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 C4::Context;
  21 use C4::Biblio;
  22 use C4::Koha;
  23 use C4::Debug;
  24 use C4::Charset;
  25 use C4::Items;
  26 use C4::MarcModificationTemplates;
  27
  28 use YAML;
  29 use Unicode::Normalize;
  30 use Time::HiRes qw(gettimeofday);
  31 use Getopt::Long;
  32 use IO::File;
  33 use Pod::Usage;
  34
  35 use Koha::Biblios;
  36 use Koha::SearchEngine;
  37 use Koha::SearchEngine::Search;
  38
  39 use open qw( :std :encoding(UTF-8) );
  40 binmode( STDOUT, ":encoding(UTF-8)" );
  41 my ( $input_marc_file, $number, $offset) = ('',0,0);
  42 my ($version, $delete, $test_parameter, $skip_marc8_conversion, $char_encoding, $verbose, $commit, $fk_off,$format,$biblios,$authorities,$keepids,$match, $isbn_check, $logfile);
  43 my ( $insert, $filters, $update, $all, $yamlfile, $authtypes, $append );
  44 my $cleanisbn = 1;
  45 my ($sourcetag,$sourcesubfield,$idmapfl, $dedup_barcode);
  46 my $framework = '';
  47 my $localcust;
  48 my $marc_mod_template = '';
  49 my $marc_mod_template_id = -1;
  50
  51 $|=1;
  52
  53 GetOptions(
  54     'commit:f'    => \$commit,
  55     'file:s'    => \$input_marc_file,
  56     'n:f' => \$number,
  57     'o|offset:f' => \$offset,
  58     'h' => \$version,
  59     'd' => \$delete,
  60     't|test' => \$test_parameter,
  61     's' => \$skip_marc8_conversion,
  62     'c:s' => \$char_encoding,
  63     'v:+' => \$verbose,
  64     'fk' => \$fk_off,
  65     'm:s' => \$format,
  66     'l:s' => \$logfile,
  67     'append' => \$append,
  68     'k|keepids:s' => \$keepids,
  69     'b|biblios' => \$biblios,
  70     'a|authorities' => \$authorities,
  71     'authtypes:s' => \$authtypes,
  72     'filter=s@'     => \$filters,
  73     'insert'        => \$insert,
  74     'update'        => \$update,
  75     'all'           => \$all,
  76     'match=s@'    => \$match,
  77     'i|isbn' => \$isbn_check,
  78     'x:s' => \$sourcetag,
  79     'y:s' => \$sourcesubfield,
  80     'idmap:s' => \$idmapfl,
  81     'cleanisbn!'     => \$cleanisbn,
  82     'yaml:s'        => \$yamlfile,
  83     'dedupbarcode' => \$dedup_barcode,
  84     'framework=s' => \$framework,
  85     'custom:s'    => \$localcust,
  86     'marcmodtemplate:s' => \$marc_mod_template,
  87 );
  88 $biblios ||= !$authorities;
  89 $insert  ||= !$update;
  90 my $writemode = ($append) ? "a" : "w";
  91
  92 pod2usage( -msg => "\nYou must specify either --biblios or --authorities, not both.\n", -exitval ) if $biblios && $authorities;
  93
  94 if ($all) {
  95     $insert = 1;
  96     $update = 1;
  97 }
  98
  99 if ($version || ($input_marc_file eq '')) {
 100     pod2usage( -verbose => 2 );
 101     exit;
 102 }
 103
 104 if(defined $localcust) { #local customize module
 105     if(!-e $localcust) {
 106         $localcust= $localcust||'LocalChanges'; #default name
 107         $localcust=~ s/^.*\/([^\/]+)$/$1/; #extract file name only
 108         $localcust=~ s/\.pm$//;           #remove extension
 109         my $fqcust= $FindBin::Bin."/$localcust.pm"; #try migration_tools dir
 110         if(-e $fqcust) {
 111             $localcust= $fqcust;
 112         }
 113         else {
 114             print "WARNING: customize module $localcust.pm not found!\n";
 115             exit 1;
 116         }
 117     }
 118     require $localcust if $localcust;
 119     $localcust=\&customize if $localcust;
 120 }
 121
 122 if($marc_mod_template ne '') {
 123     my @templates = GetModificationTemplates();
 124     foreach my $this_template (@templates) {
 125         if($this_template->{'name'} eq $marc_mod_template) {
 126             if($marc_mod_template_id < 0) {
 127                 $marc_mod_template_id = $this_template->{'template_id'};
 128             } else {
 129                 print "WARNING: MARC modification template name " .
