
  42 my ( $input_marc_file, $number, $offset) = ('',0,0);
  43 my ($version, $delete, $test_parameter, $skip_marc8_conversion, $char_encoding, $verbose, $commit, $fk_off,$format,$biblios,$authorities,$keepids,$match, $isbn_check, $logfile);
  44 my ( $insert, $filters, $update, $all, $yamlfile, $authtypes, $append );
  45 my $cleanisbn = 1;
  46 my ($sourcetag,$sourcesubfield,$idmapfl, $dedup_barcode);
  47 my $framework = '';
  48 my $localcust;
  49 my $marc_mod_template = '';
  50 my $marc_mod_template_id = -1;
  51
  52 $|=1;
  53
  54 GetOptions(
  55     'commit:f'    => \$commit,
  56     'file:s'    => \$input_marc_file,
  57     'n:f' => \$number,
  58     'o|offset:f' => \$offset,
  59     'h' => \$version,
  60     'd' => \$delete,
  61     't|test' => \$test_parameter,
  62     's' => \$skip_marc8_conversion,
  63     'c:s' => \$char_encoding,
  64     'v:+' => \$verbose,
  65     'fk' => \$fk_off,
  66     'm:s' => \$format,
  67     'l:s' => \$logfile,
  68     'append' => \$append,
  69     'k|keepids:s' => \$keepids,
  70     'b|biblios' => \$biblios,
  71     'a|authorities' => \$authorities,
  72     'authtypes:s' => \$authtypes,
  73     'filter=s@'     => \$filters,
  74     'insert'        => \$insert,
  75     'update'        => \$update,
  76     'all'           => \$all,
  77     'match=s@'    => \$match,
  78     'i|isbn' => \$isbn_check,
  79     'x:s' => \$sourcetag,
  80     'y:s' => \$sourcesubfield,
  81     'idmap:s' => \$idmapfl,
  82     'cleanisbn!'     => \$cleanisbn,
  83     'yaml:s'        => \$yamlfile,
  84     'dedupbarcode' => \$dedup_barcode,
  85     'framework=s' => \$framework,
  86     'custom:s'    => \$localcust,
  87     'marcmodtemplate:s' => \$marc_mod_template,
  88 );
  89 $biblios ||= !$authorities;
  90 $insert  ||= !$update;
  91 my $writemode = ($append) ? "a" : "w";
  92
  93 pod2usage( -msg => "\nYou must specify either --biblios or --authorities, not both.\n", -exitval ) if $biblios && $authorities;
  94
  95 if ($all) {
  96     $insert = 1;
  97     $update = 1;
  98 }
  99
 100 if ($version || ($input_marc_file eq '')) {
 101     pod2usage( -verbose => 2 );
 102     exit;
 103 }
 104 if( $update && !( $match || $isbn_check ) ) {
 105     warn "Using -update without -match or -isbn seems to be useless.\n";
 106 }
 107
 108 if(defined $localcust) { #local customize module
 109     if(!-e $localcust) {
 110         $localcust= $localcust||'LocalChanges'; #default name
 111         $localcust=~ s/^.*\/([^\/]+)$/$1/; #extract file name only
 112         $localcust=~ s/\.pm$//;           #remove extension
 113         my $fqcust= $FindBin::Bin."/$localcust.pm"; #try migration_tools dir
 114         if(-e $fqcust) {
 115             $localcust= $fqcust;
 116         }
 117         else {
 118             print "WARNING: customize module $localcust.pm not found!\n";
 119             exit 1;
 120         }
 121     }
 122     require $localcust if $localcust;
 123     $localcust=\&customize if $localcust;
 124 }
 125
 126 if($marc_mod_template ne '') {
 127     my @templates = GetModificationTemplates();
 128     foreach my $this_template (@templates) {
 129         if($this_template->{'name'} eq $marc_mod_template) {
 130             if($marc_mod_template_id < 0) {
 131                 $marc_mod_template_id = $this_template->{'template_id'};
 132             } else {
 133                 print "WARNING: MARC modification template name " .
 134                 "'$marc_mod_template' matches multiple templates. " .
