1
|
|
|
<?php |
2
|
|
|
/** |
3
|
|
|
* Zend Framework |
4
|
|
|
* |
5
|
|
|
* LICENSE |
6
|
|
|
* |
7
|
|
|
* This source file is subject to version 1.0 of the Zend Framework |
8
|
|
|
* license, that is bundled with this package in the file LICENSE, and |
9
|
|
|
* is available through the world-wide-web at the following URL: |
10
|
|
|
* http://www.zend.com/license/framework/1_0.txt. If you did not receive |
11
|
|
|
* a copy of the Zend Framework license and are unable to obtain it |
12
|
|
|
* through the world-wide-web, please send a note to [email protected] |
13
|
|
|
* so we can mail you a copy immediately. |
14
|
|
|
* |
15
|
|
|
* @package Zend_Search_Lucene |
16
|
|
|
* @subpackage Search |
17
|
|
|
* @copyright Copyright (c) 2005-2006 Zend Technologies USA Inc. (http://www.zend.com) |
18
|
|
|
* @license http://www.zend.com/license/framework/1_0.txt Zend Framework License version 1.0 |
19
|
|
|
*/ |
20
|
|
|
|
21
|
|
|
|
22
|
|
|
/** |
23
|
|
|
* Zend_Search_Lucene_Search_Query |
24
|
|
|
*/ |
25
|
|
|
require_once 'Zend/Search/Lucene/Search/Query.php'; |
26
|
|
|
|
27
|
|
|
/** |
28
|
|
|
* Zend_Search_Lucene_Search_Weight_MultiTerm |
29
|
|
|
*/ |
30
|
|
|
require_once 'Zend/Search/Lucene/Search/Weight/Phrase.php'; |
31
|
|
|
|
32
|
|
|
|
33
|
|
|
/** |
34
|
|
|
* A Query that matches documents containing a particular sequence of terms. |
35
|
|
|
* |
36
|
|
|
* @package Zend_Search_Lucene |
37
|
|
|
* @subpackage Search |
38
|
|
|
* @copyright Copyright (c) 2005-2006 Zend Technologies USA Inc. (http://www.zend.com) |
39
|
|
|
* @license http://www.zend.com/license/framework/1_0.txt Zend Framework License version 1.0 |
40
|
|
|
*/ |
41
|
|
|
class Zend_Search_Lucene_Search_Query_Phrase extends Zend_Search_Lucene_Search_Query |
42
|
|
|
{ |
43
|
|
|
/** |
44
|
|
|
* Terms to find. |
45
|
|
|
* Array of Zend_Search_Lucene_Index_Term objects. |
46
|
|
|
* |
47
|
|
|
* @var array |
48
|
|
|
*/ |
49
|
|
|
private $_terms; |
50
|
|
|
|
51
|
|
|
/** |
52
|
|
|
* Term positions (relative positions of terms within the phrase). |
53
|
|
|
* Array of integers |
54
|
|
|
* |
55
|
|
|
* @var array |
56
|
|
|
*/ |
57
|
|
|
private $_offsets; |
58
|
|
|
|
59
|
|
|
/** |
60
|
|
|
* Sets the number of other words permitted between words in query phrase. |
61
|
|
|
* If zero, then this is an exact phrase search. For larger values this works |
62
|
|
|
* like a WITHIN or NEAR operator. |
63
|
|
|
* |
64
|
|
|
* The slop is in fact an edit-distance, where the units correspond to |
65
|
|
|
* moves of terms in the query phrase out of position. For example, to switch |
66
|
|
|
* the order of two words requires two moves (the first move places the words |
67
|
|
|
* atop one another), so to permit re-orderings of phrases, the slop must be |
68
|
|
|
* at least two. |
69
|
|
|
* More exact matches are scored higher than sloppier matches, thus search |
70
|
|
|
* results are sorted by exactness. |
71
|
|
|
* |
72
|
|
|
* The slop is zero by default, requiring exact matches. |
73
|
|
|