 130                 "'$marc_mod_template' matches multiple templates. " .
 131                 "Please rename these templates\n";
 132                 exit 1;
 133             }
 134         }
 135     }
 136     if($marc_mod_template_id < 0) {
 137         die "Can't located MARC modification template '$marc_mod_template'\n";
 138     } else {
 139         print "Records will be modified using MARC modofication template: $marc_mod_template\n" if $verbose;
 140     }
 141 }
 142
 143 my $dbh = C4::Context->dbh;
 144 my $heading_fields=get_heading_fields();
 145
 146 if (defined $idmapfl) {
 147   open(IDMAP,">$idmapfl") or die "cannot open $idmapfl \n";
 148 }
 149
 150 if ((not defined $sourcesubfield) && (not defined $sourcetag)){
 151   $sourcetag="910";
 152   $sourcesubfield="a";
 153 }
 154
 155
 156 # Disable logging for the biblios and authorities import operation. It would unnecessarily
 157 # slow the import
 158
 159 # Disable the syspref cache so we can change logging settings
 160 C4::Context->disable_syspref_cache();
 161 # Save current CataloguingLog and AuthoritiesLog sysprefs values
 162 my $CataloguingLog = C4::Context->preference( 'CataloguingLog' );
 163 my $AuthoritiesLog = C4::Context->preference( 'AuthoritiesLog' );
 164 # Disable logging for both
 165 C4::Context->set_preference( 'CataloguingLog', 0 );
 166 C4::Context->set_preference( 'AuthoritiesLog', 0 );
 167
 168 if ($fk_off) {
 169         $dbh->do("SET FOREIGN_KEY_CHECKS = 0");
 170 }
 171
 172
 173 if ($delete) {
 174         if ($biblios){
 175         print "deleting biblios\n";
 176         $dbh->do("truncate biblio");
 177         $dbh->do("truncate biblioitems");
 178         $dbh->do("truncate items");
 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 print "Characteristic MARC flavour: $marcFlavour\n" if $verbose;
 196 my $starttime = gettimeofday;
 197 my $batch;
 198 my $fh = IO::File->new($input_marc_file); # don't let MARC::Batch open the file, as it applies the ':utf8' IO layer
 199 if (defined $format && $format =~ /XML/i) {
 200     # ugly hack follows -- MARC::File::XML, when used by MARC::Batch,
 201     # appears to try to convert incoming XML records from MARC-8
 202     # to UTF-8.  Setting the BinaryEncoding key turns that off
 203     # TODO: see what happens to ISO-8859-1 XML files.
 204     # TODO: determine if MARC::Batch can be fixed to handle
 205     #       XML records properly -- it probably should be
 206     #       be using a proper push or pull XML parser to
 207     #       extract the records, not using regexes to look
 208     #       for <record>.*</record>.
 209     $MARC::File::XML::_load_args{BinaryEncoding} = 'utf-8';
 210     my $recordformat= ($marcFlavour eq "MARC21"?"USMARC":uc($marcFlavour));
 211 #UNIMARC Authorities have a different way to manage encoding than UNIMARC biblios.