 135                 "Please rename these templates\n";
 136                 exit 1;
 137             }
 138         }
 139     }
 140     if($marc_mod_template_id < 0) {
 141         die "Can't located MARC modification template '$marc_mod_template'\n";
 142     } else {
 143         print "Records will be modified using MARC modofication template: $marc_mod_template\n" if $verbose;
 144     }
 145 }
 146
 147 my $dbh = C4::Context->dbh;
 148 my $heading_fields=get_heading_fields();
 149
 150 if (defined $idmapfl) {
 151   open(IDMAP,">$idmapfl") or die "cannot open $idmapfl \n";
 152 }
 153
 154 if ((not defined $sourcesubfield) && (not defined $sourcetag)){
 155   $sourcetag="910";
 156   $sourcesubfield="a";
 157 }
 158
 159
 160 # Disable logging for the biblios and authorities import operation. It would unnecessarily
 161 # slow the import
 162 $ENV{OVERRIDE_SYSPREF_CataloguingLog} = 0;
 163 $ENV{OVERRIDE_SYSPREF_AuthoritiesLog} = 0;
 164
 165 if ($fk_off) {
 166         $dbh->do("SET FOREIGN_KEY_CHECKS = 0");
 167 }
 168
 169
 170 if ($delete) {
 171         if ($biblios){
 172         print "deleting biblios\n";
 173         $dbh->do("DELETE FROM biblio");
 174         $dbh->do("ALTER TABLE biblio AUTO_INCREMENT = 1");
 175         $dbh->do("DELETE FROM biblioitems");
 176         $dbh->do("ALTER TABLE biblioitems AUTO_INCREMENT = 1");
 177         $dbh->do("DELETE FROM items");
 178         $dbh->do("ALTER TABLE items AUTO_INCREMENT = 1");
 179         }
 180         else {
 181         print "deleting authorities\n";
 182         $dbh->do("truncate auth_header");
 183         }
 184     $dbh->do("truncate zebraqueue");
 185 }
 186
 187
 188
 189 if ($test_parameter) {
 190     print "TESTING MODE ONLY\n    DOING NOTHING\n===============\n";
 191 }
 192
 193 my $marcFlavour = C4::Context->preference('marcflavour') || 'MARC21';
 194
 195 # The definition of $searcher must be before MARC::Batch->new
 196 my $searcher = Koha::SearchEngine::Search->new(
 197     {
 198         index => (
 199               $authorities
 200             ? $Koha::SearchEngine::AUTHORITIES_INDEX
 201             : $Koha::SearchEngine::BIBLIOS_INDEX
 202         )
 203     }
 204 );
 205
 206 print "Characteristic MARC flavour: $marcFlavour\n" if $verbose;
 207 my $starttime = gettimeofday;
 208 my $batch;
 209 my $fh = IO::File->new($input_marc_file); # don't let MARC::Batch open the file, as it applies the ':utf8' IO layer
 210 if (defined $format && $format =~ /XML/i) {
 211     # ugly hack follows -- MARC::File::XML, when used by MARC::Batch,
 212     # appears to try to convert incoming XML records from MARC-8
 213     # to UTF-8.  Setting the BinaryEncoding key turns that off
 214     # TODO: see what happens to ISO-8859-1 XML files.
 215     # TODO: determine if MARC::Batch can be fixed to handle
 216     #       XML records properly -- it probably should be
 217     #       be using a proper push or pull XML parser to
 218     #       extract the records, not using regexes to look
 219     #       for <record>.*</record>.
 220     $MARC::File::XML::_load_args{BinaryEncoding} = 'utf-8';
 221     my $recordformat= ($marcFlavour eq "MARC21"?"USMARC":uc($marcFlavour));
 222 #UNIMARC Authorities have a different way to manage encoding than UNIMARC biblios.
