line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
1
|
|
|
|
|
|
|
package Plucene::Plugin::WeightedQueryParser; |
2
|
1
|
|
|
1
|
|
1880
|
use 5.006; |
|
1
|
|
|
|
|
4
|
|
|
1
|
|
|
|
|
38
|
|
3
|
1
|
|
|
1
|
|
5
|
use strict; |
|
1
|
|
|
|
|
2
|
|
|
1
|
|
|
|
|
22
|
|
4
|
1
|
|
|
1
|
|
4
|
use warnings; |
|
1
|
|
|
|
|
4
|
|
|
1
|
|
|
|
|
50
|
|
5
|
|
|
|
|
|
|
our $VERSION = '1.0'; |
6
|
1
|
|
|
1
|
|
4
|
use base 'Plucene::QueryParser'; |
|
1
|
|
|
|
|
2
|
|
|
1
|
|
|
|
|
1996
|
|
7
|
|
|
|
|
|
|
__PACKAGE__->mk_accessors("weights"); |
8
|
|
|
|
|
|
|
|
9
|
|
|
|
|
|
|
=head1 NAME |
10
|
|
|
|
|
|
|
|
11
|
|
|
|
|
|
|
Plucene::Plugin::WeightedQueryParser - Specify weights for unqualified terms |
12
|
|
|
|
|
|
|
|
13
|
|
|
|
|
|
|
=head1 SYNOPSIS |
14
|
|
|
|
|
|
|
|
15
|
|
|
|
|
|
|
use Plucene::Plugin::WeightedQueryParser; |
16
|
|
|
|
|
|
|
my $parser = Plucene::WeightedQueryParser->new({ |
17
|
|
|
|
|
|
|
analyzer => Plucene::Plugin::Analyzer::PorterAnalyzer->new(), |
18
|
|
|
|
|
|
|
weights => { |
19
|
|
|
|
|
|
|
title => 5, |
20
|
|
|
|
|
|
|
subtitle => 2, |
21
|
|
|
|
|
|
|
body => 1 |
22
|
|
|
|
|
|
|
} |
23
|
|
|
|
|
|
|
}); |
24
|
|
|
|
|
|
|
$parser->parse($q); |
25
|
|
|
|
|
|
|
|
26
|
|
|
|
|
|
|
=head1 DESCRIPTION |
27
|
|
|
|
|
|
|
|
28
|
|
|
|
|
|
|
Quite often, you want unqualified search terms (C) to search in |
29
|
|
|
|
|
|
|
either the title, the body or some other part of you document. The usual |
30
|
|
|
|
|
|
|
way to do this is to create another field, C, and throw everything |
31
|
|
|
|
|
|
|
in that, then make that the default field for unqualified terms. |
32
|
|
|
|
|
|
|
|
33
|
|
|
|
|
|
|
That's fine, until you want to add different weighting for terms - so terms |
34
|
|
|
|
|
|
|
found in the title of a document come first, then those in the subtitle, |
35
|
|
|
|
|
|
|
then the body of the document. |
36
|
|
|
|
|
|
|
|
37
|
|
|
|
|
|
|
This module automatically remaps unqualified search terms, such as C |
38
|
|
|
|
|
|
|
to C<(title:hello^5 OR subtitle:hello^2 OR body:hello^1)>, based on the |
39
|
|
|
|
|
|
|
weights passed in to the constructor. |
40
|
|
|
|
|
|
|
|
41
|
|
|
|
|
|
|
=cut |
42
|
|
|
|
|
|
|
|
43
|
|
|
|
|
|
|
sub new { |
44
|
|
|
|
|
|
|
my ($self, $args) = @_; |
45
|
|
|
|
|
|
|
$args->{default} = "unused"; |
46
|
|
|
|
|
|
|
$self->SUPER::new($args); |
47
|
|
|
|
|
|
|
} |
48
|
|
|
|
|
|
|
|
49
|
|
|
|
|
|
|
sub parse { |
50
|
|
|
|
|
|
|
my ($self, $thing) = @_; |
51
|
|
|
|
|
|
|
my $ast = $self->SUPER::parse($thing, 1); |
52
|
|
|
|
|
|
|
# First, rebless the AST into our derived class. |
53
|
|
|
|
|
|
|
(ref $self)->rebless($ast); |
54
|
|
|
|
|
|
|
$ast->to_plucene($self->{weights}); |
55
|
|
|
|
|
|
|
} |
56
|
|
|
|
|
|
|
|
57
|
|
|
|
|
|
|
sub rebless { |
58
|
|
|
|
|
|
|
my ($class, $t) = @_; |
59
|
|
|
|
|
|
|
if ($t->isa("Plucene::QueryParser::Subquery")) { |
60
|
|
|
|
|
|
|
$class->rebless($_) for @{$t->{subquery}}; |
61
|
|
|
|
|
|
|
} elsif ($t->isa("Plucene::QueryParser::TopLevel")) { |
62
|
|
|
|
|
|
|
$class->rebless($_) for @$t; |
63
|
|
|
|
|
|
|
} |
64
|
|
|
|
|
|
|
my $new_class = ref $t; |
65
|
|
|
|
|
|
|
$new_class =~ s/Plucene::QueryParser/$class/; |
66
|
|
|
|
|
|
|
bless $t, $new_class; |
67
|
|
|
|
|
|
|
} |
68
|
|
|
|
|
|
|
|
69
|
|
|
|
|
|
|
package Plucene::Plugin::WeightedQueryParser::TopLevel; |
70
|
|
|
|
|
|
|
use base 'Plucene::QueryParser::TopLevel'; |
71
|
|
|
|
|
|
|
|
72
|
|
|
|
|
|
|
