misc/stage_file.pl

   1 #!/usr/bin/perl
   2
   3 # This file is part of Koha.
   4 #
   5 # Copyright (C) 2007 LibLime
   6 # Parts Copyright BSZ 2011
   7 # Parts Copyright C & P Bibliography Services 2012
   8 #
   9 # Koha is free software; you can redistribute it and/or modify it under the
  10 # terms of the GNU General Public License as published by the Free Software
  11 # Foundation; either version 2 of the License, or (at your option) any later
  12 # version.
  13 #
  14 # Koha is distributed in the hope that it will be useful, but WITHOUT ANY
  15 # WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
  16 # A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
  17 #
  18 # You should have received a copy of the GNU General Public License along
  19 # with this program; if not, write to the Free Software Foundation, Inc.,
  20 # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
  21
  22 use strict;
  23 use warnings;
  24 BEGIN {
  25     # find Koha's Perl modules
  26     # test carefully before changing this
  27     use FindBin;
  28     eval { require "$FindBin::Bin/kohalib.pl" };
  29 }
  30
  31 use C4::Context;
  32 use C4::ImportBatch;
  33 use C4::Matcher;
  34 use Getopt::Long;
  35
  36 $| = 1;
  37
  38 # command-line parameters
  39 my $record_type = "biblio";
  40 my $encoding = "UTF-8";
  41 my $authorities = 0;
  42 my $match = 0;
  43 my $add_items = 0;
  44 my $input_file = "";
  45 my $batch_comment = "";
  46 my $want_help = 0;
  47 my $no_replace;
  48 my $format = 'ISO2709';
  49 my $no_create;
  50 my $item_action = 'always_add';
  51
  52 my $result = GetOptions(
  53     'encoding:s'    => \$encoding,
  54     'file:s'        => \$input_file,
  55     'format:s'      => \$format,
  56     'match|match-bibs:s'  => \$match,
  57     'add-items'     => \$add_items,
  58     'item-action:s' => \$item_action,
  59     'no-replace'    => \$no_replace,
  60     'no-create'     => \$no_create,
  61     'comment:s'     => \$batch_comment,
  62     'authorities'   => \$authorities,
  63     'h|help'        => \$want_help
  64 );
  65
  66 $record_type = 'auth' if ($authorities);
  67
  68 if (not $result or $input_file eq "" or $want_help) {
  69     print_usage();
  70     exit 0;
  71 }
  72 if ( $format !~ /^(MARCXML|ISO2709)$/i ) {
  73     print "\n --format must be MARCXML or ISO2709\n";
  74     print_usage();
  75     exit 0;
  76 }
  77
  78 unless (-r $input_file) {
  79     die "$0: cannot open input file $input_file: $!\n";
  80 }
  81
  82 my $dbh = C4::Context->dbh;
  83 $dbh->{AutoCommit} = 0;
  84 process_batch({
  85     format        => $format,
  86     input_file    => $input_file,
  87     record_type   => $record_type,
  88     match         => $match,
  89     add_items     => $add_items,
  90     batch_comment => $batch_comment,
  91     encoding      => $encoding,
  92     no_replace    => $no_replace,
  93     no_create     => $no_create,
  94     item_action   => $item_action,
  95 });
  96 $dbh->commit();
  97
  98 exit 0;
  99
 100 sub process_batch {
 101     my ( $params ) = @_;  #Possible params are: format input_file record_type match add_items batch_comment encoding no_replace no_create item_action
 102     my $format = $params->{format} // '';
 103     my $record_type = $params->{record_type} // 'biblio';
 104
 105     my ( $errors, $marc_records );
 106     if( $format eq 'ISO2709' ) {
 107         ( $errors, $marc_records ) = C4::ImportBatch::RecordsFromISO2709File(
 108             $params->{input_file}, $record_type, $params->{encoding} );
 109     } elsif( $format eq 'MARCXML' ) {
 110         ( $errors, $marc_records ) = C4::ImportBatch::RecordsFromMARCXMLFile(
 111             $params->{input_file}, $params->{encoding} );
 112     }
 113     warn ( join ',', @$errors ) if @$errors;
 114     my $num_input_records = ($marc_records) ? scalar(@$marc_records) : 0;
 115
 116     print "... staging MARC records -- please wait\n";
 117     #FIXME: We should really allow the use of marc modification frameworks and to_marc plugins here if possible
 118     my ($batch_id, $num_valid_records, $num_items, @import_errors) =
 119         BatchStageMarcRecords($record_type, $params->{encoding}, $marc_records, $params->{input_file}, undef, $params->{batch_comment}, '', $params->{add_items}, 0,
 120                               100, \&print_progress_and_commit);
 121     print "... finished staging MARC records\n";
 122
 123     my $num_with_matches = 0;
 124     if ( $params->{match} ) {
 125         my $matcher = C4::Matcher->fetch( $params->{match} );
 126         if (defined $matcher) {
 127             SetImportBatchMatcher( $batch_id, $params->{match} );