* |
74
|
|
|
* @var unknown_type |
75
|
|
|
*/ |
76
|
|
|
private $_slop; |
77
|
|
|
|
78
|
|
|
/** |
79
|
|
|
* Result vector. |
80
|
|
|
* Bitset or array of document IDs |
81
|
|
|
* (depending from Bitset extension availability). |
82
|
|
|
* |
83
|
|
|
* @var mixed |
84
|
|
|
*/ |
85
|
|
|
private $_resVector = null; |
86
|
|
|
|
87
|
|
|
/** |
88
|
|
|
* Terms positions vectors. |
89
|
|
|
* Array of Arrays: |
90
|
|
|
* term1Id => (docId => array( pos1, pos2, ... ), ...) |
91
|
|
|
* term2Id => (docId => array( pos1, pos2, ... ), ...) |
92
|
|
|
* |
93
|
|
|
* @var array |
94
|
|
|
*/ |
95
|
|
|
private $_termsPositions = array(); |
96
|
|
|
|
97
|
|
|
/** |
98
|
|
|
* Class constructor. Create a new prase query. |
99
|
|
|
* |
100
|
|
|
* @param string $field Field to search. |
101
|
|
|
* @param array $terms Terms to search Array of strings. |
102
|
|
|
* @param array $offsets Relative term positions. Array of integers. |
103
|
|
|
* @throws Zend_Search_Lucene_Exception |
104
|
|
|
*/ |
105
|
|
|
public function __construct($terms = null, $offsets = null, $field = null) |
106
|
|
|
{ |
107
|
|
|
$this->_slop = 0; |
|
|
|
|
108
|
|
|
|
109
|
|
|
if (is_array($terms)) { |
110
|
|
|
$this->_terms = array(); |
111
|
|
|
foreach ($terms as $termId => $termText) { |
112
|
|
|
$this->_terms[$termId] = ($field !== null)? new Zend_Search_Lucene_Index_Term($termText, $field): |
113
|
|
|
new Zend_Search_Lucene_Index_Term($termText); |
114
|
|
|
} |
115
|
|
|
} else if ($terms === null) { |
116
|
|
|
$this->_terms = array(); |
117
|
|
|
} else { |
118
|
|
|
throw new Zend_Search_Lucene_Exception('terms argument must be array of strings or null'); |
119
|
|
|
} |
120
|
|
|
|
121
|
|
|
if (is_array($offsets)) { |
122
|
|
|
if (count($this->_terms) != count($offsets)) { |
123
|
|
|
throw new Zend_Search_Lucene_Exception('terms and offsets arguments must have the same size.'); |
124
|
|
|
} |
125
|
|
|
$this->_offsets = $offsets; |
126
|
|
|
} else if ($offsets === null) { |
127
|
|
|
$this->_offsets = array(); |
128
|
|
|
foreach ($this->_terms as $termId => $term) { |
129
|
|
|
$position = count($this->_offsets); |
130
|
|
|
$this->_offsets[$termId] = $position; |
131
|
|
|
} |
132
|
|
|
} else { |
133
|
|
|
throw new Zend_Search_Lucene_Exception('offsets argument must be array of strings or null'); |
134
|
|
|
} |
135
|
|
|
} |
136
|
|
|
|
137
|
|
|
/** |
138
|
|
|
* Set slop |
139
|
|
|
* |
140
|
|
|
* @param integer $slop |
141
|
|
|
*/ |
142
|
|
|
public function setSlop($slop) |
143
|
|
|
{ |
144
|
|
|
$this->_slop = $slop; |
|
|
|
|
145
|
|
|
} |
146
|
|
|
|
147
|
|
|
|
148
|
|
|
/** |
149
|
|
|
* Get slop |
150
|
|
|
* |
151
|
|
|
* @return integer |
152
|
|
|
*/ |
153
|
|
|
public function getSlop() |
154
|
|
|
{ |
155
|
|
|
return $this->_slop; |
156
|
|
|
} |
157
|
|
|
|
158
|
|
|
|
159
|
|
|
/** |
160
|
|
|
* Adds a term to the end of the query phrase. |
161
|
|
|
* The relative position of the term is specified explicitly or the one immediately |
162
|
|
|
