/[webpac2]/trunk/run.pl
This is repository of my old source code which isn't updated any more. Go to git.rot13.org for current projects!
ViewVC logotype

Diff of /trunk/run.pl

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 775 by dpavlin, Sun Nov 5 12:28:37 2006 UTC revision 864 by dpavlin, Sun May 27 22:24:30 2007 UTC
# Line 12  use WebPAC::Input 0.16; Line 12  use WebPAC::Input 0.16;
12  use WebPAC::Store 0.14;  use WebPAC::Store 0.14;
13  use WebPAC::Normalize 0.22;  use WebPAC::Normalize 0.22;
14  use WebPAC::Output::TT;  use WebPAC::Output::TT;
15  use WebPAC::Validate 0.06;  use WebPAC::Validate 0.11;
16  use WebPAC::Output::MARC;  use WebPAC::Output::MARC;
17  use WebPAC::Config;  use WebPAC::Config;
18  use Getopt::Long;  use Getopt::Long;
# Line 21  use Time::HiRes qw/time/; Line 21  use Time::HiRes qw/time/;
21  use File::Slurp;  use File::Slurp;
22  use Data::Dump qw/dump/;  use Data::Dump qw/dump/;
23  use Storable qw/dclone/;  use Storable qw/dclone/;
24    use Pod::Usage qw/pod2usage/;
25    
26  use Proc::Queue size => 1;  use Proc::Queue size => 1;
27  use POSIX ":sys_wait_h"; # imports WNOHANG  use POSIX ":sys_wait_h"; # imports WNOHANG
# Line 31  run.pl - start WebPAC indexing Line 32  run.pl - start WebPAC indexing
32    
33  B<this command will probably go away. Don't get used to it!>  B<this command will probably go away. Don't get used to it!>
34    
35  Options:  =head1 OPTIONS
36    
37  =over 4  =over 4
38    
# Line 67  and subfield usage for each input Line 68  and subfield usage for each input
68    
69  turn on extra validation of imput records, see L<WebPAC::Validation>  turn on extra validation of imput records, see L<WebPAC::Validation>
70    
71    You can use special variables C<$database> and $C<$input> in this parametar
72    like C<--validate 'conf/validate/$database-$input'> to construct filename
73    
74    =item --validate-delimiters path/to/validate_delimiters_file
75    
76    this option is used with C<--validate> to turn on extra validation of
77    delimiters. If file is non existant, it will be created on first run.
78    
79    =item --marc-generate
80    
81    Generate MARC file. This will automatically be on if file contains C<marc*> directives.
82    You can use this option as C<--no-marc-generate> to disable MARC generation.
83    
84  =item --marc-lint  =item --marc-lint
85    
86  By default turned on if normalisation file has C<marc*> directives. You can disable lint  By default turned on if normalisation file has C<marc*> directives. You can disable lint
# Line 102  my $debug = 0; Line 116  my $debug = 0;
116  my $only_filter;  my $only_filter;
117  my $stats = 0;  my $stats = 0;
118  my $validate_path;  my $validate_path;
119    my $validate_delimiters_path;
120    my $marc_generate = 1;
121  my $marc_lint = 1;  my $marc_lint = 1;
122  my $marc_dump = 0;  my $marc_dump = 0;
123  my $parallel = 0;  my $parallel = 0;
124  my $only_links = 0;  my $only_links = 0;
125  my $merge = 0;  my $merge = 0;
126    my $help;
127    
128  my $log = _new WebPAC::Common()->_get_logger();  my $log = _new WebPAC::Common()->_get_logger();
129    
# Line 120  GetOptions( Line 137  GetOptions(
137          "debug+" => \$debug,          "debug+" => \$debug,
138          "stats" => \$stats,          "stats" => \$stats,
139          "validate=s" => \$validate_path,          "validate=s" => \$validate_path,
140            "validate-delimiters=s" => \$validate_delimiters_path,
141            "marc-generate!" => \$marc_generate,
142          "marc-lint!" => \$marc_lint,          "marc-lint!" => \$marc_lint,
143          "marc-dump!" => \$marc_dump,          "marc-dump!" => \$marc_dump,
144          "parallel=i" => \$parallel,          "parallel=i" => \$parallel,
145          "only-links!" => \$only_links,          "only-links!" => \$only_links,
146          "merge" => \$merge,          "merge" => \$merge,
147            "help" => \$help,
148  );  );
149    
150    $marc_generate = 0 if ( $validate_delimiters_path );
151    
152    pod2usage(-verbose => 2) if ($help);
153    
154  my $config = new WebPAC::Config( path => $config_path );  my $config = new WebPAC::Config( path => $config_path );
155    
156  #print "config = ",dump($config) if ($debug);  #print "config = ",dump($config) if ($debug);
# Line 151  if ($merge) { Line 175  if ($merge) {
175          $log->info("created merge batch file $estcmd_path");          $log->info("created merge batch file $estcmd_path");
176  }  }
177    
   
