1 |
package Grep::Search; |
2 |
|
3 |
use strict; |
4 |
use warnings; |
5 |
|
6 |
use Data::Dump qw/dump/; |
7 |
use Lucene; |
8 |
use Jifty::Util; |
9 |
|
10 |
my $index_path = Jifty::Util->app_root . '/var/lucene'; |
11 |
|
12 |
my ( $analyzer, $store, $writer ); |
13 |
|
14 |
my $debug = 1; |
15 |
|
16 |
sub create { |
17 |
|
18 |
my $create = 0; |
19 |
if (! -e "$index_path/segments") { |
20 |
$create = 1; |
21 |
Jifty->log->debug("create index $index_path") unless ($store); |
22 |
} else { |
23 |
Jifty->log->debug("open index: $index_path") unless ($store); |
24 |
} |
25 |
return $create; |
26 |
} |
27 |
|
28 |
sub analyzer { |
29 |
my $self = shift; |
30 |
$analyzer ||= new Lucene::Analysis::Standard::StandardAnalyzer(); |
31 |
return $analyzer; |
32 |
} |
33 |
|
34 |
sub store { |
35 |
my $self = shift; |
36 |
|
37 |
$store ||= Lucene::Store::FSDirectory->getDirectory( $index_path, $self->create ); |
38 |
return $store; |
39 |
} |
40 |
|
41 |
sub writer { |
42 |
my $self = shift; |
43 |
$writer ||= new Lucene::Index::IndexWriter( $self->store, $self->analyzer, $self->create ); |
44 |
return $writer; |
45 |
} |
46 |
|
47 |
=head2 add |
48 |
|
49 |
Grep::Search->add( $record ); |
50 |
|
51 |
=cut |
52 |
|
53 |
sub add { |
54 |
my $self = shift; |
55 |
|
56 |
my $i = shift or die "no record to add"; |
57 |
|
58 |
die "record not Jifty::Record but ", ref $i unless ($i->isa('Jifty::Record')); |
59 |
|
60 |
my $pk = { $i->primary_keys }; |
61 |
|
62 |
my $doc = new Lucene::Document; |
63 |
|
64 |
my @columns = map { $_->name } $i->columns; |
65 |
|
66 |
foreach my $c ( @columns ) { |
67 |
|
68 |
my $v = $i->$c; |
69 |
|
70 |
if ( ref($v) ne '' ) { |
71 |
|
72 |
foreach my $f_c ( qw/id name title/ ) { |
73 |
if ( $i->$c->can( $f_c ) ) { |
74 |
my $f_v = $i->$c->$f_c || $i->$c->{values}->{ $f_c }; |
75 |
my $col = $c . '_' . $f_c; |
76 |
if ( $f_v ) { |
77 |
warn " # $col = $f_v\n" if ($debug); |
78 |
$doc->add(Lucene::Document::Field->Text( $col, $f_v )); |
79 |
} else { |
80 |
warn " . $col is NULL\n" if ($debug); |
81 |
} |
82 |
} |
83 |
} |
84 |
|
85 |
if ($v->isa('Jifty::DateTime')) { |
86 |
warn " d $c = $v\n" if ($debug); |
87 |
$doc->add(Lucene::Document::Field->Keyword( $c, "$v" )); |
88 |
} else { |
89 |
warn " s $c = $v [",ref($v),"]\n" if ($debug); |
90 |
} |
91 |
next; |
92 |
} |
93 |
|
94 |
next if (! defined($v) || $v eq ''); |
95 |
|
96 |
$v =~ s/<[^>]+>/ /gs; |
97 |
|
98 |
if ( defined( $pk->{$c} ) ) { |
99 |
$doc->add(Lucene::Document::Field->Keyword( $c, $v )); |
100 |
warn " * $c = $v\n" if ($debug); |
101 |
} else { |
102 |
$doc->add(Lucene::Document::Field->Text( $c, $v )); |
103 |
warn " + $c = ", $self->snippet( 50, $v ), "\n" if ($debug); |
104 |
} |
105 |
} |
106 |
|
107 |
$self->writer->addDocument($doc); |
108 |
|
109 |
Jifty->log->debug("added ", $i->id, " to index"); |
110 |
} |
111 |
|
112 |
=head2 |
113 |
|
114 |
my $ItemCollection = Grep::Search->collection( 'search query' ); |
115 |
|
116 |
=cut |
117 |
|
118 |
sub collection { |
119 |
my $self = shift; |
120 |
|
121 |
my $q = shift or die "no q?"; |
122 |
|
123 |
my $searcher = new Lucene::Search::IndexSearcher($self->store); |
124 |
my $parser = new Lucene::QueryParser("content", $self->analyzer); |
125 |
my $query = $parser->parse( $q ); |
126 |
|
127 |
Jifty->log->debug("searching for '$q'"); |
128 |
|
129 |
my $hits = $searcher->search($query); |
130 |
my $num_hits = $hits->length(); |
131 |
|
132 |
Jifty->log->debug("found $num_hits results"); |
133 |
|
134 |
my $collection = Grep::Model::ItemCollection->new(); |
135 |
|
136 |
my @results; |
137 |
|
138 |
for ( my $i = 0; $i < $num_hits; $i++ ) { |
139 |
|
140 |
my $doc = $hits->doc( $i ); |
141 |
|
142 |
my $score = $hits->score($i); |
143 |
my $title = $doc->get("title"); |
144 |
my $id = $doc->get("id"); |
145 |
|
146 |
warn "## $i $score $title\n"; |
147 |
|
148 |
my $item = Grep::Model::Item->new(); |
149 |
my ($ok,$msg) = $item->load_by_cols( id => $id ); |
150 |
|
151 |
if ( $ok ) { |
152 |
$collection->add_record( $item ); |
153 |
} else { |
154 |
warn "can't load item $id\n"; |
155 |
} |
156 |
|
157 |
} |
158 |
|
159 |
undef $hits; |
160 |
undef $query; |
161 |
undef $parser; |
162 |
undef $searcher; |
163 |
|
164 |
return $collection; |
165 |
} |
166 |
|
167 |
=head2 finish |
168 |
|
169 |
Grep::Search->finish |
170 |
|
171 |
=cut |
172 |
|
173 |
sub finish { |
174 |
my $self = shift; |
175 |
if ($writer) { |
176 |
warn "closing index\n"; |
177 |
$writer->close; |
178 |
} |
179 |
undef $writer; |
180 |
} |
181 |
|
182 |
sub _signal { |
183 |
my $s = shift; |
184 |
warn "catched SIG $s\n"; |
185 |
finish(); |
186 |
exit(0); |
187 |
} |
188 |
|
189 |
$SIG{'__DIE__'} = \&_signal; |
190 |
$SIG{'INT'} = \&_signal; |
191 |
$SIG{'QUIT'} = \&_signal; |
192 |
|
193 |
|
194 |
=head2 snippet |
195 |
|
196 |
my $short = $self->snippet( 50, $text ); |
197 |
|
198 |
|
199 |
=cut |
200 |
|
201 |
sub snippet { |
202 |
my $self = shift; |
203 |
|
204 |
my $len = shift or die "no len?"; |
205 |
my $m = join(" ", @_); |
206 |
|
207 |
$m =~ s/\s+/ /gs; |
208 |
|
209 |
if (length($m) > $len) { |
210 |
return substr($m,0,$len) . '...'; |
211 |
} else { |
212 |
return $m; |
213 |
} |
214 |
} |
215 |
|
216 |
1; |