* after the last term added. |
163
|
|
|
* |
164
|
|
|
* @param Zend_Search_Lucene_Index_Term $term |
165
|
|
|
* @param integer $position |
166
|
|
|
*/ |
167
|
|
|
public function addTerm(Zend_Search_Lucene_Index_Term $term, $position = null) { |
168
|
|
|
if ((count($this->_terms) != 0)&&(end($this->_terms)->field != $term->field)) { |
169
|
|
|
throw new Zend_Search_Lucene_Exception('All phrase terms must be in the same field: ' . |
170
|
|
|
$term->field . ':' . $term->text); |
171
|
|
|
} |
172
|
|
|
|
173
|
|
|
$this->_terms[] = $term; |
174
|
|
|
if ($position !== null) { |
175
|
|
|
$this->_offsets[] = $position; |
176
|
|
|
} else if (count($this->_offsets) != 0) { |
177
|
|
|
$this->_offsets[] = end($this->_offsets) + 1; |
178
|
|
|
} else { |
179
|
|
|
$this->_offsets[] = 0; |
180
|
|
|
} |
181
|
|
|
} |
182
|
|
|
|
183
|
|
|
|
184
|
|
|
/** |
185
|
|
|
* Returns query term |
186
|
|
|
* |
187
|
|
|
* @return array |
188
|
|
|
*/ |
189
|
|
|
public function getTerms() |
190
|
|
|
{ |
191
|
|
|
return $this->_terms; |
192
|
|
|
} |
193
|
|
|
|
194
|
|
|
|
195
|
|
|
/** |
196
|
|
|
* Set weight for specified term |
197
|
|
|
* |
198
|
|
|
* @param integer $num |
199
|
|
|
* @param Zend_Search_Lucene_Search_Weight_Term $weight |
200
|
|
|
*/ |
201
|
|
|
public function setWeight($num, $weight) |
202
|
|
|
{ |
203
|
|
|
$this->_weights[$num] = $weight; |
|
|
|
|
204
|
|
|
} |
205
|
|
|
|
206
|
|
|
|
207
|
|
|
/** |
208
|
|
|
* Constructs an appropriate Weight implementation for this query. |
209
|
|
|
* |
210
|
|
|
* @param Zend_Search_Lucene $reader |
211
|
|
|
* @return Zend_Search_Lucene_Search_Weight |
212
|
|
|
*/ |
213
|
|
|
protected function _createWeight($reader) |
214
|
|
|
{ |
215
|
|
|
return new Zend_Search_Lucene_Search_Weight_Phrase($this, $reader); |
216
|
|
|
} |
217
|
|
|
|
218
|
|
|
|
219
|
|
|
/** |
220
|
|
|
* Calculate result vector |
221
|
|
|
* |
222
|
|
|
* @param Zend_Search_Lucene $reader |
223
|
|
|
*/ |
224
|
|
View Code Duplication |
private function _calculateResult($reader) |
|
|
|
|
225
|
|
|
{ |
226
|
|
|
if (extension_loaded('bitset')) { |
227
|
|
|
foreach( $this->_terms as $termId=>$term ) { |
228
|
|
|
if($this->_resVector === null) { |
229
|
|
|
$this->_resVector = bitset_from_array($reader->termDocs($term)); |
230
|
|
|
} else { |
231
|
|
|
$this->_resVector = bitset_intersection( |
232
|
|
|
$this->_resVector, |
233
|
|
|
bitset_from_array($reader->termDocs($term)) ); |
234
|
|
|
} |
235
|
|
|
|
236
|
|
|
$this->_termsPositions[$termId] = $reader->termPositions($term); |
237
|
|
|
} |
238
|
|
|
} else { |
239
|
|
|
foreach( $this->_terms as $termId=>$term ) { |
240
|
|
|
if($this->_resVector === null) { |
241
|
|
|
$this->_resVector = array_flip($reader->termDocs($term)); |
242
|
|
|
} else { |
243
|
|
|
$termDocs = array_flip($reader->termDocs($term)); |
244
|
|
|
foreach($this->_resVector as $key=>$value) { |
245
|
|
|
if (!isset( $termDocs[$key] )) { |
246
|
|
|
unset( $this->_resVector[$key] ); |
247
|
|
|
} |
248
|
|
|
} |
249
|
|
|
