/[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!

Diff of /trunk/lib/Search/Estraier.pm

Parent Directory | Revision Log | View Patch Patch

-revision 93 by dpavlin,
Sat Jan 28 16:43:45 2006 UTC
+revision 108 by dpavlin,
Sun Feb 19 17:13:57 2006 UTC
 Line 4 
 use 5.008;
  use strict;
  use warnings;
- our $VERSION = '0.04_1';
+ our $VERSION = '0.04_2';
  =head1 NAME
 Line 17 
 Search::Estraier - pure perl module to u
          use Search::Estraier;
          # create and configure node
-         my $node = new Search::Estraier::Node;
+         my $node = new Search::Estraier::Node(
-         $node->set_url("http://localhost:1978/node/test");
+                 url => 'http://localhost:1978/node/test',
-         $node->set_auth("admin","admin");
+                 user => 'admin',
+                 passwd => 'admin'
+         );
          # create document
          my $doc = new Search::Estraier::Document;
-Line 32 
 Search::Estraier - pure perl module to u
+Line 34 
 Search::Estraier - pure perl module to u
          $doc->add_text("Somewhere over the rainbow.  Way up high.");
          $doc->add_text("There's a land that I heard of once in a lullaby.");
-         die "error: ", $node->status,"\n" unless ($node->put_doc($doc));
+         die "error: ", $node->status,"\n" unless (eval { $node->put_doc($doc) });
  =head2 Simple searcher
          use Search::Estraier;
          # create and configure node
-         my $node = new Search::Estraier::Node;
+         my $node = new Search::Estraier::Node(
-         $node->set_url("http://localhost:1978/node/test");
+                 url => 'http://localhost:1978/node/test',
-         $node->set_auth("admin","admin");
+                 user => 'admin',
+                 passwd => 'admin',
+                 croak_on_error => 1,
+         );
          # create condition
          my $cond = new Search::Estraier::Condition;
-Line 50 
 Search::Estraier - pure perl module to u
+Line 55 
 Search::Estraier - pure perl module to u
          $cond->set_phrase("rainbow AND lullaby");
          my $nres = $node->search($cond, 0);
          if (defined($nres)) {
+                 print "Got ", $nres->hits, " results\n";
                  # for each document in results
                  for my $i ( 0 ... $nres->doc_num - 1 ) {
                          # get result document
-Line 92 
 Remove multiple whitespaces from string,
+Line 100 
 Remove multiple whitespaces from string,
  =cut
  sub _s {
-         my $text = $_[1] || return;
+         my $text = $_[1];
+         return unless defined($text);
          $text =~ s/\s\s+/ /gs;
          $text =~ s/^\s+//;
          $text =~ s/\s+$//;
-Line 320 
 sub dump_draft {
+Line 329 
 sub dump_draft {
          my $draft;
          foreach my $attr_name (sort keys %{ $self->{attrs} }) {
-                 next unless(my $v = $self->{attrs}->{$attr_name});
+                 next unless defined(my $v = $self->{attrs}->{$attr_name});
                  $draft .= $attr_name . '=' . $v . "\n";
          }
-Line 369 
 sub delete {
+Line 378 
 sub delete {
  package Search::Estraier::Condition;
- use Carp qw/confess croak/;
+ use Carp qw/carp confess croak/;
  use Search::Estraier;
  our @ISA = qw/Search::Estraier/;
-Line 447 
 sub set_max {
+Line 456 
 sub set_max {
  =head2 set_options
-   $cond->set_options( SURE => 1 );
+   $cond->set_options( 'SURE' );
+   $cond->set_options( qw/AGITO NOIDF SIMPLE/ );
+ Possible options are:
+ =over 8
+ =item SURE
+ check every N-gram
+ =item USUAL
+ check every second N-gram
+ =item FAST
+ check every third N-gram
+ =item AGITO
+ check every fourth N-gram
+ =item NOIDF
+ don't perform TF-IDF tuning
+ =item SIMPLE
+ use simplified query phrase
+ =back
+ Skipping N-grams will speed up search, but reduce accuracy. Every call to C<set_options> will reset previous
+ options;
+ This option changed in version C<0.04> of this module. It's backwards compatibile.
  =cut
  my $options = {
-         # check N-gram keys skipping by three
          SURE => 1 << 0,
-         # check N-gram keys skipping by two
          USUAL => 1 << 1,
-         # without TF-IDF tuning
          FAST => 1 << 2,
-         # with the simplified phrase
          AGITO => 1 << 3,
-         # check every N-gram key
          NOIDF => 1 << 4,
-         # check N-gram keys skipping by one
          SIMPLE => 1 << 10,
  };
  sub set_options {
          my $self = shift;
-         my $option = shift;
+         my $opt = 0;
-         confess "unknown option" unless ($options->{$option});
+         foreach my $option (@_) {
-         $self->{options} ||= $options->{$option};
+                 my $mask;
+                 unless ($mask = $options->{$option}) {
+                         if ($option eq '1') {
+                                 next;
+                         } else {
+                                 croak "unknown option $option";
+                         }
+                 }
+                 $opt += $mask;
+         }
+         $self->{options} = $opt;