package Plucene::Plugin::WeightedQueryParser::Subquery; |
73
|
|
|
|
|
|
|
use base 'Plucene::QueryParser::Subquery'; |
74
|
|
|
|
|
|
|
|
75
|
|
|
|
|
|
|
package Plucene::Plugin::WeightedQueryParser::Term; |
76
|
|
|
|
|
|
|
use base 'Plucene::QueryParser::Term'; |
77
|
|
|
|
|
|
|
sub to_plucene { |
78
|
|
|
|
|
|
|
my ($self, $weights) = @_; |
79
|
|
|
|
|
|
|
return $self->SUPER::to_plucene() if exists $self->{field}; |
80
|
|
|
|
|
|
|
require Plucene::Search::BooleanQuery; |
81
|
|
|
|
|
|
|
require Plucene::Search::BooleanClause; |
82
|
|
|
|
|
|
|
my $query = new Plucene::Search::BooleanQuery; |
83
|
|
|
|
|
|
|
|
84
|
|
|
|
|
|
|
for my $field (keys %$weights) { |
85
|
|
|
|
|
|
|
my $q = $self->SUPER::to_plucene($field); |
86
|
|
|
|
|
|
|
$q->boost($weights->{$field}); |
87
|
|
|
|
|
|
|
$query->add_clause(Plucene::Search::BooleanClause->new({ |
88
|
|
|
|
|
|
|
prohibited => 0, required => 0, query => $q |
89
|
|
|
|
|
|
|
})); |
90
|
|
|
|
|
|
|
} |
91
|
|
|
|
|
|
|
$query; |
92
|
|
|
|
|
|
|
} |
93
|
|
|
|
|
|
|
|
94
|
|
|
|
|
|
|
package Plucene::Plugin::WeightedQueryParser::Phrase; |
95
|
|
|
|
|
|
|
use base 'Plucene::QueryParser::Phrase'; |
96
|
|
|
|
|
|
|
sub to_plucene { |
97
|
|
|
|
|
|
|
my ($self, $weights) = @_; |
98
|
|
|
|
|
|
|
return $self->SUPER::to_plucene() if exists $self->{field}; |
99
|
|
|
|
|
|
|
require Plucene::Search::BooleanQuery; |
100
|
|
|
|
|
|
|
require Plucene::Search::BooleanClause; |
101
|
|
|
|
|
|
|
my $query = new Plucene::Search::BooleanQuery; |
102
|
|
|
|
|
|
|
|
103
|
|
|
|
|
|
|
for my $field (keys %$weights) { |
104
|
|
|
|
|
|
|
my $q = $self->SUPER::to_plucene($field); |
105
|
|
|
|
|
|
|
$q->boost($weights->{$field}); |
106
|
|
|
|
|
|
|
$query->add_clause(Plucene::Search::BooleanClause->new({ |
107
|
|
|
|
|
|
|
prohibited => 0, required => 0, query => $q |
108
|
|
|
|
|
|
|
})); |
109
|
|
|
|
|
|
|
} |
110
|
|
|
|
|
|
|
$query; |
111
|
|
|
|
|
|
|
} |
112
|
|
|
|
|
|
|
|
113
|
|
|
|
|
|
|
package Plucene::Plugin::WeightedQueryParser::Prefix; |
114
|
|
|
|
|
|
|
use base 'Plucene::QueryParser::Prefix'; |
115
|
|
|
|
|
|
|
sub to_plucene { |
116
|
|
|
|
|
|
|
my ($self, $weights) = @_; |
117
|
|
|
|
|
|
|
return $self->SUPER::to_plucene() if exists $self->{field}; |
118
|
|
|
|
|
|
|
require Plucene::Search::BooleanQuery; |
119
|
|
|
|
|
|
|
require Plucene::Search::BooleanClause; |
120
|
|
|
|
|
|
|
my $query = new Plucene::Search::BooleanQuery; |
121
|
|
|
|
|
|
|
|
122
|
|
|
|
|
|
|
for my $field (keys %$weights) { |
123
|
|
|
|
|
|
|
my $q = $self->SUPER::to_plucene($field); |
124
|
|
|
|
|
|
|
$q->boost($weights->{$field}); |
125
|
|
|
|
|
|
|
$query->add_clause(Plucene::Search::BooleanClause->new({ |
126
|
|
|
|
|
|
|
prohibited => 0, required => 0, query => $q |
127
|
|
|
|
|
|
|
})); |
128
|
|
|
|
|
|
|
} |
129
|
|
|
|
|
|
|
$query; |
130
|
|
|
|
|
|
|
} |
131
|
|
|
|
|
|
|
|
132
|
|
|
|
|
|
|
1; |
133
|
|
|
|
|
|
|
|
134
|
|
|
|
|
|
|
=head1 AUTHOR |
135
|
|
|
|
|
|
|
|
136
|
|
|
|
|
|
|
Simon Cozens, C |
137
|
|
|
|
|
|
|
|
138
|
|
|
|
|
|
|
Development of this module was made possible by the generous sponsorship |
139
|
|
|
|
|
|
|
of Text Matters, http://www.textmatters.com/ |
140
|
|
|
|
|
|
|
|
141
|
|
|
|
|
|
|
This module may be distributed under the same terms as Plucene itself. |
142
|
|
|
|
|
|
|
|
143
|
|
|
|
|
|
|
=head1 SEE ALSO |
144
|
|
|
|
|
|
|
|
145
|
|
|
|
|
|
|
L. |
146
|
|
|
|
|
|
|
|
147
|
|
|
|
|
|
|
=cut |
148
|
|
|
|
|
|
|
|
149
|
|
|
|
|
|
|
1; |