tools/manage-marc-import.pl

   1 #!/usr/bin/perl
   2
   3 # Copyright (C) 2007 LibLime
   4 #
   5 # This file is part of Koha.
   6 #
   7 # Koha is free software; you can redistribute it and/or modify it under the
   8 # terms of the GNU General Public License as published by the Free Software
   9 # Foundation; either version 2 of the License, or (at your option) any later
  10 # version.
  11 #
  12 # Koha is distributed in the hope that it will be useful, but WITHOUT ANY
  13 # WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
  14 # A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
  15 #
  16 # You should have received a copy of the GNU General Public License along with
  17 # Koha; if not, write to the Free Software Foundation, Inc., 59 Temple Place,
  18 # Suite 330, Boston, MA  02111-1307 USA
  19
  20 use strict;
  21
  22 # standard or CPAN modules used
  23 use CGI;
  24 use CGI::Cookie;
  25 use MARC::File::USMARC;
  26
  27 # Koha modules used
  28 use C4::Context;
  29 use C4::Auth;
  30 use C4::Input;
  31 use C4::Output;
  32 use C4::Biblio;
  33 use C4::ImportBatch;
  34 use C4::Matcher;
  35 use C4::BackgroundJob;
  36 use C4::Labels qw(add_batch);
  37
  38 my $script_name = "/cgi-bin/koha/tools/manage-marc-import.pl";
  39
  40 my $input = new CGI;
  41 my $op = $input->param('op');
  42 my $completedJobID = $input->param('completedJobID');
  43 my $runinbackground = $input->param('runinbackground');
  44 my $import_batch_id = $input->param('import_batch_id');
  45
  46 # record list displays
  47 my $offset = $input->param('offset') || 0;
  48 my $results_per_page = $input->param('results_per_page') || 25;
  49
  50 my ($template, $loggedinuser, $cookie)
  51     = get_template_and_user({template_name => "tools/manage-marc-import.tmpl",
  52                  query => $input,
  53                  type => "intranet",
  54                  authnotrequired => 0,
  55                  flagsrequired => {tools => 'manage_staged_marc'},
  56                  debug => 1,
  57                  });
  58
  59 my %cookies = parse CGI::Cookie($cookie);
  60 my $sessionID = $cookies{'CGISESSID'}->value;
  61 my $dbh = C4::Context->dbh;
  62
  63 if ($op eq "create_labels") {
  64         #create a batch of labels, then lose $op & $import_batch_id so we get back to import batch list.
  65         my $label_batch_id = create_labelbatch_from_importbatch($import_batch_id);
  66         $template->param( label_batch => $label_batch_id );
  67         $op='';
  68         $import_batch_id='';
  69 }
  70 if ($op) {
  71     $template->param(script_name => $script_name, $op => 1);
  72 } else {
  73     $template->param(script_name => $script_name);
  74 }
  75
  76 if ($op eq "") {
  77     # displaying a list
  78     if ($import_batch_id eq "") {
  79         import_batches_list($template, $offset, $results_per_page);
  80     } else {
  81         import_biblios_list($template, $import_batch_id, $offset, $results_per_page);
  82     }
  83 } elsif ($op eq "commit-batch") {
  84     if ($completedJobID) {
  85         add_saved_job_results_to_template($template, $completedJobID);
  86     } else {
  87         commit_batch($template, $import_batch_id);
  88     }
  89     import_biblios_list($template, $import_batch_id, $offset, $results_per_page);
  90 } elsif ($op eq "revert-batch") {
  91     if ($completedJobID) {
  92         add_saved_job_results_to_template($template, $completedJobID);
  93     } else {
  94         revert_batch($template, $import_batch_id);
  95     }
  96     import_biblios_list($template, $import_batch_id, $offset, $results_per_page);
  97 } elsif ($op eq "clean-batch") {
  98     ;
  99 } elsif ($op eq "redo-matching") {
 100     my $new_matcher_id = $input->param('new_matcher_id');
 101     my $current_matcher_id = $input->param('current_matcher_id');
 102     my $overlay_action = $input->param('overlay_action');
 103     my $nomatch_action = $input->param('nomatch_action');
 104     my $item_action = $input->param('item_action');
 105     redo_matching($template, $import_batch_id, $new_matcher_id, $current_matcher_id,
 106                   $overlay_action, $nomatch_action, $item_action);
 107     import_biblios_list($template, $import_batch_id, $offset, $results_per_page);
