1
|
|
|
<?php |
2
|
|
|
|
3
|
|
|
|
4
|
|
|
/** |
5
|
|
|
* Provides an interface for easily building a complex search query that |
6
|
|
|
* combines multiple ranking signals. |
7
|
|
|
* |
8
|
|
|
* |
9
|
|
|
* $bldr = new Jetpack_WPES_Query_Builder(); |
10
|
|
|
* $bldr->add_filter( ... ); |
11
|
|
|
* $bldr->add_filter( ... ); |
12
|
|
|
* $bldr->add_query( ... ); |
13
|
|
|
* $es_query = $bldr->build_query(); |
14
|
|
|
* |
15
|
|
|
* |
16
|
|
|
* All ES queries take a standard form with main query (with some filters), |
17
|
|
|
* wrapped in a function_score |
18
|
|
|
* |
19
|
|
|
* Bucketed queries use an aggregation to diversify results. eg a bunch |
20
|
|
|
* of separate filters where to get different sets of results. |
21
|
|
|
* |
22
|
|
|
*/ |
23
|
|
|
|
24
|
|
|
class Jetpack_WPES_Query_Builder { |
25
|
|
|
|
26
|
|
|
public $es_filters = array(); |
27
|
|
|
|
28
|
|
|
// Custom boosting with function_score |
29
|
|
|
public $functions = array(); |
30
|
|
|
public $decays = array(); |
31
|
|
|
public $scripts = array(); |
32
|
|
|
public $functions_max_boost = 2.0; |
33
|
|
|
public $functions_score_mode = 'multiply'; |
34
|
|
|
public $query_bool_boost = null; |
35
|
|
|
|
36
|
|
|
// General aggregations for buckets and metrics |
37
|
|
|
public $aggs_query = false; |
38
|
|
|
public $aggs = array(); |
39
|
|
|
|
40
|
|
|
// The set of top level text queries to combine |
41
|
|
|
public $must_queries = array(); |
42
|
|
|
public $should_queries = array(); |
43
|
|
|
public $dis_max_queries = array(); |
44
|
|
|
|
45
|
|
|
public $diverse_buckets_query = false; |
46
|
|
|
public $bucket_filters = array(); |
47
|
|
|
public $bucket_sub_aggs = array(); |
48
|
|
|
|
49
|
|
|
//////////////////////////////////// |
50
|
|
|
// Methods for building a query |
51
|
|
|
|
52
|
|
|
public function add_filter( $filter ) { |
53
|
|
|
$this->es_filters[] = $filter; |
54
|
|
|
} |
55
|
|
|
|
56
|
|
|
public function add_query( $query, $type = 'must' ) { |
57
|
|
|
switch ( $type ) { |
58
|
|
|
case 'dis_max': |
59
|
|
|
$this->dis_max_queries[] = $query; |
60
|
|
|
break; |
61
|
|
|
|
62
|
|
|
case 'should': |
63
|
|
|
$this->should_queries[] = $query; |
64
|
|
|
break; |
65
|
|
|
|
66
|
|
|
case 'must': |
67
|
|
|
default: |
68
|
|
|
$this->must_queries[] = $query; |
69
|
|
|
break; |
70
|
|
|
} |
71
|
|
|
} |
72
|
|
|
|
73
|
|
|
/** |
74
|
|
|
* Add a scoring function to the query |
75
|
|
|
* |
76
|
|
|
* NOTE: For decays (linear, exp, or gauss), use Jetpack_WPES_Query_Builder::add_decay() instead |
77
|
|
|
* |
78
|
|
|
* @see https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-function-score-query.html |
79
|
|
|
* |
80
|
|
|
* @param $function string name of the function |
81
|
|
|
* @param $params array functions parameters |
82
|
|
|
* |
83
|
|
|
* @return void |
84
|
|
|
*/ |
85
|
|
|
public function add_function( $function, $params ) { |
86
|
|
|
$this->functions[ $function ][] = $params; |
