1
|
|
|
<?php |
2
|
|
|
|
3
|
|
|
namespace ApacheSolrForTypo3\Solr\Domain\Index; |
4
|
|
|
|
5
|
|
|
/*************************************************************** |
6
|
|
|
* Copyright notice |
7
|
|
|
* |
8
|
|
|
* (c) 2015-2016 Timo Hund <[email protected]> |
9
|
|
|
* All rights reserved |
10
|
|
|
* |
11
|
|
|
* This script is part of the TYPO3 project. The TYPO3 project is |
12
|
|
|
* free software; you can redistribute it and/or modify |
13
|
|
|
* it under the terms of the GNU General Public License as published by |
14
|
|
|
* the Free Software Foundation; either version 3 of the License, or |
15
|
|
|
* (at your option) any later version. |
16
|
|
|
* |
17
|
|
|
* The GNU General Public License can be found at |
18
|
|
|
* http://www.gnu.org/copyleft/gpl.html. |
19
|
|
|
* |
20
|
|
|
* This script is distributed in the hope that it will be useful, |
21
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
22
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
23
|
|
|
* GNU General Public License for more details. |
24
|
|
|
* |
25
|
|
|
* This copyright notice MUST APPEAR in all copies of the script! |
26
|
|
|
***************************************************************/ |
27
|
|
|
|
28
|
|
|
use ApacheSolrForTypo3\Solr\ConnectionManager; |
29
|
|
|
use ApacheSolrForTypo3\Solr\IndexQueue\Indexer; |
30
|
|
|
use ApacheSolrForTypo3\Solr\IndexQueue\Item; |
31
|
|
|
use ApacheSolrForTypo3\Solr\IndexQueue\Queue; |
32
|
|
|
use ApacheSolrForTypo3\Solr\Domain\Site\Site; |
33
|
|
|
use ApacheSolrForTypo3\Solr\System\Configuration\TypoScriptConfiguration; |
34
|
|
|
use ApacheSolrForTypo3\Solr\System\Logging\SolrLogManager; |
35
|
|
|
use ApacheSolrForTypo3\Solr\Task\IndexQueueWorkerTask; |
36
|
|
|
use Solarium\Exception\HttpException; |
37
|
|
|
use TYPO3\CMS\Backend\Utility\BackendUtility; |
38
|
|
|
use TYPO3\CMS\Core\Utility\GeneralUtility; |
39
|
|
|
use TYPO3\CMS\Extbase\SignalSlot\Dispatcher; |
40
|
|
|
|
41
|
|
|
/** |
42
|
|
|
* Service to perform indexing operations |
43
|
|
|
* |
44
|
|
|
* @author Timo Hund <[email protected]> |
45
|
|
|
*/ |
46
|
|
|
class IndexService |
47
|
|
|
{ |
48
|
|
|
/** |
49
|
|
|
* @var TypoScriptConfiguration |
50
|
|
|
*/ |
51
|
|
|
protected $configuration; |
52
|
|
|
|
53
|
|
|
/** |
54
|
|
|
* @var Site |
55
|
|
|
*/ |
56
|
|
|
protected $site; |
57
|
|
|
|
58
|
|
|
/** |
59
|
|
|
* @var IndexQueueWorkerTask |
60
|
|
|
*/ |
61
|
|
|
protected $contextTask; |
62
|
|
|
|
63
|
|
|
/** |
64
|
|
|
* @var Queue |
65
|
|
|
*/ |
66
|
|
|
protected $indexQueue; |
67
|
|
|
|
68
|
|
|
/** |
69
|
|
|
* @var Dispatcher |
70
|
|
|
*/ |
71
|
|
|
protected $signalSlotDispatcher; |
72
|
|
|
|
73
|
|
|
/** |
74
|
|
|
* @var \ApacheSolrForTypo3\Solr\System\Logging\SolrLogManager |
75
|
|
|
*/ |
76
|
|
|
protected $logger = null; |
77
|
|
|
|
78
|
|
|
/** |
79
|
|
|
* IndexService constructor. |
80
|
|
|
* @param Site $site |
81
|
|
|
* @param Queue|null $queue |
82
|
|
|
* @param Dispatcher|null $dispatcher |
83
|
|
|
* @param SolrLogManager|null $solrLogManager |
84
|
|
|
*/ |
85
|
5 |
|
public function __construct(Site $site, Queue $queue = null, Dispatcher $dispatcher = null, SolrLogManager $solrLogManager = null) |
86
|
|
|
{ |
87
|
5 |
|
$this->site = $site; |
