line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
1
|
|
|
|
|
|
|
package ElasticSearch; |
2
|
|
|
|
|
|
|
|
3
|
1
|
|
|
1
|
|
33123
|
use strict; |
|
1
|
|
|
|
|
3
|
|
|
1
|
|
|
|
|
34
|
|
4
|
1
|
|
|
1
|
|
6
|
use warnings FATAL => 'all'; |
|
1
|
|
|
|
|
2
|
|
|
1
|
|
|
|
|
41
|
|
5
|
1
|
|
|
1
|
|
803
|
use ElasticSearch::Transport(); |
|
1
|
|
|
|
|
3
|
|
|
1
|
|
|
|
|
22
|
|
6
|
1
|
|
|
1
|
|
6
|
use ElasticSearch::Error(); |
|
1
|
|
|
|
|
2
|
|
|
1
|
|
|
|
|
15
|
|
7
|
1
|
|
|
1
|
|
1142
|
use ElasticSearch::RequestParser; |
|
1
|
|
|
|
|
4
|
|
|
1
|
|
|
|
|
47
|
|
8
|
1
|
|
|
1
|
|
13
|
use ElasticSearch::Util qw(throw parse_params); |
|
1
|
|
|
|
|
3
|
|
|
1
|
|
|
|
|
2159
|
|
9
|
|
|
|
|
|
|
|
10
|
|
|
|
|
|
|
our $VERSION = '0.68'; |
11
|
|
|
|
|
|
|
our $DEBUG = 0; |
12
|
|
|
|
|
|
|
|
13
|
|
|
|
|
|
|
#=================================== |
14
|
|
|
|
|
|
|
sub new { |
15
|
|
|
|
|
|
|
#=================================== |
16
|
0
|
|
|
0
|
1
|
|
my ( $proto, $params ) = parse_params(@_); |
17
|
0
|
|
|
|
|
|
my $self = { |
18
|
|
|
|
|
|
|
_base_qs => {}, |
19
|
|
|
|
|
|
|
_default => {}, |
20
|
|
|
|
|
|
|
_builder_class => 'ElasticSearch::SearchBuilder' |
21
|
|
|
|
|
|
|
}; |
22
|
|
|
|
|
|
|
|
23
|
0
|
|
0
|
|
|
|
bless $self, ref $proto || $proto; |
24
|
0
|
|
|
|
|
|
$self->{_transport} = ElasticSearch::Transport->new($params); |
25
|
0
|
|
|
|
|
|
$self->$_( $params->{$_} ) for keys %$params; |
26
|
0
|
|
|
|
|
|
return $self; |
27
|
|
|
|
|
|
|
} |
28
|
|
|
|
|
|
|
|
29
|
|
|
|
|
|
|
#=================================== |
30
|
|
|
|
|
|
|
sub builder_class { |
31
|
|
|
|
|
|
|
#=================================== |
32
|
0
|
|
|
0
|
1
|
|
my $self = shift; |
33
|
0
|
0
|
|
|
|
|
if (@_) { |
34
|
0
|
|
|
|
|
|
$self->{_builder_class} = shift; |
35
|
0
|
|
|
|
|
|
delete $self->{_builder}; |
36
|
|
|
|
|
|
|
} |
37
|
0
|
|
|
|
|
|
return $self->{_builder_class}; |
38
|
|
|
|
|
|
|
} |
39
|
|
|
|
|
|
|
|
40
|
|
|
|
|
|
|
#=================================== |
41
|
|
|
|
|
|
|
sub builder { |
42
|
|
|
|
|
|
|
#=================================== |
43
|
0
|
|
|
0
|
1
|
|
my $self = shift; |
44
|
0
|
0
|
|
|
|
|
unless ( $self->{_builder} ) { |
45
|
0
|
0
|
|
|
|
|
my $class = $self->{_builder_class} |
46
|
|
|
|
|
|
|
or $self->throw( 'Param', "No builder_class specified" ); |
47
|
0
|
0
|
0
|
|
|
|
eval "require $class; 1" |
48
|
|
|
|
|
|
|
or $self->throw( 'Internal', |
49
|
|
|
|
|
|
|
"Couldn't load class $class: " . ( $@ || 'Unknown error' ) ); |
50
|
0
|
|
|
|
|
|
$self->{_builder} = $class->new(@_); |
51
|
|
|
|
|
|
|
} |
52
|
0
|
|
|
|
|
|
return $self->{_builder}; |
53
|
|
|
|
|
|
|
} |
54
|
|
|
|
|
|
|
|
55
|
|
|
|
|
|
|
#=================================== |
56
|
|
|
|
|
|
|
sub request { |
57
|
|
|
|
|
|
|
#=================================== |
58
|
0
|
|
|
0
|
0
|
|
my ( $self, $params ) = parse_params(@_); |
59
|
0
|
|
|
|
|
|
return $self->transport->request($params); |
60
|
|
|
|
|
|
|
} |
61
|
|
|
|
|
|
|
|
62
|
|
|
|
|
|
|
#=================================== |
63
|
|
|
|
|
|
|
sub use_index { |
64
|
|
|
|
|
|
|
#=================================== |
65
|
0
|
|
|
0
|
1
|
|
my $self = shift; |
66
|
0
|
0
|
|
|
|
|
if (@_) { |
67
|
0
|
|
|
|
|
|
$self->{_default}{index} = shift; |
68
|
|
|
|
|
|
|
} |
69
|