 108 }
 109
 110 output_html_with_http_headers $input, $cookie, $template->output;
 111
 112 exit 0;
 113
 114 sub redo_matching {
 115     my ($template, $import_batch_id, $new_matcher_id, $current_matcher_id, $overlay_action, $nomatch_action, $item_action) = @_;
 116     my $rematch_failed = 0;
 117     return if not defined $new_matcher_id and not defined $current_matcher_id;
 118     my $old_overlay_action = GetImportBatchOverlayAction($import_batch_id);
 119     my $old_nomatch_action = GetImportBatchNoMatchAction($import_batch_id);
 120     my $old_item_action = GetImportBatchItemAction($import_batch_id);
 121     return if $new_matcher_id == $current_matcher_id and
 122               $old_overlay_action eq $overlay_action and
 123               $old_nomatch_action eq $nomatch_action and
 124               $old_item_action eq $item_action;
 125
 126     if ($old_overlay_action ne $overlay_action) {
 127         SetImportBatchOverlayAction($import_batch_id, $overlay_action);
 128         $template->param('changed_overlay_action' => 1);
 129     }
 130     if ($old_nomatch_action ne $nomatch_action) {
 131         SetImportBatchNoMatchAction($import_batch_id, $nomatch_action);
 132         $template->param('changed_nomatch_action' => 1);
 133     }
 134     if ($old_item_action ne $item_action) {
 135         SetImportBatchItemAction($import_batch_id, $item_action);
 136         $template->param('changed_item_action' => 1);
 137     }
 138
 139     if ($new_matcher_id == $current_matcher_id) {
 140         return;
 141     }
 142
 143     my $num_with_matches = 0;
 144     if (defined $new_matcher_id and $new_matcher_id ne "") {
 145         my $matcher = C4::Matcher->fetch($new_matcher_id);
 146         if (defined $matcher) {
 147             $num_with_matches = BatchFindBibDuplicates($import_batch_id, $matcher);
 148             SetImportBatchMatcher($import_batch_id, $new_matcher_id);
 149         } else {
 150             $rematch_failed = 1;
 151         }
 152     } else {
 153         $num_with_matches = BatchFindBibDuplicates($import_batch_id, undef);
 154         SetImportBatchMatcher($import_batch_id, undef);
 155         SetImportBatchOverlayAction('create_new');
 156     }
 157     $template->param(rematch_failed => $rematch_failed);
 158     $template->param(rematch_attempted => 1);
 159     $template->param(num_with_matches => $num_with_matches);
 160 }
 161
 162 sub create_labelbatch_from_importbatch {
 163         my ($batch_id) = @_;
 164         my @items = GetItemNumbersFromImportBatch($batch_id);
 165         my $labelbatch = add_batch('labels',\@items);
 166         return $labelbatch;
 167 }
 168
 169 sub import_batches_list {
 170     my ($template, $offset, $results_per_page) = @_;
 171     my $batches = GetImportBatchRangeDesc($offset, $results_per_page);
 172
 173     my @list = ();
 174     foreach my $batch (@$batches) {
 175         push @list, {
 176             import_batch_id => $batch->{'import_batch_id'},
 177             num_biblios => $batch->{'num_biblios'},
 178             num_items => $batch->{'num_items'},
 179             upload_timestamp => $batch->{'upload_timestamp'},
 180             import_status => $batch->{'import_status'},
 181             file_name => $batch->{'file_name'},
 182             comments => $batch->{'comments'}
 183         };
 184     }
 185     $template->param(batch_list => \@list);
 186     my $num_batches = GetNumberOfNonZ3950ImportBatches();
 187     add_page_numbers($template, $offset, $results_per_page, $num_batches);
 188     $template->param(offset => $offset);
 189     $template->param(range_top => $offset + $results_per_page - 1);
 190     $template->param(num_results => $num_batches);
 191     $template->param(results_per_page => $results_per_page);
 192
 193 }
 194
 195 sub commit_batch {
 196     my ($template, $import_batch_id) = @_;
 197
 198     my $job = undef;
 199     $dbh->{AutoCommit} = 0;
 200     my $callback = sub {};
 201     if ($runinbackground) {
 202         $job = put_in_background($import_batch_id);
 203         $callback = progress_callback($job, $dbh);
 204     }
 205     my ($num_added, $num_updated, $num_items_added, $num_items_errored, $num_ignored) =