88
|
5 |
|
$this->indexQueue = $queue ?? GeneralUtility::makeInstance(Queue::class); |
89
|
5 |
|
$this->signalSlotDispatcher = $dispatcher ?? GeneralUtility::makeInstance(Dispatcher::class); |
90
|
5 |
|
$this->logger = $solrLogManager ?? GeneralUtility::makeInstance(SolrLogManager::class, /** @scrutinizer ignore-type */ __CLASS__); |
91
|
5 |
|
} |
92
|
|
|
|
93
|
|
|
/** |
94
|
|
|
* @param \ApacheSolrForTypo3\Solr\Task\IndexQueueWorkerTask $contextTask |
95
|
|
|
*/ |
96
|
1 |
|
public function setContextTask($contextTask) |
97
|
|
|
{ |
98
|
1 |
|
$this->contextTask = $contextTask; |
99
|
1 |
|
} |
100
|
|
|
|
101
|
|
|
/** |
102
|
|
|
* @return \ApacheSolrForTypo3\Solr\Task\IndexQueueWorkerTask |
103
|
|
|
*/ |
104
|
3 |
|
public function getContextTask() |
105
|
|
|
{ |
106
|
3 |
|
return $this->contextTask; |
107
|
|
|
} |
108
|
|
|
|
109
|
|
|
/** |
110
|
|
|
* Indexes items from the Index Queue. |
111
|
|
|
* |
112
|
|
|
* @param int $limit |
113
|
|
|
* @return bool |
114
|
|
|
*/ |
115
|
3 |
|
public function indexItems($limit) |
116
|
|
|
{ |
117
|
3 |
|
$errors = 0; |
118
|
3 |
|
$indexRunId = uniqid(); |
119
|
3 |
|
$configurationToUse = $this->site->getSolrConfiguration(); |
120
|
3 |
|
$enableCommitsSetting = $configurationToUse->getEnableCommits(); |
121
|
|
|
|
122
|
|
|
// get items to index |
123
|
3 |
|
$itemsToIndex = $this->indexQueue->getItemsToIndex($this->site, $limit); |
124
|
|
|
|
125
|
3 |
|
$this->emitSignal('beforeIndexItems', [$itemsToIndex, $this->getContextTask(), $indexRunId]); |
126
|
|
|
|
127
|
3 |
|
foreach ($itemsToIndex as $itemToIndex) { |
128
|
|
|
try { |
129
|
|
|
// try indexing |
130
|
3 |
|
$this->emitSignal('beforeIndexItem', [$itemToIndex, $this->getContextTask(), $indexRunId]); |
131
|
3 |
|
$this->indexItem($itemToIndex, $configurationToUse); |
132
|
2 |
|
$this->emitSignal('afterIndexItem', [$itemToIndex, $this->getContextTask(), $indexRunId]); |
133
|
1 |
|
} catch (\Exception $e) { |
134
|
1 |
|
$errors++; |
135
|
1 |
|
$this->indexQueue->markItemAsFailed($itemToIndex, $e->getCode() . ': ' . $e->__toString()); |
136
|
1 |
|
$this->generateIndexingErrorLog($itemToIndex, $e); |
137
|
|
|
} |
138
|
|
|
} |
139
|
|
|
|
140
|
3 |
|
$this->emitSignal('afterIndexItems', [$itemsToIndex, $this->getContextTask(), $indexRunId]); |
141
|
|
|
|
142
|
3 |
|
if ($enableCommitsSetting && count($itemsToIndex) > 0) { |
143
|
|
|
$solrServers = GeneralUtility::makeInstance(ConnectionManager::class)->getConnectionsBySite($this->site); |
144
|
|
|
foreach ($solrServers as $solrServer) { |
145
|
|
|
try { |
146
|
|
|
$solrServer->getWriteService()->commit(false, false, false); |
147
|
|
|
} catch (HttpException $e) { |
148
|
|
|
$errors++; |
149
|
|
|
} |
150
|
|
|
} |
151
|
|
|
} |
152
|
|
|
|
153
|
3 |
|
return ($errors === 0); |
154
|
|
|
} |
155
|
|
|
|
156
|
|
|
/** |
157
|
|
|
* Generates a message in the error log when an error occured. |
158
|
|
|
* |
159
|
|
|
* @param Item $itemToIndex |
160
|
|
|
* @param \Exception $e |
161
|
|
|
*/ |
162
|
1 |
|
protected function generateIndexingErrorLog(Item $itemToIndex, \Exception $e) |
163
|
|
|
{ |
164
|
1 |
|
$message = 'Failed indexing Index Queue item ' . $itemToIndex->getIndexQueueUid(); |
165
|
1 |
|
