1
|
|
|
<?php |
2
|
|
|
|
3
|
|
|
declare(strict_types=1); |
4
|
|
|
|
5
|
|
|
namespace AOE\Crawler\Domain\Repository; |
6
|
|
|
|
7
|
|
|
/*************************************************************** |
8
|
|
|
* Copyright notice |
9
|
|
|
* |
10
|
|
|
* (c) 2019 AOE GmbH <[email protected]> |
11
|
|
|
* |
12
|
|
|
* All rights reserved |
13
|
|
|
* |
14
|
|
|
* This script is part of the TYPO3 project. The TYPO3 project is |
15
|
|
|
* free software; you can redistribute it and/or modify |
16
|
|
|
* it under the terms of the GNU General Public License as published by |
17
|
|
|
* the Free Software Foundation; either version 3 of the License, or |
18
|
|
|
* (at your option) any later version. |
19
|
|
|
* |
20
|
|
|
* The GNU General Public License can be found at |
21
|
|
|
* http://www.gnu.org/copyleft/gpl.html. |
22
|
|
|
* |
23
|
|
|
* This script is distributed in the hope that it will be useful, |
24
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
25
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
26
|
|
|
* GNU General Public License for more details. |
27
|
|
|
* |
28
|
|
|
* This copyright notice MUST APPEAR in all copies of the script! |
29
|
|
|
***************************************************************/ |
30
|
|
|
|
31
|
|
|
use AOE\Crawler\Configuration\ExtensionConfigurationProvider; |
32
|
|
|
use AOE\Crawler\Domain\Model\Process; |
33
|
|
|
use AOE\Crawler\Domain\Model\ProcessCollection; |
34
|
|
|
use PDO; |
35
|
|
|
use TYPO3\CMS\Core\Database\Connection; |
36
|
|
|
use TYPO3\CMS\Core\Database\ConnectionPool; |
37
|
|
|
use TYPO3\CMS\Core\Database\Query\QueryBuilder; |
38
|
|
|
use TYPO3\CMS\Core\Utility\GeneralUtility; |
39
|
|
|
use TYPO3\CMS\Extbase\Object\ObjectManager; |
40
|
|
|
use TYPO3\CMS\Extbase\Persistence\Repository; |
41
|
|
|
|
42
|
|
|
/** |
43
|
|
|
* @internal since v9.2.5 |
44
|
|
|
*/ |
45
|
|
|
class ProcessRepository extends Repository |
46
|
|
|
{ |
47
|
|
|
/** |
48
|
|
|
* @var string |
49
|
|
|
*/ |
50
|
|
|
protected $tableName = 'tx_crawler_process'; |
51
|
|
|
|
52
|
|
|
/** |
53
|
|
|
* @var QueryBuilder |
54
|
|
|
*/ |
55
|
|
|
protected $queryBuilder; |
56
|
|
|
|
57
|
|
|
/** |
58
|
|
|
* @var array |
59
|
|
|
*/ |
60
|
|
|
protected $extensionSettings = []; |
61
|
|
|
|
62
|
|
|
public function __construct() |
63
|
66 |
|
{ |
64
|
|
|
$objectManager = GeneralUtility::makeInstance(ObjectManager::class); |
65
|
66 |
|
|
66
|
|
|
parent::__construct($objectManager); |
67
|
66 |
|
|
68
|
|
|
$this->queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable($this->tableName); |
69
|
66 |
|
|
70
|
|
|
/** @var ExtensionConfigurationProvider $configurationProvider */ |
71
|
|
|
$configurationProvider = GeneralUtility::makeInstance(ExtensionConfigurationProvider::class); |
72
|
66 |
|
$this->extensionSettings = $configurationProvider->getExtensionConfiguration(); |
73
|
66 |
|
} |
74
|
66 |
|
|
75
|
|
|
/** |
76
|
|
|
* This method is used to find all cli processes within a limit. |
77
|
|
|
*/ |
78
|
|
|
public function findAll(): ProcessCollection |
79
|
7 |
|
{ |
80
|
|
|
/** @var ProcessCollection $collection */ |
81
|
|
|
$collection = GeneralUtility::makeInstance(ProcessCollection::class); |
82
|
7 |
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable($this->tableName); |
83
|
7 |
|
|
84
|
|
|
