/[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 774 by dpavlin, Fri Nov 3 20:56:21 2006 UTC revision 850 by dpavlin, Sun May 27 10:50:39 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    =item --validate-delimiters path/to/validate_delimiters_file
72    
73    this option is used with C<--validate> to turn on extra validation of
74    delimiters. If file is non existant, it will be created on first run.
75    
76    =item --marc-generate
77    
78    Generate MARC file. This will automatically be on if file contains C<marc*> directives.
79    You can use this option as C<--no-marc-generate> to disable MARC generation.
80    
81  =item --marc-lint  =item --marc-lint
82    
83  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 113  my $debug = 0;
113  my $only_filter;  my $only_filter;
114  my $stats = 0;  my $stats = 0;
115  my $validate_path;  my $validate_path;
116    my $validate_delimiters_path;
117    my $marc_generate = 1;
118  my $marc_lint = 1;  my $marc_lint = 1;
119  my $marc_dump = 0;  my $marc_dump = 0;
120  my $parallel = 0;  my $parallel = 0;
121  my $only_links = 0;  my $only_links = 0;
122  my $merge = 0;  my $merge = 0;
123    my $help;
124    
125  my $log = _new WebPAC::Common()->_get_logger();  my $log = _new WebPAC::Common()->_get_logger();
126    
# Line 120  GetOptions( Line 134  GetOptions(
134          "debug+" => \$debug,          "debug+" => \$debug,
135          "stats" => \$stats,          "stats" => \$stats,
136          "validate=s" => \$validate_path,          "validate=s" => \$validate_path,
137            "validate-delimiters=s" => \$validate_delimiters_path,
138            "marc-generate!" => \$marc_generate,
139          "marc-lint!" => \$marc_lint,          "marc-lint!" => \$marc_lint,
140          "marc-dump!" => \$marc_dump,          "marc-dump!" => \$marc_dump,
141          "parallel=i" => \$parallel,          "parallel=i" => \$parallel,
142          "only-links!" => \$only_links,          "only-links!" => \$only_links,
143          "merge" => \$merge,          "merge" => \$merge,
144            "help" => \$help,
145  );  );
146    
147    pod2usage(-verbose => 2) if ($help);
148    
149  my $config = new WebPAC::Config( path => $config_path );  my $config = new WebPAC::Config( path => $config_path );
150    
151  #print "config = ",dump($config) if ($debug);  #print "config = ",dump($config) if ($debug);
# Line 135  die "no databases in config file!\n" unl Line 154  die "no databases in config file!\n" unl
154    
155  $log->info( "-" x 79 );  $log->info( "-" x 79 );
156    
157    my $log_file = 'log';
158    
159    if (-e $log_file ) {    # && -s $log_file > 5 * 1024 * 1024) {
160            $log->info("moved old log with ", -s $log_file, " bytes to '${log_file}.old'");
161            rename $log_file, "${log_file}.old" || $log->logwarn("can't rename $log_file to ${log_file}.old: $!");
162    }
163    
164  my $estcmd_fh;  my $estcmd_fh;
165  my $estcmd_path = './estcmd-merge.sh';  my $estcmd_path = './estcmd-merge.sh';
# Line 145  if ($merge) { Line 170  if ($merge) {
170          $log->info("created merge batch file $estcmd_path");          $log->info("created merge batch file $estcmd_path");
171  }  }
172    
   
173  my $validate;  my $validate;
174  $validate = new WebPAC::Validate(  $validate = new WebPAC::Validate(
175          path => $validate_path,          path => $validate_path,
176  ) if ($validate_path);          delimiters => $config->webpac('delimiters'),
177            delimiters_path => $validate_delimiters_path,
178    ) if ($validate_path || $validate_delimiters_path);
179    
180  my $use_indexer = $config->use_indexer;  my $use_indexer = $config->use_indexer;
181  $stats ||= $validate;  $stats ||= $validate;
# Line 420  while (my ($database, $db_config) = each Line 445  while (my ($database, $db_config) = each
445                  }                  }
446    
447                  my $marc;                  my $marc;
448                  if ($parser->have_rules( 'marc', $database, $input_name )) {                  if ($marc_generate && $parser->have_rules( 'marc', $database, $input_name )) {
449                          $marc = new WebPAC::Output::MARC(                          $marc = new WebPAC::Output::MARC(
450                                  path => "out/marc/${database}-${input_name}.marc",                                  path => "out/marc/${database}-${input_name}.marc",
451                                  lint => $marc_lint,                                  lint => $marc_lint,
# Line 526  while (my ($database, $db_config) = each Line 551  while (my ($database, $db_config) = each
551                                          $marc->add(                                          $marc->add(
552                                                  id => $mfn . ( $i ? "/$i" : '' ),                                                  id => $mfn . ( $i ? "/$i" : '' ),
553                                                  fields => $fields,                                                  fields => $fields,
554                                                  leader => WebPAC::Normalize::marc_leader(),                                                  leader => WebPAC::Normalize::_get_marc_leader(),
555                                                  row => $row,                                                  row => $row,
556                                          );                                          );
557                                          $i++;                                          $i++;
# Line 542  while (my ($database, $db_config) = each Line 567  while (my ($database, $db_config) = each
567                                  $log->info("validation errors:\n$errors\n" );                                  $log->info("validation errors:\n$errors\n" );
568                                  print $report_fh "$errors\n" if ($report_fh);                                  print $report_fh "$errors\n" if ($report_fh);
569                          }                          }
570    
571                            print $report_fh "\nAll possible subfields/delimiter templates:\n", $validate->delimiters_templates( report => 1 ), "\n\n";
572                  }                  }
573    
574                  if ($stats) {                  if ($stats) {
# Line 554  while (my ($database, $db_config) = each Line 581  while (my ($database, $db_config) = each
581                  $marc->finish if ($marc);                  $marc->finish if ($marc);
582    
583                  # close report                  # close report
584                  close($report_fh) if ($report_fh)                  close($report_fh) if ($report_fh);
585    
586                    # reset validate stats
587                    $validate->reset if ( $validate );
588          }          }
589    
590          eval { $indexer->finish } if ($indexer && $indexer->can('finish'));          eval { $indexer->finish } if ($indexer && $indexer->can('finish'));
# Line 582  if ($parallel) { Line 611  if ($parallel) {
611          $log->info("all parallel processes finished");          $log->info("all parallel processes finished");
612  }  }
613    
614    # save new delimiters if needed
615    $validate->save_delimiters_templates if ( $validate );
616    
617  #  #
618  # handle links or merge after indexing  # handle links or merge after indexing
619  #  #
Legend:
Removed from v.774  
changed lines
  Added in v.850
  ViewVC Help
Powered by ViewVC 1.1.26