0
|
|
|
|
|
|
return $self->{_default}{index}; |
70
|
|
|
|
|
|
|
} |
71
|
|
|
|
|
|
|
|
72
|
|
|
|
|
|
|
#=================================== |
73
|
|
|
|
|
|
|
sub use_type { |
74
|
|
|
|
|
|
|
#=================================== |
75
|
0
|
|
|
0
|
1
|
|
my $self = shift; |
76
|
0
|
0
|
|
|
|
|
if (@_) { |
77
|
0
|
|
|
|
|
|
$self->{_default}{type} = shift; |
78
|
|
|
|
|
|
|
} |
79
|
0
|
|
|
|
|
|
return $self->{_default}{type}; |
80
|
|
|
|
|
|
|
} |
81
|
|
|
|
|
|
|
|
82
|
|
|
|
|
|
|
#=================================== |
83
|
|
|
|
|
|
|
sub reindex { |
84
|
|
|
|
|
|
|
#=================================== |
85
|
0
|
|
|
0
|
1
|
|
my ( $self, $params ) = parse_params(@_); |
86
|
|
|
|
|
|
|
|
87
|
0
|
0
|
|
|
|
|
my $source = $params->{source} |
88
|
|
|
|
|
|
|
or $self->throw( 'Param', 'Missing source param' ); |
89
|
|
|
|
|
|
|
|
90
|
0
|
|
0
|
0
|
|
|
my $transform = $params->{transform} || sub { shift() }; |
|
0
|
|
|
|
|
|
|
91
|
0
|
|
|
|
|
|
my $verbose = !$params->{quiet}; |
92
|
0
|
|
|
|
|
|
my $dest_index = $params->{dest_index}; |
93
|
0
|
|
0
|
|
|
|
my $bulk_size = $params->{bulk_size} || 1000; |
94
|
0
|
|
0
|
|
|
|
my $method = $params->{_method_name} || 'next'; |
95
|
|
|
|
|
|
|
|
96
|
0
|
|
|
|
|
|
local $| = $verbose; |
97
|
0
|
0
|
|
|
|
|
printf( "Reindexing %d docs\n", $source->total ) |
98
|
|
|
|
|
|
|
if $verbose; |
99
|
|
|
|
|
|
|
|
100
|
0
|
|
|
|
|
|
my @docs; |
101
|
0
|
|
|
|
|
|
while (1) { |
102
|
0
|
|
|
|
|
|
my $doc = $source->$method(); |
103
|
0
|
0
|
0
|
|
|
|
if ( !$doc or @docs == $bulk_size ) { |
104
|
0
|
|
|
|
|
|
my $results = $self->bulk_index( |
105
|
|
|
|
|
|
|
docs => \@docs, |
106
|
0
|
|
|
|
|
|
map { $_ => $params->{$_} } qw(on_conflict on_error), |
107
|
|
|
|
|
|
|
); |
108
|
0
|
0
|
0
|
|
|
|
$results = $results->recv |
109
|
|
|
|
|
|
|
if ref $results ne 'HASH' |
110
|
|
|
|
|
|
|
&& $results->isa('AnyEvent::CondVar'); |
111
|
0
|
0
|
|
|
|
|
if ( my $err = $results->{errors} ) { |
112
|
0
|
|
|
|
|
|
my @errors = splice @$err, 0, 5; |
113
|
0
|
0
|
|
|
|
|
push @errors, sprintf "...and %d more", scalar @$err |
114
|
|
|
|
|
|
|
if @$err; |
115
|
0
|
|
|
|
|
|
$self->throw( 'Request', "Errors occurred while reindexing:", |
116
|
|
|
|
|
|
|
\@errors ); |
117
|
|
|
|
|
|
|
} |
118
|
0
|
|
|
|
|
|
@docs = (); |
119
|
0
|
0
|
|
|
|
|
print "." if $verbose; |
120
|
|
|
|
|
|
|
} |
121
|
0
|
0
|
|
|
|
|
last unless $doc; |
122
|
|
|
|
|
|
|
|
123
|
0
|
0
|
|
|
|
|
$doc = $transform->($doc) or next; |
124
|
0
|
0
|
|
|
|
|
$doc->{version_type} = 'external' |
125
|
|
|
|
|
|
|
if defined $doc->{_version}; |
126
|
0
|
0
|
|
|
|
|
if ( my $fields = delete $doc->{fields} ) { |
127
|
0
|
0
|
|
|
|
|
$doc->{parent} = $fields->{_parent} |
128
|
|
|
|
|
|
|
if defined $fields->{_parent}; |
129
|
|
|
|
|
|
|
} |
130
|
0
|
0
|
|
|
|
|
$doc->{_index} = $dest_index |
131
|
|
|
|
|
|
|
if $dest_index; |
132
|
0
|
|
|
|
|
|
push @docs, $doc; |
133
|
|
|
|
|
|
|
} |
134
|
|
|
|
|
|
|
|
135
|
0
|
0
|
|
|
|
|
print "\nDone\n" if $verbose; |
136
|
|
|
|
|
|
|
} |
137
|
|
|
|
|
|
|
|
138
|
|
|
|
|
|
|
#=================================== |
139
|
0
|
|
|
0
|
1
|
|
sub transport { shift()->{_transport} } |
140
|
0
|
|
|
0
|
1
|
|
sub trace_calls { shift->transport->trace_calls(@_) } |
141
|
0
|
|
|
0
|
1
|
