/[Search-Estraier]/trunk/lib/Search/Estraier.pm
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/lib/Search/Estraier.pm

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

revision 97 by dpavlin, Sat Jan 28 18:19:47 2006 UTC revision 100 by dpavlin, Sat Jan 28 19:41:59 2006 UTC
# Line 50  Search::Estraier - pure perl module to u Line 50  Search::Estraier - pure perl module to u
50          $cond->set_phrase("rainbow AND lullaby");          $cond->set_phrase("rainbow AND lullaby");
51    
52          my $nres = $node->search($cond, 0);          my $nres = $node->search($cond, 0);
53            print "Got ", $nres->hits, " results\n";
54    
55          if (defined($nres)) {          if (defined($nres)) {
56                  # for each document in results                  # for each document in results
57                  for my $i ( 0 ... $nres->doc_num - 1 ) {                  for my $i ( 0 ... $nres->doc_num - 1 ) {
# Line 370  sub delete { Line 372  sub delete {
372    
373  package Search::Estraier::Condition;  package Search::Estraier::Condition;
374    
375  use Carp qw/confess croak/;  use Carp qw/carp confess croak/;
376    
377  use Search::Estraier;  use Search::Estraier;
378  our @ISA = qw/Search::Estraier/;  our @ISA = qw/Search::Estraier/;
# Line 448  sub set_max { Line 450  sub set_max {
450    
451  =head2 set_options  =head2 set_options
452    
453    $cond->set_options( SURE => 1 );    $cond->set_options( 'SURE' );
454    
455      $cond->set_options( qw/AGITO NOIDF SIMPLE/ );
456    
457    Possible options are:
458    
459    =over 8
460    
461    =item SURE
462    
463    check every N-gram
464    
465    =item USUAL
466    
467    check every second N-gram
468    
469    =item FAST
470    
471    check every third N-gram
472    
473    =item AGITO
474    
475    check every fourth N-gram
476    
477    =item NOIDF
478    
479    don't perform TF-IDF tuning
480    
481    =item SIMPLE
482    
483    use simplified query phrase
484    
485    =back
486    
487    Skipping N-grams will speed up search, but reduce accuracy. Every call to C<set_options> will reset previous
488    options;
489    
490    This option changed in version C<0.04> of this module. It's backwards compatibile.
491    
492  =cut  =cut
493    
494  my $options = {  my $options = {
         # check N-gram keys skipping by three  
495          SURE => 1 << 0,          SURE => 1 << 0,
         # check N-gram keys skipping by two  
496          USUAL => 1 << 1,          USUAL => 1 << 1,
         # without TF-IDF tuning  
497          FAST => 1 << 2,          FAST => 1 << 2,
         # with the simplified phrase  
498          AGITO => 1 << 3,          AGITO => 1 << 3,
         # check every N-gram key  
499          NOIDF => 1 << 4,          NOIDF => 1 << 4,
         # check N-gram keys skipping by one  
500          SIMPLE => 1 << 10,          SIMPLE => 1 << 10,
501  };  };
502    
503  sub set_options {  sub set_options {
504          my $self = shift;          my $self = shift;
505          my $option = shift;          my $opt = 0;
506          confess "unknown option" unless ($options->{$option});          foreach my $option (@_) {
507          $self->{options} ||= $options->{$option};                  my $mask;
508                    unless ($mask = $options->{$option}) {
509                            if ($option eq '1') {
510                                    next;
511                            } else {
512                                    croak "unknown option $option";
513                            }
514                    }
515                    $opt += $mask;
516            }
517            $self->{options} = $opt;
518  }  }
519    
520    
# Line 693  Return number of documents Line 736  Return number of documents
736    
737    print $res->doc_num;    print $res->doc_num;
738    
739    This will return real number of documents (limited by C<max>).
740    If you want to get total number of hits, see C<hits>.
741    
742  =cut  =cut
743    
744  sub doc_num {  sub doc_num {
# Line 724  sub get_doc { Line 770  sub get_doc {
770    
771  Return specific hint from results.  Return specific hint from results.
772    
773    print $rec->hint( 'VERSION' );    print $res->hint( 'VERSION' );
774    
775  Possible hints are: C<VERSION>, C<NODE>, C<HIT>, C<HINT#n>, C<DOCNUM>, C<WORDNUM>,  Possible hints are: C<VERSION>, C<NODE>, C<HIT>, C<HINT#n>, C<DOCNUM>, C<WORDNUM>,
776  C<TIME>, C<LINK#n>, C<VIEW>.  C<TIME>, C<LINK#n>, C<VIEW>.
# Line 737  sub hint { Line 783  sub hint {
783          return $self->{hints}->{$key};          return $self->{hints}->{$key};
784  }  }
785    
786  =head2 hints  =head2 hits
787    
788  More perlish version of C<hint>. This one returns hash.  More perlish version of C<hint>. This one returns hash.
789    
790    my %hints = $rec->hints;    my %hints = $res->hints;
791    
792  =cut  =cut
793    
# Line 750  sub hints { Line 796  sub hints {
796          return $self->{hints};          return $self->{hints};
797  }  }
798    
799    =head2 hits
800    
801    Syntaxtic sugar for total number of hits for this query
802    
803      print $res->hits;
804    
805    It's same as
806    
807      print $res->hint('HIT');
808    
809    but shorter.
810    
811    =cut
812    
813    sub hits {
814            my $self = shift;
815            return $self->{hints}->{'HIT'} || 0;
816    }
817    
818  package Search::Estraier::Node;  package Search::Estraier::Node;
819    
820  use Carp qw/carp croak confess/;  use Carp qw/carp croak confess/;
Legend:
Removed from v.97  
changed lines
  Added in v.100
  ViewVC Help
Powered by ViewVC 1.1.26