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
  17 # with Koha; if not, write to the Free Software Foundation, Inc.,
  18 # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
  19
  20 use strict;
  21 use warnings;
  22
  23 # standard or CPAN modules used
  24 use CGI;
  25 use CGI::Cookie;
  26 use MARC::File::USMARC;
  27
  28 # Koha modules used
  29 use C4::Context;
  30 use C4::Koha;
  31 use C4::Auth;
  32 use C4::AuthoritiesMarc;
  33 use C4::Output;
  34 use C4::Biblio;
  35 use C4::ImportBatch;
  36 use C4::Matcher;
  37 use C4::BackgroundJob;
  38 use C4::Labels::Batch;
  39
  40 my $script_name = "/cgi-bin/koha/tools/manage-marc-import.pl";
  41
  42 my $input = new CGI;
  43 my $op = $input->param('op') || '';
  44 my $completedJobID = $input->param('completedJobID');
  45 our $runinbackground = $input->param('runinbackground');
  46 my $import_batch_id = $input->param('import_batch_id') || '';
  47
  48 # record list displays
  49 my $offset = $input->param('offset') || 0;
  50 my $results_per_page = $input->param('results_per_page') || 25;
  51
  52 my ($template, $loggedinuser, $cookie)
  53     = get_template_and_user({template_name => "tools/manage-marc-import.tmpl",
  54                  query => $input,
  55                  type => "intranet",
  56                  authnotrequired => 0,
  57                  flagsrequired => {tools => 'manage_staged_marc'},
  58                  debug => 1,
  59                  });
  60
  61 my %cookies = parse CGI::Cookie($cookie);
  62 our $sessionID = $cookies{'CGISESSID'}->value;
  63 our $dbh = C4::Context->dbh;
  64
  65 # Frameworks selection loop
  66 {
  67     my $frameworks = getframeworks;
  68     my $arrayref = [];
  69     while ( my ($key, $value) = each %$frameworks ) {
  70         push @$arrayref, { value => $key, label => $value->{frameworktext} };
  71     }
  72     $template->param( frameworks => $arrayref );
  73 }
  74
  75 if ($op eq "create_labels") {
  76         #create a batch of labels, then lose $op & $import_batch_id so we get back to import batch list.
  77         my $label_batch_id = create_labelbatch_from_importbatch($import_batch_id);
  78         if ($label_batch_id == -1) {
  79             $template->param(   label_batch_msg => "Error attempting to create label batch. Please ask your system administrator to check the log for more details.",
  80                                 message_type    => 'alert',
  81             );
  82         }
  83         else {
  84             $template->param(   label_batch_msg => "Label batch #$label_batch_id created.",
  85                                 message_type    => 'dialog',
  86             );
  87         }
  88         $op='';
  89         $import_batch_id='';
  90 }
  91 if ($op) {
  92     $template->param(script_name => $script_name, $op => 1);
  93 } else {
  94     $template->param(script_name => $script_name);
  95 }
  96
  97 if ($op eq "") {
  98     # displaying a list
  99     if ($import_batch_id eq '') {
 100         import_batches_list($template, $offset, $results_per_page);
 101     } else {
 102         import_records_list($template, $import_batch_id, $offset, $results_per_page);
 103     }
 104 } elsif ($op eq "commit-batch") {
 105     if ($completedJobID) {
 106         add_saved_job_results_to_template($template, $completedJobID);
 107     } else {
 108         my $framework = $input->param('framework');
 109         commit_batch($template, $import_batch_id, $framework);
 110     }
 111     import_records_list($template, $import_batch_id, $offset, $results_per_page);
 112 } elsif ($op eq "revert-batch") {
 113     if ($completedJobID) {
 114         add_saved_job_results_to_template($template, $completedJobID);
 115     } else {
 116         revert_batch($template, $import_batch_id);
 117     }
 118     import_records_list($template, $import_batch_id, $offset, $results_per_page);
 119 } elsif ($op eq "clean-batch") {
 120     CleanBatch($import_batch_id);