 128         } elsif ($record_type eq 'biblio')  {
 129             $matcher = C4::Matcher->new($record_type);
 130             $matcher->add_simple_matchpoint('isbn', 1000, '020', 'a', -1, 0, '');
 131             $matcher->add_simple_required_check('245', 'a', -1, 0, '',
 132                                             '245', 'a', -1, 0, '');
 133         }
 134         # set default record overlay behavior
 135         SetImportBatchOverlayAction( $batch_id, $params->{no_replace} ? 'ignore' : 'replace' );
 136         SetImportBatchNoMatchAction( $batch_id, $params->{no_create} ? 'ignore' : 'create_new' );
 137         SetImportBatchItemAction( $batch_id, $params->{item_action} );
 138         print "... looking for matches with records already in database\n";
 139         $num_with_matches = BatchFindDuplicates($batch_id, $matcher, 10, 100, \&print_progress_and_commit);
 140         print "... finished looking for matches\n";
 141     }
 142
 143     my $num_invalid_records = scalar(@import_errors);
 144     print <<_SUMMARY_;
 145
 146 MARC record staging report
 147 ------------------------------------
 148 Input file:                 $params->{input_file}
 149 Record type:                $record_type
 150 Number of input records:    $num_input_records
 151 Number of valid records:    $num_valid_records
 152 Number of invalid records:  $num_invalid_records
 153 _SUMMARY_
 154     if( $params->{match} ) {
 155         print "Number of records matched:  $num_with_matches\n";
 156     } else {
 157         print "Incoming records not matched against existing records (--match option not supplied)\n";
 158     }
 159     if ($record_type eq 'biblio') {
 160         if ( $params->{add_items} ) {
 161             print "Number of items parsed:  $num_items\n";
 162         } else {
 163             print "No items parsed (--add-items option not supplied)\n";
 164         }
 165     }
 166
 167     print "\n";
 168     print "Batch number assigned:  $batch_id\n";
 169     print "\n";
 170 }
 171
 172 sub print_progress_and_commit {
 173     my $recs = shift;
 174     $dbh->commit();
 175     print "... processed $recs records\n";
 176 }
 177
 178 sub print_usage {
 179     print <<_USAGE_;
 180 $0: stage MARC file into reservoir.
 181
 182 Use this batch job to load a file of MARC bibliographic
 183 (with optional item information) or authority records into
 184 the Koha reservoir.
 185
 186 After running this program to stage your file, you can use
 187 either the batch job commit_file.pl or the Koha
 188 Tools option "Manage Staged MARC Records" to load the
 189 records into the main Koha database.
 190
 191 Parameters:
 192     --file <file_name>      name of input MARC bib file
 193     --authorities           stage authority records instead of bibs
 194     --encoding <encoding>   encoding of MARC records, default is UTF-8.
 195                             Other possible options are: MARC-8,
 196                             ISO_5426, ISO_6937, ISO_8859-1, EUC-KR
 197     --format                The MARC transport format to use?
 198                             Defaults to ISO2709.
 199                             Available values, MARCXML, ISO2709.
 200     --match <match_id>      use this option to match records
 201                             in the file with records already in
 202                             the database for future overlay.
 203                             If <match_id> isn't defined, a default
 204                             MARC21 ISBN & title match rule will be applied
 205                             for bib imports.
 206     --add-items             use this option to specify that
 207                             item data is embedded in the MARC
 208                             bibs and should be parsed.
 209     --item-action           action to take if --add-items is specifed;
 210                             choices are 'always_add',
 211                             'add_only_for_matches', 'add_only_for_new',
 212                             'ignore', or 'replace'
 213     --no-replace            overlay action for record: default is to
 214                             replace extant with the imported record.
 215     --no-create             nomatch action for record: default is to
 216                             create new record with imported record.
 217     --comment <comment>     optional comment to describe
 218                             the record batch; if the comment
 219                             has spaces in it, surround the
 220                             comment with quotation marks.
 221     --help or -h            show this message.
 222 _USAGE_
 223 }