 212     $recordformat=$recordformat."AUTH" if ($authorities and $marcFlavour ne "MARC21");
 213     $MARC::File::XML::_load_args{RecordFormat} = $recordformat;
 214     $batch = MARC::Batch->new( 'XML', $fh );
 215 } else {
 216     $batch = MARC::Batch->new( 'USMARC', $fh );
 217 }
 218 $batch->warnings_off();
 219 $batch->strict_off();
 220 my $i=0;
 221 my $commitnum = $commit ? $commit : 50;
 222 my $yamlhash;
 223
 224 # Skip file offset
 225 if ( $offset ) {
 226     print "Skipping file offset: $offset records\n";
 227     $batch->next() while ($offset--);
 228 }
 229
 230 my ($tagid,$subfieldid);
 231 if ($authorities){
 232           $tagid='001';
 233 }
 234 else {
 235    ( $tagid, $subfieldid ) =
 236             GetMarcFromKohaField( "biblio.biblionumber", $framework );
 237         $tagid||="001";
 238 }
 239
 240 # the SQL query to search on isbn
 241 my $sth_isbn = $dbh->prepare("SELECT biblionumber,biblioitemnumber FROM biblioitems WHERE isbn=?");
 242
 243 $dbh->{AutoCommit} = 0;
 244 my $loghandle;
 245 if ($logfile){
 246    $loghandle= IO::File->new($logfile, $writemode) ;
 247    print $loghandle "id;operation;status\n";
 248 }
 249
 250 my $searcher = Koha::SearchEngine::Search->new(
 251     {
 252         index => (
 253               $authorities
 254             ? $Koha::SearchEngine::AUTHORITIES_INDEX
 255             : $Koha::SearchEngine::BIBLIOS_INDEX
 256         )
 257     }
 258 );
 259
 260 RECORD: while (  ) {
 261     my $record;
 262     # get records
 263     eval { $record = $batch->next() };
 264     if ( $@ ) {
 265         print "Bad MARC record $i: $@ skipped\n";
 266         # FIXME - because MARC::Batch->next() combines grabbing the next
 267         # blob and parsing it into one operation, a correctable condition
 268         # such as a MARC-8 record claiming that it's UTF-8 can't be recovered
 269         # from because we don't have access to the original blob.  Note
 270         # that the staging import can deal with this condition (via
 271         # C4::Charset::MarcToUTF8Record) because it doesn't use MARC::Batch.
 272         next;
 273     }
 274     # skip if we get an empty record (that is MARC valid, but will result in AddBiblio failure
 275     last unless ( $record );
 276     $i++;
 277     if( ($verbose//1)==1 ) { #no dot for verbose==2
 278         print "." . ( $i % 100==0 ? "\n$i" : '' );
 279     }
 280
 281     # transcode the record to UTF8 if needed & applicable.
 282     if ($record->encoding() eq 'MARC-8' and not $skip_marc8_conversion) {
 283         # FIXME update condition
 284         my ($guessed_charset, $charset_errors);
 285          ($record, $guessed_charset, $charset_errors) = MarcToUTF8Record($record, $marcFlavour.(($authorities and $marcFlavour ne "MARC21")?'AUTH':''));
 286         if ($guessed_charset eq 'failed') {
 287             warn "ERROR: failed to perform character conversion for record $i\n";
 288             next RECORD;
 289         }
 290     }
 291     SetUTF8Flag($record);
 292     if($marc_mod_template_id > 0) {
 293     print "Modifying MARC\n" if $verbose;
 294     ModifyRecordWithTemplate( $marc_mod_template_id, $record );
 295     }
 296     &$localcust($record) if $localcust;
 297     my $isbn;
 298     # remove trailing - in isbn (only for biblios, of course)
 299     if ($biblios && $cleanisbn) {
 300         my $tag = $marcFlavour eq 'UNIMARC' ? '010' : '020';
 301         my $field = $record->field($tag);
 302         my $isbn = $field && $field->subfield('a');
 303         if ( $isbn ) {
 304             $isbn =~ s/-//g;
 305             $field->update('a' => $isbn);
 306         }
 307     }
 308     my $id;
 309     # search for duplicates (based on Local-number)
 310     my $originalid;
 311     $originalid = GetRecordId( $record, $tagid, $subfieldid );
 312     if ($match) {
 313         require C4::Search;
 314         my $query = build_query( $match, $record );
 315         my $server = ( $authorities ? 'authorityserver' : 'biblioserver' );
 316         $debug && warn $query;
 317         my ( $error, $results, $totalhits ) = $searcher->simple_search_compat( $query, 0, 3, [$server] );
 318         # changed to warn so able to continue with one broken record
 319         if ( defined $error ) {
 320             warn "unable to search the database for duplicates : $error";
 321             printlog( { id => $id || $originalid || $match, op => "match", status => "ERROR" } ) if ($logfile);
 322             next RECORD;
 323         }
 324         $debug && warn "$query $server : $totalhits";
 325         if ( $results && scalar(@$results) == 1 ) {
 326             my $marcrecord = C4::Search::new_record_from_zebra( $server, $results->[0] );
 327             SetUTF8Flag($marcrecord);
 328             $id = GetRecordId( $marcrecord, $tagid, $subfieldid );
 329             if ( $authorities && $marcFlavour ) {
 330                 #Skip if authority in database is the same as the on in database
 331                 if ( $marcrecord->field('005') && $record->field('005') &&