} |
250
|
|
|
|
251
|
|
|
$this->_termsPositions[$termId] = $reader->termPositions($term); |
252
|
|
|
} |
253
|
|
|
} |
254
|
|
|
} |
255
|
|
|
|
256
|
|
|
|
257
|
|
|
/** |
258
|
|
|
* Score calculator for exact phrase queries (terms sequence is fixed) |
259
|
|
|
* |
260
|
|
|
* @param integer $docId |
261
|
|
|
* @return float |
262
|
|
|
*/ |
263
|
|
|
public function _exactPhraseFreq($docId) |
264
|
|
|
{ |
265
|
|
|
$freq = 0; |
266
|
|
|
|
267
|
|
|
// Term Id with lowest cardinality |
268
|
|
|
$lowCardTermId = null; |
269
|
|
|
|
270
|
|
|
// Calculate $lowCardTermId |
271
|
|
|
foreach ($this->_terms as $termId => $term) { |
272
|
|
|
if ($lowCardTermId === null || |
273
|
|
|
count($this->_termsPositions[$termId][$docId]) < |
274
|
|
|
count($this->_termsPositions[$lowCardTermId][$docId]) ) { |
275
|
|
|
$lowCardTermId = $termId; |
276
|
|
|
} |
277
|
|
|
} |
278
|
|
|
|
279
|
|
|
// Walk through positions of the term with lowest cardinality |
280
|
|
|
foreach ($this->_termsPositions[$lowCardTermId][$docId] as $lowCardPos) { |
281
|
|
|
// We expect phrase to be found |
282
|
|
|
$freq++; |
283
|
|
|
|
284
|
|
|
// Walk through other terms |
285
|
|
|
foreach ($this->_terms as $termId => $term) { |
286
|
|
|
if ($termId != $lowCardTermId) { |
287
|
|
|
$expectedPosition = $lowCardPos + |
288
|
|
|
($this->_offsets[$termId] - |
289
|
|
|
$this->_offsets[$lowCardTermId]); |
290
|
|
|
|
291
|
|
|
if (!in_array($expectedPosition, $this->_termsPositions[$termId][$docId])) { |
292
|
|
|
$freq--; // Phrase wasn't found. |
293
|
|
|
break; |
294
|
|
|
} |
295
|
|
|
} |
296
|
|
|
} |
297
|
|
|
} |
298
|
|
|
|
299
|
|
|
return $freq; |
300
|
|
|
} |
301
|
|
|
|
302
|
|
|
/** |
303
|
|
|
* Score calculator for sloppy phrase queries (terms sequence is fixed) |
304
|
|
|
* |
305
|
|
|
* @param integer $docId |
306
|
|
|
* @param Zend_Search_Lucene $reader |
307
|
|
|
* @return float |
308
|
|
|
*/ |
309
|
|
|
public function _sloppyPhraseFreq($docId, Zend_Search_Lucene $reader) |
310
|
|
|
{ |
311
|
|
|
$freq = 0; |
312
|
|
|
|
313
|
|
|
$phraseQueue = array(); |
314
|
|
|
$phraseQueue[0] = array(); // empty phrase |
315
|
|
|
$lastTerm = null; |
316
|
|
|
|
317
|
|
|
// Walk through the terms to create phrases. |
318
|
|
|
foreach ($this->_terms as $termId => $term) { |
319
|
|
|
$queueSize = count($phraseQueue); |
320
|
|
|
$firstPass = true; |
321
|
|
|
|
322
|
|
|
// Walk through the term positions. |
323
|
|
|
// Each term position produces a set of phrases. |
324
|
|
|
foreach ($this->_termsPositions[$termId][$docId] as $termPosition ) { |
325
|
|
|
if ($firstPass) { |
326
|
|
|
for ($count = 0; $count < $queueSize; $count++) { |
327
|
|
|
$phraseQueue[$count][$termId] = $termPosition; |
328
|
|
|
} |
329
|
|
|
} else { |
330
|
|
|
for ($count = 0; $count < $queueSize; $count++) { |
331
|
|
|
if ($lastTerm !== null && |
332
|
|
|
abs( $termPosition - $phraseQueue[$count][$lastTerm] - |
333
|
|
|
($this->_offsets[$termId] - $this->_offsets[$lastTerm])) > $this->_slop) { |
