|
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 2 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\IndexQueue\Indexer; |
|
29
|
|
|
use ApacheSolrForTypo3\Solr\IndexQueue\Item; |
|
30
|
|
|
use ApacheSolrForTypo3\Solr\IndexQueue\Queue; |
|
31
|
|
|
use ApacheSolrForTypo3\Solr\Site; |
|
32
|
|
|
use ApacheSolrForTypo3\Solr\System\Configuration\TypoScriptConfiguration; |
|
33
|
|
|
use ApacheSolrForTypo3\Solr\Task\IndexQueueWorkerTask; |
|
34
|
|
|
use TYPO3\CMS\Backend\Utility\BackendUtility; |
|
35
|
|
|
use TYPO3\CMS\Core\Utility\GeneralUtility; |
|
36
|
|
|
use TYPO3\CMS\Extbase\SignalSlot\Dispatcher; |
|
37
|
|
|
|
|
38
|
|
|
/** |
|
39
|
|
|
* Service to perform indexing operations |
|
40
|
|
|
* |
|
41
|
|
|
* @author Timo Hund <[email protected]> |
|
42
|
|
|
*/ |
|
43
|
|
|
class IndexService |
|
44
|
|
|
{ |
|
45
|
|
|
/** |
|
46
|
|
|
* @var TypoScriptConfiguration |
|
47
|
|
|
*/ |
|
48
|
|
|
protected $configuration; |
|
49
|
|
|
|
|
50
|
|
|
/** |
|
51
|
|
|
* @var Site |
|
52
|
|
|
*/ |
|
53
|
|
|
protected $site; |
|
54
|
|
|
|
|
55
|
|
|
/** |
|
56
|
|
|
* @var IndexQueueWorkerTask |
|
57
|
|
|
*/ |
|
58
|
|
|
protected $contextTask; |
|
59
|
|
|
|
|
60
|
|
|
/** |
|
61
|
|
|
* @var Queue |
|
62
|
|
|
*/ |
|
63
|
|
|
protected $indexQueue; |
|
64
|
|
|
|
|
65
|
|
|
/** |
|
66
|
|
|
* @var Dispatcher |
|
67
|
|
|
*/ |
|
68
|
|
|
protected $signalSlotDispatcher; |
|
69
|
|
|
|
|
70
|
|
|
/** |
|
71
|
|
|
* IndexService constructor. |
|
72
|
|
|
* @param Site $site |
|
73
|
|
|
* @param Queue|null $queue |
|
74
|
|
|
* @param Dispatcher|null $dispatcher |
|
75
|
|
|
*/ |
|
76
|
6 |
|
public function __construct(Site $site, Queue $queue = null, Dispatcher $dispatcher = null) |
|
77
|
|
|
{ |
|
78
|
6 |
|
$this->configuration = $site->getSolrConfiguration(); |
|
79
|
6 |
|
$this->site = $site; |
|
80
|
6 |
|
$this->indexQueue = is_null($queue) ? GeneralUtility::makeInstance(Queue::class) : $queue; |
|
81
|
6 |
|
$this->signalSlotDispatcher = is_null($dispatcher) ? GeneralUtility::makeInstance(Dispatcher::class) : $dispatcher; |
|
82
|
6 |
|
} |
|
83
|
|
|
|
|
84
|
|
|
/** |
|
85
|
|
|
* @param \ApacheSolrForTypo3\Solr\Task\IndexQueueWorkerTask $contextTask |
|
86
|
|
|
*/ |
|
87
|
1 |
|
public function setContextTask($contextTask) |
|
88
|
|
|
{ |
|
89
|
1 |
|
$this->contextTask = $contextTask; |
|
90
|
1 |
|
} |
|
91
|
|
|
|
|
92
|
|
|
/** |
|
93
|
|
|
* @return \ApacheSolrForTypo3\Solr\Task\IndexQueueWorkerTask |
|
94
|
|
|
*/ |
|
95
|
6 |
|
public function getContextTask() |
|
96
|
|
|
{ |
|
97
|
6 |
|
return $this->contextTask; |
|
98
|
|
|
} |
|
99
|
|
|
|
|
100
|
|
|
/** |
|
101
|
|
|
* Indexes items from the Index Queue. |
|
102
|
|
|
