688 |
|
|
689 |
package Search::Estraier::Node; |
package Search::Estraier::Node; |
690 |
|
|
691 |
use Carp qw/carp croak/; |
use Carp qw/carp croak confess/; |
692 |
use URI; |
use URI; |
693 |
use MIME::Base64; |
use MIME::Base64; |
694 |
use IO::Socket::INET; |
use IO::Socket::INET; |
854 |
|
|
855 |
Remove a registrated document using it's uri |
Remove a registrated document using it's uri |
856 |
|
|
857 |
$node->out_doc_by_uri( 'file:///document_url' ) or "can't remove document"; |
$node->out_doc_by_uri( 'file:///document/uri/42' ) or "can't remove document"; |
858 |
|
|
859 |
Return true on success or false on failture. |
Return true on success or false on failture. |
860 |
|
|
910 |
return $self->_fetch_doc( id => $id ); |
return $self->_fetch_doc( id => $id ); |
911 |
} |
} |
912 |
|
|
913 |
|
|
914 |
=head2 get_doc_by_uri |
=head2 get_doc_by_uri |
915 |
|
|
916 |
Retreive document |
Retreive document |
917 |
|
|
918 |
my $doc = $node->get_doc_by_uri( 'file:///document_uri' ) or die "can't get document"; |
my $doc = $node->get_doc_by_uri( 'file:///document/uri/42' ) or die "can't get document"; |
919 |
|
|
920 |
Return true on success or false on failture. |
Return true on success or false on failture. |
921 |
|
|
927 |
return $self->_fetch_doc( uri => $uri ); |
return $self->_fetch_doc( uri => $uri ); |
928 |
} |
} |
929 |
|
|
930 |
|
|
931 |
|
=head2 etch_doc |
932 |
|
|
933 |
|
Exctract document keywords |
934 |
|
|
935 |
|
my $keywords = $node->etch_doc( document_id ) or die "can't etch document"; |
936 |
|
|
937 |
|
=cut |
938 |
|
|
939 |
|
sub erch_doc { |
940 |
|
my $self = shift; |
941 |
|
my $id = shift || return; |
942 |
|
return $self->_fetch_doc( id => $id, etch => 1 ); |
943 |
|
} |
944 |
|
|
945 |
|
=head2 etch_doc_by_uri |
946 |
|
|
947 |
|
Retreive document |
948 |
|
|
949 |
|
my $keywords = $node->etch_doc_by_uri( 'file:///document/uri/42' ) or die "can't etch document"; |
950 |
|
|
951 |
|
Return true on success or false on failture. |
952 |
|
|
953 |
|
=cut |
954 |
|
|
955 |
|
sub etch_doc_by_uri { |
956 |
|
my $self = shift; |
957 |
|
my $uri = shift || return; |
958 |
|
return $self->_fetch_doc( uri => $uri, etch => 1 ); |
959 |
|
} |
960 |
|
|
961 |
|
|
962 |
|
=head2 uri_to_id |
963 |
|
|
964 |
|
Get ID of document specified by URI |
965 |
|
|
966 |
|
my $id = $node->uri_to_id( 'file:///document/uri/42' ); |
967 |
|
|
968 |
|
=cut |
969 |
|
|
970 |
|
sub uri_to_id { |
971 |
|
my $self = shift; |
972 |
|
my $uri = shift || return; |
973 |
|
return $self->_fetch_doc( uri => $uri, path => '/uri_to_id', chomp_resbody => 1 ); |
974 |
|
} |
975 |
|
|
976 |
|
|
977 |
=head2 _fetch_doc |
=head2 _fetch_doc |
978 |
|
|
979 |
Private function used for implementation of C<get_doc> and C<get_doc_by_uri>. |
Private function used for implementing of C<get_doc>, C<get_doc_by_uri>, |
980 |
|
C<etch_doc>, C<etch_doc_by_uri>. |
981 |
|
|
982 |
my $doc = $node->fetch_doc( id => 42 ); |
# this will decode received draft into Search::Estraier::Document object |
983 |
my $doc = $node->fetch_doc( uri => 'file://uri/42' ); |
my $doc = $node->_fetch_doc( id => 42 ); |
984 |
|
my $doc = $node->_fetch_doc( uri => 'file:///document/uri/42' ); |
985 |
|
|
986 |
|
# to extract keywords, add etch |
987 |
|
my $doc = $node->_fetch_doc( id => 42, etch => 1 ); |
988 |
|
my $doc = $node->_fetch_doc( uri => 'file:///document/uri/42', etch => 1 ); |
989 |
|
|
990 |
|
# more general form which allows implementation of |
991 |
|
# uri_to_id |
992 |
|
my $id = $node->_fetch_doc( |
993 |
|
uri => 'file:///document/uri/42', |
994 |
|
path => '/uri_to_id', |
995 |
|
chomp_resbody => 1 |
996 |
|
); |
997 |
|
|
998 |
=cut |
=cut |
999 |
|
|
1000 |
sub _fetch_doc { |
sub _fetch_doc { |
1001 |
my $self = shift; |
my $self = shift; |
1002 |
my ($name,$val) = @_; |
my $a = {@_}; |
1003 |
return unless ($name && defined($val) && $self->{url}); |
return unless ( ($a->{id} || $a->{uri}) && $self->{url} ); |
1004 |
if ($name eq 'id') { |
|
1005 |
croak "id must be numberm not '$val'" unless ($val =~ m/^\d+$/); |
my ($arg, $resbody); |
1006 |
|
|
1007 |
|
my $path = $a->{path} || '/get_doc'; |
1008 |
|
$path = '/etch_doc' if ($a->{etch}); |
1009 |
|
|
1010 |
|
if ($a->{id}) { |
1011 |
|
croak "id must be numberm not '$a->{id}'" unless ($a->{id} =~ m/^\d+$/); |
1012 |
|
$arg = 'id=' . $a->{id}; |
1013 |
|
} elsif ($a->{uri}) { |
1014 |
|
$arg = 'uri=' . $a->{uri}; |
1015 |
|
} else { |
1016 |
|
confess "unhandled argument. Need id or uri."; |
1017 |
} |
} |
1018 |
my $rv = $self->shuttle_url( $self->{url} . '/get_doc', |
|
1019 |
|
my $rv = $self->shuttle_url( $self->{url} . $path, |
1020 |
'application/x-www-form-urlencoded', |
'application/x-www-form-urlencoded', |
1021 |
"$name=$val", |
$arg, |
1022 |
my $draft, |
\$resbody, |
1023 |
); |
); |
1024 |
|
|
1025 |
return if ($rv != 200); |
return if ($rv != 200); |
1026 |
return new Search::Estraier::Document($draft); |
|
1027 |
|
if ($a->{etch}) { |
1028 |
|
$self->{kwords} = {}; |
1029 |
|
return +{} unless ($resbody); |
1030 |
|
foreach my $l (split(/\n/, $resbody)) { |
1031 |
|
my ($k,$v) = split(/\t/, $l, 2); |
1032 |
|
$self->{kwords}->{$k} = $v if ($v); |
1033 |
|
} |
1034 |
|
return $self->{kwords}; |
1035 |
|
} elsif ($a->{chomp_resbody}) { |
1036 |
|
return unless (defined($resbody)); |
1037 |
|
chomp($resbody); |
1038 |
|
return $resbody; |
1039 |
|
} else { |
1040 |
|
return new Search::Estraier::Document($resbody); |
1041 |
|
} |
1042 |
} |
} |
1043 |
|
|
1044 |
|
|