$data = ['code' => $e->getCode(), 'message' => $e->getMessage(), 'trace' => $e->getTraceAsString(), 'item' => (array)$itemToIndex]; |
166
|
|
|
|
167
|
1 |
|
$this->logger->log( |
168
|
1 |
|
SolrLogManager::ERROR, |
169
|
|
|
$message, |
170
|
|
|
$data |
171
|
|
|
); |
172
|
1 |
|
} |
173
|
|
|
|
174
|
|
|
/** |
175
|
|
|
* Builds an emits a singal for the IndexService. |
176
|
|
|
* |
177
|
|
|
* @param string $name |
178
|
|
|
* @param array $arguments |
179
|
|
|
* @return mixed |
180
|
|
|
*/ |
181
|
3 |
|
protected function emitSignal($name, $arguments) |
182
|
|
|
{ |
183
|
3 |
|
return $this->signalSlotDispatcher->dispatch(__CLASS__, $name, $arguments); |
184
|
|
|
} |
185
|
|
|
|
186
|
|
|
/** |
187
|
|
|
* Indexes an item from the Index Queue. |
188
|
|
|
* |
189
|
|
|
* @param Item $item An index queue item to index |
190
|
|
|
* @param TypoScriptConfiguration $configuration |
191
|
|
|
* @return bool TRUE if the item was successfully indexed, FALSE otherwise |
192
|
|
|
*/ |
193
|
2 |
|
protected function indexItem(Item $item, TypoScriptConfiguration $configuration) |
194
|
|
|
{ |
195
|
2 |
|
$indexer = $this->getIndexerByItem($item->getIndexingConfigurationName(), $configuration); |
196
|
|
|
|
197
|
|
|
// Remember original http host value |
198
|
2 |
|
$originalHttpHost = isset($_SERVER['HTTP_HOST']) ? $_SERVER['HTTP_HOST'] : null; |
199
|
|
|
|
200
|
2 |
|
$itemChangedDate = $item->getChanged(); |
201
|
2 |
|
$itemChangedDateAfterIndex = 0; |
202
|
|
|
|
203
|
|
|
try { |
204
|
2 |
|
$this->initializeHttpServerEnvironment($item); |
205
|
2 |
|
$itemIndexed = $indexer->index($item); |
206
|
|
|
|
207
|
|
|
// update IQ item so that the IQ can determine what's been indexed already |
208
|
1 |
|
if ($itemIndexed) { |
209
|
1 |
|
$this->indexQueue->updateIndexTimeByItem($item); |
210
|
1 |
|
$itemChangedDateAfterIndex = $item->getChanged(); |
211
|
|
|
} |
212
|
|
|
|
213
|
1 |
|
if ($itemChangedDateAfterIndex > $itemChangedDate && $itemChangedDateAfterIndex > time()) { |
214
|
1 |
|
$this->indexQueue->setForcedChangeTimeByItem($item, $itemChangedDateAfterIndex); |
215
|
|
|
} |
216
|
1 |
|
} catch (\Exception $e) { |
217
|
1 |
|
$this->restoreOriginalHttpHost($originalHttpHost); |
218
|
1 |
|
throw $e; |
219
|
|
|
} |
220
|
|
|
|
221
|
1 |
|
$this->restoreOriginalHttpHost($originalHttpHost); |
222
|
|
|
|
223
|
1 |
|
return $itemIndexed; |
224
|
|
|
} |
225
|
|
|
|
226
|
|
|
/** |
227
|
|
|
* A factory method to get an indexer depending on an item's configuration. |
228
|
|
|
* |
229
|
|
|
* By default all items are indexed using the default indexer |
230
|
|
|
* (ApacheSolrForTypo3\Solr\IndexQueue\Indexer) coming with EXT:solr. Pages by default are |
231
|
|
|
* configured to be indexed through a dedicated indexer |
232
|
|
|
* (ApacheSolrForTypo3\Solr\IndexQueue\PageIndexer). In all other cases a dedicated indexer |
233
|
|
|
* can be specified through TypoScript if needed. |
234
|
|
|
* |
235
|
|
|
* @param string $indexingConfigurationName Indexing configuration name. |
236
|
|
|
* @param TypoScriptConfiguration $configuration |
237
|
|
|
* @return Indexer |
238
|
|
|
*/ |
239
|
|
|
protected function getIndexerByItem($indexingConfigurationName, TypoScriptConfiguration $configuration) |
240