 332                      $marcrecord->field('005')->data && $record->field('005')->data &&
 333                      $marcrecord->field('005')->data >= $record->field('005')->data ) {
 334                     if ($yamlfile) {
 335                         $yamlhash->{$originalid}->{'authid'} = $id;
 336
 337                         # we recover all subfields of the heading authorities
 338                         my @subfields;
 339                         foreach my $field ( $marcrecord->field("2..") ) {
 340                             push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
 341                         }
 342                         $yamlhash->{$originalid}->{'subfields'} = \@subfields;
 343                     }
 344                     next;
 345                 }
 346             }
 347         } elsif ( $results && scalar(@$results) > 1 ) {
 348             $debug && warn "more than one match for $query";
 349         } else {
 350             $debug && warn "nomatch for $query";
 351         }
 352     }
 353     if ($keepids && $originalid) {
 354             my $storeidfield;
 355             if ( length($keepids) == 3 ) {
 356                 $storeidfield = MARC::Field->new( $keepids, $originalid );
 357             } else {
 358                 $storeidfield = MARC::Field->new( substr( $keepids, 0, 3 ), "", "", substr( $keepids, 3, 1 ), $originalid );
 359             }
 360             $record->insert_fields_ordered($storeidfield);
 361             $record->delete_field( $record->field($tagid) );
 362     }
 363     foreach my $stringfilter (@$filters) {
 364         if ( length($stringfilter) == 3 ) {
 365             foreach my $field ( $record->field($stringfilter) ) {
 366                 $record->delete_field($field);
 367                 $debug && warn "removed : ", $field->as_string;
 368             }
 369         } elsif ($stringfilter =~ /([0-9]{3})([a-z0-9])(.*)/) {
 370             my $removetag = $1;
 371             my $removesubfield = $2;
 372             my $removematch = $3;
 373             if ( ( $removetag > "010" ) && $removesubfield ) {
 374                 foreach my $field ( $record->field($removetag) ) {
 375                     $field->delete_subfield( code => "$removesubfield", match => $removematch );
 376                     $debug && warn "Potentially removed : ", $field->subfield($removesubfield);
 377                 }
 378             }
 379         }
 380     }
 381     unless ($test_parameter) {
 382         if ($authorities){
 383             use C4::AuthoritiesMarc;
 384             my $authtypecode=GuessAuthTypeCode($record, $heading_fields);
 385             my $authid= ($id?$id:GuessAuthId($record));
 386             if ($authid && GetAuthority($authid) && $update ){
 387             ## Authority has an id and is in database : Replace
 388                 eval { ( $authid ) = ModAuthority($authid,$record, $authtypecode) };
 389                 if ($@){
 390                     warn "Problem with authority $authid Cannot Modify";
 391                                         printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ERROR"}) if ($logfile);
 392                 }
 393                                 else{
 394                                         printlog({id=>$originalid||$id||$authid, op=>"edit",status=>"ok"}) if ($logfile);
 395                                 }
 396             }
 397             elsif (defined $authid) {
 398             ## An authid is defined but no authority in database : add
 399                 eval { ( $authid ) = AddAuthority($record,$authid, $authtypecode) };
 400                 if ($@){
 401                     warn "Problem with authority $authid Cannot Add ".$@;
 402                                         printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
 403                 }
 404                                 else{
 405                                         printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
 406                                 }
 407             }
 408                 else {
 409             ## True insert in database
 410                 eval { ( $authid ) = AddAuthority($record,"", $authtypecode) };
 411                 if ($@){
 412                     warn "Problem with authority $authid Cannot Add".$@;
 413                                         printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ERROR"}) if ($logfile);
 414                 }
 415                                 else{
 416                                         printlog({id=>$originalid||$id||$authid, op=>"insert",status=>"ok"}) if ($logfile);
 417                                 }
 418                 }
 419             if ($yamlfile) {
 420             $yamlhash->{$originalid}->{'authid'} = $authid;
 421             my @subfields;
 422             foreach my $field ( $record->field("2..") ) {
 423                 push @subfields, map { ( $_->[0] =~ /[a-z]/ ? $_->[1] : () ) } $field->subfields();
