misc/migration_tools/bulkmarcimport.pl

   1 #!/usr/bin/perl
   2 # Import an iso2709 file into Koha 3
   3
   4 use strict;
   5 # use warnings;
   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 # Uncomment the line below and use MARC::File::XML again when it works better.
  16 # -- thd
  17 # use MARC::File::XML;
  18 use MARC::Record;
  19 use MARC::Batch;
  20 use MARC::Charset;
  21
  22 use C4::Context;
  23 use C4::Biblio;
  24 use C4::Charset;
  25 use C4::Items;
  26 use Unicode::Normalize;
  27 use Time::HiRes qw(gettimeofday);
  28 use Getopt::Long;
  29 binmode(STDOUT, ":utf8");
  30
  31 use Getopt::Long;
  32
  33 my ( $input_marc_file, $number) = ('',0);
  34 my ($version, $delete, $test_parameter, $skip_marc8_conversion, $char_encoding, $verbose, $commit, $fk_off,$format);
  35
  36 $|=1;
  37
  38 GetOptions(
  39     'commit:f'    => \$commit,
  40     'file:s'    => \$input_marc_file,
  41     'n:f' => \$number,
  42     'h' => \$version,
  43     'd' => \$delete,
  44     't' => \$test_parameter,
  45     's' => \$skip_marc8_conversion,
  46     'c:s' => \$char_encoding,
  47     'v:s' => \$verbose,
  48     'fk' => \$fk_off,
  49     'm:s' => \$format,
  50 );
  51
  52 if ($version || ($input_marc_file eq '')) {
  53     print <<EOF
  54 small script to import an iso2709 file into Koha.
  55 parameters :
  56 \th : this version/help screen
  57 \tfile /path/to/file/to/dump : the file to import
  58 \tv : verbose mode. 1 means "some infos", 2 means "MARC dumping"
  59 \tfk : Turn off foreign key checks during import.
  60 \tn : the number of records to import. If missing, all the file is imported
  61 \tcommit : the number of records to wait before performing a 'commit' operation
  62 \tt : test mode : parses the file, saying what he would do, but doing nothing.
  63 \ts : skip automatic conversion of MARC-8 to UTF-8.  This option is
  64 \t    provided for debugging.
  65 \tc : the characteristic MARC flavour. At the moment, only MARC21 and UNIMARC
  66 \tsupported. MARC21 by default.
  67 \td : delete EVERYTHING related to biblio in koha-DB before import  :tables :
  68 \t\tbiblio, \tbiblioitems,\titems
  69 IMPORTANT : don't use this script before you've entered and checked your MARC parameters tables twice (or more!).
  70 Otherwise, the import won't work correctly and you will get invalid data.
  71
  72 SAMPLE :
  73 \t\$ export KOHA_CONF=/etc/koha.conf
  74 \t\$ perl misc/migration_tools/bulkmarcimport.pl -d -commit 1000 -file /home/jmf/koha.mrc -n 3000
  75 EOF
  76 ;#'
  77 exit;
  78 }
  79
  80 my $dbh = C4::Context->dbh;
  81
  82 # save the CataloguingLog property : we don't want to log a bulkmarcimport. It will slow the import &
  83 # will create problems in the action_logs table, that can't handle more than 1 entry per second per user.
  84 my $CataloguingLog = C4::Context->preference('CataloguingLog');
  85 $dbh->do("UPDATE systempreferences SET value=0 WHERE variable='CataloguingLog'");
  86
  87 if ($fk_off) {
  88         $dbh->do("SET FOREIGN_KEY_CHECKS = 0");
  89 }
  90
  91
  92 if ($delete) {
  93     print "deleting biblios\n";
  94     $dbh->do("truncate biblio");
  95     $dbh->do("truncate biblioitems");
  96     $dbh->do("truncate items");
  97     $dbh->do("truncate zebraqueue");
  98 }
  99
 100
 101
 102 if ($test_parameter) {
 103     print "TESTING MODE ONLY\n    DOING NOTHING\n===============\n";
 104 }
 105
 106 my $marcFlavour = C4::Context->preference('marcflavour') || 'MARC21';
 107
 108 print "Characteristic MARC flavour: $marcFlavour\n" if $verbose;
 109 my $starttime = gettimeofday;
 110 my $batch;
 111 if ($format =~ /XML/i) {
 112     $batch = MARC::Batch->new( 'XML', $input_marc_file );
 113 } else {
 114     $batch = MARC::Batch->new( 'USMARC', $input_marc_file );
 115 }
 116 $batch->warnings_off();
 117 $batch->strict_off();
 118 my $i=0;
 119 my $commitnum = 50;
 120
 121 if ($commit) {
 122
 123 $commitnum = $commit;
 124
 125 }
 126
 127 $dbh->{AutoCommit} = 0;
 128 RECORD: while ( my $record = $batch->next() ) {
 129     $i++;
 130     print ".";
 131     print "\r$i" unless $i % 100;
 132
 133     if ($record->encoding() eq 'MARC-8' and not $skip_marc8_conversion) {
 134         # FIXME update condition
 135         my ($guessed_charset, $charset_errors);
 136         ($record, $guessed_charset, $charset_errors) = MarcToUTF8Record($record, $marcFlavour);
 137         if ($guessed_charset eq 'failed') {
 138             warn "ERROR: failed to perform character conversion for record $i\n";
 139             next RECORD;
 140         }
 141     }
 142
 143     unless ($test_parameter) {
 144         my ( $biblionumber, $biblioitemnumber, $itemnumbers_ref, $errors_ref );
 145         eval { ( $biblionumber, $biblioitemnumber ) = AddBiblio($record, '', { defer_marc_save => 1 }) };
 146         if ( $@ ) {
 147             warn "ERROR: Adding biblio $biblionumber failed: $@\n";
 148             next RECORD;
 149         }
 150         eval { ( $itemnumbers_ref, $errors_ref ) = AddItemBatchFromMarc( $record, $biblionumber, $biblioitemnumber, '' ); };
 151         if ( $@ ) {
 152             warn "ERROR: Adding items to bib $biblionumber failed: $@\n";
 153             # if we failed because of an exception, assume that
 154             # the MARC columns in biblioitems were not set.
 155             ModBiblioMarc( $record, $biblionumber, '' );
 156             next RECORD;
 157         }
 158         if ($#{ $errors_ref } > -1) {
 159             report_item_errors($biblionumber, $errors_ref);
 160         }
 161
 162         $dbh->commit() if (0 == $i % $commitnum);
 163     }
 164     last if $i == $number;
 165 }
 166 $dbh->commit();
 167
 168
 169 if ($fk_off) {
 170         $dbh->do("SET FOREIGN_KEY_CHECKS = 1");
 171 }
 172
 173 # restore CataloguingLog
 174 $dbh->do("UPDATE systempreferences SET value=$CataloguingLog WHERE variable='CataloguingLog'");
 175
 176 my $timeneeded = gettimeofday - $starttime;
 177 print "\n$i MARC records done in $timeneeded seconds\n";
 178
 179 exit 0;
 180
 181 sub report_item_errors {
 182     my $biblionumber = shift;
 183     my $errors_ref = shift;
 184
 185     foreach my $error (@{ $errors_ref }) {
 186         my $msg = "Item not added (bib $biblionumber, item tag #$error->{'item_sequence'}, barcode $error->{'item_barcode'}): ";
 187         my $error_code = $error->{'error_code'};
 188         $error_code =~ s/_/ /g;
 189         $msg .= "$error_code $error->{'error_information'}";
 190         print $msg, "\n";
 191     }
 192 }