 121     import_batches_list($template, $offset, $results_per_page);
 122     $template->param(
 123         did_clean       => 1,
 124         import_batch_id => $import_batch_id,
 125     );
 126 } elsif ($op eq "redo-matching") {
 127     my $new_matcher_id = $input->param('new_matcher_id');
 128     my $current_matcher_id = $input->param('current_matcher_id');
 129     my $overlay_action = $input->param('overlay_action');
 130     my $nomatch_action = $input->param('nomatch_action');
 131     my $item_action = $input->param('item_action');
 132     redo_matching($template, $import_batch_id, $new_matcher_id, $current_matcher_id,
 133                   $overlay_action, $nomatch_action, $item_action);
 134     import_records_list($template, $import_batch_id, $offset, $results_per_page);
 135 }
 136
 137 output_html_with_http_headers $input, $cookie, $template->output;
 138
 139 exit 0;
 140
 141 sub redo_matching {
 142     my ($template, $import_batch_id, $new_matcher_id, $current_matcher_id, $overlay_action, $nomatch_action, $item_action) = @_;
 143     my $rematch_failed = 0;
 144     return if not defined $new_matcher_id and not defined $current_matcher_id;
 145     my $old_overlay_action = GetImportBatchOverlayAction($import_batch_id);
 146     my $old_nomatch_action = GetImportBatchNoMatchAction($import_batch_id);
 147     my $old_item_action = GetImportBatchItemAction($import_batch_id);
 148     return if $new_matcher_id eq $current_matcher_id and
 149               $old_overlay_action eq $overlay_action and
 150               $old_nomatch_action eq $nomatch_action and
 151               $old_item_action eq $item_action;
 152
 153     if ($old_overlay_action ne $overlay_action) {
 154         SetImportBatchOverlayAction($import_batch_id, $overlay_action);
 155         $template->param('changed_overlay_action' => 1);
 156     }
 157     if ($old_nomatch_action ne $nomatch_action) {
 158         SetImportBatchNoMatchAction($import_batch_id, $nomatch_action);
 159         $template->param('changed_nomatch_action' => 1);
 160     }
 161     if ($old_item_action ne $item_action) {
 162         SetImportBatchItemAction($import_batch_id, $item_action);
 163         $template->param('changed_item_action' => 1);
 164     }
 165
 166     my $num_with_matches = 0;
 167     if (defined $new_matcher_id and $new_matcher_id ne "") {
 168         my $matcher = C4::Matcher->fetch($new_matcher_id);
 169         if (defined $matcher) {
 170             $num_with_matches = BatchFindDuplicates($import_batch_id, $matcher);
 171             SetImportBatchMatcher($import_batch_id, $new_matcher_id);
 172         } else {
 173             $rematch_failed = 1;
 174         }
 175     } else {
 176         $num_with_matches = BatchFindDuplicates($import_batch_id, undef);
 177         SetImportBatchMatcher($import_batch_id, undef);
 178         SetImportBatchOverlayAction('create_new');
 179     }
 180     $template->param(rematch_failed => $rematch_failed);
 181     $template->param(rematch_attempted => 1);
 182     $template->param(num_with_matches => $num_with_matches);
 183 }
 184
 185 sub create_labelbatch_from_importbatch {
 186         my ($batch_id) = @_;
 187         my $err = undef;
 188         my $branch_code = C4::Context->userenv->{'branch'};
 189         my $batch = C4::Labels::Batch->new(branch_code => $branch_code);
 190         my @items = GetItemNumbersFromImportBatch($batch_id);
 191         if (grep{$_ == 0} @items) {
 192             warn sprintf('create_labelbatch_from_importbatch() : Call to C4::ImportBatch::GetItemNumbersFromImportBatch returned no item number(s) from import batch #%s.', $batch_id);
 193             return -1;
 194         }
 195         foreach my $item_number (@items) {
 196             $err = $batch->add_item($item_number);
 197             if ($err == -1) {
 198                 warn sprintf('create_labelbatch_from_importbatch() : Error attempting to add item #%s of import batch #%s to label batch.', $item_number, $batch_id);
