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                     }
 345                     next;
 346                 }
 347             }
 348         } elsif ( $results && scalar(@$results) > 1 ) {
 349             $debug && warn "more than one match for $query";
 350         } else {
 351             $debug && warn "nomatch for $query";
 352         }
 353     }
 354     if ($keepids && $originalid) {
 355             my $storeidfield;
 356             if ( length($keepids) == 3 ) {
 357                 $storeidfield = MARC::Field->new( $keepids, $originalid );
 358             } else {
 359                 $storeidfield = MARC::Field->new( substr( $keepids, 0, 3 ), "", "", substr( $keepids, 3, 1 ), $originalid );
 360             }
 361             $record->insert_fields_ordered($storeidfield);
 362             $record->delete_field( $record->field($tagid) );
 363     }
 364     foreach my $stringfilter (@$filters) {
 365         if ( length($stringfilter) == 3 ) {
 366             foreach my $field ( $record->field($stringfilter) ) {
 367                 $record->delete_field($field);
 368                 $debug && warn "removed : ", $field->as_string;
 369             }
 370         } elsif ($stringfilter =~ /([0-9]{3})([a-z0-9])(.*)/) {
 371             my $removetag = $1;
 372             my $removesubfield = $2;
 373             my $removematch = $3;
 374             if ( ( $removetag > "010" ) && $removesubfield ) {
 375                 foreach my $field ( $record->field($removetag) ) {
 376                     $field->delete_subfield( code => "$removesubfield", match => $removematch );
 377                     $debug && warn "Potentially removed : ", $field->subfield($removesubfield);
 378                 }
 379             }
 380         }
 381     }
 382     unless ($test_parameter) {
 383         if ($authorities){
 384             use C4::AuthoritiesMarc;
 385             my $authtypecode=GuessAuthTypeCode($record, $heading_fields);
 386             my $authid= ($id?$id:GuessAuthId($record));
 387             if ($authid && GetAuthority($authid) && $update ){
 388             ## Authority has an id and is in database : Replace
 389                 eval { ( $authid ) = ModAuthority($authid,$record, $authtypecode) };
 390                 if ($@){
 391                     warn "Problem with authority $authid Cannot Modify";
 392                                         printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ERROR"}) if ($logfile);
 393                 }
 394                                 else{
 395                                         printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ok"}) if ($logfile);
 396                                 }
 397             }
 398             elsif (defined $authid) {
 399             ## An authid is defined but no authority in database : add
 400                 eval { ( $authid ) = AddAuthority($record,$authid, $authtypecode) };
 401                 if ($@){
 402                     warn "Problem with authority $authid Cannot Add ".$@;
 403                                         printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
 404                 }
 405                                 else{
 406                                         printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
 407                                 }
 408             }
 409                 else {
 410             ## True insert in database
 411                 eval { ( $authid ) = AddAuthority($record,"", $authtypecode) };
 412                 if ($@){
 413                     warn "Problem with authority $authid Cannot Add".$@;
 414                                         printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
 415                 }
 416                                 else{
 417                                         printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
 418                                 }
 419                 }
 420             if ($yamlfile) {
 421             $yamlhash->{$originalid}->{'authid'} = $authid;
 422             my @subfields;
 423             foreach my $field ( $record->field("2..") ) {
 424                 push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
 425             }
 426             $yamlhash->{$originalid}->{'subfields'} = \@subfields;
 427             }
 428         }
 429         else {
 430             my ( $biblionumber, $biblioitemnumber, $itemnumbers_ref, $errors_ref );
 431             $biblionumber = $id;
 432             # check for duplicate, based on ISBN (skip it if we already have found a duplicate with match parameter
 433             if (!$biblionumber && $isbn_check && $isbn) {
 434     #         warn "search ISBN : $isbn";
 435                 $sth_isbn->execute($isbn);
 436                 ($biblionumber,$biblioitemnumber) = $sth_isbn->fetchrow;
 437             }
 438                 if (defined $idmapfl) {
 439                                 if ($sourcetag < "010"){
 440                                         if ($record->field($sourcetag)){
 441                                           my $source = $record->field($sourcetag)->data();
 442                                           printf(IDMAP "%s|%s\n",$source,$biblionumber);
 443                                         }
 444                             } else {
 445                                         my $source=$record->subfield($sourcetag,$sourcesubfield);
 446                                         printf(IDMAP "%s|%s\n",$source,$biblionumber);
 447                           }
 448                         }
 449                                         # create biblio, unless we already have it ( either match or isbn )
 450             if ($biblionumber) {
 451                 eval{
 452                     $biblioitemnumber = Koha::Biblios->find( $biblionumber )->biblioitem->biblioitemnumber;
 453                 };
 454                 if ($update) {
 455                     eval { ModBiblio( $record, $biblionumber, GetFrameworkCode($biblionumber) ) };
 456                     if ($@) {
 457                         warn "ERROR: Edit biblio $biblionumber failed: $@\n";
 458                         printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ERROR" } ) if ($logfile);
 459                         next RECORD;
 460                     } else {
 461                         printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ok" } ) if ($logfile);
 462                     }
 463                 } else {
 464                     printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "warning : already in database" } ) if ($logfile);
 465                 }
 466             } else {
 467                 if ($insert) {
 468                     eval { ( $biblionumber, $biblioitemnumber ) = AddBiblio( $record, '', { defer_marc_save => 1 } ) };
 469                     if ($@) {
 470                         warn "ERROR: Adding biblio $biblionumber failed: $@\n";
 471                         printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ERROR" } ) if ($logfile);
 472                         next RECORD;
 473                     } else {
 474                         printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ok" } ) if ($logfile);
 475                     }
 476                 } else {
 477                     warn "WARNING: Updating record ".($id||$originalid)." failed";
 478                     printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "warning : not in database" } ) if ($logfile);
 479                     next RECORD;
 480                 }
 481             }
 482             eval { ( $itemnumbers_ref, $errors_ref ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
 483             my $error_adding = $@;
 484             # Work on a clone so that if there are real errors, we can maybe
 485             # fix them up later.