  }
-Line 692 
 Return number of documents
+Line 742 
 Return number of documents
    print $res->doc_num;
+ This will return real number of documents (limited by C<max>).
+ If you want to get total number of hits, see C<hits>.
  =cut
  sub doc_num {
-Line 723 
 sub get_doc {
+Line 776 
 sub get_doc {
  Return specific hint from results.
-   print $rec->hint( 'VERSION' );
+   print $res->hint( 'VERSION' );
  Possible hints are: C<VERSION>, C<NODE>, C<HIT>, C<HINT#n>, C<DOCNUM>, C<WORDNUM>,
  C<TIME>, C<LINK#n>, C<VIEW>.
-Line 740 
 sub hint {
+Line 793 
 sub hint {
  More perlish version of C<hint>. This one returns hash.
-   my %hints = $rec->hints;
+   my %hints = $res->hints;
  =cut
-Line 749 
 sub hints {
+Line 802 
 sub hints {
          return $self->{hints};
  }
+ =head2 hits
+ Syntaxtic sugar for total number of hits for this query
+   print $res->hits;
+ It's same as
+   print $res->hint('HIT');
+ but shorter.
+ =cut
+ sub hits {
+         my $self = shift;
+         return $self->{hints}->{'HIT'} || 0;
+ }
  package Search::Estraier::Node;
  use Carp qw/carp croak confess/;
-Line 1100 
 Get ID of document specified by URI
+Line 1172 
 Get ID of document specified by URI
    my $id = $node->uri_to_id( 'file:///document/uri/42' );
+ This method won't croak, even if using C<croak_on_error>.
  =cut
  sub uri_to_id {
          my $self = shift;
          my $uri = shift || return;
-         return $self->_fetch_doc( uri => $uri, path => '/uri_to_id', chomp_resbody => 1 );
+         return $self->_fetch_doc( uri => $uri, path => '/uri_to_id', chomp_resbody => 1, croak_on_error => 0 );
  }
-Line 1165 
 sub _fetch_doc {
+Line 1239 
 sub _fetch_doc {
                  'application/x-www-form-urlencoded',
                  $arg,
                  \$resbody,
+                 $a->{croak_on_error},
          );
          return if ($rv != 200);
-Line 1433 
 use LWP::UserAgent;
+Line 1508 
 use LWP::UserAgent;
  sub shuttle_url {
          my $self = shift;
-         my ($url, $content_type, $reqbody, $resbody) = @_;
+         my ($url, $content_type, $reqbody, $resbody, $croak_on_error) = @_;
+         $croak_on_error = $self->{croak_on_error} unless defined($croak_on_error);
          $self->{status} = -1;
-Line 1477 
 sub shuttle_url {
+Line 1554 
 sub shuttle_url {
          ($self->{status}, $self->{status_message}) = split(/\s+/, $res->status_line, 2);
          if (! $res->is_success) {
-                 if ($self->{croak_on_error}) {
+                 if ($croak_on_error) {
                          croak("can't get $url: ",$res->status_line);
                  } else {
                          return -1;
-Line 1582 
 sub set_link {
+Line 1659 
 sub set_link {
          my $reqbody = 'url=' . uri_escape($url) . '&label=' . uri_escape($label);
          $reqbody .= '&credit=' . $credit if ($credit > 0);
-         $self->shuttle_url( $self->{url} . '/_set_link',
+         if ($self->shuttle_url( $self->{url} . '/_set_link',
                  'application/x-www-form-urlencoded',
                  $reqbody,
                  undef
-         ) == 200;
+         ) == 200) {
+                 # refresh node info after adding link
+                 $self->_set_info;
+                 return 1;
+         }
+ }
+ =head2 admins
+  my @admins = @{ $node->admins };
+ Return array of users with admin rights on node
+ =cut
+ sub admins {
+         my $self = shift;
+         $self->_set_info unless ($self->{name});
+         return $self->{admins};
+ }
+ =head2 guests
+  my @guests = @{ $node->guests };
+ Return array of users with guest rights on node
+ =cut
+ sub guests {
+         my $self = shift;
+         $self->_set_info unless ($self->{name});
+         return $self->{guests};
+ }
+ =head2 links
+  my $links = @{ $node->links };
+ Return array of links for this node
+ =cut
+ sub links {
+         my $self = shift;
+         $self->_set_info unless ($self->{name});
+         return $self->{links};
  }
-Line 1617 
 sub _set_info {
+Line 1740 
 sub _set_info {
          return if ($rv != 200 || !$resbody);
-         # it seems that response can have multiple line endings
+         my @lines = split(/[\r\n]/,$resbody);
-         $resbody =~ s/[\r\n]+$//;
          ( $self->{name}, $self->{label}, $self->{dnum}, $self->{wnum}, $self->{size} ) =
-                 split(/\t/, $resbody, 5);
+                 split(/\t/, shift @lines, 5);
+         return $resbody unless (@lines);
+         shift @lines;
+         while(my $admin = shift @lines) {
+                 push @{$self->{admins}}, $admin;
+         }
+         while(my $guest = shift @lines) {
+                 push @{$self->{guests}}, $guest;
+         }
+         while(my $link = shift @lines) {
+                 push @{$self->{links}}, $link;
+         }
+         return $resbody;
  }

 Legend:



Removed from v.93
 


changed lines


 
Added in v.108
 Legend:



Removed from v.93
 


changed lines


 
Added in v.108
-Removed from v.93
+Added in v.108

	ViewVC Help
Powered by ViewVC 1.1.26