 199                 return -1;
 200             }
 201         }
 202         return $batch->get_attr('batch_id');
 203 }
 204
 205 sub import_batches_list {
 206     my ($template, $offset, $results_per_page) = @_;
 207     my $batches = GetImportBatchRangeDesc($offset, $results_per_page);
 208
 209     my @list = ();
 210     foreach my $batch (@$batches) {
 211         push @list, {
 212             import_batch_id => $batch->{'import_batch_id'},
 213             num_records => $batch->{'num_records'},
 214             num_items => $batch->{'num_items'},
 215             upload_timestamp => $batch->{'upload_timestamp'},
 216             import_status => $batch->{'import_status'},
 217             file_name => $batch->{'file_name'} || "($batch->{'batch_type'})",
 218             comments => $batch->{'comments'},
 219             can_clean => ($batch->{'import_status'} ne 'cleaned') ? 1 : 0,
 220             record_type => $batch->{'record_type'},
 221         };
 222     }
 223     $template->param(batch_list => \@list);
 224     my $num_batches = GetNumberOfNonZ3950ImportBatches();
 225     add_page_numbers($template, $offset, $results_per_page, $num_batches);
 226     $template->param(offset => $offset);
 227     $template->param(range_top => $offset + $results_per_page - 1);
 228     $template->param(num_results => $num_batches);
 229     $template->param(results_per_page => $results_per_page);
 230
 231 }
 232
 233 sub commit_batch {
 234     my ($template, $import_batch_id, $framework) = @_;
 235
 236     my $job = undef;
 237     $dbh->{AutoCommit} = 0;
 238     my $callback = sub {};
 239     if ($runinbackground) {
 240         $job = put_in_background($import_batch_id);
 241         $callback = progress_callback($job, $dbh);
 242     }
 243     my ($num_added, $num_updated, $num_items_added, $num_items_replaced, $num_items_errored, $num_ignored) =
 244         BatchCommitRecords($import_batch_id, $framework, 50, $callback);
 245     $dbh->commit();
 246
 247     my $results = {
 248         did_commit => 1,
 249         num_added => $num_added,
 250         num_updated => $num_updated,
 251         num_items_added => $num_items_added,
 252         num_items_replaced => $num_items_replaced,
 253         num_items_errored => $num_items_errored,
 254         num_ignored => $num_ignored
 255     };
 256     if ($runinbackground) {
 257         $job->finish($results);
 258     } else {
 259         add_results_to_template($template, $results);
 260     }
 261 }
 262
 263 sub revert_batch {
 264     my ($template, $import_batch_id) = @_;
 265
 266     $dbh->{AutoCommit} = 0;
 267     my $job = undef;
 268     my $callback = sub {};
 269     if ($runinbackground) {
 270         $job = put_in_background($import_batch_id);
 271         $callback = progress_callback($job, $dbh);
 272     }
 273     my ($num_deleted, $num_errors, $num_reverted, $num_items_deleted, $num_ignored) =
 274         BatchRevertRecords($import_batch_id, 50, $callback);
 275     $dbh->commit();
 276
 277     my $results = {
 278         did_revert => 1,
 279         num_deleted => $num_deleted,
 280         num_items_deleted => $num_items_deleted,
 281         num_errors => $num_errors,
 282         num_reverted => $num_reverted,
 283         num_ignored => $num_ignored,
 284     };
 285     if ($runinbackground) {
 286         $job->finish($results);
 287     } else {
 288         add_results_to_template($template, $results);
 289     }
 290 }
 291
 292 sub put_in_background {
 293     my $import_batch_id = shift;
 294
 295     my $batch = GetImportBatch($import_batch_id);
 296     my $job = C4::BackgroundJob->new($sessionID, $batch->{'file_name'}, $ENV{'SCRIPT_NAME'}, $batch->{'num_records'});
 297     my $jobID = $job->id();
 298
 299     # fork off
 300     if (my $pid = fork) {
 301         # parent
 302         # return job ID as JSON
 303
 304         # prevent parent exiting from