334
|
|
|
continue; |
335
|
|
|
} |
336
|
|
|
|
337
|
|
|
$newPhraseId = count($phraseQueue); |
338
|
|
|
$phraseQueue[$newPhraseId] = $phraseQueue[$count]; |
339
|
|
|
$phraseQueue[$newPhraseId][$termId] = $termPosition; |
340
|
|
|
} |
341
|
|
|
|
342
|
|
|
} |
343
|
|
|
|
344
|
|
|
$firstPass = false; |
345
|
|
|
} |
346
|
|
|
$lastTerm = $termId; |
347
|
|
|
} |
348
|
|
|
|
349
|
|
|
|
350
|
|
|
foreach ($phraseQueue as $phrasePos) { |
351
|
|
|
$minDistance = null; |
352
|
|
|
|
353
|
|
|
for ($shift = -$this->_slop; $shift <= $this->_slop; $shift++) { |
354
|
|
|
$distance = 0; |
355
|
|
|
$start = reset($phrasePos) - reset($this->_offsets) + $shift; |
356
|
|
|
|
357
|
|
|
foreach ($this->_terms as $termId => $term) { |
358
|
|
|
$distance += abs($phrasePos[$termId] - $this->_offsets[$termId] - $start); |
359
|
|
|
|
360
|
|
|
if($distance > $this->_slop) { |
361
|
|
|
break; |
362
|
|
|
} |
363
|
|
|
} |
364
|
|
|
|
365
|
|
|
if ($minDistance === null || $distance < $minDistance) { |
366
|
|
|
$minDistance = $distance; |
367
|
|
|
} |
368
|
|
|
} |
369
|
|
|
|
370
|
|
|
if ($minDistance <= $this->_slop) { |
371
|
|
|
$freq += $reader->getSimilarity()->sloppyFreq($minDistance); |
372
|
|
|
} |
373
|
|
|
} |
374
|
|
|
|
375
|
|
|
return $freq; |
376
|
|
|
} |
377
|
|
|
|
378
|
|
|
|
379
|
|
|
/** |
380
|
|
|
* Score specified document |
381
|
|
|
* |
382
|
|
|
* @param integer $docId |
383
|
|
|
* @param Zend_Search_Lucene $reader |
384
|
|
|
* @return float |
385
|
|
|
*/ |
386
|
|
|
public function score($docId, $reader) |
387
|
|
|
{ |
388
|
|
|
// optimize zero-term case |
389
|
|
|
if (count($this->_terms) == 0) { |
390
|
|
|
return 0; |
391
|
|
|
} |
392
|
|
|
|
393
|
|
|
if($this->_resVector === null) { |
394
|
|
|
$this->_calculateResult($reader); |
395
|
|
|
$this->_initWeight($reader); |
396
|
|
|
} |
397
|
|
|
|
398
|
|
|
if ( (extension_loaded('bitset')) ? |
399
|
|
|
bitset_in($this->_resVector, $docId) : |
400
|
|
|
isset($this->_resVector[$docId]) ) { |
401
|
|
|
if ($this->_slop == 0) { |
402
|
|
|
$freq = $this->_exactPhraseFreq($docId); |
403
|
|
|
} else { |
404
|
|
|
$freq = $this->_sloppyPhraseFreq($docId, $reader); |
405
|
|
|
} |
406
|
|
|
|
407
|
|
|
/* |
408
|
|
|
return $reader->getSimilarity()->tf($freq) * |
409
|
|
|
$this->_weight->getValue() * |
410
|
|
|
$reader->norm($docId, reset($this->_terms)->field); |
411
|
|
|
*/ |
412
|
|
|
if ($freq != 0) { |
413
|
|
|
$tf = $reader->getSimilarity()->tf($freq); |
414
|
|
|
$weight = $this->_weight->getValue(); |
415
|
|
|
$norm = $reader->norm($docId, reset($this->_terms)->field); |
416
|
|
|
|
417
|
|
|
return $tf*$weight*$norm; |
418
|
|
|
} |
419
|
|
|
} else { |
420
|
|
|
return 0; |
421
|
|
|
} |
422
|
|
|
} |
423
|
|
|
} |
424
|
|
|
|
425
|
|
|
|
Our type inference engine has found an assignment to a property that is incompatible with the declared type of that property.
Either this assignment is in error or the assigned type should be added to the documentation/type hint for that property..