$statement = $queryBuilder |
85
|
|
|
->select('*') |
86
|
7 |
|
->from($this->tableName) |
87
|
7 |
|
->orderBy('ttl', 'DESC') |
88
|
7 |
|
->execute(); |
89
|
7 |
|
|
90
|
|
|
while ($row = $statement->fetch()) { |
91
|
7 |
|
$process = GeneralUtility::makeInstance(Process::class); |
92
|
7 |
|
$process->setProcessId($row['process_id']); |
93
|
7 |
|
$process->setActive($row['active']); |
94
|
7 |
|
$process->setTtl($row['ttl']); |
95
|
7 |
|
$process->setAssignedItemsCount($row['assigned_items_count']); |
96
|
7 |
|
$process->setDeleted($row['deleted']); |
97
|
7 |
|
$process->setSystemProcessId($row['system_process_id']); |
98
|
7 |
|
$collection->append($process); |
99
|
7 |
|
} |
100
|
|
|
|
101
|
|
|
return $collection; |
|
|
|
|
102
|
7 |
|
} |
103
|
|
|
|
104
|
|
|
/** |
105
|
|
|
* @param string $processId |
106
|
|
|
*/ |
107
|
|
|
public function findByProcessId($processId) |
108
|
1 |
|
{ |
109
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable($this->tableName); |
110
|
1 |
|
|
111
|
|
|
return $queryBuilder |
112
|
|
|
->select('*') |
113
|
1 |
|
->from($this->tableName) |
114
|
1 |
|
->where( |
115
|
1 |
|
$queryBuilder->expr()->eq('process_id', $queryBuilder->createNamedParameter($processId, PDO::PARAM_STR)) |
116
|
1 |
|
)->execute()->fetch(0); |
117
|
1 |
|
} |
118
|
|
|
|
119
|
|
|
public function findAllActive(): ProcessCollection |
120
|
5 |
|
{ |
121
|
|
|
/** @var ProcessCollection $collection */ |
122
|
|
|
$collection = GeneralUtility::makeInstance(ProcessCollection::class); |
123
|
5 |
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable($this->tableName); |
124
|
5 |
|
|
125
|
|
|
$statement = $queryBuilder |
126
|
|
|
->select('*') |
127
|
5 |
|
->from($this->tableName) |
128
|
5 |
|
->where( |
129
|
5 |
|
$queryBuilder->expr()->eq('active', 1), |
130
|
5 |
|
$queryBuilder->expr()->eq('deleted', 0) |
131
|
5 |
|
) |
132
|
|
|
->orderBy('ttl', 'DESC') |
133
|
5 |
|
->execute(); |
134
|
5 |
|
|
135
|
|
|
while ($row = $statement->fetch()) { |
136
|
5 |
|
$process = new Process(); |
137
|
3 |
|
$process->setProcessId($row['process_id']); |
138
|
3 |
|
$process->setActive($row['active']); |
139
|
3 |
|
$process->setTtl($row['ttl']); |
140
|
3 |
|
$process->setAssignedItemsCount($row['assigned_items_count']); |
141
|
3 |
|
$process->setDeleted($row['deleted']); |
142
|
3 |
|
$process->setSystemProcessId($row['system_process_id']); |
143
|
3 |
|
$collection->append($process); |
144
|
3 |
|
} |
145
|
|
|
|
146
|
|
|
return $collection; |
147
|
5 |
|
} |
148
|
|
|
|
149
|
|
|
/** |
150
|
|
|
* @param string $processId |
151
|
|
|
*/ |
152
|
|
|
public function removeByProcessId($processId): void |
153
|
4 |
|
{ |
154
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable($this->tableName); |
155
|
4 |
|
|
156
|
|
|
$queryBuilder |
157
|
|
|
->delete($this->tableName) |
158
|
4 |
|
->where( |
159
|
4 |
|
$queryBuilder->expr()->eq('process_id', $queryBuilder->createNamedParameter($processId, PDO::PARAM_STR)) |
160
|
4 |
|
)->execute(); |
161
|
4 |
|
} |
162
|
4 |
|
|
163
|
|
|
/** |
164
|
|
|
* Returns the number of active processes. |
165
|
|
|
* |
166
|
|
|
* @return int |
167
|
|
|
* @deprecated Using ProcessRepository->countActive() is deprecated since 9.1.1 and will be removed in v11.x, please use ProcessRepository->findAllActive->count() instead |