|
sub timeout { shift->transport->timeout(@_) } |
142
|
0
|
|
|
0
|
1
|
|
sub refresh_servers { shift->transport->refresh_servers(@_) } |
143
|
|
|
|
|
|
|
#=================================== |
144
|
|
|
|
|
|
|
|
145
|
|
|
|
|
|
|
#=================================== |
146
|
|
|
|
|
|
|
sub query_parser { |
147
|
|
|
|
|
|
|
#=================================== |
148
|
0
|
|
|
0
|
1
|
|
require ElasticSearch::QueryParser; |
149
|
0
|
|
|
|
|
|
shift; # drop class/$self |
150
|
0
|
|
|
|
|
|
ElasticSearch::QueryParser->new(@_); |
151
|
|
|
|
|
|
|
} |
152
|
|
|
|
|
|
|
|
153
|
|
|
|
|
|
|
=head1 NAME |
154
|
|
|
|
|
|
|
|
155
|
|
|
|
|
|
|
ElasticSearch - DEPRECATED: An API for communicating with ElasticSearch |
156
|
|
|
|
|
|
|
|
157
|
|
|
|
|
|
|
=head1 VERSION |
158
|
|
|
|
|
|
|
|
159
|
|
|
|
|
|
|
Version 0.68, tested against ElasticSearch server version 0.90.0. |
160
|
|
|
|
|
|
|
|
161
|
|
|
|
|
|
|
=head1 DEPRECATION |
162
|
|
|
|
|
|
|
|
163
|
|
|
|
|
|
|
This module is being deprecated in favour of the new official client |
164
|
|
|
|
|
|
|
L |
165
|
|
|
|
|
|
|
and will be removed from CPAN in 2015. |
166
|
|
|
|
|
|
|
|
167
|
|
|
|
|
|
|
=head1 DESCRIPTION |
168
|
|
|
|
|
|
|
|
169
|
|
|
|
|
|
|
ElasticSearch is an Open Source (Apache 2 license), distributed, RESTful |
170
|
|
|
|
|
|
|
Search Engine based on Lucene, and built for the cloud, with a JSON API. |
171
|
|
|
|
|
|
|
|
172
|
|
|
|
|
|
|
Check out its features: L |
173
|
|
|
|
|
|
|
|
174
|
|
|
|
|
|
|
This module is a thin API which makes it easy to communicate with an |
175
|
|
|
|
|
|
|
ElasticSearch cluster. |
176
|
|
|
|
|
|
|
|
177
|
|
|
|
|
|
|
It maintains a list of all servers/nodes in the ElasticSearch cluster, and |
178
|
|
|
|
|
|
|
spreads the load across these nodes in round-robin fashion. |
179
|
|
|
|
|
|
|
If the current active node disappears, then it attempts to connect to another |
180
|
|
|
|
|
|
|
node in the list. |
181
|
|
|
|
|
|
|
|
182
|
|
|
|
|
|
|
Forking a process triggers a server list refresh, and a new connection to |
183
|
|
|
|
|
|
|
a randomly chosen node in the list. |
184
|
|
|
|
|
|
|
|
185
|
|
|
|
|
|
|
=cut |
186
|
|
|
|
|
|
|
|
187
|
|
|
|
|
|
|
=head1 SYNOPSIS |
188
|
|
|
|
|
|
|
|
189
|
|
|
|
|
|
|
|
190
|
|
|
|
|
|
|
use ElasticSearch; |
191
|
|
|
|
|
|
|
my $es = ElasticSearch->new( |
192
|
|
|
|
|
|
|
servers => 'search.foo.com:9200', # default '127.0.0.1:9200' |
193
|
|
|
|
|
|
|
transport => 'http' # default 'http' |
194
|
|
|
|
|
|
|
| 'httplite' |
195
|
|
|
|
|
|
|
| 'httptiny' |
196
|
|
|
|
|
|
|
| 'curl' |
197
|
|
|
|
|
|
|
| 'aehttp' |
198
|
|
|
|
|
|
|
| 'aecurl' |
199
|
|
|
|
|
|
|
| 'thrift', |
200
|
|
|
|
|
|
|
max_requests => 10_000, # default 10_000 |
201
|
|
|
|
|
|
|
trace_calls => 'log_file', |
202
|
|
|
|
|
|
|
no_refresh => 0 | 1, |
203
|
|
|
|
|
|
|
); |
204
|
|
|
|
|
|
|
|
205
|
|
|
|
|
|
|
$es->index( |
206
|
|
|
|
|
|
|
index => 'twitter', |
207
|
|
|
|
|
|
|
type => 'tweet', |
208
|
|
|
|
|
|
|
id => 1, |
209
|
|
|
|
|
|
|
data => { |
210
|
|
|
|
|
|
|
user => 'kimchy', |
211
|
|
|
|
|
|
|
post_date => '2009-11-15T14:12:12', |
212
|
|
|
|
|
|
|
message => 'trying out Elastic Search' |
213
|
|
|
|
|
|
|
} |
214
|
|
|
|
|
|
|
); |
215
|
|
|
|
|
|
|
|
216
|
|
|
|
|
|
|
$data = $es->get( |
217
|
|
|
|
|
|
|