87
|
|
|
} |
88
|
|
|
|
89
|
|
|
/** |
90
|
|
|
* Add a decay function to score results |
91
|
|
|
* |
92
|
|
|
* This method should be used instead of Jetpack_WPES_Query_Builder::add_function() for decays, as the internal ES structure |
93
|
|
|
* is slightly different for them. |
94
|
|
|
* |
95
|
|
|
* @see https://www.elastic.co/guide/en/elasticsearch/guide/current/decay-functions.html |
96
|
|
|
* |
97
|
|
|
* @param $function string name of the decay function - linear, exp, or gauss |
98
|
|
|
* @param $params array The decay functions parameters, passed to ES directly |
99
|
|
|
* |
100
|
|
|
* @return void |
101
|
|
|
*/ |
102
|
|
|
public function add_decay( $function, $params ) { |
103
|
|
|
$this->decays[ $function ][] = $params; |
104
|
|
|
} |
105
|
|
|
|
106
|
|
|
/** |
107
|
|
|
* Add a scoring mode to the query |
108
|
|
|
* |
109
|
|
|
* @see https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-function-score-query.html |
110
|
|
|
* |
111
|
|
|
* @param $mode string name of how to score |
112
|
|
|
* |
113
|
|
|
* @return void |
114
|
|
|
*/ |
115
|
|
|
public function add_score_mode_to_functions( $mode='multiply' ) { |
116
|
|
|
$this->functions_score_mode = $mode; |
117
|
|
|
} |
118
|
|
|
|
119
|
|
|
public function add_max_boost_to_functions( $boost ) { |
120
|
|
|
$this->functions_max_boost = $boost; |
121
|
|
|
} |
122
|
|
|
|
123
|
|
|
public function add_boost_to_query_bool( $boost ) { |
124
|
|
|
$this->query_bool_boost = $boost; |
125
|
|
|
} |
126
|
|
|
|
127
|
|
|
public function add_aggs( $aggs_name, $aggs ) { |
128
|
|
|
$this->aggs_query = true; |
129
|
|
|
$this->aggs[$aggs_name] = $aggs; |
130
|
|
|
} |
131
|
|
|
|
132
|
|
|
public function add_aggs_sub_aggs( $aggs_name, $sub_aggs ) { |
133
|
|
|
if ( ! array_key_exists( 'aggs', $this->aggs[$aggs_name] ) ) { |
134
|
|
|
$this->aggs[$aggs_name]['aggs'] = array(); |
135
|
|
|
} |
136
|
|
|
$this->aggs[$aggs_name]['aggs'] = $sub_aggs; |
137
|
|
|
} |
138
|
|
|
|
139
|
|
|
public function add_bucketed_query( $name, $query ) { |
140
|
|
|
$this->_add_bucket_filter( $name, $query ); |
141
|
|
|
|
142
|
|
|
$this->add_query( $query, 'dis_max' ); |
143
|
|
|
} |
144
|
|
|
|
145
|
|
|
public function add_bucketed_terms( $name, $field, $terms, $boost = 1 ) { |
146
|
|
|
if ( ! is_array( $terms ) ) { |
147
|
|
|
$terms = array( $terms ); |
148
|
|
|
} |
149
|
|
|
|
150
|
|
|
$this->_add_bucket_filter( $name, array( |
151
|
|
|
'terms' => array( |
152
|
|
|
$field => $terms, |
153
|
|
|
), |
154
|
|
|
)); |
155
|
|
|
|
156
|
|
|
$this->add_query( array( |
157
|
|
|
'constant_score' => array( |
158
|
|
|
'filter' => array( |
159
|
|
|
'terms' => array( |
160
|
|
|
$field => $terms, |
161
|
|
|
), |
162
|
|
|
), |
163
|
|
|
'boost' => $boost, |
164
|
|
|
), |
165
|
|
|
), 'dis_max' ); |
166
|
|
|
} |
167
|
|
|
|
168
|
|
|
public function add_bucket_sub_aggs( $agg ) { |
169
|
|
|
$this->bucket_sub_aggs = array_merge( $this->bucket_sub_aggs, $agg ); |