168
|
|
|
* @codeCoverageIgnore |
169
|
|
|
*/ |
170
|
|
|
public function countActive() |
171
|
|
|
{ |
172
|
|
|
return $this->findAllActive()->count(); |
173
|
|
|
} |
174
|
|
|
|
175
|
|
|
/** |
176
|
|
|
* @return array|null |
177
|
|
|
* |
178
|
|
|
* Function is moved from ProcessCleanUpHook |
179
|
|
|
* TODO: Check why we need both getActiveProcessesOlderThanOneHour and getActiveOrphanProcesses, the get getActiveOrphanProcesses does not really check for Orphan in this implementation. |
180
|
|
|
*/ |
181
|
|
|
public function getActiveProcessesOlderThanOneHour() |
182
|
3 |
|
{ |
183
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable($this->tableName); |
184
|
3 |
|
$activeProcesses = []; |
185
|
3 |
|
$statement = $queryBuilder |
186
|
|
|
->select('process_id', 'system_process_id') |
187
|
3 |
|
->from($this->tableName) |
188
|
3 |
|
->where( |
189
|
3 |
|
$queryBuilder->expr()->lte('ttl', intval(time() - $this->extensionSettings['processMaxRunTime'] - 3600)), |
190
|
3 |
|
$queryBuilder->expr()->eq('active', 1) |
191
|
3 |
|
) |
192
|
|
|
->execute(); |
193
|
3 |
|
|
194
|
|
|
while ($row = $statement->fetch()) { |
195
|
3 |
|
$activeProcesses[] = $row; |
196
|
1 |
|
} |
197
|
|
|
|
198
|
|
|
return $activeProcesses; |
199
|
3 |
|
} |
200
|
|
|
|
201
|
|
|
/** |
202
|
|
|
* Function is moved from ProcessCleanUpHook |
203
|
|
|
* |
204
|
|
|
* @return array |
205
|
|
|
* @see getActiveProcessesOlderThanOneHour |
206
|
|
|
*/ |
207
|
|
|
public function getActiveOrphanProcesses() |
208
|
3 |
|
{ |
209
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable($this->tableName); |
210
|
3 |
|
|
211
|
|
|
return $queryBuilder |
212
|
|
|
->select('process_id', 'system_process_id') |
213
|
3 |
|
->from($this->tableName) |
214
|
3 |
|
->where( |
215
|
3 |
|
$queryBuilder->expr()->lte('ttl', intval(time() - $this->extensionSettings['processMaxRunTime'])), |
216
|
3 |
|
$queryBuilder->expr()->eq('active', 1) |
217
|
3 |
|
) |
218
|
|
|
->execute()->fetchAll(); |
219
|
3 |
|
} |
220
|
|
|
|
221
|
|
|
/** |
222
|
|
|
* Returns the number of processes that live longer than the given timestamp. |
223
|
|
|
*/ |
224
|
|
|
public function countNotTimeouted(int $ttl): int |
225
|
1 |
|
{ |
226
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable($this->tableName); |
227
|
1 |
|
|
228
|
|
|
return $queryBuilder |
229
|
|
|
->count('*') |
230
|
1 |
|
->from($this->tableName) |
231
|
1 |
|
->where( |
232
|
1 |
|
$queryBuilder->expr()->eq('deleted', 0), |
233
|
1 |
|
$queryBuilder->expr()->gt('ttl', intval($ttl)) |
234
|
1 |
|
) |
235
|
|
|
->execute() |
236
|
1 |
|
->fetchColumn(0); |
237
|
1 |
|
} |
238
|
|
|
|
239
|
|
|
/** |
240
|
|
|
* Get limit clause |
241
|
|
|
* @deprecated Using ProcessRepository::getLimitFromItemCountAndOffset() is deprecated since 9.1.1 and will be removed in v11.x, was not used, so will be removed |
242
|
|
|
*/ |
243
|
|
|
public static function getLimitFromItemCountAndOffset(int $itemCount, int $offset): string |
244
|
6 |
|
{ |
245
|
|
|
$itemCount = filter_var($itemCount, FILTER_VALIDATE_INT, ['options' => ['min_range' => 1, 'default' => 20]]); |
246
|
6 |
|
$offset = filter_var($offset, FILTER_VALIDATE_INT, ['options' => ['min_range' => 0, 'default' => 0]]); |
247
|
6 |
|
|
248
|
|
|
return $offset . ', ' . $itemCount; |
249
|
6 |
|