 424             }
 425             $yamlhash->{$originalid}->{'subfields'} = \@subfields;
 426             }
 427         }
 428         else {
 429             my ( $biblionumber, $biblioitemnumber, $itemnumbers_ref, $errors_ref );
 430             $biblionumber = $id;
 431             # check for duplicate, based on ISBN (skip it if we already have found a duplicate with match parameter
 432             if (!$biblionumber && $isbn_check && $isbn) {
 433     #         warn "search ISBN : $isbn";
 434                 $sth_isbn->execute($isbn);
 435                 ($biblionumber,$biblioitemnumber) = $sth_isbn->fetchrow;
 436             }
 437                 if (defined $idmapfl) {
 438                                 if ($sourcetag < "010"){
 439                                         if ($record->field($sourcetag)){
 440                                           my $source = $record->field($sourcetag)->data();
 441                                           printf(IDMAP "%s|%s\n",$source,$biblionumber);
 442                                         }
 443                             } else {
 444                                         my $source=$record->subfield($sourcetag,$sourcesubfield);
 445                                         printf(IDMAP "%s|%s\n",$source,$biblionumber);
 446                           }
 447                         }
 448                                         # create biblio, unless we already have it ( either match or isbn )
 449             if ($biblionumber) {
 450                 eval{
 451                     $biblioitemnumber = Koha::Biblios->find( $biblionumber )->biblioitem->biblioitemnumber;
 452                 };
 453                 if ($update) {
 454                     eval { ( $biblionumber, $biblioitemnumber ) = ModBiblio( $record, $biblionumber, GetFrameworkCode($biblionumber) ) };
 455                     if ($@) {
 456                         warn "ERROR: Edit biblio $biblionumber failed: $@\n";
 457                         printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ERROR" } ) if ($logfile);
 458                         next RECORD;
 459                     } else {
 460                         printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "ok" } ) if ($logfile);
 461                     }
 462                 } else {
 463                     printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "warning : already in database" } ) if ($logfile);
 464                 }
 465             } else {
 466                 if ($insert) {
 467                     eval { ( $biblionumber, $biblioitemnumber ) = AddBiblio( $record, '', { defer_marc_save => 1 } ) };
 468                     if ($@) {
 469                         warn "ERROR: Adding biblio $biblionumber failed: $@\n";
 470                         printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ERROR" } ) if ($logfile);
 471                         next RECORD;
 472                     } else {
 473                         printlog( { id => $id || $originalid || $biblionumber, op => "insert", status => "ok" } ) if ($logfile);
 474                     }
 475                 } else {
 476                     printlog( { id => $id || $originalid || $biblionumber, op => "update", status => "warning : not in database" } ) if ($logfile);
 477                 }
 478             }
 479             eval { ( $itemnumbers_ref, $errors_ref ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
 480             my $error_adding = $@;
 481             # Work on a clone so that if there are real errors, we can maybe
 482             # fix them up later.
 483                         my $clone_record = $record->clone();
 484             C4::Biblio::_strip_item_fields($clone_record, '');
 485             # This sets the marc fields if there was an error, and also calls
 486             # defer_marc_save.
 487             ModBiblioMarc( $clone_record, $biblionumber, $framework );
 488             if ( $error_adding ) {
 489                 warn "ERROR: Adding items to bib $biblionumber failed: $error_adding";
 490                                 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
 491                 # if we failed because of an exception, assume that
 492                 # the MARC columns in biblioitems were not set.
 493                 next RECORD;
 494             }
 495                         else{
 496                                 printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ok"}) if ($logfile);
 497                         }
 498             if ($dedup_barcode && grep { exists $_->{error_code} && $_->{error_code} eq 'duplicate_barcode' } @$errors_ref) {
 499                 # Find the record called 'barcode'
 500                 my ($tag, $sub) = C4::Biblio::GetMarcFromKohaField('items.barcode', $framework);
 501                 # Now remove any items that didn't have a duplicate_barcode error,
 502                 # erase the barcodes on items that did, and re-add those items.