index => 'twitter', |
218
|
|
|
|
|
|
|
type => 'tweet', |
219
|
|
|
|
|
|
|
id => 1 |
220
|
|
|
|
|
|
|
); |
221
|
|
|
|
|
|
|
|
222
|
|
|
|
|
|
|
# native elasticsearch query language |
223
|
|
|
|
|
|
|
$results = $es->search( |
224
|
|
|
|
|
|
|
index => 'twitter', |
225
|
|
|
|
|
|
|
type => 'tweet', |
226
|
|
|
|
|
|
|
query => { |
227
|
|
|
|
|
|
|
text => { user => 'kimchy' } |
228
|
|
|
|
|
|
|
} |
229
|
|
|
|
|
|
|
); |
230
|
|
|
|
|
|
|
|
231
|
|
|
|
|
|
|
# ElasticSearch::SearchBuilder Perlish query language |
232
|
|
|
|
|
|
|
$results = $es->search( |
233
|
|
|
|
|
|
|
index => 'twitter', |
234
|
|
|
|
|
|
|
type => 'tweet', |
235
|
|
|
|
|
|
|
queryb => { |
236
|
|
|
|
|
|
|
message => 'Perl API', |
237
|
|
|
|
|
|
|
user => 'kimchy', |
238
|
|
|
|
|
|
|
post_date => { |
239
|
|
|
|
|
|
|
'>' => '2010-01-01', |
240
|
|
|
|
|
|
|
'<=' => '2011-01-01', |
241
|
|
|
|
|
|
|
} |
242
|
|
|
|
|
|
|
} |
243
|
|
|
|
|
|
|
); |
244
|
|
|
|
|
|
|
|
245
|
|
|
|
|
|
|
|
246
|
|
|
|
|
|
|
$dodgy_qs = "foo AND AND bar"; |
247
|
|
|
|
|
|
|
$results = $es->search( |
248
|
|
|
|
|
|
|
index => 'twitter', |
249
|
|
|
|
|
|
|
type => 'tweet', |
250
|
|
|
|
|
|
|
query => { |
251
|
|
|
|
|
|
|
query_string => { |
252
|
|
|
|
|
|
|
query => $es->query_parser->filter($dodgy_qs) |
253
|
|
|
|
|
|
|
}, |
254
|
|
|
|
|
|
|
} |
255
|
|
|
|
|
|
|
); |
256
|
|
|
|
|
|
|
|
257
|
|
|
|
|
|
|
See the C directory for a simple working example. |
258
|
|
|
|
|
|
|
|
259
|
|
|
|
|
|
|
=cut |
260
|
|
|
|
|
|
|
|
261
|
|
|
|
|
|
|
=head1 GETTING ElasticSearch |
262
|
|
|
|
|
|
|
|
263
|
|
|
|
|
|
|
You can download the latest released version of ElasticSearch from |
264
|
|
|
|
|
|
|
L. |
265
|
|
|
|
|
|
|
|
266
|
|
|
|
|
|
|
See here for setup instructions: |
267
|
|
|
|
|
|
|
L |
268
|
|
|
|
|
|
|
|
269
|
|
|
|
|
|
|
=cut |
270
|
|
|
|
|
|
|
|
271
|
|
|
|
|
|
|
=head1 CALLING CONVENTIONS |
272
|
|
|
|
|
|
|
|
273
|
|
|
|
|
|
|
I've tried to follow the same terminology as used in the ElasticSearch docs |
274
|
|
|
|
|
|
|
when naming methods, so it should be easy to tie the two together. |
275
|
|
|
|
|
|
|
|
276
|
|
|
|
|
|
|
Some methods require a specific C and a specific C, while others |
277
|
|
|
|
|
|
|
allow a list of indices or types, or allow you to specify all indices or |
278
|
|
|
|
|
|
|
types. I distinguish between them as follows: |
279
|
|
|
|
|
|
|
|
280
|
|
|
|
|
|
|
$es->method( index => multi, type => single, ...) |
281
|
|
|
|
|
|
|
|
282
|
|
|
|
|
|
|
C values must be a scalar, and are required parameters |
283
|
|
|
|
|
|
|
|
284
|
|
|
|
|
|
|
type => 'tweet' |
285
|
|
|
|
|
|
|
|
286
|
|
|
|
|
|
|
C values can be: |
287
|
|
|
|
|
|
|
|
288
|
|
|
|
|
|
|
index => 'twitter' # specific index |
289
|
|
|
|
|
|
|
index => ['twitter','user'] # list of indices |
290
|
|
|
|
|
|
|
index => undef # (or not specified) = all indices |
291
|
|
|
|
|
|
|
|
292
|
|
|
|
|
|
|
C values work like C values, but at least one value is |
293
|
|
|
|
|
|
|
required, so: |
294
|
|
|
|
|
|
|
|
295
|
|
|
|
|
|
|
index => 'twitter' # specific index |
296
|
|
|
|
|
|
|
index => ['twitter','user'] # list of indices |
297
|
|
|
|
|
|
|
index => '_all' # all indices |
298
|
|
|
|
|
|
|
|
299
|
|
|
|
|
|
|
index => [] # error |
300
|
|
|
|
|
|
|
index => undef # error |
301
