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 Modern::Perl;
  23 BEGIN {
  24     # find Koha's Perl modules
  25     # test carefully before changing this
  26     use FindBin;
  27     eval { require "$FindBin::Bin/kohalib.pl" };
  28 }
  29
  30 use Koha::Script;
  31 use C4::Context;
  32 use C4::ImportBatch;
  33 use C4::Matcher;
  34 use C4::MarcModificationTemplates;
  35 use Getopt::Long;
  36
  37 $| = 1;
  38
  39 # command-line parameters
  40 my $record_type = "biblio";
  41 my $encoding = "UTF-8";
  42 my $authorities = 0;
  43 my $match = 0;
  44 my $add_items = 0;
  45 my $input_file = "";
  46 my $batch_comment = "";
  47 my $want_help = 0;
  48 my $no_replace;
  49 my $format = 'ISO2709';
  50 my $no_create;
  51 my $item_action = 'always_add';
  52 my $marc_mod_template = '';
  53 my $marc_mod_template_id = undef;
  54
  55 my $result = GetOptions(
  56     'encoding:s'         => \$encoding,
  57     'file:s'             => \$input_file,
  58     'format:s'           => \$format,
  59     'match|match-bibs:s' => \$match,
  60     'add-items'          => \$add_items,
  61     'item-action:s'      => \$item_action,
  62     'no-replace'         => \$no_replace,
  63     'no-create'          => \$no_create,
  64     'comment:s'          => \$batch_comment,
  65     'authorities'        => \$authorities,
  66     'marcmodtemplate:s'  => \$marc_mod_template,
  67     'h|help'             => \$want_help
  68 );
  69
  70 if($marc_mod_template ne '') {
  71    my @templates = GetModificationTemplates();
  72    foreach my $this_template (@templates) {
  73        if($this_template->{'name'} eq $marc_mod_template) {
  74           if(!defined $marc_mod_template_id) {
  75                $marc_mod_template_id = $this_template->{'template_id'};
  76            } else {
  77                print "WARNING: MARC modification template name " .
  78                    "'$marc_mod_template' matches multiple templates. " .
  79                    "Please fix this issue before proceeding.\n";
  80                exit 1;
  81            }
  82        }
  83    }
  84
  85    if(!defined $marc_mod_template_id ) {
  86        die "Can't locate MARC modification template '$marc_mod_template'\n";
  87    }
  88 }
  89
  90 $record_type = 'auth' if ($authorities);
  91
  92 if (not $result or $input_file eq "" or $want_help) {
  93     print_usage();
  94     exit 0;
  95 }
  96 if ( $format !~ /^(MARCXML|ISO2709)$/i ) {
  97     print "\n --format must be MARCXML or ISO2709\n";
  98     print_usage();
  99     exit 0;
 100 }
 101
 102 unless (-r $input_file) {
 103     die "$0: cannot open input file $input_file: $!\n";
 104 }
 105
 106 my $dbh = C4::Context->dbh;
 107 $dbh->{AutoCommit} = 0;
 108 process_batch(
 109     {
 110         format               => $format,
 111         input_file           => $input_file,
 112         record_type          => $record_type,
 113         match                => $match,
 114         add_items            => $add_items,
 115         batch_comment        => $batch_comment,
 116         encoding             => $encoding,
 117         no_replace           => $no_replace,
 118         no_create            => $no_create,
 119         item_action          => $item_action,
 120         marc_mod_template_id => $marc_mod_template_id,
 121     }
 122 );
 123 $dbh->commit();
 124
 125 exit 0;
 126
 127 sub process_batch {
 128     my ( $params ) = @_;  #Possible params are: format input_file record_type match add_items batch_comment encoding no_replace no_create item_action
 129     my $format = $params->{format} // '';
 130     my $record_type = $params->{record_type} // 'biblio';
 131
 132     my ( $errors, $marc_records );
 133     if( $format eq 'ISO2709' ) {
 134         ( $errors, $marc_records ) = C4::ImportBatch::RecordsFromISO2709File(
 135             $params->{input_file}, $record_type, $params->{encoding} );
 136     } elsif( $format eq 'MARCXML' ) {
 137         ( $errors, $marc_records ) = C4::ImportBatch::RecordsFromMARCXMLFile(
 138             $params->{input_file}, $params->{encoding} );
 139     }
 140     warn ( join ',', @$errors ) if @$errors;
 141     my $num_input_records = ($marc_records) ? scalar(@$marc_records) : 0;
 142
 143     print "... staging MARC records -- please wait\n";
 144     #FIXME: We should really allow the use of marc modification frameworks and to_marc plugins here if possible
 145     my ( $batch_id, $num_valid_records, $num_items, @import_errors ) =
 146       BatchStageMarcRecords(
 147         $record_type,                      $params->{encoding},
 148         $marc_records,                     $params->{input_file},
 149         $params->{'marc_mod_template_id'}, $params->{batch_comment},
 150         '',                                $params->{add_items},
 151         0,                                 100,
 152         \&print_progress_and_commit
 153       );
 154     print "... finished staging MARC records\n";
 155
 156     my $num_with_matches = 0;
 157     if ( $params->{match} ) {
 158         my $matcher = C4::Matcher->fetch( $params->{match} );
 159         if (defined $matcher) {
 160             SetImportBatchMatcher( $batch_id, $params->{match} );