 503                 my %dupes;
 504                 foreach my $i (0 .. $#{$errors_ref}) {
 505                     my $ref = $errors_ref->[$i];
 506                     if ($ref && ($ref->{error_code} eq 'duplicate_barcode')) {
 507                         $dupes{$ref->{item_sequence}} = 1;
 508                         # Delete the error message because we're going to
 509                         # retry this one.
 510                         delete $errors_ref->[$i];
 511                     }
 512                 }
 513                 my $seq = 0;
 514                 foreach my $field ($record->field($tag)) {
 515                     $seq++;
 516                     if ($dupes{$seq}) {
 517                         # Here we remove the barcode
 518                         $field->delete_subfield(code => $sub);
 519                     } else {
 520                         # otherwise we delete the field because we don't want
 521                         # two of them
 522                         $record->delete_fields($field);
 523                     }
 524                 }
 525                 # Now re-add the record as before, adding errors to the prev list
 526                 my $more_errors;
 527                 eval { ( $itemnumbers_ref, $more_errors ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
 528                 if ( $@ ) {
 529                     warn "ERROR: Adding items to bib $biblionumber failed: $@\n";
 530                     printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ERROR"}) if ($logfile);
 531                     # if we failed because of an exception, assume that
 532                     # the MARC columns in biblioitems were not set.
 533                     ModBiblioMarc( $record, $biblionumber, $framework );
 534                     next RECORD;
 535                 } else {
 536                     printlog({id=>$id||$originalid||$biblionumber, op=>"insertitem",status=>"ok"}) if ($logfile);
 537                 }
 538                 push @$errors_ref, @{ $more_errors };
 539             }
 540             if ($#{ $errors_ref } > -1) {
 541                 report_item_errors($biblionumber, $errors_ref);
 542             }
 543             $yamlhash->{$originalid} = $biblionumber if ($yamlfile);
 544         }
 545         $dbh->commit() if (0 == $i % $commitnum);
 546     }
 547     print $record->as_formatted()."\n" if ($verbose//0)==2;
 548     last if $i == $number;
 549 }
 550 $dbh->commit();
 551 $dbh->{AutoCommit} = 1;
 552
 553
 554 if ($fk_off) {
 555         $dbh->do("SET FOREIGN_KEY_CHECKS = 1");
 556 }
 557
 558 # Restore CataloguingLog
 559 C4::Context->set_preference( 'CataloguingLog', $CataloguingLog );
 560 # Restore AuthoritiesLog
 561 C4::Context->set_preference( 'AuthoritiesLog', $AuthoritiesLog );
 562
 563 my $timeneeded = gettimeofday - $starttime;
 564 print "\n$i MARC records done in $timeneeded seconds\n";
 565 if ($logfile){
 566   print $loghandle "file : $input_marc_file\n";
 567   print $loghandle "$i MARC records done in $timeneeded seconds\n";
 568   $loghandle->close;
 569 }
 570 if ($yamlfile) {
 571     open my $yamlfileout, q{>}, "$yamlfile" or die "cannot open $yamlfile \n";
 572     print $yamlfileout Dump($yamlhash);
 573 }
 574 exit 0;
 575
 576 sub GetRecordId{
 577         my $marcrecord=shift;
 578         my $tag=shift;
 579         my $subfield=shift;
 580         my $id;
 581         if ($tag lt "010"){
 582                 return $marcrecord->field($tag)->data() if $marcrecord->field($tag);
 583         }
 584         elsif ($subfield){
 585                 if ($marcrecord->field($tag)){
 586                         return $marcrecord->subfield($tag,$subfield);
 587                 }
 588         }
 589         return $id;
 590 }
 591 sub build_query {
 592         my $match = shift;
 593         my $record=shift;
 594         my @searchstrings;
 595         foreach my $matchingpoint (@$match){
 596           my $string = build_simplequery($matchingpoint,$record);
 597           push @searchstrings,$string if (length($string)>0);
 598         }
 599     my $QParser;
