/[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 47 by dpavlin,
Fri Jan  6 01:51:28 2006 UTC
+revision 58 by dpavlin,
Fri Jan  6 21:05:05 2006 UTC
 Line 645 
 Return number of documents
  sub doc_num {
          my $self = shift;
-         return $#{$self->{docs}};
+         return $#{$self->{docs}} + 1;
  }
 Line 692 
 use Carp qw/carp croak confess/;
  use URI;
  use MIME::Base64;
  use IO::Socket::INET;
+ use URI::Escape qw/uri_escape/;
  =head1 Search::Estraier::Node
-Line 716 
 sub new {
+Line 717 
 sub new {
          };
          bless($self, $class);
-         if (@_) {
+         my $args = {@_};
-                 $self->{debug} = shift;
-                 warn "## Node debug on\n";
+         $self->{debug} = $args->{debug};
-         }
+         warn "## Node debug on\n" if ($self->{debug});
          $self ? return $self : return undef;
  }
-Line 866 
 sub out_doc_by_uri {
+Line 867 
 sub out_doc_by_uri {
          return unless ($self->{url});
          $self->shuttle_url( $self->{url} . '/out_doc',
                  'application/x-www-form-urlencoded',
-                 "uri=$uri",
+                 "uri=" . uri_escape($uri),
                  undef
          ) == 200;
  }
-Line 928 
 sub get_doc_by_uri {
+Line 929 
 sub get_doc_by_uri {
  }
+ =head2 get_doc_attr
+ Retrieve the value of an atribute from object
+   my $val = $node->get_doc_attr( document_id, 'attribute_name' ) or
+         die "can't get document attribute";
+ =cut
+ sub get_doc_attr {
+         my $self = shift;
+         my ($id,$name) = @_;
+         return unless ($id && $name);
+         return $self->_fetch_doc( id => $id, attr => $name );
+ }
+ =head2 get_doc_attr_by_uri
+ Retrieve the value of an atribute from object
+   my $val = $node->get_doc_attr_by_uri( document_id, 'attribute_name' ) or
+         die "can't get document attribute";
+ =cut
+ sub get_doc_attr_by_uri {
+         my $self = shift;
+         my ($uri,$name) = @_;
+         return unless ($uri && $name);
+         return $self->_fetch_doc( uri => $uri, attr => $name );
+ }
  =head2 etch_doc
  Exctract document keywords
-Line 936 
 Exctract document keywords
+Line 971 
 Exctract document keywords
  =cut
- sub erch_doc {
+ sub etch_doc {
          my $self = shift;
          my $id = shift || return;
          return $self->_fetch_doc( id => $id, etch => 1 );
-Line 987 
 C<etch_doc>, C<etch_doc_by_uri>.
+Line 1022 
 C<etch_doc>, C<etch_doc_by_uri>.
   my $doc = $node->_fetch_doc( id => 42, etch => 1 );
   my $doc = $node->_fetch_doc( uri => 'file:///document/uri/42', etch => 1 );
+  # to get document attrubute add attr
+  my $doc = $node->_fetch_doc( id => 42, attr => '@mdate' );
+  my $doc = $node->_fetch_doc( uri => 'file:///document/uri/42', attr => '@mdate' );
   # more general form which allows implementation of
   # uri_to_id
   my $id = $node->_fetch_doc(
-Line 1011 
 sub _fetch_doc {
+Line 1050 
 sub _fetch_doc {
                  croak "id must be numberm not '$a->{id}'" unless ($a->{id} =~ m/^\d+$/);
                  $arg = 'id=' . $a->{id};
          } elsif ($a->{uri}) {
-                 $arg = 'uri=' . $a->{uri};
+                 $arg = 'uri=' . uri_escape($a->{uri});
          } else {
                  confess "unhandled argument. Need id or uri.";
          }
+         if ($a->{attr}) {
+                 $path = '/get_doc_attr';
+                 $arg .= '&attr=' . uri_escape($a->{attr});
+                 $a->{chomp_resbody} = 1;
+         }
          my $rv = $self->shuttle_url( $self->{url} . $path,
                  'application/x-www-form-urlencoded',
                  $arg,
-Line 1042 
 sub _fetch_doc {
+Line 1087 
 sub _fetch_doc {
  }
+ =head2 name
+   my $node_name = $node->name;
+ =cut
+ sub name {
+         my $self = shift;
+         $self->_set_info unless ($self->{name});
+         return $self->{name};
+ }
+ =head2 label
+   my $node_label = $node->label;
+ =cut
+ sub label {
+         my $self = shift;
+         $self->_set_info unless ($self->{label});
+         return $self->{label};
+ }
+ =head2 doc_num
+   my $documents_in_node = $node->doc_num;
+ =cut
+ sub doc_num {
+         my $self = shift;
+         $self->_set_info if ($self->{dnum} < 0);
+         return $self->{dnum};
+ }
+ =head2 word_num
+   my $words_in_node = $node->word_num;
+ =cut
+ sub word_num {
+         my $self = shift;
+         $self->_set_info if ($self->{wnum} < 0);
+         return $self->{wnum};
+ }
+ =head2 size
+   my $node_size = $node->size;
+ =cut
+ sub size {
+         my $self = shift;
+         $self->_set_info if ($self->{size} < 0);
+         return $self->{size};
+ }
+ =head2 search
+ Search documents which match condition