* |
|
103
|
|
|
* @param int $limit |
|
104
|
|
|
* @return bool |
|
105
|
|
|
*/ |
|
106
|
6 |
|
public function indexItems($limit) |
|
107
|
|
|
{ |
|
108
|
6 |
|
$errors = 0; |
|
109
|
6 |
|
$indexRunId = uniqid(); |
|
110
|
|
|
|
|
111
|
|
|
// get items to index |
|
112
|
6 |
|
$itemsToIndex = $this->indexQueue->getItemsToIndex($this->site, $limit); |
|
113
|
|
|
|
|
114
|
6 |
|
$this->emitSignal('beforeIndexItems', [$itemsToIndex, $this->getContextTask(), $indexRunId]); |
|
115
|
|
|
|
|
116
|
6 |
|
foreach ($itemsToIndex as $itemToIndex) { |
|
117
|
|
|
try { |
|
118
|
|
|
// try indexing |
|
119
|
6 |
|
$this->emitSignal('beforeIndexItem', [$itemToIndex, $this->getContextTask(), $indexRunId]); |
|
120
|
6 |
|
$this->indexItem($itemToIndex); |
|
121
|
6 |
|
$this->emitSignal('afterIndexItem', [$itemToIndex, $this->getContextTask(), $indexRunId]); |
|
122
|
|
|
} catch (\Exception $e) { |
|
123
|
|
|
$errors++; |
|
124
|
|
|
$this->indexQueue->markItemAsFailed($itemToIndex, $e->getCode() . ': ' . $e->__toString()); |
|
125
|
6 |
|
$this->generateIndexingErrorLog($itemToIndex, $e); |
|
126
|
|
|
} |
|
127
|
|
|
} |
|
128
|
|
|
|
|
129
|
6 |
|
$this->emitSignal('afterIndexItems', [$itemsToIndex, $this->getContextTask(), $indexRunId]); |
|
130
|
|
|
|
|
131
|
6 |
|
return ($errors === 0); |
|
132
|
|
|
} |
|
133
|
|
|
|
|
134
|
|
|
/** |
|
135
|
|
|
* Generates a message in the error log when an error occured. |
|
136
|
|
|
* |
|
137
|
|
|
* @param Item $itemToIndex |
|
138
|
|
|
* @param \Exception $e |
|
139
|
|
|
*/ |
|
140
|
|
|
protected function generateIndexingErrorLog(Item $itemToIndex, \Exception $e) |
|
141
|
|
|
{ |
|
142
|
|
|
$message = 'Failed indexing Index Queue item ' . $itemToIndex->getIndexQueueUid(); |
|
143
|
|
|
$data = [ 'code' => $e->getCode(), |
|
144
|
|
|
'message' => $e->getMessage(), |
|
145
|
|
|
'trace' => $e->getTrace(), |
|
146
|
|
|
'item' => (array)$itemToIndex]; |
|
147
|
|
|
GeneralUtility::devLog($message, 'solr', 3, $data); |
|
148
|
|
|
} |
|
149
|
|
|
|
|
150
|
|
|
/** |
|
151
|
|
|
* Builds an emits a singal for the IndexService. |
|
152
|
|
|
* |
|
153
|
|
|
* @param string $name |
|
154
|
|
|
* @param array $arguments |
|
155
|
|
|
* @return mixed |
|
156
|
|
|
*/ |
|
157
|
6 |
|
protected function emitSignal($name, $arguments) |
|
158
|
|
|
{ |
|
159
|
6 |
|
return $this->signalSlotDispatcher->dispatch(__CLASS__, $name, $arguments); |
|
160
|
|
|
} |
|
161
|
|
|
|
|
162
|
|
|
/** |
|
163
|
|
|
* Indexes an item from the Index Queue. |
|
164
|
|
|
* |
|
165
|
|
|
* @param Item $item An index queue item to index |
|
166
|
|
|
* @return bool TRUE if the item was successfully indexed, FALSE otherwise |
|
167
|
|
|
*/ |
|
