/[webpac2]/trunk/lib/WebPAC/Output/KinoSearch.pm
This is repository of my old source code which isn't updated any more. Go to git.rot13.org for current projects!
ViewVC logotype

Annotation of /trunk/lib/WebPAC/Output/KinoSearch.pm

Parent Directory Parent Directory | Revision Log Revision Log


Revision 914 - (hide annotations)
Tue Oct 30 20:11:04 2007 UTC (16 years, 6 months ago) by dpavlin
File size: 2897 byte(s)
 r1373@llin:  dpavlin | 2007-10-30 21:11:01 +0100
 use KinoSearch::Simple and convert to new Output API

1 dpavlin 431 package WebPAC::Output::KinoSearch;
2    
3     use warnings;
4     use strict;
5    
6 dpavlin 914 use base qw/WebPAC::Common WebPAC::Output Class::Accessor/;
7     __PACKAGE__->mk_accessors(qw(
8     path
9     database
10     encoding
11     clean
12 dpavlin 431
13 dpavlin 914 index
14     ));
15    
16     use KinoSearch::Simple;
17     use File::Path;
18 dpavlin 431 use Encode qw/from_to/;
19 dpavlin 887 use Data::Dump qw/dump/;
20 dpavlin 536 use Storable;
21 dpavlin 431
22     =head1 NAME
23    
24     WebPAC::Output::KinoSearch - Create KinoSearch full text index
25    
26     =head1 VERSION
27    
28 dpavlin 914 Version 0.04
29 dpavlin 431
30     =cut
31    
32 dpavlin 914 our $VERSION = '0.04';
33 dpavlin 431
34     =head1 SYNOPSIS
35    
36     Create full text index using KinoSearch index from data with
37     type C<search>.
38    
39     =head1 FUNCTIONS
40    
41     =head2 new
42    
43     Open KinoSearch index
44    
45 dpavlin 914 my $est = new WebPAC::Output::KinoSearch({
46     path => '/path/to/invindex',
47 dpavlin 431 database => 'demo',
48     encoding => 'iso-8859-2',
49     clean => 1,
50 dpavlin 914 });
51 dpavlin 431
52     Options are:
53    
54     =over 4
55    
56 dpavlin 914 =item path
57 dpavlin 431
58     path to KinoSearch index to use
59    
60     =item database
61    
62     name of database from which data comes
63    
64     =item encoding
65    
66     character encoding of C<data_structure> if it's differenet than C<ISO-8859-2>
67     (and it probably is). This encoding will be converted to C<UTF-8> for
68     index.
69    
70     =back
71    
72     =cut
73    
74 dpavlin 914 sub init {
75     my $self = shift;
76 dpavlin 431
77     my $log = $self->_get_logger;
78    
79 dpavlin 887 #$log->debug("self: ", sub { dump($self) });
80 dpavlin 431
81 dpavlin 914 foreach my $p (qw/path database/) {
82     $log->logdie("need $p") unless ($self->$p);
83 dpavlin 431 }
84    
85 dpavlin 914 # $log->logdie("fields is not ARRAY") unless (ref($self->{fields}) eq 'ARRAY');
86 dpavlin 431
87 dpavlin 914 $self->encoding( 'ISO-8859-2' ) unless $self->encoding;
88 dpavlin 431
89 dpavlin 914 if ( ! -e $self->path ) {
90     mkpath $self->path || $log->logdie("can't create ", $self->path,": $!");
91     $log->info("created ", $self->path);
92     }
93 dpavlin 610
94 dpavlin 914 my $path = $self->path . '/' . $self->database;
95 dpavlin 431
96 dpavlin 914 $log->info("using index $path with encoding ", $self->encoding);
97 dpavlin 609
98 dpavlin 914 my $index = KinoSearch::Simple->new(
99     path => $path,
100     language => 'en',
101 dpavlin 431 );
102    
103 dpavlin 914 $log->logdie("can't open $path: $!") unless $index;
104 dpavlin 536
105 dpavlin 914 $self->index( $index );
106 dpavlin 431
107     }
108    
109    
110     =head2 add
111    
112 dpavlin 914 Adds one entry
113 dpavlin 431
114 dpavlin 914 $est->add( 42, $ds );
115 dpavlin 431
116     =cut
117    
118     sub add {
119     my $self = shift;
120    
121 dpavlin 914 my ( $id, $ds ) = @_;
122 dpavlin 431
123     my $log = $self->_get_logger;
124 dpavlin 914 $log->logdie("need id") unless defined $id;
125     $log->logdie("need ds") unless $ds;
126 dpavlin 431
127 dpavlin 914 $log->debug("id: $id ds = ",dump($ds));
128 dpavlin 431
129 dpavlin 914 my $hash = $self->ds_to_hash( $ds, 'search' ) || return;
130 dpavlin 431
131 dpavlin 914 warn "add( $id, ",dump($ds)," ) => ", dump( $hash );
132 dpavlin 431
133 dpavlin 914 $self->index->add_doc( $hash );
134 dpavlin 431
135     return 1;
136     }
137    
138 dpavlin 434 =head2 finish
139 dpavlin 431
140 dpavlin 434 Close index
141    
142     $index->finish;
143    
144     =cut
145    
146     sub finish {
147     my $self = shift;
148    
149 dpavlin 536 my $log = $self->_get_logger();
150    
151 dpavlin 914 $log->info("dummy finish");
152 dpavlin 536
153 dpavlin 434 }
154    
155 dpavlin 431 =head2 convert
156    
157     my $utf8_string = $self->convert('string in codepage');
158    
159     =cut
160    
161     sub convert {
162     my $self = shift;
163    
164     my $text = shift || return;
165     from_to($text, $self->{encoding}, 'UTF-8');
166     return $text;
167     }
168    
169     =head1 AUTHOR
170    
171     Dobrica Pavlinusic, C<< <dpavlin@rot13.org> >>
172    
173     =head1 COPYRIGHT & LICENSE
174    
175 dpavlin 914 Copyright 2005-2007 Dobrica Pavlinusic, All Rights Reserved.
176 dpavlin 431
177     This program is free software; you can redistribute it and/or modify it
178     under the same terms as Perl itself.
179    
180     =cut
181    
182     1; # End of WebPAC::Output::Estraier

  ViewVC Help
Powered by ViewVC 1.1.26