popfile-commit

[<ssc...@us...>]

Update of /cvsroot/popfile/engine
In directory sc8-pr-cvs1:/tmp/cvs-serv12282

Modified Files:
	traintest.pl 
Log Message:
update for message queue changes to allow traintest.pl to run

Index: traintest.pl
===================================================================
RCS file: /cvsroot/popfile/engine/traintest.pl,v
retrieving revision 1.5
retrieving revision 1.6
diff -C2 -d -r1.5 -r1.6
*** traintest.pl	29 May 2003 05:49:42 -0000	1.5
--- traintest.pl	8 Jul 2003 01:54:37 -0000	1.6
***************
*** 11,14 ****
--- 11,15 ----
  use Classifier::Bayes;
  use POPFile::Configuration;
+ use POPFile::MQ;
  
  my $debug = 0;
***************
*** 64,68 ****
  {
      my ($config) = @_;
!     
      $config->parameter("csv",$DEFAULT_CSV);
      $config->parameter("toe",$DEFAULT_TOE);
--- 65,69 ----
  {
      my ($config) = @_;
! 
      $config->parameter("csv",$DEFAULT_CSV);
      $config->parameter("toe",$DEFAULT_TOE);
***************
*** 79,84 ****
  sub cvs_out
  {
!     my ($config, $file, @log) = @_;    
!  
      if ($file eq 'auto') {
          if ($config->parameter('toe') == 1) {
--- 80,85 ----
  sub cvs_out
  {
!     my ($config, $file, @log) = @_;
! 
      if ($file eq 'auto') {
          if ($config->parameter('toe') == 1) {
***************
*** 90,113 ****
               ( $config->parameter('window2') ne $DEFAULT_WINDOW2 ) ) {
              $file .= "$config->parameter('window1')and$config->parameter('window2')";
!                 
          }
!         
          if ( $config->parameter('stopwords') ne $DEFAULT_STOP ) {
              $file .= "_";
!             if ($config->parameter('stopwords') != 1) {                
                  $file .= "no";
              }
!             $file .= "stop";                
          }
!         
!         
          $file .= ".csv";
      }
!     
      print STDERR "Printing data to $file\n";
!     
      open CSV, ">$file";
!     
!     
      $, = ",";
  
--- 91,114 ----
               ( $config->parameter('window2') ne $DEFAULT_WINDOW2 ) ) {
              $file .= "$config->parameter('window1')and$config->parameter('window2')";
! 
          }
! 
          if ( $config->parameter('stopwords') ne $DEFAULT_STOP ) {
              $file .= "_";
!             if ($config->parameter('stopwords') != 1) {
                  $file .= "no";
              }
!             $file .= "stop";
          }
! 
! 
          $file .= ".csv";
      }
! 
      print STDERR "Printing data to $file\n";
! 
      open CSV, ">$file";
! 
! 
      $, = ",";
  
***************
*** 122,130 ****
          print CSV "\n";
      }
!     
      print STDERR "Data saved to $file\n";
!     
      $, = "";
!     
  }
  
--- 123,131 ----
          print CSV "\n";
      }
! 
      print STDERR "Data saved to $file\n";
! 
      $, = "";
! 
  }
  
***************
*** 151,155 ****
  {
      my ($toe,$predicted,$actual) = @_;
!     
      if ($toe == 1) {
          return ($predicted ne $actual);
--- 152,156 ----
  {
      my ($toe,$predicted,$actual) = @_;
! 
      if ($toe == 1) {
          return ($predicted ne $actual);
***************
*** 167,175 ****
      my $dir = $self->{configuration__}->parameter('corpus_out');
      mkdir($dir);
!     
      foreach my $abucket ( keys %{$self->{total__}} ) {
!         
          print "saving $abucket corpus.\n";
!         
          my $subdir = $dir;
          $subdir .= "/$abucket";
--- 168,176 ----
      my $dir = $self->{configuration__}->parameter('corpus_out');
      mkdir($dir);
! 
      foreach my $abucket ( keys %{$self->{total__}} ) {
! 
          print "saving $abucket corpus.\n";
! 
          my $subdir = $dir;
          $subdir .= "/$abucket";
***************
*** 185,190 ****
                  }
              }
!         }        
!     }    
  }
  
--- 186,191 ----
                  }
              }
!         }
!     }
  }
  
***************
*** 272,278 ****
--- 273,285 ----
      my $b = new Classifier::Bayes;
      my $c = new POPFile::Configuration;
+     my $mq = new POPFile::MQ;
  
      $b->configuration( $c );
      $c->configuration( $c );
+     $mq->configuration( $c);
+ 
+     $b->mq( $mq );
+     $c->mq( $mq );
+     $mq->mq( $mq );
  
      $c->initialize();
***************
*** 281,296 ****
  
      $c->load_configuration();
!     
      $c->parse_command_line();
!     
  #    $b->{unclassified} = ($c->parameter('unclassified_probability') || 0.0001);
      $b->{unclassified__} = log($c->parameter("bayes_unclassified_probability") || 0.5);
!     
!     # test with or without stop-words    
       if ( $c->parameter("stopwords") eq 0 ) {
          $b->{parser__}->{mangle__}->{stop__} = {};
          $b->{mangler__}->{stop__} = {};
      }
!     
  
      my $archive = $c->parameter("html_archive_dir");
--- 288,303 ----
  
      $c->load_configuration();
! 
      $c->parse_command_line();
! 
  #    $b->{unclassified} = ($c->parameter('unclassified_probability') || 0.0001);
      $b->{unclassified__} = log($c->parameter("bayes_unclassified_probability") || 0.5);
! 
!     # test with or without stop-words
       if ( $c->parameter("stopwords") eq 0 ) {
          $b->{parser__}->{mangle__}->{stop__} = {};
          $b->{mangler__}->{stop__} = {};
      }
! 
  
      my $archive = $c->parameter("html_archive_dir");
***************
*** 355,362 ****
          my $message_count = $index + 1;
          $index = $#sorted_messages - $index;
!         print "\n$self->{messages}->{messages}{ @sorted_messages[$index] }{long}:" if ($debug);        
          $bucket_class = $b->classify_file( $self->{messages}->{messages}{ @sorted_messages[$index] }{long});
          $bucket_true = $self->{messages}->{messages}{ @sorted_messages[$index] }{bucket};
!         
          $b->{parser__}->{subject__} =~ s/(\r\n|\n|\r)[ \t]+//gm;
  
--- 362,369 ----
          my $message_count = $index + 1;
          $index = $#sorted_messages - $index;
!         print "\n$self->{messages}->{messages}{ @sorted_messages[$index] }{long}:" if ($debug);
          $bucket_class = $b->classify_file( $self->{messages}->{messages}{ @sorted_messages[$index] }{long});
          $bucket_true = $self->{messages}->{messages}{ @sorted_messages[$index] }{bucket};
! 
          $b->{parser__}->{subject__} =~ s/(\r\n|\n|\r)[ \t]+//gm;
  
***************
*** 425,434 ****
  
      my $end_time = time;
!     
      if ($c->parameter("dump")) {
          dump_corpus($b);
!         
      }
!     
  
      my $total_messages = $#sorted_messages + 1;
--- 432,441 ----
  
      my $end_time = time;
! 
      if ($c->parameter("dump")) {
          dump_corpus($b);
! 
      }
! 
  
      my $total_messages = $#sorted_messages + 1;