} |
250
|
|
|
|
251
|
|
|
public function deleteProcessesWithoutItemsAssigned(): void |
252
|
3 |
|
{ |
253
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable($this->tableName); |
254
|
3 |
|
$queryBuilder |
255
|
|
|
->delete($this->tableName) |
256
|
3 |
|
->where( |
257
|
3 |
|
$queryBuilder->expr()->eq('assigned_items_count', 0) |
258
|
3 |
|
) |
259
|
|
|
->execute(); |
260
|
3 |
|
} |
261
|
3 |
|
|
262
|
|
|
public function deleteProcessesMarkedAsDeleted(): void |
263
|
3 |
|
{ |
264
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable($this->tableName); |
265
|
3 |
|
$queryBuilder |
266
|
|
|
->delete($this->tableName) |
267
|
3 |
|
->where( |
268
|
3 |
|
$queryBuilder->expr()->eq('deleted', 1) |
269
|
3 |
|
) |
270
|
|
|
->execute(); |
271
|
3 |
|
} |
272
|
3 |
|
|
273
|
|
|
public function isProcessActive(string $processId): bool |
274
|
2 |
|
{ |
275
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable($this->tableName); |
276
|
2 |
|
$isActive = $queryBuilder |
277
|
|
|
->select('active') |
278
|
2 |
|
->from($this->tableName) |
279
|
2 |
|
->where( |
280
|
2 |
|
$queryBuilder->expr()->eq('process_id', $queryBuilder->createNamedParameter($processId)) |
281
|
2 |
|
) |
282
|
|
|
->orderBy('ttl') |
283
|
2 |
|
->execute() |
284
|
2 |
|
->fetchColumn(0); |
285
|
2 |
|
|
286
|
|
|
return (bool) $isActive; |
287
|
2 |
|
} |
288
|
|
|
|
289
|
|
|
/** |
290
|
|
|
* @param $numberOfAffectedRows |
291
|
|
|
*/ |
292
|
|
|
public function updateProcessAssignItemsCount($numberOfAffectedRows, string $processId): void |
293
|
3 |
|
{ |
294
|
|
|
GeneralUtility::makeInstance(ConnectionPool::class)->getConnectionForTable($this->tableName) |
295
|
3 |
|
->update( |
296
|
3 |
|
'tx_crawler_process', |
297
|
3 |
|
['assigned_items_count' => (int) $numberOfAffectedRows], |
298
|
3 |
|
['process_id' => $processId] |
299
|
3 |
|
); |
300
|
|
|
} |
301
|
3 |
|
|
302
|
|
|
public function markRequestedProcessesAsNotActive(array $processIds): void |
303
|
3 |
|
{ |
304
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable($this->tableName); |
305
|
3 |
|
$queryBuilder->update('tx_crawler_process') |
306
|
3 |
|
->where( |
307
|
3 |
|
'NOT EXISTS ( |
308
|
3 |
|
SELECT * FROM tx_crawler_queue |
309
|
|
|
WHERE tx_crawler_queue.process_id = tx_crawler_process.process_id |
310
|
|
|
AND tx_crawler_queue.exec_time = 0 |
311
|
|
|
)', |
312
|
|
|
$queryBuilder->expr()->in('process_id', $queryBuilder->createNamedParameter($processIds, Connection::PARAM_STR_ARRAY)), |
313
|
3 |
|
$queryBuilder->expr()->eq('deleted', 0) |
314
|
3 |
|
) |
315
|
|
|
->set('active', 0) |
316
|
3 |
|
->execute(); |
317
|
3 |
|
} |
318
|
3 |
|
|
319
|
|
|
public function addProcess(string $processId, int $systemProcessId): void |
320
|
|
|
{ |
321
|
|
|
GeneralUtility::makeInstance(ConnectionPool::class)->getConnectionForTable($this->tableName)->insert( |
322
|
|
|
'tx_crawler_process', |
323
|
|
|
[ |
324
|
|
|
'process_id' => $processId, |
325
|
|
|
'active' => 1, |
326
|
|
|
'ttl' => time() + (int) $this->extensionSettings['processMaxRunTime'], |
327
|
|
|
'system_process_id' => $systemProcessId, |
328
|
|
|
] |
329
|
|
|
); |
330
|
|
|
} |
331
|
|
|
} |
332
|
|
|
|
In the issue above, the returned value is violating the contract defined by the mentioned interface.
Let's take a look at an example: