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