+   my $nres = $node->search( $cond, $depth );
+ C<$cond> is C<Search::Estraier::Condition> object, while <$depth> specifies
+ depth for meta search.
+ Function results C<Search::Estraier::NodeResult> object.
+ =cut
+ sub search {
+         my $self = shift;
+         my ($cond, $depth) = @_;
+         return unless ($cond && defined($depth) && $self->{url});
+         croak "cond mush be Search::Estraier::Condition, not '$cond->isa'" unless ($cond->isa('Search::Estraier::Condition'));
+         croak "depth needs number, not '$depth'" unless ($depth =~ m/^\d+$/);
+         my $resbody;
+         my $rv = $self->shuttle_url( $self->{url} . '/search',
+                 'application/x-www-form-urlencoded',
+                 $self->cond_to_query( $cond ),
+                 \$resbody,
+         );
+         return if ($rv != 200);
+         my (@docs, $hints);
+         my @lines = split(/\n/, $resbody);
+         return unless (@lines);
+         my $border = $lines[0];
+         my $isend = 0;
+         my $lnum = 1;
+         while ( $lnum <= $#lines ) {
+                 my $line = $lines[$lnum];
+                 $lnum++;
+                 #warn "## $line\n";
+                 if ($line && $line =~ m/^\Q$border\E(:END)*$/) {
+                         $isend = $1;
+                         last;
+                 }
+                 if ($line =~ /\t/) {
+                         my ($k,$v) = split(/\t/, $line, 2);
+                         $hints->{$k} = $v;
+                 }
+         }
+         my $snum = $lnum;
+         while( ! $isend && $lnum <= $#lines ) {
+                 my $line = $lines[$lnum];
+                 #warn "# $lnum: $line\n";
+                 $lnum++;
+                 if ($line && $line =~ m/^\Q$border\E/) {
+                         if ($lnum > $snum) {
+                                 my $rdattrs;
+                                 my $rdvector;
+                                 my $rdsnippet;
+                                 my $rlnum = $snum;
+                                 while ($rlnum < $lnum - 1 ) {
+                                         #my $rdline = $self->_s($lines[$rlnum]);
+                                         my $rdline = $lines[$rlnum];
+                                         $rlnum++;
+                                         last unless ($rdline);
+                                         if ($rdline =~ /^%/) {
+                                                 $rdvector = $1 if ($rdline =~ /^%VECTOR\t(.+)$/);
+                                         } elsif($rdline =~ /=/) {
+                                                 $rdattrs->{$1} = $2 if ($rdline =~ /^(.+)=(.+)$/);
+                                         } else {
+                                                 confess "invalid format of response";
+                                         }
+                                 }
+                                 while($rlnum < $lnum - 1) {
+                                         my $rdline = $lines[$rlnum];
+                                         $rlnum++;
+                                         $rdsnippet .= "$rdline\n";
+                                 }
+                                 #warn Dumper($rdvector, $rdattrs, $rdsnippet);
+                                 if (my $rduri = $rdattrs->{'@uri'}) {
+                                         push @docs, new Search::Estraier::ResultDocument(
+                                                 uri => $rduri,
+                                                 attrs => $rdattrs,
+                                                 snippet => $rdsnippet,
+                                                 keywords => $rdvector,
+                                         );
+                                 }
+                         }
+                         $snum = $lnum;
+                         #warn "### $line\n";
+                         $isend = 1 if ($line =~ /:END$/);
+                 }
+         }
+         if (! $isend) {
+                 warn "received result doesn't have :END\n$resbody";
+                 return;
+         }
+         #warn Dumper(\@docs, $hints);
+         return new Search::Estraier::NodeResult( docs => \@docs, hints => $hints );
+ }
+ =head2 cond_to_query
+ Return URI encoded string generated from Search::Estraier::Condition
+   my $args = $node->cond_to_query( $cond );
+ =cut
+ sub cond_to_query {
+         my $self = shift;
+         my $cond = shift || return;
+         croak "condition must be Search::Estraier::Condition, not '$cond->isa'" unless ($cond->isa('Search::Estraier::Condition'));
+         my @args;
+         if (my $phrase = $cond->phrase) {
