1
|
|
|
<?php |
2
|
|
|
namespace ApacheSolrForTypo3\Solr; |
3
|
|
|
|
4
|
|
|
/*************************************************************** |
5
|
|
|
* Copyright notice |
6
|
|
|
* |
7
|
|
|
* (c) 2009-2015 Ingo Renner <[email protected]> |
8
|
|
|
* All rights reserved |
9
|
|
|
* |
10
|
|
|
* This script is part of the TYPO3 project. The TYPO3 project is |
11
|
|
|
* free software; you can redistribute it and/or modify |
12
|
|
|
* it under the terms of the GNU General Public License as published by |
13
|
|
|
* the Free Software Foundation; either version 2 of the License, or |
14
|
|
|
* (at your option) any later version. |
15
|
|
|
* |
16
|
|
|
* The GNU General Public License can be found at |
17
|
|
|
* http://www.gnu.org/copyleft/gpl.html. |
18
|
|
|
* |
19
|
|
|
* This script is distributed in the hope that it will be useful, |
20
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
21
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
22
|
|
|
* GNU General Public License for more details. |
23
|
|
|
* |
24
|
|
|
* This copyright notice MUST APPEAR in all copies of the script! |
25
|
|
|
***************************************************************/ |
26
|
|
|
|
27
|
|
|
use ApacheSolrForTypo3\Solr\Search\FacetsModifier; |
28
|
|
|
use ApacheSolrForTypo3\Solr\System\Configuration\TypoScriptConfiguration; |
29
|
|
|
use ApacheSolrForTypo3\Solr\System\Logging\SolrLogManager; |
30
|
|
|
use TYPO3\CMS\Core\SingletonInterface; |
31
|
|
|
use TYPO3\CMS\Core\Utility\GeneralUtility; |
32
|
|
|
|
33
|
|
|
/** |
34
|
|
|
* Class to handle solr search requests |
35
|
|
|
* |
36
|
|
|
* @author Ingo Renner <[email protected]> |
37
|
|
|
*/ |
38
|
|
|
class Search |
39
|
|
|
{ |
40
|
|
|
|
41
|
|
|
/** |
42
|
|
|
* An instance of the Solr service |
43
|
|
|
* |
44
|
|
|
* @var SolrService |
45
|
|
|
*/ |
46
|
|
|
protected $solr = null; |
47
|
|
|
|
48
|
|
|
/** |
49
|
|
|
* The search query |
50
|
|
|
* |
51
|
|
|
* @var Query |
52
|
|
|
*/ |
53
|
|
|
protected $query = null; |
54
|
|
|
|
55
|
|
|
/** |
56
|
|
|
* The search response |
57
|
|
|
* |
58
|
|
|
* @var \Apache_Solr_Response |
59
|
|
|
*/ |
60
|
|
|
protected $response = null; |
61
|
|
|
|
62
|
|
|
/** |
63
|
|
|
* Flag for marking a search |
64
|
|
|
* |
65
|
|
|
* @var bool |
66
|
|
|
*/ |
67
|
|
|
protected $hasSearched = false; |
68
|
|
|
|
69
|
|
|
/** |
70
|
|
|
* @var TypoScriptConfiguration |
71
|
|
|
*/ |
72
|
|
|
protected $configuration; |
73
|
|
|
|
74
|
|
|
// TODO Override __clone to reset $response and $hasSearched |
75
|
|
|
|
76
|
|
|
/** |
77
|
|
|
* @var \ApacheSolrForTypo3\Solr\System\Logging\SolrLogManager |
78
|
|
|
*/ |
79
|
|
|
protected $logger = null; |
80
|
|
|
|
81
|
|
|
/** |
82
|
|
|
* Constructor |
83
|
|
|
* |
84
|
|
|
* @param SolrService $solrConnection The Solr connection to use for searching |
85
|
|
|
*/ |
86
|
|
|
public function __construct(SolrService $solrConnection = null) |
87
|
|
|
{ |
88
|
|
|
$this->logger = GeneralUtility::makeInstance(SolrLogManager::class, __CLASS__); |
89
|
36 |
|
|
90
|
|
|
$this->solr = $solrConnection; |
91
|
36 |
|
|
92
|
|
|
if (is_null($solrConnection)) { |
93
|
36 |
|
/** @var $connectionManager ConnectionManager */ |
94
|
|
|
$connectionManager = GeneralUtility::makeInstance(ConnectionManager::class); |
95
|
36 |
|