168
|
5 |
|
protected function indexItem(Item $item) |
|
169
|
|
|
{ |
|
170
|
5 |
|
$indexer = $this->getIndexerByItem($item->getIndexingConfigurationName()); |
|
171
|
|
|
|
|
172
|
|
|
// Remember original http host value |
|
173
|
5 |
|
$originalHttpHost = isset($_SERVER['HTTP_HOST']) ? $_SERVER['HTTP_HOST'] : null; |
|
174
|
|
|
|
|
175
|
5 |
|
$this->initializeHttpServerEnvironment($item); |
|
176
|
5 |
|
$itemIndexed = $indexer->index($item); |
|
177
|
|
|
|
|
178
|
|
|
// update IQ item so that the IQ can determine what's been indexed already |
|
179
|
5 |
|
if ($itemIndexed) { |
|
180
|
5 |
|
$item->updateIndexedTime(); |
|
181
|
|
|
} |
|
182
|
|
|
|
|
183
|
5 |
|
if (!is_null($originalHttpHost)) { |
|
184
|
|
|
$_SERVER['HTTP_HOST'] = $originalHttpHost; |
|
185
|
|
|
} else { |
|
186
|
5 |
|
unset($_SERVER['HTTP_HOST']); |
|
187
|
|
|
} |
|
188
|
|
|
|
|
189
|
|
|
// needed since TYPO3 7.5 |
|
190
|
5 |
|
GeneralUtility::flushInternalRuntimeCaches(); |
|
191
|
|
|
|
|
192
|
5 |
|
return $itemIndexed; |
|
193
|
|
|
} |
|
194
|
|
|
|
|
195
|
|
|
/** |
|
196
|
|
|
* A factory method to get an indexer depending on an item's configuration. |
|
197
|
|
|
* |
|
198
|
|
|
* By default all items are indexed using the default indexer |
|
199
|
|
|
* (ApacheSolrForTypo3\Solr\IndexQueue\Indexer) coming with EXT:solr. Pages by default are |
|
200
|
|
|
* configured to be indexed through a dedicated indexer |
|
201
|
|
|
* (ApacheSolrForTypo3\Solr\IndexQueue\PageIndexer). In all other cases a dedicated indexer |
|
202
|
|
|
* can be specified through TypoScript if needed. |
|
203
|
|
|
* |
|
204
|
|
|
* @param string $indexingConfigurationName Indexing configuration name. |
|
205
|
|
|
* @throws \RuntimeException |
|
206
|
|
|
* @return Indexer An instance of ApacheSolrForTypo3\Solr\IndexQueue\Indexer or a sub class of it. |
|
207
|
|
|
*/ |
|
208
|
5 |
|
protected function getIndexerByItem($indexingConfigurationName) |
|
209
|
|
|
{ |
|
210
|
5 |
|
$indexerClass = $this->configuration->getIndexQueueIndexerByConfigurationName($indexingConfigurationName); |
|
211
|
5 |
|
$indexerConfiguration = $this->configuration->getIndexQueueIndexerConfigurationByConfigurationName($indexingConfigurationName); |
|
212
|
|
|
|
|
213
|
5 |
|
$indexer = GeneralUtility::makeInstance($indexerClass, $indexerConfiguration); |
|
214
|
5 |
|
if (!($indexer instanceof Indexer)) { |
|
215
|
|
|
throw new \RuntimeException( |
|
216
|
|
|
'The indexer class "' . $indexerClass . '" for indexing configuration "' . $indexingConfigurationName . '" is not a valid indexer. Must be a subclass of ApacheSolrForTypo3\Solr\IndexQueue\Indexer.', |
|
217
|
|
|
1260463206 |
|
218
|
|
|
); |
|
219
|
|
|
} |
|