 223     $recordformat=$recordformat."AUTH" if ($authorities and $marcFlavour ne "MARC21");
 224     $MARC::File::XML::_load_args{RecordFormat} = $recordformat;
 225     $batch = MARC::Batch->new( 'XML', $fh );
 226 } else {
 227     $batch = MARC::Batch->new( 'USMARC', $fh );
 228 }
 229 $batch->warnings_off();
 230 $batch->strict_off();
 231 my $i=0;
 232 my $commitnum = $commit ? $commit : 50;
 233 my $yamlhash;
 234
 235 # Skip file offset
 236 if ( $offset ) {
 237     print "Skipping file offset: $offset records\n";
 238     $batch->next() while ($offset--);
 239 }
 240
 241 my ($tagid,$subfieldid);
 242 if ($authorities){
 243           $tagid='001';
 244 }
 245 else {
 246    ( $tagid, $subfieldid ) =
 247             GetMarcFromKohaField( "biblio.biblionumber" );
 248         $tagid||="001";
 249 }
 250
 251 # the SQL query to search on isbn
 252 my $sth_isbn = $dbh->prepare("SELECT biblionumber,biblioitemnumber FROM biblioitems WHERE isbn=?");
 253
 254 $dbh->{AutoCommit} = 0;
 255 my $loghandle;
 256 if ($logfile){
 257    $loghandle= IO::File->new($logfile, $writemode) ;
 258    print $loghandle "id;operation;status\n";
 259 }
 260
 261 RECORD: while (  ) {
 262     my $record;
 263     # get records
 264     eval { $record = $batch->next() };
 265     if ( $@ ) {
 266         print "Bad MARC record $i: $@ skipped\n";
 267         # FIXME - because MARC::Batch->next() combines grabbing the next
 268         # blob and parsing it into one operation, a correctable condition
 269         # such as a MARC-8 record claiming that it's UTF-8 can't be recovered
 270         # from because we don't have access to the original blob.  Note
 271         # that the staging import can deal with this condition (via
 272         # C4::Charset::MarcToUTF8Record) because it doesn't use MARC::Batch.
 273         next;
 274     }
 275     # skip if we get an empty record (that is MARC valid, but will result in AddBiblio failure
 276     last unless ( $record );
 277     $i++;
 278     if( ($verbose//1)==1 ) { #no dot for verbose==2
 279         print "." . ( $i % 100==0 ? "\n$i" : '' );
 280     }
 281
 282     # transcode the record to UTF8 if needed & applicable.
 283     if ($record->encoding() eq 'MARC-8' and not $skip_marc8_conversion) {
 284         # FIXME update condition
 285         my ($guessed_charset, $charset_errors);
 286          ($record, $guessed_charset, $charset_errors) = MarcToUTF8Record($record, $marcFlavour.(($authorities and $marcFlavour ne "MARC21")?'AUTH':''));
 287         if ($guessed_charset eq 'failed') {
 288             warn "ERROR: failed to perform character conversion for record $i\n";
 289             next RECORD;
 290         }
 291     }
 292     SetUTF8Flag($record);
 293     if($marc_mod_template_id > 0) {
 294     print "Modifying MARC\n" if $verbose;
 295     ModifyRecordWithTemplate( $marc_mod_template_id, $record );
 296     }
 297     &$localcust($record) if $localcust;
 298     my $isbn;
 299     # remove trailing - in isbn (only for biblios, of course)
 300     if( $biblios ) {
 301         my $tag = $marcFlavour eq 'UNIMARC' ? '010' : '020';
 302         my $field = $record->field($tag);
 303         $isbn = $field && $field->subfield('a');
 304         if ( $isbn && $cleanisbn ) {
 305             $isbn =~ s/-//g;
 306             $field->update('a' => $isbn);
 307         }
 308     }
 309     my $id;
 310     # search for duplicates (based on Local-number)
 311     my $originalid;
 312     $originalid = GetRecordId( $record, $tagid, $subfieldid );
 313     if ($match) {
 314         require C4::Search;
 315         my $query = build_query( $match, $record );
 316         my $server = ( $authorities ? 'authorityserver' : 'biblioserver' );
 317         $debug && warn $query;
 318         my ( $error, $results, $totalhits ) = $searcher->simple_search_compat( $query, 0, 3, [$server] );
 319         # changed to warn so able to continue with one broken record
 320         if ( defined $error ) {
 321             warn "unable to search the database for duplicates : $error";
 322             printlog( { id => $id || $originalid || $match, op => "match", status => "ERROR" } ) if ($logfile);
 323             next RECORD;
 324         }
 325         $debug && warn "$query $server : $totalhits";
 326         if ( $results && scalar(@$results) == 1 ) {
 327             my $marcrecord = C4::Search::new_record_from_zebra( $server, $results->[0] );
 328             SetUTF8Flag($marcrecord);
 329             $id = GetRecordId( $marcrecord, $tagid, $subfieldid );
 330             if ( $authorities && $marcFlavour ) {
 331                 #Skip if authority in database is the same as the on in database
