--- trunk/Estraier.pm	2006/05/10 14:57:50	142
+++ trunk/Estraier.pm	2006/11/05 16:26:57	191
@@ -4,7 +4,7 @@
 use strict;
 use warnings;
 
-our $VERSION = '0.06';
+our $VERSION = '0.08';
 
 =head1 NAME
 
@@ -120,9 +120,34 @@
 
 =head1 Search::Estraier::Document
 
-This class implements Document which is collection of attributes
-(key=value), vectors (also key value) display text and hidden text.
+This class implements Document which is single item in Hyper Estraier.
 
+It's is collection of:
+
+=over 4
+
+=item attributes
+
+C<< 'key' => 'value' >> pairs which can later be used for filtering of results
+
+You can add common filters to C<attrindex> in estmaster's C<_conf>
+file for better performance. See C<attrindex> in
+L<Hyper Estraier P2P Guide|http://hyperestraier.sourceforge.net/nguide-en.html>.
+
+=item vectors
+
+also C<< 'key' => 'value' >> pairs
+
+=item display text
+
+Text which will be used to create searchable corpus of your index and
+included in snippet output.
+
+=item hidden text
+
+Text which will be searchable, but will not be included in snippet.
+
+=back
 
 =head2 new
 
@@ -157,11 +182,15 @@
 
 			if ($line =~ m/^%VECTOR\t(.+)$/) {
 				my @fields = split(/\t/, $1);
-				for my $i ( 0 .. ($#fields - 1) ) {
-					$self->{kwords}->{ $fields[ $i ] } = $fields[ $i + 1 ];
-					$i++;
+				if ($#fields % 2 == 1) {
+					$self->{kwords} = { @fields };
+				} else {
+					warn "can't decode $line\n";
 				}
 				next;
+			} elsif ($line =~ m/^%SCORE\t(.+)$/) {
+			    $self->{score} = $1;
+			    next;
 			} elsif ($line =~ m/^%/) {
 				# What is this? comment?
 				#warn "$line\n";
@@ -243,6 +272,53 @@
 	push @{ $self->{htexts} }, $self->_s($text);
 }
 
+=head2 add_vectors
+
+Add a vectors
+
+  $doc->add_vector(
+  	'vector_name' => 42,
+	'another' => 12345,
+  );
+
+=cut
+
+sub add_vectors {
+	my $self = shift;
+	return unless (@_);
+
+	# this is ugly, but works
+	die "add_vector needs HASH as argument" unless ($#_ % 2 == 1);
+
+	$self->{kwords} = {@_};
+}
+
+=head2 set_score
+
+Set the substitute score
+
+  $doc->set_score(12345);
+
+=cut
+
+sub set_score {
+    my $self = shift;
+    my $score = shift;
+    return unless (defined($score));
+    $self->{score} = $score;
+}
+
+=head2 score
+
+Get the substitute score
+
+=cut
+
+sub score {
+    my $self = shift;
+    return -1 unless (defined($self->{score}));
+    return $self->{score};
+}
 
 =head2 id
 
@@ -337,13 +413,17 @@
 	}
 
 	if ($self->{kwords}) {
-		$draft .= '%%VECTOR';
+		$draft .= '%VECTOR';
 		while (my ($key, $value) = each %{ $self->{kwords} }) {
 			$draft .= "\t$key\t$value";
 		}
 		$draft .= "\n";
 	}
 
+	if (defined($self->{score}) && $self->{score} >= 0) {
+	    $draft .= "%SCORE\t" . $self->{score} . "\n";
+	}
+
 	$draft .= "\n";
 
 	$draft .= join("\n", @{ $self->{dtexts} }) . "\n" if ($self->{dtexts});
@@ -631,6 +711,47 @@
 }
 
 
+=head2 set_distinct
+
+  $cond->set_distinct('@author');
+
+=cut
+
+sub set_distinct {
+	my $self = shift;
+	$self->{distinct} = shift;
+}
+
+=head2 distinct
+
+Return distinct attribute
+
+  print $cond->distinct;
+
+=cut
+
+sub distinct {
+	my $self = shift;
+	return $self->{distinct};
+}
+
+=head2 set_mask
+
+Filter out some links when searching.
+
+Argument array of link numbers, starting with 0 (current node).
+
+  $cond->set_mask(qw/0 1 4/);
+
+=cut
+
+sub set_mask {
+	my $self = shift;
+	return unless (@_);
+	$self->{mask} = \@_;
+}
+
+
 package Search::Estraier::ResultDocument;
 
 use Carp qw/croak/;
@@ -1053,7 +1174,7 @@
 
   $node->put_doc( $document_draft ) or die "can't add document";
 
-Return true on success or false on failture.
+Return true on success or false on failure.
 
 =cut
 
@@ -1061,11 +1182,15 @@
 	my $self = shift;
 	my $doc = shift || return;
 	return unless ($self->{url} && $doc->isa('Search::Estraier::Document'));
-	$self->shuttle_url( $self->{url} . '/put_doc',
+	if ($self->shuttle_url( $self->{url} . '/put_doc',
 		'text/x-estraier-draft',
 		$doc->dump_draft,
 		undef
-	) == 200;
+	) == 200) {
+		$self->_clear_info;
+		return 1;
+	}
+	return undef;
 }
 
 
@@ -1084,11 +1209,15 @@
 	my $id = shift || return;
 	return unless ($self->{url});
 	croak "id must be number, not '$id'" unless ($id =~ m/^\d+$/);
-	$self->shuttle_url( $self->{url} . '/out_doc',
+	if ($self->shuttle_url( $self->{url} . '/out_doc',
 		'application/x-www-form-urlencoded',
 		"id=$id",
 		undef
-	) == 200;
+	) == 200) {
+		$self->_clear_info;
+		return 1;
+	}
+	return undef;
 }
 
