| 1 |
#!/usr/bin/perl -w |
| 2 |
|
| 3 |
use strict; |
| 4 |
|
| 5 |
# |
| 6 |
# simple implementation to use SWISH-e with queryies like |
| 7 |
# Lucene (subject:something) |
| 8 |
# |
| 9 |
|
| 10 |
use SWISH::API; |
| 11 |
use Text::Iconv; |
| 12 |
use Data::Dumper; |
| 13 |
|
| 14 |
my $iso2utf = Text::Iconv->new('ISO-8859-2','UTF-8'); |
| 15 |
my $utf2iso = Text::Iconv->new('UTF-8','ISO-8859-2'); |
| 16 |
|
| 17 |
sub open_index { |
| 18 |
my $self = shift; |
| 19 |
|
| 20 |
my $swish = $self->{index}; |
| 21 |
|
| 22 |
if (! $swish) { |
| 23 |
|
| 24 |
my $index_file = $self->{index_file} || croak "open_index needs index filename"; |
| 25 |
$index_file .= "/swish-e"; |
| 26 |
print STDERR "opening index '$index_file'\n"; |
| 27 |
$swish = SWISH::API->new($index_file); |
| 28 |
$swish->AbortLastError if $swish->Error; |
| 29 |
|
| 30 |
$self->{index} = $swish; |
| 31 |
} |
| 32 |
|
| 33 |
return $swish; |
| 34 |
} |
| 35 |
|
| 36 |
sub search_index { |
| 37 |
my $self = shift; |
| 38 |
|
| 39 |
croak "search_index needs query" if (! @_); |
| 40 |
|
| 41 |
my $index = $self->open_index; |
| 42 |
|
| 43 |
my $sw; |
| 44 |
|
| 45 |
foreach my $s (@_) { |
| 46 |
|
| 47 |
if ($s =~ /^\s*(\w+):(.+)\s*$/) { |
| 48 |
$sw .= "$1=($2)"; |
| 49 |
} else { |
| 50 |
# and/or/not operators |
| 51 |
$sw .= " $s "; |
| 52 |
} |
| 53 |
} |
| 54 |
|
| 55 |
print STDERR "swish search: $sw\n"; |
| 56 |
|
| 57 |
# convert to UTF-8 |
| 58 |
$sw = $iso2utf->convert($sw) || $sw; |
| 59 |
my $results = $index->Query($sw); |
| 60 |
|
| 61 |
# store total number of hits |
| 62 |
$self->{'total_hits'} = $results->Hits; |
| 63 |
|
| 64 |
my @res_ids; |
| 65 |
|
| 66 |
my $count = 1; |
| 67 |
|
| 68 |
while ( my $r = $results->NextResult ) { |
| 69 |
|
| 70 |
sub p($$) { |
| 71 |
my ($r,$prop) = @_; |
| 72 |
$prop = $r->Property($prop); |
| 73 |
$prop =~ s/##lf##/\n/gs; |
| 74 |
return $utf2iso->convert($prop); |
| 75 |
} |
| 76 |
|
| 77 |
my $id = p($r,"swishdocpath"); |
| 78 |
push @res_ids, $id; |
| 79 |
|
| 80 |
foreach my $p (qw(from to cc bcc)) { |
| 81 |
@{$self->{cache}->{$id}->{$p}} = (); |
| 82 |
foreach my $v (split(/##/, p($r,$p.'_phrase'))) { |
| 83 |
push @{$self->{cache}->{$id}->{$p}}, $v; |
| 84 |
$self->add_counter($p,$v); |
| 85 |
} |
| 86 |
} |
| 87 |
|
| 88 |
foreach my $p (qw(subject body date)) { |
| 89 |
$self->{cache}->{$id}->{$p} = p($r,$p); |
| 90 |
} |
| 91 |
|
| 92 |
$self->add_counter_calendar(p($r,'date_utime')); |
| 93 |
|
| 94 |
# this is redundant, but needed for templates later... |
| 95 |
$self->{cache}->{$id}->{'id'} = $id; |
| 96 |
|
| 97 |
last if (++$count > $self->{max_results}); |
| 98 |
} |
| 99 |
|
| 100 |
return @res_ids; |
| 101 |
} |
| 102 |
|
| 103 |
sub add_index { |
| 104 |
my $self = shift; |
| 105 |
|
| 106 |
my $mbox_id = shift || croak "add_index needs mbox_id"; |
| 107 |
my $document = shift || croak "add_index needs document"; |
| 108 |
|
| 109 |
my ($mbox,$id) = split(/\s/,$mbox_id,2); |
| 110 |
|
| 111 |
my $xml = qq{<message>}; |
| 112 |
foreach my $tag (keys %$document) { |
| 113 |
my $data = $document->{$tag}; |
| 114 |
next if (! $data || $data eq ''); |
| 115 |
# save [cr/]lf before conversion to XML |
| 116 |
$data =~ s/\n\r/##lf##/gs; |
| 117 |
$data =~ s/\n/##lf##/gs; |
| 118 |
$xml .= "<$tag><![CDATA[".$data."]]></$tag>\n"; |
| 119 |
} |
| 120 |
$xml .= qq{</message>}; |
| 121 |
|
| 122 |
$xml = $iso2utf->convert($xml); |
| 123 |
use bytes; # as opposed to chars |
| 124 |
print "Path-Name: $mbox $id\n"; |
| 125 |
print "Content-Length: ".(length($xml)+1)."\n"; |
| 126 |
print "Document-Type: XML\n\n$xml\n"; |
| 127 |
|
| 128 |
} |
| 129 |
|
| 130 |
sub close_index { |
| 131 |
my $self = shift; |
| 132 |
|
| 133 |
} |
| 134 |
|
| 135 |
1; |
Parent Directory
| 
Revision Log