 332                 if ( $marcrecord->field('005') && $record->field('005') &&
 333                      $marcrecord->field('005')->data && $record->field('005')->data &&
 334                      $marcrecord->field('005')->data >= $record->field('005')->data ) {
 335                     if ($yamlfile) {
 336                         $yamlhash->{$originalid}->{'authid'} = $id;
 337
 338                         # we recover all subfields of the heading authorities
 339                         my @subfields;
 340                         foreach my $field ( $marcrecord->field("2..") ) {
 341                             push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
 342                         }
 343                         $yamlhash->{$originalid}->{'subfields'} = \@subfields;
 344                         $yamlhash->{$originalid}->{'updated'} = 0;
 345                     }
 346                     next;
 347                 }
 348             }
 349         } elsif ( $results && scalar(@$results) > 1 ) {
 350             $debug && warn "more than one match for $query";
 351         } else {
 352             $debug && warn "nomatch for $query";
 353         }
 354     }
 355     if ($keepids && $originalid) {
 356             my $storeidfield;
 357             if ( length($keepids) == 3 ) {
 358                 $storeidfield = MARC::Field->new( $keepids, $originalid );
 359             } else {
 360                 $storeidfield = MARC::Field->new( substr( $keepids, 0, 3 ), "", "", substr( $keepids, 3, 1 ), $originalid );
 361             }
 362             $record->insert_fields_ordered($storeidfield);
 363             $record->delete_field( $record->field($tagid) );
 364     }
 365     foreach my $stringfilter (@$filters) {
 366         if ( length($stringfilter) == 3 ) {
 367             foreach my $field ( $record->field($stringfilter) ) {
 368                 $record->delete_field($field);
 369                 $debug && warn "removed : ", $field->as_string;
 370             }
 371         } elsif ($stringfilter =~ /([0-9]{3})([a-z0-9])(.*)/) {
 372             my $removetag = $1;
 373             my $removesubfield = $2;
 374             my $removematch = $3;
 375             if ( ( $removetag > "010" ) && $removesubfield ) {
 376                 foreach my $field ( $record->field($removetag) ) {
 377                     $field->delete_subfield( code => "$removesubfield", match => $removematch );
 378                     $debug && warn "Potentially removed : ", $field->subfield($removesubfield);
 379                 }
 380             }
 381         }
 382     }
 383     unless ($test_parameter) {
 384         if ($authorities){
 385             use C4::AuthoritiesMarc;
 386             my $authtypecode=GuessAuthTypeCode($record, $heading_fields);
 387             my $authid= ($id?$id:GuessAuthId($record));
 388             if ($authid && GetAuthority($authid) && $update ){
 389             ## Authority has an id and is in database : Replace
 390                 eval { ( $authid ) = ModAuthority($authid,$record, $authtypecode) };
 391                 if ($@){
 392                     warn "Problem with authority $authid Cannot Modify";
 393                                         printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ERROR"}) if ($logfile);
 394                 }
 395                                 else{
 396                                         printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ok"}) if ($logfile);
 397                                 }
 398             }
 399             elsif (defined $authid) {
 400             ## An authid is defined but no authority in database : add
 401                 eval { ( $authid ) = AddAuthority($record,$authid, $authtypecode) };
 402                 if ($@){
 403                     warn "Problem with authority $authid Cannot Add ".$@;
 404                                         printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
 405                 }
 406                                 else{
 407                                         printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
 408                                 }
 409             }
 410                 else {
 411             ## True insert in database
 412                 eval { ( $authid ) = AddAuthority($record,"", $authtypecode) };
 413                 if ($@){
 414                     warn "Problem with authority $authid Cannot Add".$@;
 415                                         printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
 416                 }
 417                                 else{
 418                                         printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
 419                                 }