|
|
|
|
|
|
|
|
302
|
|
|
|
|
|
|
|
303
|
|
|
|
|
|
|
Also, see L"use_index()/use_type()">. |
304
|
|
|
|
|
|
|
|
305
|
|
|
|
|
|
|
=head2 as_json |
306
|
|
|
|
|
|
|
|
307
|
|
|
|
|
|
|
If you pass C<< as_json => 1 >> to any request to the ElasticSearch server, |
308
|
|
|
|
|
|
|
it will return the raw UTF8-decoded JSON response, rather than a Perl |
309
|
|
|
|
|
|
|
datastructure. |
310
|
|
|
|
|
|
|
|
311
|
|
|
|
|
|
|
=cut |
312
|
|
|
|
|
|
|
|
313
|
|
|
|
|
|
|
=head1 RETURN VALUES AND EXCEPTIONS |
314
|
|
|
|
|
|
|
|
315
|
|
|
|
|
|
|
Methods that query the ElasticSearch cluster return the raw data structure |
316
|
|
|
|
|
|
|
that the cluster returns. This may change in the future, but as these |
317
|
|
|
|
|
|
|
data structures are still in flux, I thought it safer not to try to interpret. |
318
|
|
|
|
|
|
|
|
319
|
|
|
|
|
|
|
Anything that is known to be an error throws an exception, eg trying to delete |
320
|
|
|
|
|
|
|
a non-existent index. |
321
|
|
|
|
|
|
|
|
322
|
|
|
|
|
|
|
=cut |
323
|
|
|
|
|
|
|
|
324
|
|
|
|
|
|
|
=head1 INTEGRATION WITH ElasticSearch::SearchBuilder |
325
|
|
|
|
|
|
|
|
326
|
|
|
|
|
|
|
L provides a concise Perlish |
327
|
|
|
|
|
|
|
L-style query language, which gets translated into the native |
328
|
|
|
|
|
|
|
L that |
329
|
|
|
|
|
|
|
ElasticSearch uses. |
330
|
|
|
|
|
|
|
|
331
|
|
|
|
|
|
|
For instance: |
332
|
|
|
|
|
|
|
|
333
|
|
|
|
|
|
|
{ |
334
|
|
|
|
|
|
|
content => 'search keywords', |
335
|
|
|
|
|
|
|
-filter => { |
336
|
|
|
|
|
|
|
tags => ['perl','ruby'], |
337
|
|
|
|
|
|
|
date => { |
338
|
|
|
|
|
|
|
'>' => '2010-01-01', |
339
|
|
|
|
|
|
|
'<=' => '2011-01-01' |
340
|
|
|
|
|
|
|
}, |
341
|
|
|
|
|
|
|
} |
342
|
|
|
|
|
|
|
} |
343
|
|
|
|
|
|
|
|
344
|
|
|
|
|
|
|
Would be translated to: |
345
|
|
|
|
|
|
|
|
346
|
|
|
|
|
|
|
{ query => { |
347
|
|
|
|
|
|
|
filtered => { |
348
|
|
|
|
|
|
|
query => { text => { content => "search keywords" } }, |
349
|
|
|
|
|
|
|
filter => { |
350
|
|
|
|
|
|
|
and => [ |
351
|
|
|
|
|
|
|
{ terms => { tags => ["perl", "ruby"] } }, |
352
|
|
|
|
|
|
|
{ numeric_range => { |
353
|
|
|
|
|
|
|
date => { |
354
|
|
|
|
|
|
|
gt => "2010-01-01", |
355
|
|
|
|
|
|
|
lte => "2011-01-01" |
356
|
|
|
|
|
|
|
}}}, |
357
|
|
|
|
|
|
|
], |
358
|
|
|
|
|
|
|
} |
359
|
|
|
|
|
|
|
}}} |
360
|
|
|
|
|
|
|
|
361
|
|
|
|
|
|
|
All you have to do to start using L is to change |
362
|
|
|
|
|
|
|
your C or C parameter to C or C (where the |
363
|
|
|
|
|
|
|
extra C stands for C): |
364
|
|
|
|
|
|
|
|
365
|
|
|
|
|
|
|
$es->search( |
366
|
|
|
|
|
|
|
queryb => { content => 'keywords' } |
367
|
|
|
|
|
|
|
) |
368
|
|
|
|
|
|
|
|
369
|
|
|
|
|
|
|
If you want to see what your SearchBuilder-style query is being converted into, |
370
|
|
|
|
|
|
|
you can either use L"trace_calls()"> or access it directly with: |
371
|
|
|
|
|
|
|
|
372
|
|
|
|
|
|
|
$native_query = $es->builder->query( $query ) |
373
|
|
|
|
|
|
|
$native_filter = $es->builder->filter( $filter ) |
374
|
|
|
|
|
|
|
|
375
|
|
|
|
|
|
|
See the L docs for more information about |
376
|
|
|
|
|
|
|
the syntax. |
377
|
|
|
|
|
|
|
|
378
|
|
|
|
|
|
|
=head1 METHODS |
379
|
|
|
|
|
|
|
|
380
|
|
|
|
|
|
|
=head2 Creating a new ElasticSearch instance |
381
|
|
|
|
|
|
|
|
382
|
|
|
|
|
|
