1 |
package WebPAC::Output::KinoSearch; |
2 |
|
3 |
use warnings; |
4 |
use strict; |
5 |
|
6 |
use base qw/WebPAC::Common/; |
7 |
|
8 |
use KinoSearch::InvIndexer; |
9 |
use KinoSearch::Analysis::PolyAnalyzer; |
10 |
use Encode qw/from_to/; |
11 |
use Data::Dumper; |
12 |
|
13 |
=head1 NAME |
14 |
|
15 |
WebPAC::Output::KinoSearch - Create KinoSearch full text index |
16 |
|
17 |
=head1 VERSION |
18 |
|
19 |
Version 0.01 |
20 |
|
21 |
=cut |
22 |
|
23 |
our $VERSION = '0.01'; |
24 |
|
25 |
=head1 SYNOPSIS |
26 |
|
27 |
Create full text index using KinoSearch index from data with |
28 |
type C<search>. |
29 |
|
30 |
=head1 FUNCTIONS |
31 |
|
32 |
=head2 new |
33 |
|
34 |
Open KinoSearch index |
35 |
|
36 |
my $est = new WebPAC::Output::KinoSearch( |
37 |
index_path => '/path/to/invindex', |
38 |
fields => qw/name of all filelds used/, |
39 |
database => 'demo', |
40 |
label => 'node label', |
41 |
encoding => 'iso-8859-2', |
42 |
clean => 1, |
43 |
); |
44 |
|
45 |
Options are: |
46 |
|
47 |
=over 4 |
48 |
|
49 |
=item index_path |
50 |
|
51 |
path to KinoSearch index to use |
52 |
|
53 |
=item fields |
54 |
|
55 |
name of all fields used in this index |
56 |
|
57 |
=item database |
58 |
|
59 |
name of database from which data comes |
60 |
|
61 |
=item label |
62 |
|
63 |
label for node (optional) |
64 |
|
65 |
=item encoding |
66 |
|
67 |
character encoding of C<data_structure> if it's differenet than C<ISO-8859-2> |
68 |
(and it probably is). This encoding will be converted to C<UTF-8> for |
69 |
index. |
70 |
|
71 |
=back |
72 |
|
73 |
=cut |
74 |
|
75 |
sub new { |
76 |
my $class = shift; |
77 |
my $self = {@_}; |
78 |
bless($self, $class); |
79 |
|
80 |
my $log = $self->_get_logger; |
81 |
|
82 |
#$log->debug("self: ", sub { Dumper($self) }); |
83 |
|
84 |
foreach my $p (qw/index_path fields database/) { |
85 |
$log->logdie("need $p") unless ($self->{$p}); |
86 |
} |
87 |
|
88 |
$log->logdie("fields is not ARRAY") unless (ref($self->{fields}) eq 'ARRAY'); |
89 |
|
90 |
$self->{encoding} ||= 'ISO-8859-2'; |
91 |
|
92 |
$log->info("using index $self->{index_path} with encoding $self->{encoding}"); |
93 |
|
94 |
my $analyzer = KinoSearch::Analysis::PolyAnalyzer->new( language => 'en' ); |
95 |
|
96 |
$self->{invindex} = KinoSearch::InvIndexer->new( |
97 |
invindex => $self->{index_path}, |
98 |
create => $self->{clean}, |
99 |
analyzer => $analyzer, |
100 |
); |
101 |
|
102 |
foreach my $f (@{ $self->{fields} }) { |
103 |
$self->{invindex}->spec_field( |
104 |
name => $f, |
105 |
# boost => 10, |
106 |
stored => 1, |
107 |
indexed => 1, |
108 |
vectorized => 0, |
109 |
); |
110 |
} |
111 |
|
112 |
$self ? return $self : return undef; |
113 |
} |
114 |
|
115 |
|
116 |
=head2 add |
117 |
|
118 |
Adds one entry to database. |
119 |
|
120 |
$est->add( |
121 |
id => 42, |
122 |
ds => $ds, |
123 |
type => 'display', |
124 |
text => 'optional text from which snippet is created', |
125 |