$this->solr = $connectionManager->getConnectionByPageId($GLOBALS['TSFE']->id, $GLOBALS['TSFE']->sys_language_uid); |
96
|
|
|
} |
97
|
1 |
|
|
98
|
1 |
|
$this->configuration = Util::getSolrConfiguration(); |
99
|
|
|
} |
100
|
|
|
|
101
|
36 |
|
/** |
102
|
36 |
|
* Gets the Solr connection used by this search. |
103
|
|
|
* |
104
|
|
|
* @return SolrService Solr connection |
105
|
|
|
*/ |
106
|
|
|
public function getSolrConnection() |
107
|
|
|
{ |
108
|
|
|
return $this->solr; |
109
|
|
|
} |
110
|
|
|
|
111
|
|
|
/** |
112
|
|
|
* Sets the Solr connection used by this search. |
113
|
|
|
* |
114
|
|
|
* Since ApacheSolrForTypo3\Solr\Search is a \TYPO3\CMS\Core\SingletonInterface, this is needed to |
115
|
|
|
* be able to switch between multiple cores/connections during |
116
|
|
|
* one request |
117
|
|
|
* |
118
|
|
|
* @param SolrService $solrConnection |
119
|
|
|
*/ |
120
|
|
|
public function setSolrConnection(SolrService $solrConnection) |
121
|
|
|
{ |
122
|
|
|
$this->solr = $solrConnection; |
123
|
|
|
} |
124
|
|
|
|
125
|
|
|
/** |
126
|
|
|
* Executes a query against a Solr server. |
127
|
|
|
* |
128
|
|
|
* 1) Gets the query string |
129
|
|
|
* 2) Conducts the actual search |
130
|
|
|
* 3) Checks debug settings |
131
|
|
|
* |
132
|
|
|
* @param Query $query The query with keywords, filters, and so on. |
133
|
|
|
* @param int $offset Result offset for pagination. |
134
|
|
|
* @param int $limit Maximum number of results to return. If set to NULL, this value is taken from the query object. |
135
|
|
|
* @return \Apache_Solr_Response Solr response |
136
|
|
|
*/ |
137
|
|
|
public function search(Query $query, $offset = 0, $limit = 10) |
|
|
|
|
138
|
|
|
{ |
139
|
|
|
$this->query = $query; |
140
|
30 |
|
|
141
|
|
|
if (empty($limit)) { |
142
|
30 |
|
$limit = $query->getResultsPerPage(); |
143
|
30 |
|
} |
144
|
|
|
|
145
|
30 |
|
try { |
146
|
26 |
|
$response = $this->solr->search( |
147
|
|
|
$query->getQueryString(), |
148
|
|
|
$offset, |
149
|
|
|
$limit, |
150
|
30 |
|
$query->getQueryParameters() |
151
|
30 |
|
); |
152
|
|
|
|
153
|
|
|
if ($this->configuration->getLoggingQueryQueryString()) { |
154
|
30 |
|
$this->logger->log( |
155
|
|
|
SolrLogManager::INFO, |
156
|
|
|
'Querying Solr, getting result', |
157
|
29 |
|
[ |
158
|
|
|
'query string' => $query->getQueryString(), |
159
|
|
|
'query parameters' => $query->getQueryParameters(), |
160
|
|
|
'response' => json_decode($response->getRawResponse(), |
161
|
|
|
true) |
162
|
|
|
] |
163
|
|
|
); |
164
|
|
|
} |
165
|
29 |
|
} catch (\RuntimeException $e) { |
166
|
|
|
$response = $this->solr->getResponse(); |
167
|
|
|
|
168
|
|
|
if ($this->configuration->getLoggingExceptions()) { |
169
|
1 |
|
$this->logger->log( |
170
|
1 |
|
SolrLogManager::ERROR, |
171
|
|
|
'Exception while querying Solr', |
172
|
1 |
|
[ |
173
|
1 |
|
'exception' => $e->__toString(), |
174
|
1 |
|
'query' => (array)$query, |
175
|
1 |
|
'offset' => $offset, |
176
|
|
|
'limit' => $limit |
177
|
1 |
|
] |
178
|
1 |
|
); |
179
|
1 |
|
} |
180
|
1 |
|
} |
181
|
|
|
|
182
|
|
|
$this->response = $response; |
183
|
|
|
$this->hasSearched = true; |
184
|
|
|
|
185
|
|
|
return $this->response; |
186
|
30 |
|
} |
187
|
30 |
|
|
188
|
30 |
|
/** |
189
|
|
|
* Sends a ping to the solr server to see whether it is available. |
190
|
30 |
|
* |
191
|
|
|