 206         BatchCommitBibRecords($import_batch_id, 50, $callback);
 207     $dbh->commit();
 208
 209     my $results = {
 210         did_commit => 1,
 211         num_added => $num_added,
 212         num_updated => $num_updated,
 213         num_items_added => $num_items_added,
 214         num_items_errored => $num_items_errored,
 215         num_ignored => $num_ignored
 216     };
 217     if ($runinbackground) {
 218         $job->finish($results);
 219     } else {
 220         add_results_to_template($template, $results);
 221     }
 222 }
 223
 224 sub revert_batch {
 225     my ($template, $import_batch_id) = @_;
 226
 227     $dbh->{AutoCommit} = 0;
 228     my $job = undef;
 229     my $callback = sub {};
 230     if ($runinbackground) {
 231         $job = put_in_background($import_batch_id);
 232         $callback = progress_callback($job, $dbh);
 233     }
 234     my ($num_deleted, $num_errors, $num_reverted, $num_items_deleted, $num_ignored) =
 235         BatchRevertBibRecords($import_batch_id, 50, $callback);
 236     $dbh->commit();
 237
 238     my $results = {
 239         did_revert => 1,
 240         num_deleted => $num_deleted,
 241         num_items_deleted => $num_items_deleted,
 242         num_errors => $num_errors,
 243         num_reverted => $num_reverted,
 244         num_ignored => $num_ignored,
 245     };
 246     if ($runinbackground) {
 247         $job->finish($results);
 248     } else {
 249         add_results_to_template($template, $results);
 250     }
 251 }
 252
 253 sub put_in_background {
 254     my $import_batch_id = shift;
 255
 256     my $batch = GetImportBatch($import_batch_id);
 257     my $job = C4::BackgroundJob->new($sessionID, $batch->{'file_name'}, $ENV{'SCRIPT_NAME'}, $batch->{'num_biblios'});
 258     my $jobID = $job->id();
 259
 260     # fork off
 261     if (my $pid = fork) {
 262         # parent
 263         # return job ID as JSON
 264
 265         # prevent parent exiting from
 266         # destroying the kid's database handle
 267         # FIXME: according to DBI doc, this may not work for Oracle
 268         $dbh->{InactiveDestroy}  = 1;
 269
 270         my $reply = CGI->new("");
 271         print $reply->header(-type => 'text/html');
 272         print "{ jobID: '$jobID' }";
 273         exit 0;
 274     } elsif (defined $pid) {
 275         # child
 276         # close STDOUT to signal to Apache that
 277         # we're now running in the background
 278         close STDOUT;
 279         close STDERR;
 280     } else {
 281         # fork failed, so exit immediately
 282         warn "fork failed while attempting to run $ENV{'SCRIPT_NAME'} as a background job";
 283         exit 0;
 284     }
 285     return $job;
 286 }
 287
 288 sub progress_callback {
 289     my $job = shift;
 290     my $dbh = shift;
 291     return sub {
 292         my $progress = shift;
 293         $job->progress($progress);
 294         $dbh->commit();
 295     }
 296 }
 297
 298 sub add_results_to_template {
 299     my $template = shift;
 300     my $results = shift;
 301     $template->param(map { $_ => $results->{$_} } keys %{ $results });
 302 }
 303
 304 sub add_saved_job_results_to_template {
 305     my $template = shift;
 306     my $completedJobID = shift;
 307     my $job = C4::BackgroundJob->fetch($sessionID, $completedJobID);
 308     my $results = $job->results();
 309     add_results_to_template($template, $results);
 310 }
 311
 312 sub import_biblios_list {
 313     my ($template, $import_batch_id, $offset, $results_per_page) = @_;
 314
 315     my $batch = GetImportBatch($import_batch_id);
 316     my $biblios = GetImportBibliosRange($import_batch_id, $offset, $results_per_page);
 317     my @list = ();
 318     foreach my $biblio (@$biblios) {
 319         my $citation = $biblio->{'title'};
 320         $citation .= " $biblio->{'author'}" if $biblio->{'author'};
 321         $citation .= " (" if $biblio->{'issn'} or $biblio->{'isbn'};
 322         $citation .= $biblio->{'isbn'} if $biblio->{'isbn'};
 323         $citation .= ", " if $biblio->{'issn'} and $biblio->{'isbn'};
 324         $citation .= $biblio->{'issn'} if $biblio->{'issn'};
 325         $citation .= ")" if $biblio->{'issn'} or $biblio->{'isbn'};