); |
126 |
|
127 |
This function will create entries in index using following URI format: |
128 |
|
129 |
C<file:///type/database%20name/000> |
130 |
|
131 |
Each tag in C<data_structure> with specified C<type> will create one |
132 |
attribute and corresponding hidden text (used for search). |
133 |
|
134 |
=cut |
135 |
|
136 |
sub add { |
137 |
my $self = shift; |
138 |
|
139 |
my $args = {@_}; |
140 |
|
141 |
my $log = $self->_get_logger; |
142 |
|
143 |
my $database = $self->{'database'} || $log->logconfess('no database in $self'); |
144 |
$log->logconfess('need invindex in object') unless ($self->{'invindex'}); |
145 |
|
146 |
foreach my $p (qw/id ds type/) { |
147 |
$log->logdie("need $p") unless ($args->{$p}); |
148 |
} |
149 |
|
150 |
my $type = $args->{'type'}; |
151 |
my $id = $args->{'id'}; |
152 |
|
153 |
my $uri = "file:///$type/$database/$id"; |
154 |
$log->debug("creating $uri"); |
155 |
|
156 |
my $doc = $self->{invindex}->new_doc( $uri ) || $log->logdie("can't create new_doc( $uri )"); |
157 |
eval { $doc->set_value('uri', $self->convert($uri) ) }; |
158 |
|
159 |
$log->debug("ds = ", sub { Dumper($args->{'ds'}) } ); |
160 |
|
161 |
# filter all tags which have type defined |
162 |
my @tags = grep { |
163 |
ref($args->{'ds'}->{$_}) eq 'HASH' && defined( $args->{'ds'}->{$_}->{$type} ) |
164 |
} keys %{ $args->{'ds'} }; |
165 |
|
166 |
$log->debug("tags = ", join(",", @tags)); |
167 |
|
168 |
return unless (@tags); |
169 |
|
170 |
foreach my $tag (@tags) { |
171 |
|
172 |
my $vals = join(" ", @{ $args->{'ds'}->{$tag}->{$type} }); |
173 |
|
174 |
next if (! $vals); |
175 |
|
176 |
$vals = $self->convert( $vals ) or |
177 |
$log->logdie("can't convert '$vals' to UTF-8"); |
178 |
|
179 |
eval { $doc->set_value( $tag, $vals ) }; |
180 |
} |
181 |
|
182 |
my $text = $args->{'text'}; |
183 |
if ( $text ) { |
184 |
$text = $self->convert( $text ) or |
185 |
$log->logdie("can't convert '$text' to UTF-8"); |
186 |
eval { $doc->set_value( bodytext => $text ) }; |
187 |
} |
188 |
|
189 |
#$log->debug("adding ", sub { $doc->dump_draft } ); |
190 |
$self->{invindex}->add_doc($doc) || $log->warn("can't add document $uri"); |
191 |
|
192 |
return 1; |
193 |
} |
194 |
|
195 |
|
196 |
=head2 convert |
197 |
|
198 |
my $utf8_string = $self->convert('string in codepage'); |
199 |
|
200 |
=cut |
201 |
|
202 |
sub convert { |
203 |
my $self = shift; |
204 |
|
205 |
my $text = shift || return; |
206 |
from_to($text, $self->{encoding}, 'UTF-8'); |
207 |
return $text; |
208 |
} |
209 |
|
210 |
=head1 AUTHOR |
211 |
|
212 |
Dobrica Pavlinusic, C<< <dpavlin@rot13.org> >> |
213 |
|
214 |
=head1 COPYRIGHT & LICENSE |
215 |
|
216 |
Copyright 2005 Dobrica Pavlinusic, All Rights Reserved. |
217 |
|
218 |
This program is free software; you can redistribute it and/or modify it |
219 |
under the same terms as Perl itself. |
220 |
|
221 |
=cut |
222 |
|
223 |
1; # End of WebPAC::Output::Estraier |