 305         # destroying the kid's database handle
 306         # FIXME: according to DBI doc, this may not work for Oracle
 307         $dbh->{InactiveDestroy}  = 1;
 308
 309         my $reply = CGI->new("");
 310         print $reply->header(-type => 'text/html');
 311         print '{"jobID":"' . $jobID . '"}';
 312         exit 0;
 313     } elsif (defined $pid) {
 314         # child
 315         # close STDOUT to signal to Apache that
 316         # we're now running in the background
 317         close STDOUT;
 318         close STDERR;
 319     } else {
 320         # fork failed, so exit immediately
 321         warn "fork failed while attempting to run $ENV{'SCRIPT_NAME'} as a background job";
 322         exit 0;
 323     }
 324     return $job;
 325 }
 326
 327 sub progress_callback {
 328     my $job = shift;
 329     my $dbh = shift;
 330     return sub {
 331         my $progress = shift;
 332         $job->progress($progress);
 333         $dbh->commit();
 334     }
 335 }
 336
 337 sub add_results_to_template {
 338     my $template = shift;
 339     my $results = shift;
 340     $template->param(map { $_ => $results->{$_} } keys %{ $results });
 341 }
 342
 343 sub add_saved_job_results_to_template {
 344     my $template = shift;
 345     my $completedJobID = shift;
 346     my $job = C4::BackgroundJob->fetch($sessionID, $completedJobID);
 347     my $results = $job->results();
 348     add_results_to_template($template, $results);
 349 }
 350
 351 sub import_records_list {
 352     my ($template, $import_batch_id, $offset, $results_per_page) = @_;
 353
 354     my $batch = GetImportBatch($import_batch_id);
 355     my $records = GetImportRecordsRange($import_batch_id, $offset, $results_per_page);
 356     my @list = ();
 357     foreach my $record (@$records) {
 358         my $citation = $record->{'title'} || $record->{'authorized_heading'};
 359         $citation .= " $record->{'author'}" if $record->{'author'};
 360         $citation .= " (" if $record->{'issn'} or $record->{'isbn'};
 361         $citation .= $record->{'isbn'} if $record->{'isbn'};
 362         $citation .= ", " if $record->{'issn'} and $record->{'isbn'};
 363         $citation .= $record->{'issn'} if $record->{'issn'};
 364         $citation .= ")" if $record->{'issn'} or $record->{'isbn'};
 365
 366         my $match = GetImportRecordMatches($record->{'import_record_id'}, 1);
 367         my $match_citation = '';
 368         my $match_id;
 369         if ($#$match > -1) {
 370             if ($match->[0]->{'record_type'} eq 'biblio') {
 371                 $match_citation .= $match->[0]->{'title'} if defined($match->[0]->{'title'});
 372                 $match_citation .= ' ' . $match->[0]->{'author'} if defined($match->[0]->{'author'});
 373                 $match_id = $match->[0]->{'biblionumber'};
 374             } elsif ($match->[0]->{'record_type'} eq 'auth') {
 375                 if (defined($match->[0]->{'authorized_heading'})) {
 376                     $match_citation .= $match->[0]->{'authorized_heading'};
 377                     $match_id = $match->[0]->{'candidate_match_id'};
 378                 }
 379             }
 380         }
 381
 382         push @list,
 383           { import_record_id         => $record->{'import_record_id'},
 384             final_match_id           => $record->{'matched_biblionumber'} || $record->{'matched_authid'},
 385             citation                 => $citation,
 386             status                   => $record->{'status'},
 387             record_sequence          => $record->{'record_sequence'},
 388             overlay_status           => $record->{'overlay_status'},
 389             # Sorry about the match_id being from the "biblionumber" field;
 390             # as it turns out, any match id will go in biblionumber
 391             match_id                 => $match_id,
 392             match_citation           => $match_citation,