178  my $validate;  my $validate;
179  $validate = new WebPAC::Validate(  $validate = new WebPAC::Validate(
180          path => $validate_path,          delimiters => $config->webpac('delimiters'),
181  ) if ($validate_path);  ) if ($validate_path || $validate_delimiters_path);
   
182    
183  my $use_indexer = $config->use_indexer;  my $use_indexer = $config->use_indexer;
184  $stats ||= $validate;  $stats ||= $validate;
# Line 190  sub create_ds_config { Line 212  sub create_ds_config {
212          return $c;          return $c;
213  }  }
214    
215  while (my ($database, $db_config) = each %{ $config->databases }) {  foreach my $database ( sort keys %{ $config->databases } ) {
216            my $db_config = $config->databases->{$database};
217    
218          my ($only_database,$only_input) = split(m#/#, $only_filter) if ($only_filter);          my ($only_database,$only_input) = split(m#/#, $only_filter) if ($only_filter);
219          next if ($only_database && $database !~ m/$only_database/i);          next if ($only_database && $database !~ m/$only_database/i);
# Line 423  while (my ($database, $db_config) = each Line 446  while (my ($database, $db_config) = each
446                          print $report_fh "Report for database '$database' input '$input_name' records ",                          print $report_fh "Report for database '$database' input '$input_name' records ",
447                                  $offset || 1, "-", $limit || $input->{limit} || $maxmfn, "\n\n";                                  $offset || 1, "-", $limit || $input->{limit} || $maxmfn, "\n\n";
448                          $log->info("Generating report file $path");                          $log->info("Generating report file $path");
449    
450                            if ( $validate ) {
451                                    $validate->read_validate_file( $validate->fill_in( $validate_path, database => $database, input => $input_name ) ) if ( $validate_path );
452                                    $validate->read_validate_delimiters_file( $validate->fill_in( $validate_delimiters_path, database => $database, input => $input_name ) ) if ( $validate_delimiters_path );
453                            }
454                  }                  }
455    
456                  my $marc;                  my $marc;
457                  if ($parser->have_rules( 'marc', $database, $input_name )) {                  if ($marc_generate && $parser->have_rules( 'marc', $database, $input_name )) {
458                          $marc = new WebPAC::Output::MARC(                          $marc = new WebPAC::Output::MARC(
459                                  path => "out/marc/${database}-${input_name}.marc",                                  path => "out/marc/${database}-${input_name}.marc",
460                                  lint => $marc_lint,                                  lint => $marc_lint,
# Line 532  while (my ($database, $db_config) = each Line 560  while (my ($database, $db_config) = each
560                                          $marc->add(                                          $marc->add(
561                                                  id => $mfn . ( $i ? "/$i" : '' ),                                                  id => $mfn . ( $i ? "/$i" : '' ),
562                                                  fields => $fields,                                                  fields => $fields,
563                                                  leader => WebPAC::Normalize::marc_leader(),                                                  leader => WebPAC::Normalize::_get_marc_leader(),
564                                                  row => $row,                                                  row => $row,
565                                          );                                          );
566                                          $i++;                                          $i++;
# Line 548  while (my ($database, $db_config) = each Line 576  while (my ($database, $db_config) = each
576                                  $log->info("validation errors:\n$errors\n" );                                  $log->info("validation errors:\n$errors\n" );
577                                  print $report_fh "$errors\n" if ($report_fh);                                  print $report_fh "$errors\n" if ($report_fh);
578                          }                          }
579    
580                            print $report_fh "\nAll possible subfields/delimiter templates:\n", $validate->delimiters_templates( report => 1, current_input => 1 ), "\n\n";
581    
582                            # must be last thing that touches $validate for this input
583                            $validate->reset;
584                  }                  }
585    
586                  if ($stats) {                  if ($stats) {
# Line 560  while (my ($database, $db_config) = each Line 593  while (my ($database, $db_config) = each
593                  $marc->finish if ($marc);                  $marc->finish if ($marc);
594    
595                  # close report                  # close report
596                  close($report_fh) if ($report_fh)                  close($report_fh) if ($report_fh);
   
597          }          }
598    
599          eval { $indexer->finish } if ($indexer && $indexer->can('finish'));          eval { $indexer->finish } if ($indexer && $indexer->can('finish'));
# Line 588  if ($parallel) { Line 620  if ($parallel) {
620          $log->info("all parallel processes finished");          $log->info("all parallel processes finished");
621  }  }
622    
623    # save new delimiters if needed
624    $validate->save_delimiters_templates if ( $validate_delimiters_path );
625    
626  #  #
627  # handle links or merge after indexing  # handle links or merge after indexing
628  #  #
Legend:
Removed from v.775  
changed lines
  Added in v.864
  ViewVC Help
Powered by ViewVC 1.1.26