 486                         my $clone_record = $record->clone();
 487             C4::Biblio::_strip_item_fields($clone_record, '');
 488             # This sets the marc fields if there was an error, and also calls
 489             # defer_marc_save.
 490             ModBiblioMarc( $clone_record, $biblionumber, $framework );
 491             if ( $error_adding ) {
 492                 warn "ERROR: Adding items to bib $biblionumber failed: $error_adding";
 493                                 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
 494                 # if we failed because of an exception, assume that
 495                 # the MARC columns in biblioitems were not set.
 496                 next RECORD;
 497             }
 498                         else{
 499                                 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ok"}) if ($logfile);
 500                         }
 501             if ($dedup_barcode && grep { exists $_->{error_code} && $_->{error_code} eq 'duplicate_barcode' } @$errors_ref) {
 502                 # Find the record called 'barcode'
 503                 my ($tag, $sub) = C4::Biblio::GetMarcFromKohaField( 'items.barcode' );
 504                 # Now remove any items that didn't have a duplicate_barcode error,
 505                 # erase the barcodes on items that did, and re-add those items.
 506                 my %dupes;
 507                 foreach my $i (0 .. $#{$errors_ref}) {
 508                     my $ref = $errors_ref->[$i];
 509                     if ($ref && ($ref->{error_code} eq 'duplicate_barcode')) {
 510                         $dupes{$ref->{item_sequence}} = 1;
 511                         # Delete the error message because we're going to
 512                         # retry this one.
 513                         delete $errors_ref->[$i];
 514                     }
 515                 }
 516                 my $seq = 0;
 517                 foreach my $field ($record->field($tag)) {
 518                     $seq++;
 519                     if ($dupes{$seq}) {
 520                         # Here we remove the barcode
 521                         $field->delete_subfield(code => $sub);
 522                     } else {
 523                         # otherwise we delete the field because we don't want
 524                         # two of them
 525                         $record->delete_fields($field);
 526                     }
 527                 }
 528                 # Now re-add the record as before, adding errors to the prev list
 529                 my $more_errors;
 530                 eval { ( $itemnumbers_ref, $more_errors ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
 531                 if ( $@ ) {
 532                     warn "ERROR: Adding items to bib $biblionumber failed: $@\n";
 533                     printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
 534                     # if we failed because of an exception, assume that
 535                     # the MARC columns in biblioitems were not set.