|
|
|
{ |
241
|
|
|
$indexerClass = $configuration->getIndexQueueIndexerByConfigurationName($indexingConfigurationName); |
242
|
|
|
$indexerConfiguration = $configuration->getIndexQueueIndexerConfigurationByConfigurationName($indexingConfigurationName); |
243
|
|
|
|
244
|
|
|
$indexer = GeneralUtility::makeInstance($indexerClass, /** @scrutinizer ignore-type */ $indexerConfiguration); |
245
|
|
|
if (!($indexer instanceof Indexer)) { |
246
|
|
|
throw new \RuntimeException( |
247
|
|
|
'The indexer class "' . $indexerClass . '" for indexing configuration "' . $indexingConfigurationName . '" is not a valid indexer. Must be a subclass of ApacheSolrForTypo3\Solr\IndexQueue\Indexer.', |
248
|
|
|
1260463206 |
249
|
|
|
); |
250
|
|
|
} |
251
|
|
|
|
252
|
|
|
return $indexer; |
253
|
|
|
} |
254
|
|
|
|
255
|
|
|
/** |
256
|
|
|
* Gets the indexing progress. |
257
|
|
|
* |
258
|
|
|
* @return float Indexing progress as a two decimal precision float. f.e. 44.87 |
259
|
|
|
*/ |
260
|
1 |
|
public function getProgress() |
261
|
|
|
{ |
262
|
1 |
|
return $this->indexQueue->getStatisticsBySite($this->site)->getSuccessPercentage(); |
263
|
|
|
} |
264
|
|
|
|
265
|
|
|
/** |
266
|
|
|
* Returns the amount of failed queue items for the current site. |
267
|
|
|
* |
268
|
|
|
* @return int |
269
|
|
|
*/ |
270
|
1 |
|
public function getFailCount() |
271
|
|
|
{ |
272
|
1 |
|
return $this->indexQueue->getStatisticsBySite($this->site)->getFailedCount(); |
273
|
|
|
} |
274
|
|
|
|
275
|
|
|
/** |
276
|
|
|
* Initializes the $_SERVER['HTTP_HOST'] environment variable in CLI |
277
|
|
|
* environments dependent on the Index Queue item's root page. |
278
|
|
|
* |
279
|
|
|
* When the Index Queue Worker task is executed by a cron job there is no |
280
|
|
|
* HTTP_HOST since we are in a CLI environment. RealURL needs the host |
281
|
|
|
* information to generate a proper URL though. Using the Index Queue item's |
282
|
|
|
* root page information we can determine the correct host although being |
283
|
|
|
* in a CLI environment. |
284
|
|
|
* |
285
|
|
|
* @param Item $item Index Queue item to use to determine the host. |
286
|
|
|
* @param |
287
|
|
|
*/ |
288
|
2 |
|
protected function initializeHttpServerEnvironment(Item $item) |
289
|
|
|
{ |
290
|
2 |
|
static $hosts = []; |
291
|
2 |
|
$rootpageId = $item->getRootPageUid(); |
292
|
2 |
|
$hostFound = !empty($hosts[$rootpageId]); |
293
|
|
|
|
294
|
2 |
|
if (!$hostFound) { |
295
|
2 |
|
$hosts[$rootpageId] = $item->getSite()->getDomain(); |
296
|
|
|
} |
297
|
|
|
|
298
|
2 |
|
$_SERVER['HTTP_HOST'] = $hosts[$rootpageId]; |
299
|
|
|
|
300
|
|
|
// needed since TYPO3 7.5 |
301
|
2 |
|
GeneralUtility::flushInternalRuntimeCaches(); |
302
|
2 |
|
} |
303
|
|
|
|
304
|
|
|
/** |
305
|
|
|
* @param string|null $originalHttpHost |
306
|
|
|
*/ |
307
|
1 |
|
protected function restoreOriginalHttpHost($originalHttpHost) |
308
|
|
|
{ |
309
|
1 |
|
if (!is_null($originalHttpHost)) { |
310
|
1 |
|
$_SERVER['HTTP_HOST'] = $originalHttpHost; |
311
|
|
|
} else { |
312
|
|
|
unset($_SERVER['HTTP_HOST']); |
313
|
|
|
} |
314
|
|
|
|
315
|
|
|
// needed since TYPO3 7.5 |
316
|
1 |
|
GeneralUtility::flushInternalRuntimeCaches(); |
317
|
1 |
|
} |
318
|
|
|
} |
319
|
|
|
|