|
=head3 new() |
383
|
|
|
|
|
|
|
|
384
|
|
|
|
|
|
|
$es = ElasticSearch->new( |
385
|
|
|
|
|
|
|
transport => 'http', |
386
|
|
|
|
|
|
|
servers => '127.0.0.1:9200' # single server |
387
|
|
|
|
|
|
|
| ['es1.foo.com:9200', |
388
|
|
|
|
|
|
|
'es2.foo.com:9200'], # multiple servers |
389
|
|
|
|
|
|
|
trace_calls => 1 | '/path/to/log/file' | $fh |
390
|
|
|
|
|
|
|
timeout => 30, |
391
|
|
|
|
|
|
|
max_requests => 10_000, # refresh server list |
392
|
|
|
|
|
|
|
# after max_requests |
393
|
|
|
|
|
|
|
|
394
|
|
|
|
|
|
|
no_refresh => 0 | 1 # don't retrieve the live |
395
|
|
|
|
|
|
|
# server list. Instead, use |
396
|
|
|
|
|
|
|
# just the servers specified |
397
|
|
|
|
|
|
|
); |
398
|
|
|
|
|
|
|
|
399
|
|
|
|
|
|
|
C can be either a single server or an ARRAY ref with a list of servers. |
400
|
|
|
|
|
|
|
If not specified, then it defaults to C and the port for the |
401
|
|
|
|
|
|
|
specified transport (eg C<9200> for C or C<9500> for C). |
402
|
|
|
|
|
|
|
|
403
|
|
|
|
|
|
|
These servers are used in a round-robin fashion. If any server fails to |
404
|
|
|
|
|
|
|
connect, then the other servers in the list are tried, and if any |
405
|
|
|
|
|
|
|
succeeds, then a list of all servers/nodes currently known to the |
406
|
|
|
|
|
|
|
ElasticSearch cluster are retrieved and stored. |
407
|
|
|
|
|
|
|
|
408
|
|
|
|
|
|
|
Every C (default 10,000) this list of known nodes is refreshed |
409
|
|
|
|
|
|
|
automatically. To disable this automatic refresh, you can set C |
410
|
|
|
|
|
|
|
to C<0>. |
411
|
|
|
|
|
|
|
|
412
|
|
|
|
|
|
|
To force a lookup of live nodes, you can do: |
413
|
|
|
|
|
|
|
|
414
|
|
|
|
|
|
|
$es->refresh_servers(); |
415
|
|
|
|
|
|
|
|
416
|
|
|
|
|
|
|
=head4 no_refresh() |
417
|
|
|
|
|
|
|
|
418
|
|
|
|
|
|
|
Regardless of the C setting, a list of live nodes will still be |
419
|
|
|
|
|
|
|
retrieved on the first request. This may not be desirable behaviour |
420
|
|
|
|
|
|
|
if, for instance, you are connecting to remote servers which use internal |
421
|
|
|
|
|
|
|
IP addresses, or which don't allow remote C requests. |
422
|
|
|
|
|
|
|
|
423
|
|
|
|
|
|
|
If you want to disable this behaviour completely, set C to C<1>, |
424
|
|
|
|
|
|
|
in which case the transport module will round robin through the |
425
|
|
|
|
|
|
|
C list only. Failed nodes will be removed from the list |
426
|
|
|
|
|
|
|
(but added back in every C or when all nodes have failed). |
427
|
|
|
|
|
|
|
|
428
|
|
|
|
|
|
|
=head4 Transport Backends |
429
|
|
|
|
|
|
|
|
430
|
|
|
|
|
|
|
There are various C backends that ElasticSearch can use: |
431
|
|
|
|
|
|
|
C (the default, based on LWP), C (based on L), |
432
|
|
|
|
|
|
|
C (based on L), C (based on L), |
433
|
|
|
|
|
|
|
C (based on L), C (based on |
434
|
|
|
|
|
|
|
L) and C (which uses the Thrift protocol). |
435
|
|
|
|
|
|
|
|
436
|
|
|
|
|
|
|
Although the C interface has the right buzzwords (binary, compact, |
437
|
|
|
|
|
|
|
sockets), the generated Perl code is very slow. Until that is improved, I |
438
|
|
|
|
|
|
|
recommend one of the C backends instead. |
439
|
|
|
|
|
|
|
|
440
|
|
|
|
|
|
|
The C backend is about 30% faster than the default C backend, |
441
|
|
|
|
|
|
|
and will probably become the default after more testing in production. |
442
