line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
1
|
|
|
|
|
|
|
package Plucene::Plugin::FileDocument; |
2
|
|
|
|
|
|
|
|
3
|
|
|
|
|
|
|
our $VERSION = '1.01'; |
4
|
|
|
|
|
|
|
|
5
|
1
|
|
|
1
|
|
25453
|
use strict; |
|
1
|
|
|
|
|
3
|
|
|
1
|
|
|
|
|
48
|
|
6
|
1
|
|
|
1
|
|
6
|
use warnings; |
|
1
|
|
|
|
|
1
|
|
|
1
|
|
|
|
|
41
|
|
7
|
|
|
|
|
|
|
|
8
|
|
|
|
|
|
|
=head1 NAME |
9
|
|
|
|
|
|
|
|
10
|
|
|
|
|
|
|
Plucene::Plugin::FileDocument - turn a file into a document |
11
|
|
|
|
|
|
|
|
12
|
|
|
|
|
|
|
=head1 SYNOPSIS |
13
|
|
|
|
|
|
|
|
14
|
|
|
|
|
|
|
my $writer = Plucene::Index::Writer->new(...); |
15
|
|
|
|
|
|
|
$writer->add_document(Plucene::Plugin::FileDocument->new($filename)); |
16
|
|
|
|
|
|
|
|
17
|
|
|
|
|
|
|
=head1 DESCRIPTION |
18
|
|
|
|
|
|
|
|
19
|
|
|
|
|
|
|
This is a Plucene::Document subclass for indexing an entire file on disk. |
20
|
|
|
|
|
|
|
|
21
|
|
|
|
|
|
|
=head1 CONSTRUCTOR |
22
|
|
|
|
|
|
|
|
23
|
|
|
|
|
|
|
=head2 new |
24
|
|
|
|
|
|
|
|
25
|
|
|
|
|
|
|
$writer->add_document(Plucene::Plugin::FileDocument->new($filename)); |
26
|
|
|
|
|
|
|
|
27
|
|
|
|
|
|
|
This loads the file with the given name, and turns it into a |
28
|
|
|
|
|
|
|
Plucene::Document. |
29
|
|
|
|
|
|
|
|
30
|
|
|
|
|
|
|
By default two Text fields will be created: 'path', which is the full |
31
|
|
|
|
|
|
|
name of the file, and 'text' which is the content of the file. |
32
|
|
|
|
|
|
|
|
33
|
|
|
|
|
|
|
Extra fields can can be added by passing a hash of { fieldname => subref } |
34
|
|
|
|
|
|
|
pairs to the constructor: |
35
|
|
|
|
|
|
|
|
36
|
|
|
|
|
|
|
$writer->add_document(Plucene::Plugin::FileDocument->new($filename, |
37
|
|
|
|
|
|
|
lastmod => sub { (stat(+shift->{path}))[9] }, |
38
|
|
|
|
|
|
|
); |
39
|
|
|
|
|
|
|
|
40
|
|
|
|
|
|
|
These subrefs will be passed a hashref of the 'path' and 'text'. |
41
|
|
|
|
|
|
|
|
42
|
|
|
|
|
|
|
=head1 CUSTOMISATION |
43
|
|
|
|
|
|
|
|
44
|
|
|
|
|
|
|
=head2 document_class / field_class |
45
|
|
|
|
|
|
|
|
46
|
|
|
|
|
|
|
By default we assume you are using Plucene::Document and |
47
|
|
|
|
|
|
|
Plucene::Document::Field. If you are using subclasses of these instead, |
48
|
|
|
|
|
|
|
then you should override these methods accordingly. |
49
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
=cut |
51
|
|
|
|
|
|
|
|
52
|
1
|
|
|
1
|
|
940
|
use File::Slurp 'read_file'; |
|
1
|
|
|
|
|
18116
|
|
|
1
|
|
|
|
|
73
|
|
53
|
|
|
|
|
|
|
|
54
|
1
|
|
|
1
|
|
963
|
use Plucene::Document; |
|
1
|
|
|
|
|
5333
|
|
|
1
|
|
|
|
|
40
|
|
55
|
1
|
|
|
1
|
|
936
|
use Plucene::Document::Field; |
|
1
|
|
|
|
|
769
|
|
|
1
|
|
|
|
|
9
|
|
56
|
|
|
|
|
|
|
|
57
|
|
|
|
|
|
|
sub new { |
58
|
2
|
|
|
2
|
1
|
5980
|
my ($self, $file, %args) = @_; |
59
|
2
|
|
|
|
|
9
|
my $doc = $self->document_class->new; |
60
|
2
|
|
|
|
|
32
|
my $field_class = $self->field_class; |
61
|
2
|
|
|
|
|
12
|
my $hash = { |
62
|
|
|
|
|
|
|
path => $file, |
63
|
|
|
|
|
|
|
text => scalar read_file($file), |
64
|
|
|
|
|
|
|
}; |
65
|
2
|
|
|
|
|
405
|
while (my ($field, $subref) = each %args) { |
66
|
1
|
|
|
|
|
4
|
$hash->{$field} = $subref->($hash); |
67
|
|
|
|
|
|
|
} |
68
|
2
|
|
|
|
|
38
|
while (my ($key, $val) = each %$hash) { |
69
|
5
|
|
|
|
|
153
|
$doc->add($field_class->Text($key => $val)); |
70
|
|
|
|
|
|
|
} |
71
|
2
|
|
|
|
|
67
|
return $doc; |
72
|
|
|
|
|
|
|
} |
73
|
|
|
|
|
|
|
|
74
|
2
|
|
|
2
|
1
|
24
|
sub document_class { "Plucene::Document" } |
75
|
2
|
|
|
2
|
1
|
5
|
sub field_class { "Plucene::Document::Field" } |
76
|
|
|
|
|
|
|
|
77
|
|
|
|
|
|
|
=head1 AUTHOR |
78
|
|
|
|
|
|
|
|
79
|
|
|
|
|
|
|
Tony Bowden, based on example code from the Lucene benchmark suite. |
80
|
|
|
|
|
|
|
|
81
|
|
|
|
|
|
|
=head1 BUGS and QUERIES |
82
|
|
|
|
|
|
|
|
83
|
|
|
|
|
|
|
Please direct all correspondence regarding this module to: |
84
|
|
|
|
|
|
|
bug-Plucene-Plugin-FileDocument@rt.cpan.org |
85
|
|
|
|
|
|
|
|
86
|
|
|
|
|
|
|
=head1 COPYRIGHT AND LICENSE |
87
|
|
|
|
|
|
|
|
88
|
|
|
|
|
|
|
Copyright (C) 2004-2005 Kasei |
89
|
|
|
|
|
|
|
|
90
|
|
|
|
|
|
|
This program is free software; you can redistribute it and/or modify |
91
|
|
|
|
|
|
|
it under the terms of the GNU General Public License; either version |
92
|
|
|
|
|
|
|
2 of the License, or (at your option) any later version. |
93
|
|
|
|
|
|
|
|
94
|
|
|
|
|
|
|
This program is distributed in the hope that it will be useful, |
95
|
|
|
|
|
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of |
96
|
|
|
|
|
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. |
97
|
|
|
|
|
|
|
|
98
|
|
|
|
|
|
|
=cut |
99
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
1; |
101
|
|
|
|
|
|
|
|