170
|
|
|
} |
171
|
|
|
|
172
|
|
|
public function _add_bucket_filter( $name, $filter ) { |
173
|
|
|
$this->diverse_buckets_query = true; |
174
|
|
|
$this->bucket_filters[ $name ] = $filter; |
175
|
|
|
} |
176
|
|
|
|
177
|
|
|
//////////////////////////////////// |
178
|
|
|
// Building Final Query |
179
|
|
|
|
180
|
|
|
/** |
181
|
|
|
* Combine all the queries, functions, decays, scripts, and max_boost into an ES query |
182
|
|
|
* |
183
|
|
|
* @return array Array representation of the built ES query |
184
|
|
|
*/ |
185
|
|
|
public function build_query() { |
186
|
|
|
$query = array(); |
|
|
|
|
187
|
|
|
|
188
|
|
|
//dis_max queries just become a single must query |
189
|
|
|
if ( ! empty( $this->dis_max_queries ) ) { |
190
|
|
|
$this->must_queries[] = array( |
191
|
|
|
'dis_max' => array( |
192
|
|
|
'queries' => $this->dis_max_queries, |
193
|
|
|
), |
194
|
|
|
); |
195
|
|
|
} |
196
|
|
|
|
197
|
|
|
if ( empty( $this->must_queries ) ) { |
198
|
|
|
$this->must_queries = array( |
199
|
|
|
array( |
200
|
|
|
'match_all' => array(), |
201
|
|
|
), |
202
|
|
|
); |
203
|
|
|
} |
204
|
|
|
|
205
|
|
|
if ( empty( $this->should_queries ) ) { |
206
|
|
|
if ( 1 == count( $this->must_queries ) ) { |
207
|
|
|
$query = $this->must_queries[0]; |
208
|
|
|
} else { |
209
|
|
|
$query = array( |
210
|
|
|
'bool' => array( |
211
|
|
|
'must' => $this->must_queries, |
212
|
|
|
), |
213
|
|
|
); |
214
|
|
|
} |
215
|
|
|
} else { |
216
|
|
|
$query = array( |
217
|
|
|
'bool' => array( |
218
|
|
|
'must' => $this->must_queries, |
219
|
|
|
'should' => $this->should_queries, |
220
|
|
|
), |
221
|
|
|
); |
222
|
|
|
} |
223
|
|
|
|
224
|
|
|
if ( ! is_null( $this->query_bool_boost ) && isset( $query['bool'] ) ) { |
225
|
|
|
$query['bool']['boost'] = $this->query_bool_boost; |
226
|
|
|
} |
227
|
|
|
|
228
|
|
|
// If there are any function score adjustments, then combine those |
229
|
|
|
if ( $this->functions || $this->decays || $this->scripts ) { |
|
|
|
|
230
|
|
|
$weighting_functions = array(); |
231
|
|
|
|
232
|
|
|
if ( $this->functions ) { |
|
|
|
|
233
|
|
|
foreach ( $this->functions as $function_type => $configs ) { |
234
|
|
|
foreach ( $configs as $config ) { |
235
|
|
|
foreach ( $config as $field => $params ) { |
236
|
|
|
$func_arr = $params; |
237
|
|
|
|
238
|
|
|
$func_arr['field'] = $field; |
239
|
|
|
|
240
|
|
|
$weighting_functions[] = array( |
241
|
|
|
$function_type => $func_arr, |
242
|
|
|
); |
243
|
|
|
} |
244
|
|
|
} |
245
|
|
|
} |
246
|
|
|
} |
247
|
|
|
|
248
|
|
|
if ( $this->decays ) { |
|
|
|
|
249
|
|
|
foreach ( $this->decays as $decay_type => $configs ) { |
250
|
|
|
foreach ( $configs as $config ) { |
251
|
|
|
foreach ( $config as $field => $params ) { |
252
|
|
|
$weighting_functions[] = array( |
253
|
|
|
$decay_type => array( |
254
|
|
|
$field => $params, |
255
|
|
|
), |
256
|
|
|
); |
257
|
|
|
} |
258
|
|
|
} |
259
|
|