 
@@ -1106,11 +1235,15 @@
 	my $self = shift;
 	my $uri = shift || return;
 	return unless ($self->{url});
-	$self->shuttle_url( $self->{url} . '/out_doc',
+	if ($self->shuttle_url( $self->{url} . '/out_doc',
 		'application/x-www-form-urlencoded',
 		"uri=" . uri_escape($uri),
 		undef
-	) == 200;
+	) == 200) {
+		$self->_clear_info;
+		return 1;
+	}
+	return undef;
 }
 
 
@@ -1128,11 +1261,15 @@
 	my $self = shift;
 	my $doc = shift || return;
 	return unless ($self->{url} && $doc->isa('Search::Estraier::Document'));
-	$self->shuttle_url( $self->{url} . '/edit_doc',
+	if ($self->shuttle_url( $self->{url} . '/edit_doc',
 		'text/x-estraier-draft',
 		$doc->dump_draft,
 		undef
-	) == 200;
+	) == 200) {
+		$self->_clear_info;
+		return 1;
+	}
+	return undef;
 }
 
 
@@ -1290,7 +1427,7 @@
 	$path = '/etch_doc' if ($a->{etch});
 
 	if ($a->{id}) {
-		croak "id must be numberm not '$a->{id}'" unless ($a->{id} =~ m/^\d+$/);
+		croak "id must be number not '$a->{id}'" unless ($a->{id} =~ m/^\d+$/);
 		$arg = 'id=' . $a->{id};
 	} elsif ($a->{uri}) {
 		$arg = 'uri=' . uri_escape($a->{uri});
@@ -1499,7 +1636,18 @@
 	push @args, 'wwidth=' . $self->{wwidth};
 	push @args, 'hwidth=' . $self->{hwidth};
 	push @args, 'awidth=' . $self->{awidth};
-	push @args, 'skip=' . $self->{skip} if ($self->{skip});
+	push @args, 'skip=' . $cond->{skip} if ($cond->{skip});
+
+	if (my $distinct = $cond->distinct) {
+		push @args, 'distinct=' . uri_escape($distinct);
+	}
+
+	if ($cond->{mask}) {
+		my $mask = 0;
+		map { $mask += ( 2 ** $_ ) } @{ $cond->{mask} };
+
+		push @args, 'mask=' . $mask if ($mask);
+	}
 
 	return join('&', @args);
 }
@@ -1646,7 +1794,7 @@
 	croak "mode must be number, not '$mode'" unless ($mode =~ m/^\d+$/);
 
 	$self->shuttle_url( $self->{url} . '/_set_user',
-		'text/plain',
+		'application/x-www-form-urlencoded',
 		'name=' . uri_escape($name) . '&mode=' . $mode,
 		undef
 	) == 200;
@@ -1679,9 +1827,10 @@
 		undef
 	) == 200) {
 		# refresh node info after adding link
-		$self->_set_info;
+		$self->_clear_info;
 		return 1;
 	}
+	return undef;
 }
 
 =head2 admins
@@ -1726,6 +1875,31 @@
 	return $self->{inform}->{links};
 }
 
+=head2 cacheusage
+
+Return cache usage for a node
+
+  my $cache = $node->cacheusage;
+
+=cut
+
+sub cacheusage {
+	my $self = shift;
+
+	return unless ($self->{url});
+
+	my $resbody;
+	my $rv = $self->shuttle_url( $self->{url} . '/cacheusage',
+		'text/plain',
+		undef,
+		\$resbody,
+	);
+
+	return if ($rv != 200 || !$resbody);
+
+	return $resbody;
+}
+
 =head2 master
 
 Set actions on Hyper Estraier node master (C<estmaster> process)
@@ -1828,6 +2002,10 @@
 	) or confess "shuttle_url failed";
 
 	if ($status == $rest->{status}) {
+
+		# refresh node info after sync
+		$self->_clear_info if ($action eq 'sync' || $action =~ m/^node(?:add|del|clr)$/);
+
 		if ($rest->{returns} && wantarray) {
 
 			my @results;
@@ -1885,7 +2063,7 @@
 
 	my @lines = split(/[\r\n]/,$resbody);
 
-	$self->{inform} = {};
+	$self->_clear_info;
 
 	( $self->{inform}->{name}, $self->{inform}->{label}, $self->{inform}->{dnum},
 		$self->{inform}->{wnum}, $self->{inform}->{size} ) = split(/\t/, shift @lines, 5);
@@ -1910,6 +2088,25 @@
 
 }
 
+=head2 _clear_info
+
+Clear information for node
+
+  $node->_clear_info;
+
+On next call to C<name>, C<label>, C<doc_num>, C<word_num> or C<size> node
+info will be fetch again from Hyper Estraier.
+
+=cut
+sub _clear_info {
+	my $self = shift;
+	$self->{inform} = {
+		dnum => -1,
+		wnum => -1,
+		size => -1.0,
+	};
+}
+
 ###
 
 =head1 EXPORT
@@ -1922,6 +2119,10 @@
 
 Hyper Estraier Ruby interface on which this module is based.
 
+Hyper Estraier now also has pure-perl binding included in distribution. It's
+a faster way to access databases directly if you are not running
+C<estmaster> P2P server.
+
 =head1 AUTHOR
 
 Dobrica Pavlinusic, E<lt>dpavlin@rot13.orgE<gt>