 393             match_score              => $#$match > -1 ? $match->[0]->{'score'} : 0,
 394             record_type              => $record->{'record_type'},
 395           };
 396     }
 397     my $num_records = $batch->{'num_records'};
 398     $template->param(record_list => \@list);
 399     add_page_numbers($template, $offset, $results_per_page, $num_records);
 400     $template->param(offset => $offset);
 401     $template->param(range_top => $offset + $results_per_page - 1);
 402     $template->param(num_results => $num_records);
 403     $template->param(results_per_page => $results_per_page);
 404     $template->param(import_batch_id => $import_batch_id);
 405     my $overlay_action = GetImportBatchOverlayAction($import_batch_id);
 406     $template->param("overlay_action_${overlay_action}" => 1);
 407     $template->param(overlay_action => $overlay_action);
 408     my $nomatch_action = GetImportBatchNoMatchAction($import_batch_id);
 409     $template->param("nomatch_action_${nomatch_action}" => 1);
 410     $template->param(nomatch_action => $nomatch_action);
 411     my $item_action = GetImportBatchItemAction($import_batch_id);
 412     $template->param("item_action_${item_action}" => 1);
 413     $template->param(item_action => $item_action);
 414     batch_info($template, $batch);
 415
 416 }
 417
 418 sub batch_info {
 419     my ($template, $batch) = @_;
 420     $template->param(batch_info => 1);
 421     $template->param(file_name => $batch->{'file_name'});
 422     $template->param(comments => $batch->{'comments'});
 423     $template->param(import_status => $batch->{'import_status'});
 424     $template->param(upload_timestamp => $batch->{'upload_timestamp'});
 425     $template->{VARS}->{'record_type'} = $batch->{'record_type'};
 426     $template->param(num_records => $batch->{'num_records'});
 427     $template->param(num_items => $batch->{'num_items'});
 428     if ($batch->{'import_status'} ne 'cleaned') {
 429         $template->param(can_clean => 1);
 430     }
 431     if ($batch->{'num_records'} > 0) {
 432         if ($batch->{'import_status'} eq 'staged' or $batch->{'import_status'} eq 'reverted') {
 433             $template->param(can_commit => 1);
 434         }
 435         if ($batch->{'import_status'} eq 'imported') {
 436             $template->param(can_revert => 1);
 437         }
 438     }
 439     if (defined $batch->{'matcher_id'}) {
 440         my $matcher = C4::Matcher->fetch($batch->{'matcher_id'});
 441         if (defined $matcher) {
 442             $template->param('current_matcher_id' => $batch->{'matcher_id'});
 443             $template->param('current_matcher_code' => $matcher->code());
 444             $template->param('current_matcher_description' => $matcher->description());
 445         }
 446     }
 447     add_matcher_list($template,$batch->{'matcher_id'});
 448 }
 449
 450 sub add_matcher_list {
 451     my ($template,$current_matcher_id) = @_;
 452     my @matchers = C4::Matcher::GetMatcherList();
 453     if (defined $current_matcher_id) {
 454         for (my $i = 0; $i <= $#matchers; $i++) {
 455             if ($matchers[$i]->{'matcher_id'} eq $current_matcher_id) {
 456                 $matchers[$i]->{'selected'} = 1;
 457             }
 458         }
 459     }
 460     $template->param(available_matchers => \@matchers);
 461 }
 462
 463 sub add_page_numbers {
 464     my ($template, $offset, $results_per_page, $total_results) = @_;
 465     my $max_pages = POSIX::ceil($total_results / $results_per_page);
 466     return if $max_pages < 2;
 467     my $current_page = int($offset / $results_per_page) + 1;
 468     my @pages = ();
 469     for (my $i = 1; $i <= $max_pages; $i++) {
 470         push @pages, {
 471             page_number => $i,
 472             current_page => ($current_page == $i) ? 1 : 0,
 473             offset => ($i - 1) * $results_per_page
 474         }
 475     }
 476     $template->param(pages => \@pages);
 477 }
 478