 420                 }
 421             if ($yamlfile) {
 422             $yamlhash->{$originalid}->{'authid'} = $authid;
 423             my @subfields;
 424             foreach my $field ( $record->field("2..") ) {
 425                 push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
 426             }
 427             $yamlhash->{$originalid}->{'subfields'} = \@subfields;
 428             $yamlhash->{$originalid}->{'updated'} = 1;
 429             }
 430         }
 431         else {
 432             my ( $biblionumber, $biblioitemnumber, $itemnumbers_ref, $errors_ref );
 433             $biblionumber = $id;
 434             # check for duplicate, based on ISBN (skip it if we already have found a duplicate with match parameter
 435             if (!$biblionumber && $isbn_check && $isbn) {
 436     #         warn "search ISBN : $isbn";
 437                 $sth_isbn->execute($isbn);
 438                 ($biblionumber,$biblioitemnumber) = $sth_isbn->fetchrow;
 439             }
 440                 if (defined $idmapfl) {
 441                                 if ($sourcetag < "010"){
 442                                         if ($record->field($sourcetag)){
 443                                           my $source = $record->field($sourcetag)->data();
 444                                           printf(IDMAP "%s|%s\n",$source,$biblionumber);
 445                                         }
 446                             } else {
 447                                         my $source=$record->subfield($sourcetag,$sourcesubfield);
 448                                         printf(IDMAP "%s|%s\n",$source,$biblionumber);
 449                           }
 450                         }
 451                                         # create biblio, unless we already have it ( either match or isbn )
 452             if ($biblionumber) {
 453                 eval{
 454                     $biblioitemnumber = Koha::Biblios->find( $biblionumber )->biblioitem->biblioitemnumber;
 455                 };
 456                 if ($update) {
 457                     eval { ModBiblio( $record, $biblionumber, GetFrameworkCode($biblionumber) ) };
 458                     if ($@) {
 459                         warn "ERROR: Edit biblio $biblionumber failed: $@\n";
 460                         printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ERROR" } ) if ($logfile);
 461                         next RECORD;
 462                     } else {
 463                         printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ok" } ) if ($logfile);
 464                     }
 465                 } else {
 466                     printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "warning : already in database" } ) if ($logfile);
 467                 }
 468             } else {
 469                 if ($insert) {
 470                     eval { ( $biblionumber, $biblioitemnumber ) = AddBiblio( $record, '', { defer_marc_save => 1 } ) };
 471                     if ($@) {
 472                         warn "ERROR: Adding biblio $biblionumber failed: $@\n";
 473                         printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ERROR" } ) if ($logfile);
 474                         next RECORD;
 475                     } else {
 476                         printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ok" } ) if ($logfile);
 477                     }
 478                 } else {
 479                     warn "WARNING: Updating record ".($id||$originalid)." failed";
 480                     printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "warning : not in database" } ) if ($logfile);
 481                     next RECORD;
 482                 }
 483             }
 484             eval { ( $itemnumbers_ref, $errors_ref ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
 485             my $error_adding = $@;
 486             # Work on a clone so that if there are real errors, we can maybe
 487             # fix them up later.
 488                         my $clone_record = $record->clone();
 489             C4::Biblio::_strip_item_fields($clone_record, '');
 490             # This sets the marc fields if there was an error, and also calls
 491             # defer_marc_save.
 492             ModBiblioMarc( $clone_record, $biblionumber, $framework );
 493             if ( $error_adding ) {
 494                 warn "ERROR: Adding items to bib $biblionumber failed: $error_adding";
 495                                 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
 496                 # if we failed because of an exception, assume that
 497                 # the MARC columns in biblioitems were not set.