|
} |
260
|
|
|
} |
261
|
|
|
|
262
|
|
|
if ( $this->scripts ) { |
|
|
|
|
263
|
|
|
foreach ( $this->scripts as $script ) { |
264
|
|
|
$weighting_functions[] = array( |
265
|
|
|
'script_score' => array( |
266
|
|
|
'script' => $script, |
267
|
|
|
), |
268
|
|
|
); |
269
|
|
|
} |
270
|
|
|
} |
271
|
|
|
|
272
|
|
|
$query = array( |
273
|
|
|
'function_score' => array( |
274
|
|
|
'query' => $query, |
275
|
|
|
'functions' => $weighting_functions, |
276
|
|
|
'max_boost' => $this->functions_max_boost, |
277
|
|
|
'score_mode' => $this->functions_score_mode, |
278
|
|
|
), |
279
|
|
|
); |
280
|
|
|
} // End if(). |
281
|
|
|
|
282
|
|
|
return $query; |
283
|
|
|
} |
284
|
|
|
|
285
|
|
|
/** |
286
|
|
|
* Assemble the 'filter' portion of an ES query, from all registered filters |
287
|
|
|
* |
288
|
|
|
* @return array|null Combiled ES filters, or null if none have been defined |
289
|
|
|
*/ |
290
|
|
|
public function build_filter() { |
291
|
|
|
if ( empty( $this->es_filters ) ) { |
292
|
|
|
$filter = null; |
293
|
|
|
} elseif ( 1 == count( $this->es_filters ) ) { |
294
|
|
|
$filter = $this->es_filters[0]; |
295
|
|
|
} else { |
296
|
|
|
$filter = array( |
297
|
|
|
'and' => $this->es_filters, |
298
|
|
|
); |
299
|
|
|
} |
300
|
|
|
|
301
|
|
|
return $filter; |
302
|
|
|
} |
303
|
|
|
|
304
|
|
|
/** |
305
|
|
|
* Assemble the 'aggregation' portion of an ES query, from all general aggregations. |
306
|
|
|
* |
307
|
|
|
* @return an aggregation query as an array of topics, filters, and bucket names |
308
|
|
|
*/ |
309
|
|
|
function build_aggregation() { |
310
|
|
|
if ( empty( $this->bucket_sub_aggs ) && empty( $this->aggs_query ) ) { |
311
|
|
|
return array(); |
312
|
|
|
} |
313
|
|
|
|
314
|
|
|
if ( ! $this->diverse_buckets_query && empty( $this->aggs_query ) ) { |
315
|
|
|
return $this->bucket_sub_aggs; |
316
|
|
|
} |
317
|
|
|
|
318
|
|
|
$aggregations = array( |
319
|
|
|
'topics' => array( |
320
|
|
|
'filters' => array( |
321
|
|
|
'filters' => array(), |
322
|
|
|
), |
323
|
|
|
), |
324
|
|
|
); |
325
|
|
|
|
326
|
|
|
if ( ! empty( $this->bucket_sub_aggs ) ) { |
327
|
|
|
$aggregations['topics']['aggs'] = $this->bucket_sub_aggs; |
328
|
|
|
} |
329
|
|
|
|
330
|
|
|
foreach ( $this->bucket_filters as $bucket_name => $filter ) { |
331
|
|
|
$aggregations['topics']['filters']['filters'][ $bucket_name ] = $filter; |
332
|
|
|
} |
333
|
|
|
|
334
|
|
|
if ( ! empty( $this->aggs_query ) ) { |
335
|
|
|
$aggregations = $this->aggs; |
336
|
|
|
} |
337
|
|
|
|
338
|
|
|
return $aggregations; |
339
|
|
|
} |
340
|
|
|
|
341
|
|
|
} |
342
|
|
|
|
This check looks for variable assignements that are either overwritten by other assignments or where the variable is not used subsequently.
Both the
$myVar
assignment in line 1 and the$higher
assignment in line 2 are dead. The first because$myVar
is never used and the second because$higher
is always overwritten for every possible time line.