 600     $QParser = C4::Context->queryparser if (C4::Context->preference('UseQueryParser'));
 601     my $op;
 602     if ($QParser) {
 603         $op = '&&';
 604     } else {
 605         $op = 'and';
 606     }
 607     return join(" $op ",@searchstrings);
 608 }
 609 sub build_simplequery {
 610         my $element=shift;
 611         my $record=shift;
 612     my @searchstrings;
 613     my ($index,$recorddata)=split /,/,$element;
 614     if ($recorddata=~/(\d{3})(.*)/) {
 615         my ($tag,$subfields) =($1,$2);
 616         foreach my $field ($record->field($tag)){
 617                   if (length($field->as_string("$subfields"))>0){
 618               push @searchstrings,"$index:\"".$field->as_string("$subfields")."\"";
 619                   }
 620         }
 621     }
 622     my $QParser;
 623     $QParser = C4::Context->queryparser if (C4::Context->preference('UseQueryParser'));
 624     my $op;
 625     if ($QParser) {
 626         $op = '&&';
 627     } else {
 628         $op = 'and';
 629     }
 630     return join(" $op ",@searchstrings);
 631 }
 632 sub report_item_errors {
 633     my $biblionumber = shift;
 634     my $errors_ref = shift;
 635
 636     foreach my $error (@{ $errors_ref }) {
 637         next if !$error;
 638         my $msg = "Item not added (bib $biblionumber, item tag #$error->{'item_sequence'}, barcode $error->{'item_barcode'}): ";
 639         my $error_code = $error->{'error_code'};
 640         $error_code =~ s/_/ /g;
 641         $msg .= "$error_code $error->{'error_information'}";
 642         print $msg, "\n";
 643     }
 644 }
 645 sub printlog{
 646         my $logelements=shift;
 647     print $loghandle join( ";", map { defined $_ ? $_ : "" } @$logelements{qw<id op status>} ), "\n";
 648 }
 649 sub get_heading_fields{
 650     my $headingfields;
 651     if ($authtypes){
 652         $headingfields=YAML::LoadFile($authtypes);
 653         $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
 654         $debug && warn YAML::Dump($headingfields);
 655     }
 656     unless ($headingfields){
 657         $headingfields=$dbh->selectall_hashref("SELECT auth_tag_to_report, authtypecode from auth_types",'auth_tag_to_report',{Slice=>{}});
 658         $headingfields={C4::Context->preference('marcflavour')=>$headingfields};
 659     }
 660     return $headingfields;
 661 }
 662
 663 =head1 NAME
 664
 665 bulkmarcimport.pl - Import bibliographic/authority records into Koha
 666
 667 =head1 USAGE
 668
 669  $ export KOHA_CONF=/etc/koha.conf
 670  $ perl misc/migration_tools/bulkmarcimport.pl -d -commit 1000 \\
 671     -file /home/jmf/koha.mrc -n 3000
 672
 673 =head1 WARNING
 674
 675 Don't use this script before you've entered and checked your MARC parameters
 676 tables twice (or more!). Otherwise, the import won't work correctly and you
 677 will get invalid data.
 678
 679 =head1 DESCRIPTION
 680
 681 =over
 682
 683 =item  B<-h>
 684
 685 This version/help screen
 686
 687 =item B<-b, -biblios>
 688
 689 Type of import: bibliographic records
 690
 691 =item B<-a, -authorities>
 692
 693 Type of import: authority records
 694
 695 =item B<-file>=I<FILE>
 696
 697 The I<FILE> to import
 698
 699 =item  B<-v>
 700
 701 Verbose mode. 1 means "some infos", 2 means "MARC dumping"
 702
 703 =item B<-fk>
 704
 705 Turn off foreign key checks during import.
 706
 707 =item B<-n>=I<NUMBER>
 708
 709 The I<NUMBER> of records to import. If missing, all the file is imported
 710
 711 =item B<-o, -offset>=I<NUMBER>
 712
 713 File offset before importing, ie I<NUMBER> of records to skip.
 714
 715 =item B<-commit>=I<NUMBER>
 716
 717 The I<NUMBER> of records to wait before performing a 'commit' operation
 718
 719 =item B<-l>
 720
 721 File logs actions done for each record and their status into file
 722
 723 =item B<-append>
 724
 725 If specified, data will be appended to the logfile. If not, the logfile will be erased for each execution.
 726
 727 =item B<-t, -test>
 728
 729 Test mode: parses the file, saying what it would do, but doing nothing.
 730
 731 =item B<-s>
 732
 733 Skip automatic conversion of MARC-8 to UTF-8.  This option is provided for
 734 debugging.