 498                 next RECORD;
 499             }
 500                         else{
 501                                 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ok"}) if ($logfile);
 502                         }
 503             if ($dedup_barcode && grep { exists $_->{error_code} && $_->{error_code} eq 'duplicate_barcode' } @$errors_ref) {
 504                 # Find the record called 'barcode'
 505                 my ($tag, $sub) = C4::Biblio::GetMarcFromKohaField( 'items.barcode' );
 506                 # Now remove any items that didn't have a duplicate_barcode error,
 507                 # erase the barcodes on items that did, and re-add those items.
 508                 my %dupes;
 509                 foreach my $i (0 .. $#{$errors_ref}) {
 510                     my $ref = $errors_ref->[$i];
 511                     if ($ref && ($ref->{error_code} eq 'duplicate_barcode')) {
 512                         $dupes{$ref->{item_sequence}} = 1;
 513                         # Delete the error message because we're going to
 514                         # retry this one.
 515                         delete $errors_ref->[$i];
 516                     }
 517                 }
 518                 my $seq = 0;
 519                 foreach my $field ($record->field($tag)) {
 520                     $seq++;
 521                     if ($dupes{$seq}) {
 522                         # Here we remove the barcode
 523                         $field->delete_subfield(code => $sub);
 524                     } else {
 525                         # otherwise we delete the field because we don't want
 526                         # two of them
 527                         $record->delete_fields($field);
 528                     }
 529                 }
 530                 # Now re-add the record as before, adding errors to the prev list
 531                 my $more_errors;
 532                 eval { ( $itemnumbers_ref, $more_errors ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
 533                 if ( $@ ) {
 534                     warn "ERROR: Adding items to bib $biblionumber failed: $@\n";
 535                     printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
 536                     # if we failed because of an exception, assume that
 537                     # the MARC columns in biblioitems were not set.
 538                     ModBiblioMarc( $record, $biblionumber, $framework );
 539                     next RECORD;
 540                 } else {
 541                     printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ok"}) if ($logfile);
 542                 }
 543                 push @$errors_ref, @{ $more_errors };
 544             }
 545             if ($#{ $errors_ref } > -1) {
 546                 report_item_errors($biblionumber, $errors_ref);
 547             }
 548             $yamlhash->{$originalid} = $biblionumber if ($yamlfile);
 549         }
 550         $dbh->commit() if (0 == $i % $commitnum);
 551     }
 552     print $record->as_formatted()."\n" if ($verbose//0)==2;
 553     last if $i == $number;
 554 }
 555 $dbh->commit();
 556 $dbh->{AutoCommit} = 1;
 557
 558
 559 if ($fk_off) {
 560         $dbh->do("SET FOREIGN_KEY_CHECKS = 1");
 561 }
 562
 563 # Restore CataloguingLog and AuthoritiesLog
 564 delete $ENV{OVERRIDE_SYSPREF_CataloguingLog};
 565 delete $ENV{OVERRIDE_SYSPREF_AuthoritiesLog};
 566
 567 my $timeneeded = gettimeofday - $starttime;
 568 print "\n$i MARC records done in $timeneeded seconds\n";
 569 if ($logfile){
 570   print $loghandle "file : $input_marc_file\n";
 571   print $loghandle "$i MARC records done in $timeneeded seconds\n";
 572   $loghandle->close;
 573 }
 574 if ($yamlfile) {
 575     open my $yamlfileout, q{>}, "$yamlfile" or die "cannot open $yamlfile \n";
 576     print $yamlfileout Dump($yamlhash);
 577 }
 578 exit 0;
 579
 580 sub GetRecordId{
 581         my $marcrecord=shift;
 582         my $tag=shift;
 583         my $subfield=shift;
 584         my $id;
 585         if ($tag lt "010"){
 586                 return $marcrecord->field($tag)->data() if $marcrecord->field($tag);
 587         }
 588         elsif ($subfield){
 589                 if ($marcrecord->field($tag)){