|
|
|
|
|
|
|
|
443
|
|
|
|
|
|
|
The C backend is 1% faster again than C. |
444
|
|
|
|
|
|
|
|
445
|
|
|
|
|
|
|
See also: L, L"timeout()">, L"trace_calls()">, |
446
|
|
|
|
|
|
|
L |
447
|
|
|
|
|
|
|
and L |
448
|
|
|
|
|
|
|
|
449
|
|
|
|
|
|
|
=cut |
450
|
|
|
|
|
|
|
|
451
|
|
|
|
|
|
|
=head2 Document-indexing methods |
452
|
|
|
|
|
|
|
|
453
|
|
|
|
|
|
|
=head3 index() |
454
|
|
|
|
|
|
|
|
455
|
|
|
|
|
|
|
$result = $es->index( |
456
|
|
|
|
|
|
|
index => single, |
457
|
|
|
|
|
|
|
type => single, |
458
|
|
|
|
|
|
|
id => $document_id, # optional, otherwise auto-generated |
459
|
|
|
|
|
|
|
data => { |
460
|
|
|
|
|
|
|
key => value, |
461
|
|
|
|
|
|
|
... |
462
|
|
|
|
|
|
|
}, |
463
|
|
|
|
|
|
|
|
464
|
|
|
|
|
|
|
# optional |
465
|
|
|
|
|
|
|
consistency => 'quorum' | 'one' | 'all', |
466
|
|
|
|
|
|
|
create => 0 | 1, |
467
|
|
|
|
|
|
|
parent => $parent, |
468
|
|
|
|
|
|
|
percolate => $percolate, |
469
|
|
|
|
|
|
|
refresh => 0 | 1, |
470
|
|
|
|
|
|
|
replication => 'sync' | 'async', |
471
|
|
|
|
|
|
|
routing => $routing, |
472
|
|
|
|
|
|
|
timeout => eg '1m' or '10s' |
473
|
|
|
|
|
|
|
version => int, |
474
|
|
|
|
|
|
|
version_type => 'internal' | 'external', |
475
|
|
|
|
|
|
|
); |
476
|
|
|
|
|
|
|
|
477
|
|
|
|
|
|
|
eg: |
478
|
|
|
|
|
|
|
|
479
|
|
|
|
|
|
|
$result = $es->index( |
480
|
|
|
|
|
|
|
index => 'twitter', |
481
|
|
|
|
|
|
|
type => 'tweet', |
482
|
|
|
|
|
|
|
id => 1, |
483
|
|
|
|
|
|
|
data => { |
484
|
|
|
|
|
|
|
user => 'kimchy', |
485
|
|
|
|
|
|
|
post_date => '2009-11-15T14:12:12', |
486
|
|
|
|
|
|
|
message => 'trying out Elastic Search' |
487
|
|
|
|
|
|
|
}, |
488
|
|
|
|
|
|
|
); |
489
|
|
|
|
|
|
|
|
490
|
|
|
|
|
|
|
Used to add a document to a specific C as a specific C with |
491
|
|
|
|
|
|
|
a specific C. If the C combination already exists, |
492
|
|
|
|
|
|
|
then that document is updated, otherwise it is created. |
493
|
|
|
|
|
|
|
|
494
|
|
|
|
|
|
|
Note: |
495
|
|
|
|
|
|
|
|
496
|
|
|
|
|
|
|
=over |
497
|
|
|
|
|
|
|
|
498
|
|
|
|
|
|
|
=item * |
499
|
|
|
|
|
|
|
|
500
|
|
|
|
|
|
|
If the C is not specified, then ElasticSearch autogenerates a unique |
501
|
|
|
|
|
|
|
ID and a new document is always created. |
502
|
|
|
|
|
|
|
|
503
|
|
|
|
|
|
|
=item * |
504
|
|
|
|
|
|
|
|
505
|
|
|
|
|
|
|
If C is passed, and the current version in ElasticSearch is |
506
|
|
|
|
|
|
|
different, then a C error will be thrown. |
507
|
|
|
|
|
|
|
|
508
|
|
|
|
|
|
|
=item * |
509
|
|
|
|
|
|
|
|
510
|
|
|
|
|
|
|
C can also be a raw JSON encoded string (but ensure that it is correctly |
511
|
|
|
|
|
|
|
encoded, otherwise you see errors when trying to retrieve it from ElasticSearch). |
512
|
|
|
|
|
|
|
|
513
|
|
|
|
|
|
|
$es->index( |
514
|
|
|
|
|
|
|
index => 'foo', |
515
|
|
|
|
|
|
|
type => 'bar', |
516
|
|
|
|
|
|
|
id => 1, |
517
|
|
|
|
|
|
|
data => '{"foo":"bar"}' |
518
|
|
|
|
|
|
|
); |
519
|
|
|
|
|
|
|
|
520
|
|
|
|
|
|
|
=item * |
521
|
|
|
|
|
|
|
|
522
|
|
|
|
|
|
|
C for all CRUD methods and L"search()"> is a query timeout, |
523
|
|
|
|
|
|
|
specifying the amount of time ElasticSearch will spend (roughly) processing a |
524
|
|
|
|
|
|
|
query. Units can be concatenated with the integer value, e.g., C<500ms> or |
525
|
|
|
|
|
|
|
C<1s>. |
526
|
|