 326         my $match = GetImportRecordMatches($biblio->{'import_record_id'}, 1);
 327         push @list, {
 328             import_record_id => $biblio->{'import_record_id'},
 329             citation => $citation,
 330             status => $biblio->{'status'},
 331             record_sequence => $biblio->{'record_sequence'},
 332             overlay_status => $biblio->{'overlay_status'},
 333             match_biblionumber => $#$match > -1 ? $match->[0]->{'biblionumber'} : 0,
 334             match_citation => $#$match > -1 ? $match->[0]->{'title'} . ' ' . $match->[0]->{'author'} : '',
 335             match_score => $#$match > -1 ? $match->[0]->{'score'} : 0,
 336         };
 337     }
 338     my $num_biblios = $batch->{'num_biblios'};
 339     $template->param(biblio_list => \@list);
 340     add_page_numbers($template, $offset, $results_per_page, $num_biblios);
 341     $template->param(offset => $offset);
 342     $template->param(range_top => $offset + $results_per_page - 1);
 343     $template->param(num_results => $num_biblios);
 344     $template->param(results_per_page => $results_per_page);
 345     $template->param(import_batch_id => $import_batch_id);
 346     my $overlay_action = GetImportBatchOverlayAction($import_batch_id);
 347     $template->param("overlay_action_${overlay_action}" => 1);
 348     $template->param(overlay_action => $overlay_action);
 349     my $nomatch_action = GetImportBatchNoMatchAction($import_batch_id);
 350     $template->param("nomatch_action_${nomatch_action}" => 1);
 351     $template->param(nomatch_action => $nomatch_action);
 352     my $item_action = GetImportBatchItemAction($import_batch_id);
 353     $template->param("item_action_${item_action}" => 1);
 354     $template->param(item_action => $item_action);
 355     batch_info($template, $batch);
 356
 357 }
 358
 359 sub batch_info {
 360     my ($template, $batch) = @_;
 361     $template->param(batch_info => 1);
 362     $template->param(file_name => $batch->{'file_name'});
 363     $template->param(comments => $batch->{'comments'});
 364     $template->param(import_status => $batch->{'import_status'});
 365     $template->param(upload_timestamp => $batch->{'upload_timestamp'});
 366     $template->param(num_biblios => $batch->{'num_biblios'});
 367     $template->param(num_items => $batch->{'num_biblios'});
 368     if ($batch->{'num_biblios'} > 0) {
 369         if ($batch->{'import_status'} eq 'staged' or $batch->{'import_status'} eq 'reverted') {
 370             $template->param(can_commit => 1);
 371         }
 372         if ($batch->{'import_status'} eq 'imported') {
 373             $template->param(can_revert => 1);
 374         }
 375     }
 376     if (defined $batch->{'matcher_id'}) {
 377         my $matcher = C4::Matcher->fetch($batch->{'matcher_id'});
 378         if (defined $matcher) {
 379             $template->param('current_matcher_id' => $batch->{'matcher_id'});
 380             $template->param('current_matcher_code' => $matcher->code());
 381             $template->param('current_matcher_description' => $matcher->description());
 382         }
 383     }
 384     add_matcher_list($batch->{'matcher_id'});
 385 }
 386
 387 sub add_matcher_list {
 388     my $current_matcher_id = shift;
 389     my @matchers = C4::Matcher::GetMatcherList();
 390     if (defined $current_matcher_id) {
 391         for (my $i = 0; $i <= $#matchers; $i++) {
 392             if ($matchers[$i]->{'matcher_id'} == $current_matcher_id) {
 393                 $matchers[$i]->{'selected'} = 1;
 394             }
 395         }
 396     }
 397     $template->param(available_matchers => \@matchers);
 398 }
 399
 400 sub add_page_numbers {
 401     my ($template, $offset, $results_per_page, $total_results) = @_;
 402     my $max_pages = POSIX::ceil($total_results / $results_per_page);
 403     return if $max_pages < 2;
 404     my $current_page = int($offset / $results_per_page) + 1;
 405     my @pages = ();
 406     for (my $i = 1; $i <= $max_pages; $i++) {
 407         push @pages, {
 408             page_number => $i,
 409             current_page => ($current_page == $i) ? 1 : 0,
 410             offset => ($i - 1) * $results_per_page
 411         }
 412     }
 413     $template->param(pages => \@pages);
 414 }
 415