 735
 736 =item B<-c>=I<CHARACTERISTIC>
 737
 738 The I<CHARACTERISTIC> MARC flavour. At the moment, only I<MARC21> and
 739 I<UNIMARC> are supported. MARC21 by default.
 740
 741 =item B<-d>
 742
 743 Delete EVERYTHING related to biblio in koha-DB before import. Tables: biblio,
 744 biblioitems, items
 745
 746 =item B<-m>=I<FORMAT>
 747
 748 Input file I<FORMAT>: I<MARCXML> or I<ISO2709> (defaults to ISO2709)
 749
 750 =item B<-authtypes>
 751
 752 file yamlfile with authoritiesTypes and distinguishable record field in order
 753 to store the correct authtype
 754
 755 =item B<-yaml>
 756
 757 yaml file  format a yaml file with ids
 758
 759 =item B<-filter>
 760
 761 list of fields that will not be imported. Can be any from 000 to 999 or field,
 762 subfield and subfield's matching value such as 200avalue
 763
 764 =item B<-insert>
 765
 766 if set, only insert when possible
 767
 768 =item B<-update>
 769
 770 if set, only updates (any biblio should have a matching record)
 771
 772 =item B<-all>
 773
 774 if set, do whatever is required
 775
 776 =item B<-k, -keepids>=<FIELD>
 777
 778 Field store ids in I<FIELD> (useful for authorities, where 001 contains the
 779 authid for Koha, that can contain a very valuable info for authorities coming
 780 from LOC or BNF. useless for biblios probably)
 781
 782 =item B<-match>=<FIELD>
 783
 784 I<FIELD> matchindex,fieldtomatch matchpoint to use to deduplicate fieldtomatch
 785 can be either 001 to 999 or field and list of subfields as such 100abcde
 786
 787 =item B<-i,-isbn>
 788
 789 If set, a search will be done on isbn, and, if the same isbn is found, the
 790 biblio is not added. It's another method to deduplicate.  B<-match> & B<-isbn>
 791 can be both set.
 792
 793 =item B<-cleanisbn>
 794
 795 Clean ISBN fields from entering biblio records, ie removes hyphens. By default,
 796 ISBN are cleaned. --nocleanisbn will keep ISBN unchanged.
 797
 798 =item B<-x>=I<TAG>
 799
 800 Source bib I<TAG> for reporting the source bib number
 801
 802 =item B<-y>=I<SUBFIELD>
 803
 804 Source I<SUBFIELD> for reporting the source bib number
 805
 806 =item B<-idmap>=I<FILE>
 807
 808 I<FILE> for the koha bib and source id
 809
 810 =item B<-keepids>
 811
 812 Store ids in 009 (useful for authorities, where 001 contains the authid for
 813 Koha, that can contain a very valuable info for authorities coming from LOC or
 814 BNF. useless for biblios probably)
 815
 816 =item B<-dedupbarcode>
 817
 818 If set, whenever a duplicate barcode is detected, it is removed and the attempt
 819 to add the record is retried, thereby giving the record a blank barcode. This
 820 is useful when something has set barcodes to be a biblio ID, or similar
 821 (usually other software.)
 822
 823 =item B<-framework>
 824
 825 This is the code for the framework that the requested records will have attached
 826 to them when they are created. If not specified, then the default framework
 827 will be used.
 828
 829 =item B<-custom>=I<MODULE>
 830
 831 This parameter allows you to use a local module with a customize subroutine
 832 that is called for each MARC record.
 833 If no filename is passed, LocalChanges.pm is assumed to be in the
 834 migration_tools subdirectory. You may pass an absolute file name or a file name
 835 from the migration_tools directory.
 836
 837 =item B<-marcmodtemplate>=I<TEMPLATE>
 838
 839 This parameter allows you to specify the name of an existing MARC
 840 modification template to apply as the MARC records are imported (these
 841 templates are created in the "MARC modification templates" tool in Koha).
 842 If not specified, no MARC modification templates are used (default).
 843
 844 =back
 845
 846 =cut
 847