220
|
|
|
|
|
221
|
5 |
|
return $indexer; |
|
222
|
|
|
} |
|
223
|
|
|
|
|
224
|
|
|
/** |
|
225
|
|
|
* Gets the indexing progress. |
|
226
|
|
|
* |
|
227
|
|
|
* @return float Indexing progress as a two decimal precision float. f.e. 44.87 |
|
228
|
|
|
*/ |
|
229
|
1 |
|
public function getProgress() |
|
230
|
|
|
{ |
|
231
|
1 |
|
$itemsIndexedPercentage = 0.0; |
|
232
|
|
|
|
|
233
|
1 |
|
$totalItemsCount = $GLOBALS['TYPO3_DB']->exec_SELECTcountRows( |
|
234
|
1 |
|
'uid', |
|
235
|
1 |
|
'tx_solr_indexqueue_item', |
|
236
|
1 |
|
'root = ' . $this->site->getRootPageId() |
|
237
|
|
|
); |
|
238
|
1 |
|
$remainingItemsCount = $GLOBALS['TYPO3_DB']->exec_SELECTcountRows( |
|
239
|
1 |
|
'uid', |
|
240
|
1 |
|
'tx_solr_indexqueue_item', |
|
241
|
1 |
|
'changed > indexed AND root = ' . $this->site->getRootPageId() |
|
242
|
|
|
); |
|
243
|
1 |
|
$itemsIndexedCount = $totalItemsCount - $remainingItemsCount; |
|
244
|
|
|
|
|
245
|
1 |
|
if ($totalItemsCount > 0) { |
|
246
|
1 |
|
$itemsIndexedPercentage = $itemsIndexedCount * 100 / $totalItemsCount; |
|
247
|
1 |
|
$itemsIndexedPercentage = round($itemsIndexedPercentage, 2); |
|
248
|
|
|
} |
|
249
|
|
|
|
|
250
|
1 |
|
return $itemsIndexedPercentage; |
|
251
|
|
|
} |
|
252
|
|
|
|
|
253
|
|
|
/** |
|
254
|
|
|
* Initializes the $_SERVER['HTTP_HOST'] environment variable in CLI |
|
255
|
|
|
* environments dependent on the Index Queue item's root page. |
|
256
|
|
|
* |
|
257
|
|
|
* When the Index Queue Worker task is executed by a cron job there is no |
|
258
|
|
|
* HTTP_HOST since we are in a CLI environment. RealURL needs the host |
|
259
|
|
|
* information to generate a proper URL though. Using the Index Queue item's |
|
260
|
|
|
* root page information we can determine the correct host although being |
|
261
|
|
|
* in a CLI environment. |
|
262
|
|
|
* |
|
263
|
|
|
* @param Item $item Index Queue item to use to determine the host. |
|
264
|
|
|
* @param |
|
265
|
|
|
*/ |
|
266
|
5 |
|
protected function initializeHttpServerEnvironment(Item $item) |
|
267
|
|
|
{ |
|
268
|
5 |
|
static $hosts = array(); |
|
269
|
5 |
|
$rootpageId = $item->getRootPageUid(); |
|
270
|
5 |
|
$hostFound = !empty($hosts[$rootpageId]); |
|
271
|
|
|
|
|
272
|
5 |
|
if (!$hostFound) { |
|
273
|
5 |
|
$rootline = BackendUtility::BEgetRootLine($rootpageId); |
|
274
|
5 |
|
$host = BackendUtility::firstDomainRecord($rootline); |
|
275
|
5 |
|
$hosts[$rootpageId] = $host; |
|
276
|
|
|
} |
|
277
|
|
|
|
|
278
|
5 |
|
$_SERVER['HTTP_HOST'] = $hosts[$rootpageId]; |
|
279
|
|
|
|
|
280
|
|
|
// needed since TYPO3 7.5 |
|
281
|
5 |
|
GeneralUtility::flushInternalRuntimeCaches(); |
|
282
|
5 |
|
} |
|
283
|
|
|
} |
|
284
|
|
|
|