 536                     ModBiblioMarc( $record, $biblionumber, $framework );
 537                     next RECORD;
 538                 } else {
 539                     printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ok"}) if ($logfile);
 540                 }
 541                 push @$errors_ref, @{ $more_errors };
 542             }
 543             if ($#{ $errors_ref } > -1) {
 544                 report_item_errors($biblionumber, $errors_ref);
 545             }
 546             $yamlhash->{$originalid} = $biblionumber if ($yamlfile);
 547         }
 548         $dbh->commit() if (0 == $i % $commitnum);
 549     }
 550     print $record->as_formatted()."\n" if ($verbose//0)==2;
 551     last if $i == $number;
 552 }
 553 $dbh->commit();
 554 $dbh->{AutoCommit} = 1;
 555
 556
 557 if ($fk_off) {
 558         $dbh->do("SET FOREIGN_KEY_CHECKS = 1");
 559 }
 560
 561 # Restore CataloguingLog and AuthoritiesLog
 562 delete $ENV{OVERRIDE_SYSPREF_CataloguingLog};
 563 delete $ENV{OVERRIDE_SYSPREF_AuthoritiesLog};
 564
 565 my $timeneeded = gettimeofday - $starttime;
 566 print "\n$i MARC records done in $timeneeded seconds\n";
 567 if ($logfile){
 568   print $loghandle "file : $input_marc_file\n";
 569   print $loghandle "$i MARC records done in $timeneeded seconds\n";
 570   $loghandle->close;
 571 }
 572 if ($yamlfile) {
 573     open my $yamlfileout, q{>}, "$yamlfile" or die "cannot open $yamlfile \n";
 574     print $yamlfileout Dump($yamlhash);
 575 }
 576 exit 0;
 577
 578 sub GetRecordId{
 579         my $marcrecord=shift;
 580         my $tag=shift;
 581         my $subfield=shift;
 582         my $id;
 583         if ($tag lt "010"){
 584                 return $marcrecord->field($tag)->data() if $marcrecord->field($tag);
 585         }
 586         elsif ($subfield){
 587                 if ($marcrecord->field($tag)){
 588                         return $marcrecord->subfield($tag,$subfield);
 589                 }
 590         }
 591         return $id;
 592 }
 593 sub build_query {
 594         my $match = shift;
 595         my $record=shift;
 596         my @searchstrings;
 597         foreach my $matchingpoint (@$match){
 598           my $string = build_simplequery($matchingpoint,$record);
 599           push @searchstrings,$string if (length($string)>0);
 600         }
 601     my $op = 'and';
 602     return join(" $op ",@searchstrings);
 603 }
 604 sub build_simplequery {
 605         my $element=shift;
 606         my $record=shift;
 607     my @searchstrings;
 608     my ($index,$recorddata)=split /,/,$element;
 609     if ($recorddata=~/(\d{3})(.*)/) {
 610         my ($tag,$subfields) =($1,$2);
 611         foreach my $field ($record->field($tag)){
 612                   if (length($field->as_string("$subfields"))>0){
 613               push @searchstrings,"$index:\"".$field->as_string("$subfields")."\"";
 614                   }
 615         }
 616     }
 617     my $op = 'and';
 618     return join(" $op ",@searchstrings);
 619 }
 620 sub report_item_errors {
 621     my $biblionumber = shift;
 622     my $errors_ref = shift;
 623
 624     foreach my $error (@{ $errors_ref }) {
 625         next if !$error;
 626         my $msg = "Item not added (bib $biblionumber, item tag #$error->{'item_sequence'}, barcode $error->{'item_barcode'}): ";
 627         my $error_code = $error->{'error_code'};
 628         $error_code =~ s/_/ /g;
 629         $msg .= "$error_code $error->{'error_information'}";
 630         print $msg, "\n";
 631     }
 632 }
 633 sub printlog{
 634         my $logelements=shift;
 635     print $loghandle join( ";", map { defined $_ ? $_ : "" } @$logelements{qw<id op status>} ), "\n";
 636 }
 637 sub get_heading_fields{
 638     my $headingfields;
 639     if ($authtypes){
 640         $headingfields=YAML::LoadFile($authtypes);
 641         $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
 642         $debug && warn YAML::Dump($headingfields);
 643     }
 644     unless ($headingfields){
 645         $headingfields=$dbh->selectall_hashref("SELECT auth_tag_to_report, authtypecode from auth_types",'auth_tag_to_report',{Slice=>{}});
 646         $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
 647     }
 648     return $headingfields;
 649 }
 650
 651 =head1 NAME
 652
 653 bulkmarcimport.pl - Import bibliographic/authority records into Koha
 654
 655 =head1 USAGE
 656
 657  $ export KOHA_CONF=/etc/koha.conf
 658  $ perl misc/migration_tools/bulkmarcimport.pl -d -commit 1000 \\
 659     -file /home/jmf/koha.mrc -n 3000
 660
 661 =head1 WARNING
 662
 663 Don't use this script before you've entered and checked your MARC parameters
 664 tables twice (or more!). Otherwise, the import won't work correctly and you
 665 will get invalid data.