 590                         return $marcrecord->subfield($tag,$subfield);
 591                 }
 592         }
 593         return $id;
 594 }
 595 sub build_query {
 596         my $match = shift;
 597         my $record=shift;
 598         my @searchstrings;
 599         foreach my $matchingpoint (@$match){
 600           my $string = build_simplequery($matchingpoint,$record);
 601           push @searchstrings,$string if (length($string)>0);
 602         }
 603     my $op = 'and';
 604     return join(" $op ",@searchstrings);
 605 }
 606 sub build_simplequery {
 607         my $element=shift;
 608         my $record=shift;
 609     my @searchstrings;
 610     my ($index,$recorddata)=split /,/,$element;
 611     if ($recorddata=~/(\d{3})(.*)/) {
 612         my ($tag,$subfields) =($1,$2);
 613         foreach my $field ($record->field($tag)){
 614                   if (length($field->as_string("$subfields"))>0){
 615               push @searchstrings,"$index:\"".$field->as_string("$subfields")."\"";
 616                   }
 617         }
 618     }
 619     my $op = 'and';
 620     return join(" $op ",@searchstrings);
 621 }
 622 sub report_item_errors {
 623     my $biblionumber = shift;
 624     my $errors_ref = shift;
 625
 626     foreach my $error (@{ $errors_ref }) {
 627         next if !$error;
 628         my $msg = "Item not added (bib $biblionumber, item tag #$error->{'item_sequence'}, barcode $error->{'item_barcode'}): ";
 629         my $error_code = $error->{'error_code'};
 630         $error_code =~ s/_/ /g;
 631         $msg .= "$error_code $error->{'error_information'}";
 632         print $msg, "\n";
 633     }
 634 }
 635 sub printlog{
 636         my $logelements=shift;
 637     print $loghandle join( ";", map { defined $_ ? $_ : "" } @$logelements{qw<id op status>} ), "\n";
 638 }
 639 sub get_heading_fields{
 640     my $headingfields;
 641     if ($authtypes){
 642         $headingfields=YAML::LoadFile($authtypes);
 643         $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
 644         $debug && warn YAML::Dump($headingfields);
 645     }
 646     unless ($headingfields){
 647         $headingfields=$dbh->selectall_hashref("SELECT auth_tag_to_report, authtypecode from auth_types",'auth_tag_to_report',{Slice=>{}});
 648         $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
 649     }
 650     return $headingfields;
 651 }
 652
 653 =head1 NAME
 654
 655 bulkmarcimport.pl - Import bibliographic/authority records into Koha
 656
 657 =head1 USAGE
 658
 659  $ export KOHA_CONF=/etc/koha.conf
 660  $ perl misc/migration_tools/bulkmarcimport.pl -d -commit 1000 \\
 661     -file /home/jmf/koha.mrc -n 3000
 662
 663 =head1 WARNING
 664
 665 Don't use this script before you've entered and checked your MARC parameters
 666 tables twice (or more!). Otherwise, the import won't work correctly and you
 667 will get invalid data.
 668
 669 =head1 DESCRIPTION
 670
 671 =over
 672
 673 =item  B<-h>
 674
 675 This version/help screen
 676
 677 =item B<-b, -biblios>
 678
 679 Type of import: bibliographic records
 680
 681 =item B<-a, -authorities>
 682
 683 Type of import: authority records
 684
 685 =item B<-file>=I<FILE>
 686
 687 The I<FILE> to import
 688
 689 =item  B<-v>
 690
 691 Verbose mode. 1 means "some infos", 2 means "MARC dumping"
 692
 693 =item B<-fk>
 694
 695 Turn off foreign key checks during import.
 696
 697 =item B<-n>=I<NUMBER>
 698
 699 The I<NUMBER> of records to import. If missing, all the file is imported
 700
 701 =item B<-o, -offset>=I<NUMBER>
 702
 703 File offset before importing, ie I<NUMBER> of records to skip.
 704
 705 =item B<-commit>=I<NUMBER>
 706
 707 The I<NUMBER> of records to wait before performing a 'commit' operation
 708
 709 =item B<-l>
 710
 711 File logs actions done for each record and their status into file
 712
 713 =item B<-append>
 714
 715 If specified, data will be appended to the logfile. If not, the logfile will be erased for each execution.
 716
 717 =item B<-t, -test>
 718
 719 Test mode: parses the file, saying what it would do, but doing nothing.
 720
 721 =item B<-s>
 722
 723 Skip automatic conversion of MARC-8 to UTF-8.  This option is provided for
 724 debugging.