 161         } elsif ($record_type eq 'biblio')  {
 162             $matcher = C4::Matcher->new($record_type);
 163             $matcher->add_simple_matchpoint('isbn', 1000, '020', 'a', -1, 0, '');
 164             $matcher->add_simple_required_check('245', 'a', -1, 0, '',
 165                                             '245', 'a', -1, 0, '');
 166         }
 167         # set default record overlay behavior
 168         SetImportBatchOverlayAction( $batch_id, $params->{no_replace} ? 'ignore' : 'replace' );
 169         SetImportBatchNoMatchAction( $batch_id, $params->{no_create} ? 'ignore' : 'create_new' );
 170         SetImportBatchItemAction( $batch_id, $params->{item_action} );
 171         print "... looking for matches with records already in database\n";
 172         $num_with_matches = BatchFindDuplicates($batch_id, $matcher, 10, 100, \&print_progress_and_commit);
 173         print "... finished looking for matches\n";
 174     }
 175
 176     my $num_invalid_records = scalar(@import_errors);
 177     print <<_SUMMARY_;
 178
 179 MARC record staging report
 180 ------------------------------------
 181 Input file:                 $params->{input_file}
 182 Record type:                $record_type
 183 Number of input records:    $num_input_records
 184 Number of valid records:    $num_valid_records
 185 Number of invalid records:  $num_invalid_records
 186 _SUMMARY_
 187     if( $params->{match} ) {
 188         print "Number of records matched:  $num_with_matches\n";
 189     } else {
 190         print "Incoming records not matched against existing records (--match option not supplied)\n";
 191     }
 192     if ($record_type eq 'biblio') {
 193         if ( $params->{add_items} ) {
 194             print "Number of items parsed:  $num_items\n";
 195         } else {
 196             print "No items parsed (--add-items option not supplied)\n";
 197         }
 198     }
 199
 200     print "\n";
 201     print "Batch number assigned:  $batch_id\n";
 202     print "\n";
 203 }
 204
 205 sub print_progress_and_commit {
 206     my $recs = shift;
 207     $dbh->commit();
 208     print "... processed $recs records\n";
 209 }
 210
 211 sub print_usage {
 212     print <<_USAGE_;
 213 $0: stage MARC file into reservoir.
 214
 215 Use this batch job to load a file of MARC bibliographic
 216 (with optional item information) or authority records into
 217 the Koha reservoir.
 218
 219 After running this program to stage your file, you can use
 220 either the batch job commit_file.pl or the Koha
 221 Tools option "Manage Staged MARC Records" to load the
 222 records into the main Koha database.
 223
 224 Parameters:
 225     --file <file_name>      name of input MARC bib file
 226     --authorities           stage authority records instead of bibs
 227     --encoding <encoding>   encoding of MARC records, default is UTF-8.
 228                             Other possible options are: MARC-8,
 229                             ISO_5426, ISO_6937, ISO_8859-1, EUC-KR
 230     --format                The MARC transport format to use?
 231                             Defaults to ISO2709.
 232                             Available values, MARCXML, ISO2709.
 233     --match <match_id>      use this option to match records
 234                             in the file with records already in
 235                             the database for future overlay.
 236                             If <match_id> isn't defined, a default
 237                             MARC21 ISBN & title match rule will be applied
 238                             for bib imports.
 239     --add-items             use this option to specify that
 240                             item data is embedded in the MARC
 241                             bibs and should be parsed.
 242     --item-action           action to take if --add-items is specifed;
 243                             choices are 'always_add',
 244                             'add_only_for_matches', 'add_only_for_new',
 245                             'ignore', or 'replace'
 246     --no-replace            overlay action for record: default is to
 247                             replace extant with the imported record.
 248     --no-create             nomatch action for record: default is to
 249                             create new record with imported record.
 250     --comment <comment>     optional comment to describe
 251                             the record batch; if the comment
 252                             has spaces in it, surround the
 253                             comment with quotation marks.
 254     --marcmodtemplate <TEMPLATE>
 255                             This parameter allows you to specify the
 256                             name of an existing MARC modification
 257                             template to apply as the MARC records are
 258                             imported (these templates are created in
 259                             the "MARC modification templates" tool in
 260                             Koha). If not specified, no MARC modification
 261                             templates are used (default).
 262     --help or -h            show this message.
 263 _USAGE_
 264 }