* @param bool $useCache Set to true if the cache should be used. |
192
|
|
|
* @return bool Returns TRUE on successful ping. |
193
|
|
|
* @throws \Exception Throws an exception in case ping was not successful. |
194
|
|
|
*/ |
195
|
|
|
public function ping($useCache = true) |
196
|
|
|
{ |
197
|
|
|
$solrAvailable = false; |
198
|
|
|
|
199
|
30 |
|
try { |
200
|
|
|
if (!$this->solr->ping(2, $useCache)) { |
201
|
|
|
throw new \Exception('Solr Server not responding.', 1237475791); |
202
|
30 |
|
} |
203
|
26 |
|
|
204
|
26 |
|
$solrAvailable = true; |
205
|
|
|
} catch (\Exception $e) { |
206
|
26 |
|
if ($this->configuration->getLoggingExceptions()) { |
207
|
26 |
|
$this->logger->log( |
208
|
|
|
SolrLogManager::ERROR, |
209
|
|
|
'Exception while trying to ping the solr server', |
210
|
|
|
[ |
211
|
26 |
|
$e->__toString() |
212
|
|
|
] |
213
|
|
|
); |
214
|
|
|
} |
215
|
26 |
|
} |
216
|
|
|
|
217
|
|
|
return $solrAvailable; |
218
|
|
|
} |
219
|
|
|
|
220
|
|
|
/** |
221
|
30 |
|
* checks whether a search has been executed |
222
|
|
|
* |
223
|
|
|
* @return bool TRUE if there was a search, FALSE otherwise (if the user just visited the search page f.e.) |
224
|
|
|
*/ |
225
|
|
|
public function hasSearched() |
226
|
|
|
{ |
227
|
|
|
return $this->hasSearched; |
228
|
|
|
} |
229
|
|
|
|
230
|
|
|
/** |
231
|
|
|
* Gets the query object. |
232
|
30 |
|
* |
233
|
|
|
* @return Query Query |
234
|
|
|
*/ |
235
|
30 |
|
public function getQuery() |
236
|
|
|
{ |
237
|
|
|
return $this->query; |
238
|
|
|
} |
239
|
|
|
|
240
|
|
|
/** |
241
|
|
|
* Gets the Solr response |
242
|
|
|
* |
243
|
|
|
* @return \Apache_Solr_Response |
244
|
|
|
*/ |
245
|
|
|
public function getResponse() |
246
|
|
|
{ |
247
|
|
|
return $this->response; |
248
|
|
|
} |
249
|
|
|
|
250
|
|
|
public function getRawResponse() |
251
|
|
|
{ |
252
|
|
|
return $this->response->getRawResponse(); |
253
|
|
|
} |
254
|
|
|
|
255
|
|
|
public function getResponseHeader() |
256
|
|
|
{ |
257
|
30 |
|
return $this->getResponse()->responseHeader; |
258
|
|
|
} |
259
|
|
|
|
260
|
|
|
public function getResponseBody() |
261
|
|
|
{ |
262
|
|
|
return $this->getResponse()->response; |
|
|
|
|
263
|
|
|
} |
264
|
|
|
|
265
|
|
|
/** |
266
|
|
|
* Returns all results documents raw. Use with caution! |
267
|
25 |
|
* |
268
|
|
|
* @return \Apache_Solr_Document[] |
269
|
25 |
|
*/ |
270
|
|
|
public function getResultDocumentsRaw() |
271
|
|
|
{ |
272
|
25 |
|
return $this->getResponseBody()->docs; |
273
|
|
|
} |
274
|
|
|
|
275
|
|
|
/** |
276
|
25 |
|
* Returns all result documents but applies htmlspecialchars() on all fields retrieved |
277
|
|
|
* from solr except the configured fields in plugin.tx_solr.search.trustedFields |
278
|
|
|
* |
279
|
|
|
* @return \Apache_Solr_Document[] |
280
|
|
|
*/ |
281
|
|
|
public function getResultDocumentsEscaped() |
282
|
|
|
{ |
283
|
|
|
return $this->applyHtmlSpecialCharsOnAllFields($this->getResponseBody()->docs); |
284
|
|
|
} |
285
|
|
|
|
286
|
|
|
/** |
287
|
|
|
* This method is used to apply htmlspecialchars on all document fields that |
288
|
|
|
* are not configured to be secure. Secure mean that we know where the content is coming from. |
289
|
25 |
|
* |
290
|
|
|
* @param array $documents |
291
|
|
|
* @return \Apache_Solr_Document[] |
292
|
|
|
*/ |
293
|
|
|