+                 push @args, 'phrase=' . uri_escape($phrase);
+         }
+         if (my @attrs = $cond->attrs) {
+                 for my $i ( 0 .. $#attrs ) {
+                         push @args,'attr' . ($i+1) . '=' . uri_escape( $attrs[$i] );
+                 }
+         }
+         if (my $order = $cond->order) {
+                 push @args, 'order=' . uri_escape($order);
+         }
+         if (my $max = $cond->max) {
+                 push @args, 'max=' . $max;
+         } else {
+                 push @args, 'max=' . (1 << 30);
+         }
+         if (my $options = $cond->options) {
+                 push @args, 'options=' . $options;
+         }
+         push @args, 'depth=' . $self->{depth} if ($self->{depth});
+         push @args, 'wwidth=' . $self->{wwidth};
+         push @args, 'hwidth=' . $self->{hwidth};
+         push @args, 'awidth=' . $self->{awidth};
+         return join('&', @args);
+ }
  =head2 shuttle_url
-Line 1049 
 sub _fetch_doc {
+Line 1320 
 sub _fetch_doc {
  This is method which uses C<IO::Socket::INET> to communicate with Hyper Estraier node
  master.
-   my $rv = shuttle_url( $url, $content_type, \$req_body, \$resbody );
+   my $rv = shuttle_url( $url, $content_type, $req_body, \$resbody );
  C<$resheads> and C<$resbody> booleans controll if response headers and/or response
  body will be saved within object.
-Line 1155 
 sub shuttle_url {
+Line 1426 
 sub shuttle_url {
          return $self->{status};
  }
+ =head2 set_snippet_width
+ Set width of snippets in results
+   $node->set_snippet_width( $wwidth, $hwidth, $awidth );
+ C<$wwidth> specifies whole width of snippet. It's C<480> by default. If it's C<0> snippet
+ is not sent with results. If it is negative, whole document text is sent instead of snippet.
+ C<$hwidth> specified width of strings from beginning of string. Default
+ value is C<96>. Negative or zero value keep previous value.
+ C<$awidth> specifies width of strings around each highlighted word. It's C<96> by default.
+ If negative of zero value is provided previous value is kept unchanged.
+ =cut
+ sub set_snippet_width {
+         my $self = shift;
+         my ($wwidth, $hwidth, $awidth) = @_;
+         $self->{wwidth} = $wwidth;
+         $self->{hwidth} = $hwidth if ($hwidth >= 0);
+         $self->{awidth} = $awidth if ($awidth >= 0);
+ }
+ =head2 set_user
+ Manage users of node
+   $node->set_user( 'name', $mode );
+ C<$mode> can be one of:
+ =over 4
+ =item 0
+ delete account
+ =item 1
+ set administrative right for user
+ =item 2
+ set user account as guest
+ =back
+ Return true on success, otherwise false.
+ =cut
+ sub set_user {
+         my $self = shift;
+         my ($name, $mode) = @_;
+         return unless ($self->{url});
+         croak "mode must be number, not '$mode'" unless ($mode =~ m/^\d+$/);
+         $self->shuttle_url( $self->{url} . '/_set_user',
+                 'text/plain',
+                 'name=' . uri_escape($name) . '&mode=' . $mode,
+                 undef
+         ) == 200;
+ }
+ =head2 set_link
+ Manage node links
+   $node->set_link('http://localhost:1978/node/another', 'another node label', $credit);
+ If C<$credit> is negative, link is removed.
+ =cut
+ sub set_link {
+         my $self = shift;
+         my ($url, $label, $credit) = @_;
+         return unless ($self->{url});
+         croak "mode credit be number, not '$credit'" unless ($credit =~ m/^\d+$/);
+         my $reqbody = 'url=' . uri_escape($url) . '&label=' . uri_escape($label);
+         $reqbody .= '&credit=' . $credit if ($credit > 0);
+         $self->shuttle_url( $self->{url} . '/_set_link',
+                 'text/plain',
+                 $reqbody,
+                 undef
+         ) == 200;
+ }
+ =head1 PRIVATE METHODS
+ You could call those directly, but you don't have to. I hope.
+ =head2 _set_info
+ Set information for node
+   $node->_set_info;
+ =cut
+ sub _set_info {
+         my $self = shift;
+         $self->{status} = -1;
+         return unless ($self->{url});
+         my $resbody;
+         my $rv = $self->shuttle_url( $self->{url} . '/inform',
+                 'text/plain',
+                 undef,
+                 \$resbody,
+         );
+         return if ($rv != 200 || !$resbody);
+         # it seems that response can have multiple line endings
+         $resbody =~ s/[\r\n]+$//;
+         ( $self->{name}, $self->{label}, $self->{dnum}, $self->{wnum}, $self->{size} ) =
+                 split(/\t/, $resbody, 5);
+ }
  ###
  =head1 EXPORT

 Legend:



Removed from v.47
 


changed lines


 
Added in v.58
 Legend:



Removed from v.47
 


changed lines


 
Added in v.58
-Removed from v.47
+Added in v.58

	ViewVC Help
Powered by ViewVC 1.1.26