|
|
|
|
|
|
527
|
|
|
|
|
|
|
See also: L |
528
|
|
|
|
|
|
|
|
529
|
|
|
|
|
|
|
Note: this is distinct from the transport timeout, see L"timeout()">. |
530
|
|
|
|
|
|
|
|
531
|
|
|
|
|
|
|
=back |
532
|
|
|
|
|
|
|
|
533
|
|
|
|
|
|
|
See also: L, |
534
|
|
|
|
|
|
|
L"bulk()"> and L"put_mapping()"> |
535
|
|
|
|
|
|
|
|
536
|
|
|
|
|
|
|
=head3 set() |
537
|
|
|
|
|
|
|
|
538
|
|
|
|
|
|
|
C is a synonym for L"index()"> |
539
|
|
|
|
|
|
|
|
540
|
|
|
|
|
|
|
|
541
|
|
|
|
|
|
|
=head3 create() |
542
|
|
|
|
|
|
|
|
543
|
|
|
|
|
|
|
$result = $es->create( |
544
|
|
|
|
|
|
|
index => single, |
545
|
|
|
|
|
|
|
type => single, |
546
|
|
|
|
|
|
|
id => $document_id, # optional, otherwise auto-generated |
547
|
|
|
|
|
|
|
data => { |
548
|
|
|
|
|
|
|
key => value, |
549
|
|
|
|
|
|
|
... |
550
|
|
|
|
|
|
|
}, |
551
|
|
|
|
|
|
|
|
552
|
|
|
|
|
|
|
# optional |
553
|
|
|
|
|
|
|
consistency => 'quorum' | 'one' | 'all', |
554
|
|
|
|
|
|
|
parent => $parent, |
555
|
|
|
|
|
|
|
percolate => $percolate, |
556
|
|
|
|
|
|
|
refresh => 0 | 1, |
557
|
|
|
|
|
|
|
replication => 'sync' | 'async', |
558
|
|
|
|
|
|
|
routing => $routing, |
559
|
|
|
|
|
|
|
timeout => eg '1m' or '10s', |
560
|
|
|
|
|
|
|
version => int, |
561
|
|
|
|
|
|
|
version_type => 'internal' | 'external', |
562
|
|
|
|
|
|
|
); |
563
|
|
|
|
|
|
|
|
564
|
|
|
|
|
|
|
eg: |
565
|
|
|
|
|
|
|
|
566
|
|
|
|
|
|
|
$result = $es->create( |
567
|
|
|
|
|
|
|
index => 'twitter', |
568
|
|
|
|
|
|
|
type => 'tweet', |
569
|
|
|
|
|
|
|
id => 1, |
570
|
|
|
|
|
|
|
data => { |
571
|
|
|
|
|
|
|
user => 'kimchy', |
572
|
|
|
|
|
|
|
post_date => '2009-11-15T14:12:12', |
573
|
|
|
|
|
|
|
message => 'trying out Elastic Search' |
574
|
|
|
|
|
|
|
}, |
575
|
|
|
|
|
|
|
); |
576
|
|
|
|
|
|
|
|
577
|
|
|
|
|
|
|
Used to add a NEW document to a specific C as a specific C with |
578
|
|
|
|
|
|
|
a specific C. If the C combination already exists, |
579
|
|
|
|
|
|
|
then a C error is thrown. |
580
|
|
|
|
|
|
|
|
581
|
|
|
|
|
|
|
If the C is not specified, then ElasticSearch autogenerates a unique |
582
|
|
|
|
|
|
|
ID. |
583
|
|
|
|
|
|
|
|
584
|
|
|
|
|
|
|
If you pass a C parameter to C, then it must be C<0> unless |
585
|
|
|
|
|
|
|
you also set C to C. |
586
|
|
|
|
|
|
|
|
587
|
|
|
|
|
|
|
See also: L"index()"> |
588
|
|
|
|
|
|
|
|
589
|
|
|
|
|
|
|
=head3 update() |
590
|
|
|
|
|
|
|
|
591
|
|
|
|
|
|
|
$result = $es->update( |
592
|
|
|
|
|
|
|
index => single, |
593
|
|
|
|
|
|
|
type => single, |
594
|
|
|
|
|
|
|
id => single, |
595
|
|
|
|
|
|
|
|
596
|
|
|
|
|
|
|
# required |
597
|
|
|
|
|
|
|
script => $script, |
598
|
|
|
|
|
|
|
| doc => $doc |
599
|
|
|
|
|
|
|
|
600
|
|
|
|
|
|
|
# optional |
601
|
|
|
|
|
|
|
params => { params }, |
602
|
|
|
|
|
|
|
upsert => { new_doc }, |
603
|
|
|
|
|
|
|
consistency => 'quorum' | 'one' | 'all', |
604
|
|
|
|
|
|
|
fields => ['_source'], |
605
|
|
|
|
|
|
|
ignore_missing => 0 | 1, |
606
|
|
|
|
|
|
|
parent => $parent, |
607
|
|
|
|
|
|
|
percolate => $percolate, |
608
|
|
|
|
|
|
|
retry_on_conflict => 2, |
609
|
|
|
|
|
|
|
routing => $routing, |
610
|
|
|
|
|
|
|
timeout => '10s', |
611
|
|
|
|
|
|
|
replication => 'sync' | 'async' |
612
|
|
|
|
|
|
|
) |
613
|
|
|
|
|
|
|
|
614
|
|
|
|
|
|
|
The C method accepts a C |