protected function applyHtmlSpecialCharsOnAllFields(array $documents) |
294
|
|
|
{ |
295
|
|
|
$trustedSolrFields = $this->configuration->getSearchTrustedFieldsArray(); |
296
|
|
|
|
297
|
25 |
|
foreach ($documents as $key => $document) { |
298
|
|
|
$fieldNames = $document->getFieldNames(); |
299
|
25 |
|
|
300
|
|
|
foreach ($fieldNames as $fieldName) { |
301
|
|
|
if (in_array($fieldName, $trustedSolrFields)) { |
302
|
|
|
// we skip this field, since it was marked as secure |
303
|
|
|
continue; |
304
|
|
|
} |
305
|
|
|
|
306
|
|
|
$document->{$fieldName} = $this->applyHtmlSpecialCharsOnSingleFieldValue($document->{$fieldName}); |
307
|
22 |
|
} |
308
|
|
|
|
309
|
22 |
|
$documents[$key] = $document; |
310
|
|
|
} |
311
|
|
|
|
312
|
|
|
return $documents; |
313
|
|
|
} |
314
|
|
|
|
315
|
|
|
/** |
316
|
|
|
* Applies htmlspecialchars on all items of an array of a single value. |
317
|
23 |
|
* |
318
|
|
|
* @param $fieldValue |
319
|
23 |
|
* @return array|string |
320
|
|
|
*/ |
321
|
|
|
protected function applyHtmlSpecialCharsOnSingleFieldValue($fieldValue) |
322
|
|
|
{ |
323
|
|
|
if (is_array($fieldValue)) { |
324
|
|
|
foreach ($fieldValue as $key => $fieldValueItem) { |
325
|
|
|
$fieldValue[$key] = htmlspecialchars($fieldValueItem, null, null, false); |
326
|
|
|
} |
327
|
20 |
|
} else { |
328
|
|
|
$fieldValue = htmlspecialchars($fieldValue, null, null, false); |
329
|
20 |
|
} |
330
|
|
|
|
331
|
|
|
return $fieldValue; |
332
|
23 |
|
} |
333
|
|
|
|
334
|
23 |
|
/** |
335
|
|
|
* Gets the time Solr took to execute the query and return the result. |
336
|
|
|
* |
337
|
|
|
* @return int Query time in milliseconds |
338
|
|
|
*/ |
339
|
|
|
public function getQueryTime() |
340
|
|
|
{ |
341
|
|
|
return $this->getResponseHeader()->QTime; |
342
|
22 |
|
} |
343
|
|
|
|
344
|
22 |
|
/** |
345
|
|
|
* Gets the number of results per page. |
346
|
|
|
* |
347
|
|
|
* @return int Number of results per page |
348
|
|
|
*/ |
349
|
|
|
public function getResultsPerPage() |
350
|
|
|
{ |
351
|
|
|
return $this->getResponseHeader()->params->rows; |
352
|
|
|
} |
353
|
3 |
|
|
354
|
|
|
/** |
355
|
3 |
|
* Gets all facets with their fields, options, and counts. |
356
|
|
|
* |
357
|
|
|
* @return array |
358
|
|
|
*/ |
359
|
|
|
public function getFacetCounts() |
360
|
|
|
{ |
361
|
|
|
static $facetCountsModified = false; |
362
|
|
|
static $facetCounts = null; |
363
|
|
|
|
364
|
|
|
$unmodifiedFacetCounts = $this->response->facet_counts; |
|
|
|
|
365
|
3 |
|
|
366
|
|
|
if (is_array($GLOBALS['TYPO3_CONF_VARS']['EXTCONF']['solr']['modifyFacets'])) { |
367
|
3 |
|
if (!$facetCountsModified) { |
368
|
|
|
$facetCounts = $unmodifiedFacetCounts; |
369
|
3 |
|
|
370
|
3 |
|
foreach ($GLOBALS['TYPO3_CONF_VARS']['EXTCONF']['solr']['modifyFacets'] as $classReference) { |
371
|
|
|
$facetsModifier = GeneralUtility::getUserObj($classReference); |
372
|
3 |
|
|
373
|
3 |
|
if ($facetsModifier instanceof FacetsModifier) { |
374
|
|
|
$facetCounts = $facetsModifier->modifyFacets($facetCounts); |
375
|
3 |
|
$facetCountsModified = true; |
376
|
|
|
} else { |
377
|
|
|
throw new \UnexpectedValueException( |
378
|
3 |
|
get_class($facetsModifier) . ' must implement interface ' . FacetsModifier::class, |
379
|
|
|
1310387526 |
380
|
|
|
); |
381
|
3 |
|
} |
382
|
|
|
} |
383
|
|
|
} |
384
|
3 |
