[popfile-commit] engine/Classifier Bayes.pm,1.195,1.196

[<jgr...@us...>]

Update of /cvsroot/popfile/engine/Classifier
In directory sc8-pr-cvs1:/tmp/cvs-serv23438/Classifier

Modified Files:
	Bayes.pm 
Log Message:
Filter out special values from bucket data, start counting upgrade counts from 0, fix untie warnings

Index: Bayes.pm
===================================================================
RCS file: /cvsroot/popfile/engine/Classifier/Bayes.pm,v
retrieving revision 1.195
retrieving revision 1.196
diff -C2 -d -r1.195 -r1.196
*** Bayes.pm	10 Sep 2003 22:42:48 -0000	1.195
--- Bayes.pm	11 Sep 2003 08:06:03 -0000	1.196
***************
*** 294,300 ****
  
      for my $bucket (keys %{$self->{matrix__}})  {
          untie %{$self->{matrix__}{$bucket}};
          delete $self->{matrix__}{$bucket};
-         delete $self->{db__}{$bucket};
      }
  }
--- 294,301 ----
  
      for my $bucket (keys %{$self->{matrix__}})  {
+         undef $self->{db__}{$bucket};
+         delete $self->{db__}{$bucket};
          untie %{$self->{matrix__}{$bucket}};
          delete $self->{matrix__}{$bucket};
      }
  }
***************
*** 633,637 ****
  	    print "\nUpgrading bucket $bucket...";
              flush STDOUT;
!             my $wc = 1;
  
              my $first = <WORDS>;
--- 634,638 ----
  	    print "\nUpgrading bucket $bucket...";
              flush STDOUT;
!             my $wc = 0;
  
              my $first = <WORDS>;
***************
*** 1447,1451 ****
      my ( $self, $bucket, $prefix ) = @_;
  
!     return grep {/^$prefix/} keys %{$self->{matrix__}{$bucket}};
  }
  
--- 1448,1452 ----
      my ( $self, $bucket, $prefix ) = @_;
  
!     return grep {/^$prefix/} grep {!/__POPFILE__(UNIQUE|TOTAL)__/} keys %{$self->{matrix__}{$bucket}};
  }
  
***************
*** 1464,1468 ****
  
      my $prev = '';
!     return grep {$_ ne $prev && ($prev = $_, 1)} sort map {substr($_,0,1)} keys %{$self->{matrix__}{$bucket}};
  }
  
--- 1465,1469 ----
  
      my $prev = '';
!     return grep {$_ ne $prev && ($prev = $_, 1)} sort map {substr($_,0,1)} grep {!/__POPFILE__(UNIQUE|TOTAL)__/} keys %{$self->{matrix__}{$bucket}};
  }
  
***************
*** 1883,1886 ****
--- 1884,1888 ----
      my $bucket_directory = $self->config_( 'corpus' ) . "/$bucket";
  
+     undef $self->{db__}{$bucket};
      untie %{$self->{matrix__}{$bucket}};
      unlink( "$bucket_directory/table.db" );