 666
 667 =head1 DESCRIPTION
 668
 669 =over
 670
 671 =item  B<-h>
 672
 673 This version/help screen
 674
 675 =item B<-b, -biblios>
 676
 677 Type of import: bibliographic records
 678
 679 =item B<-a, -authorities>
 680
 681 Type of import: authority records
 682
 683 =item B<-file>=I<FILE>
 684
 685 The I<FILE> to import
 686
 687 =item  B<-v>
 688
 689 Verbose mode. 1 means "some infos", 2 means "MARC dumping"
 690
 691 =item B<-fk>
 692
 693 Turn off foreign key checks during import.
 694
 695 =item B<-n>=I<NUMBER>
 696
 697 The I<NUMBER> of records to import. If missing, all the file is imported
 698
 699 =item B<-o, -offset>=I<NUMBER>
 700
 701 File offset before importing, ie I<NUMBER> of records to skip.
 702
 703 =item B<-commit>=I<NUMBER>
 704
 705 The I<NUMBER> of records to wait before performing a 'commit' operation
 706
 707 =item B<-l>
 708
 709 File logs actions done for each record and their status into file
 710
 711 =item B<-append>
 712
 713 If specified, data will be appended to the logfile. If not, the logfile will be erased for each execution.
 714
 715 =item B<-t, -test>
 716
 717 Test mode: parses the file, saying what it would do, but doing nothing.
 718
 719 =item B<-s>
 720
 721 Skip automatic conversion of MARC-8 to UTF-8.  This option is provided for
 722 debugging.
 723
 724 =item B<-c>=I<CHARACTERISTIC>
 725
 726 The I<CHARACTERISTIC> MARC flavour. At the moment, only I<MARC21> and
 727 I<UNIMARC> are supported. MARC21 by default.
 728
 729 =item B<-d>
 730
 731 Delete EVERYTHING related to biblio in koha-DB before import. Tables: biblio,
 732 biblioitems, items
 733
 734 =item B<-m>=I<FORMAT>
 735
 736 Input file I<FORMAT>: I<MARCXML> or I<ISO2709> (defaults to ISO2709)
 737
 738 =item B<-authtypes>
 739
 740 file yamlfile with authoritiesTypes and distinguishable record field in order
 741 to store the correct authtype
 742
 743 =item B<-yaml>
 744
 745 yaml file  format a yaml file with ids
 746
 747 =item B<-filter>
 748
 749 list of fields that will not be imported. Can be any from 000 to 999 or field,
 750 subfield and subfield's matching value such as 200avalue
 751
 752 =item B<-insert>
 753
 754 if set, only insert when possible
 755
 756 =item B<-update>
 757
 758 if set, only updates (any biblio should have a matching record)
 759
 760 =item B<-all>
 761
 762 if set, do whatever is required
 763
 764 =item B<-k, -keepids>=<FIELD>
 765
 766 Field store ids in I<FIELD> (useful for authorities, where 001 contains the
 767 authid for Koha, that can contain a very valuable info for authorities coming
 768 from LOC or BNF. useless for biblios probably)
 769
 770 =item B<-match>=<FIELD>
 771
 772 I<FIELD> matchindex,fieldtomatch matchpoint to use to deduplicate fieldtomatch
 773 can be either 001 to 999 or field and list of subfields as such 100abcde
 774
 775 =item B<-i,-isbn>
 776
 777 If set, a search will be done on isbn, and, if the same isbn is found, the
 778 biblio is not added. It's another method to deduplicate.  B<-match> & B<-isbn>
 779 can be both set.
 780
 781 =item B<-cleanisbn>
 782
 783 Clean ISBN fields from entering biblio records, ie removes hyphens. By default,
 784 ISBN are cleaned. --nocleanisbn will keep ISBN unchanged.
 785
 786 =item B<-x>=I<TAG>
 787
 788 Source bib I<TAG> for reporting the source bib number
 789
 790 =item B<-y>=I<SUBFIELD>
 791
 792 Source I<SUBFIELD> for reporting the source bib number
 793
 794 =item B<-idmap>=I<FILE>
 795
 796 I<FILE> for the koha bib and source id
 797
 798 =item B<-keepids>
 799
 800 Store ids in 009 (useful for authorities, where 001 contains the authid for
 801 Koha, that can contain a very valuable info for authorities coming from LOC or
 802 BNF. useless for biblios probably)
 803
 804 =item B<-dedupbarcode>
 805
 806 If set, whenever a duplicate barcode is detected, it is removed and the attempt
 807 to add the record is retried, thereby giving the record a blank barcode. This
 808 is useful when something has set barcodes to be a biblio ID, or similar
 809 (usually other software.)
 810
 811 =item B<-framework>
 812
 813 This is the code for the framework that the requested records will have attached
 814 to them when they are created. If not specified, then the default framework
 815 will be used.
 816
 817 =item B<-custom>=I<MODULE>
 818
 819 This parameter allows you to use a local module with a customize subroutine
 820 that is called for each MARC record.
 821 If no filename is passed, LocalChanges.pm is assumed to be in the
 822 migration_tools subdirectory. You may pass an absolute file name or a file name
 823 from the migration_tools directory.
 824
 825 =item B<-marcmodtemplate>=I<TEMPLATE>
 826
 827 This parameter allows you to specify the name of an existing MARC
 828 modification template to apply as the MARC records are imported (these
 829 templates are created in the "MARC modification templates" tool in Koha).
 830 If not specified, no MARC modification templates are used (default).
 831
 832 =back
 833
 834 =cut
 835