|
} else { |
385
|
|
|
$facetCounts = $unmodifiedFacetCounts; |
386
|
|
|
} |
387
|
|
|
|
388
|
|
|
return $facetCounts; |
389
|
|
|
} |
390
|
|
|
|
391
|
|
|
public function getFacetFieldOptions($facetField) |
392
|
|
|
{ |
393
|
3 |
|
$facetOptions = null; |
394
|
|
|
|
395
|
3 |
|
if (property_exists($this->getFacetCounts()->facet_fields, |
396
|
3 |
|
$facetField)) { |
397
|
3 |
|
$facetOptions = get_object_vars($this->getFacetCounts()->facet_fields->$facetField); |
398
|
|
|
} |
399
|
|
|
|
400
|
3 |
|
return $facetOptions; |
401
|
|
|
} |
402
|
|
|
|
403
|
3 |
|
public function getFacetQueryOptions($facetField) |
404
|
|
|
{ |
405
|
|
|
$options = []; |
406
|
|
|
|
407
|
|
|
$facetQueries = get_object_vars($this->getFacetCounts()->facet_queries); |
408
|
|
|
foreach ($facetQueries as $facetQuery => $numberOfResults) { |
409
|
|
|
// remove tags from the facet.query response, for facet.field |
410
|
|
|
// and facet.range Solr does that on its own automatically |
411
|
20 |
|
$facetQuery = preg_replace('/^\{!ex=[^\}]*\}(.*)/', '\\1', |
412
|
|
|
$facetQuery); |
413
|
20 |
|
|
414
|
|
|
if (GeneralUtility::isFirstPartOfStr($facetQuery, $facetField)) { |
415
|
|
|
$options[$facetQuery] = $numberOfResults; |
416
|
|
|
} |
417
|
|
|
} |
418
|
|
|
|
419
|
|
|
// filter out queries with no results |
420
|
|
|
$options = array_filter($options); |
421
|
|
|
|
422
|
|
|
return $options; |
423
|
|
|
} |
424
|
|
|
|
425
|
|
|
public function getFacetRangeOptions($rangeFacetField) |
426
|
|
|
{ |
427
|
|
|
return get_object_vars($this->getFacetCounts()->facet_ranges->$rangeFacetField); |
428
|
|
|
} |
429
|
|
|
|
430
|
|
|
public function getNumberOfResults() |
431
|
|
|
{ |
432
|
|
|
return $this->response->response->numFound; |
|
|
|
|
433
|
|
|
} |
434
|
|
|
|
435
|
|
|
/** |
436
|
|
|
* Gets the result offset. |
437
|
|
|
* |
438
|
|
|
* @return int Result offset |
439
|
|
|
*/ |
440
|
|
|
public function getResultOffset() |
441
|
|
|
{ |
442
|
|
|
return $this->response->response->start; |
|
|
|
|
443
|
|
|
} |
444
|
|
|
|
445
|
|
|
public function getMaximumResultScore() |
446
|
|
|
{ |
447
|
|
|
return $this->response->response->maxScore; |
|
|
|
|
448
|
|
|
} |
449
|
|
|
|
450
|
|
|
public function getDebugResponse() |
451
|
|
|
{ |
452
|
|
|
return $this->response->debug; |
453
|
|
|
} |
454
|
|
|
|
455
|
|
|
public function getHighlightedContent() |
456
|
|
|
{ |
457
|
|
|
$highlightedContent = false; |
458
|
|
|
|
459
|
|
|
if ($this->response->highlighting) { |
460
|
|
|
$highlightedContent = $this->response->highlighting; |
|
|
|
|
461
|
|
|
} |
462
|
|
|
|
463
|
|
|
return $highlightedContent; |
464
|
|
|
} |
465
|
|
|
|
466
|
|
|
public function getSpellcheckingSuggestions() |
467
|
|
|
{ |
468
|
|
|
$spellcheckingSuggestions = false; |
469
|
|
|
|
470
|
|
|
$suggestions = (array)$this->response->spellcheck->suggestions; |
471
|
|
|
|
472
|
|
|
if (!empty($suggestions)) { |
473
|
|
|
$spellcheckingSuggestions = $suggestions; |
474
|
|
|
|
475
|
|
|
if (isset($this->response->spellcheck->collations)) { |
476
|
|
|
$collections = (array)$this->response->spellcheck->collations; |
477
|
|
|
$spellcheckingSuggestions['collation'] = $collections['collation']; |
478
|
|
|
} |
479
|
|
|
} |
480
|
|
|
|
481
|
|
|
return $spellcheckingSuggestions; |
482
|
|
|
} |
483
|
|
|
} |
484
|
|
|
|