 725
 726 =item B<-c>=I<CHARACTERISTIC>
 727
 728 The I<CHARACTERISTIC> MARC flavour. At the moment, only I<MARC21> and
 729 I<UNIMARC> are supported. MARC21 by default.
 730
 731 =item B<-d>
 732
 733 Delete EVERYTHING related to biblio in koha-DB before import. Tables: biblio,
 734 biblioitems, items
 735
 736 =item B<-m>=I<FORMAT>
 737
 738 Input file I<FORMAT>: I<MARCXML> or I<ISO2709> (defaults to ISO2709)
 739
 740 =item B<-authtypes>
 741
 742 file yamlfile with authoritiesTypes and distinguishable record field in order
 743 to store the correct authtype
 744
 745 =item B<-yaml>
 746
 747 yaml file  format a yaml file with ids
 748
 749 =item B<-filter>
 750
 751 list of fields that will not be imported. Can be any from 000 to 999 or field,
 752 subfield and subfield's matching value such as 200avalue
 753
 754 =item B<-insert>
 755
 756 if set, only insert when possible
 757
 758 =item B<-update>
 759
 760 if set, only updates (any biblio should have a matching record)
 761
 762 =item B<-all>
 763
 764 if set, do whatever is required
 765
 766 =item B<-k, -keepids>=<FIELD>
 767
 768 Field store ids in I<FIELD> (useful for authorities, where 001 contains the
 769 authid for Koha, that can contain a very valuable info for authorities coming
 770 from LOC or BNF. useless for biblios probably)
 771
 772 =item B<-match>=<FIELD>
 773
 774 I<FIELD> matchindex,fieldtomatch matchpoint to use to deduplicate fieldtomatch
 775 can be either 001 to 999 or field and list of subfields as such 100abcde
 776
 777 =item B<-i,-isbn>
 778
 779 If set, a search will be done on isbn, and, if the same isbn is found, the
 780 biblio is not added. It's another method to deduplicate.  B<-match> & B<-isbn>
 781 can be both set.
 782
 783 =item B<-cleanisbn>
 784
 785 Clean ISBN fields from entering biblio records, ie removes hyphens. By default,
 786 ISBN are cleaned. --nocleanisbn will keep ISBN unchanged.
 787
 788 =item B<-x>=I<TAG>
 789
 790 Source bib I<TAG> for reporting the source bib number
 791
 792 =item B<-y>=I<SUBFIELD>
 793
 794 Source I<SUBFIELD> for reporting the source bib number
 795
 796 =item B<-idmap>=I<FILE>
 797
 798 I<FILE> for the koha bib and source id
 799
 800 =item B<-keepids>
 801
 802 Store ids in 009 (useful for authorities, where 001 contains the authid for
 803 Koha, that can contain a very valuable info for authorities coming from LOC or
 804 BNF. useless for biblios probably)
 805
 806 =item B<-dedupbarcode>
 807
 808 If set, whenever a duplicate barcode is detected, it is removed and the attempt
 809 to add the record is retried, thereby giving the record a blank barcode. This
 810 is useful when something has set barcodes to be a biblio ID, or similar
 811 (usually other software.)
 812
 813 =item B<-framework>
 814
 815 This is the code for the framework that the requested records will have attached
 816 to them when they are created. If not specified, then the default framework
 817 will be used.
 818
 819 =item B<-custom>=I<MODULE>
 820
 821 This parameter allows you to use a local module with a customize subroutine
 822 that is called for each MARC record.
 823 If no filename is passed, LocalChanges.pm is assumed to be in the
 824 migration_tools subdirectory. You may pass an absolute file name or a file name
 825 from the migration_tools directory.
 826
 827 =item B<-marcmodtemplate>=I<TEMPLATE>
 828
 829 This parameter allows you to specify the name of an existing MARC
 830 modification template to apply as the MARC records are imported (these
 831 templates are created in the "MARC modification templates" tool in Koha).
 832 If not specified, no MARC modification templates are used (default).
 833
 834 =back
 835
 836 =cut
 837