|
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
|
|
|
public const TABLE_NAME = 'tx_crawler_process'; |
|
48
|
|
|
|
|
49
|
|
|
/** |
|
50
|
|
|
* @var string |
|
51
|
|
|
* @deprecated Since v9.2.5 - This will be remove in v10 |
|
52
|
|
|
*/ |
|
53
|
|
|
protected $tableName = 'tx_crawler_process'; |
|
54
|
|
|
|
|
55
|
|
|
/** |
|
56
|
|
|
* @var QueryBuilder |
|
57
|
|
|
*/ |
|
58
|
|
|
protected $queryBuilder; |
|
59
|
|
|
|
|
60
|
|
|
/** |
|
61
|
|
|
* @var array |
|
62
|
|
|
*/ |
|
63
|
|
|
protected $extensionSettings = []; |
|
64
|
|
|
|
|
65
|
|
|
public function __construct() |
|
66
|
|
|
{ |
|
67
|
|
|
$objectManager = GeneralUtility::makeInstance(ObjectManager::class); |
|
68
|
|
|
|
|
69
|
|
|
parent::__construct($objectManager); |
|
70
|
|
|
|
|
71
|
|
|
$this->queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable(self::TABLE_NAME); |
|
72
|
|
|
|
|
73
|
|
|
/** @var ExtensionConfigurationProvider $configurationProvider */ |
|
74
|
|
|
$configurationProvider = GeneralUtility::makeInstance(ExtensionConfigurationProvider::class); |
|
75
|
|
|
$this->extensionSettings = $configurationProvider->getExtensionConfiguration(); |
|
76
|
|
|
} |
|
77
|
|
|
|
|
78
|
|
|
/** |
|
79
|
|
|
* This method is used to find all cli processes within a limit. |
|
80
|
|
|
*/ |
|
81
|
|
|
public function findAll(): ProcessCollection |
|
82
|
|
|
{ |
|
83
|
|
|
/** @var ProcessCollection $collection */ |
|
84
|
|
|
$collection = GeneralUtility::makeInstance(ProcessCollection::class); |
|
85
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable(self::TABLE_NAME); |
|
86
|
|
|
|
|
87
|
|
|
$statement = $queryBuilder |
|
88
|
|
|
->select('*') |
|
89
|
|
|
->from(self::TABLE_NAME) |
|
90
|
|
|
->orderBy('ttl', 'DESC') |
|
91
|
|
|
->execute(); |
|
92
|
|
|
|
|
93
|
|
|
while ($row = $statement->fetch()) { |
|
94
|
|
|
$process = GeneralUtility::makeInstance(Process::class); |
|
95
|
|
|
$process->setProcessId($row['process_id']); |
|
96
|
|
|
$process->setActive($row['active']); |
|
97
|
|
|
$process->setTtl($row['ttl']); |
|
98
|
|
|
$process->setAssignedItemsCount($row['assigned_items_count']); |
|
99
|
|
|
$process->setDeleted($row['deleted']); |
|
100
|
|
|
$process->setSystemProcessId($row['system_process_id']); |
|
101
|
|
|
$collection->append($process); |
|
102
|
|
|
} |
|
103
|
|
|
|
|
104
|
|
|
return $collection; |
|
|
|
|
|
|
105
|
|
|
} |
|
106
|
|
|
|
|
107
|
|
|
/** |
|
108
|
|
|
* @param string $processId |
|
109
|
|
|
*/ |
|
110
|
|
|
public function findByProcessId($processId) |
|
111
|
|
|
{ |
|
112
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable(self::TABLE_NAME); |
|
113
|
|
|
|
|
114
|
|
|
return $queryBuilder |
|
115
|
|
|
->select('*') |
|
116
|
|
|
->from(self::TABLE_NAME) |
|
117
|
|
|
->where( |
|
118
|
|
|
$queryBuilder->expr()->eq('process_id', $queryBuilder->createNamedParameter($processId, PDO::PARAM_STR)) |
|
119
|
|
|
)->execute()->fetch(0); |
|
120
|
|
|
} |
|
121
|
|
|
|
|
122
|
|
|
public function findAllActive(): ProcessCollection |
|
123
|
|
|
{ |
|
124
|
|
|
/** @var ProcessCollection $collection */ |
|
125
|
|
|
$collection = GeneralUtility::makeInstance(ProcessCollection::class); |
|
126
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable(self::TABLE_NAME); |
|
127
|
|
|
|
|
128
|
|
|
$statement = $queryBuilder |
|
129
|
|
|
->select('*') |
|
130
|
|
|
->from(self::TABLE_NAME) |
|
131
|
|
|
->where( |
|
132
|
|
|
$queryBuilder->expr()->eq('active', 1), |
|
133
|
|
|
$queryBuilder->expr()->eq('deleted', 0) |
|
134
|
|
|
) |
|
135
|
|
|
->orderBy('ttl', 'DESC') |
|
136
|
|
|
->execute(); |
|
137
|
|
|
|
|
138
|
|
|
while ($row = $statement->fetch()) { |
|
139
|
|
|
$process = new Process(); |
|
140
|
|
|
$process->setProcessId($row['process_id']); |
|
141
|
|
|
$process->setActive($row['active']); |
|
142
|
|
|
$process->setTtl($row['ttl']); |
|
143
|
|
|
$process->setAssignedItemsCount($row['assigned_items_count']); |
|
144
|
|
|
$process->setDeleted($row['deleted']); |
|
145
|
|
|
$process->setSystemProcessId($row['system_process_id']); |
|
146
|
|
|
$collection->append($process); |
|
147
|
|
|
} |
|
148
|
|
|
|
|
149
|
|
|
return $collection; |
|
150
|
|
|
} |
|
151
|
|
|
|
|
152
|
|
|
/** |
|
153
|
|
|
* @param string $processId |
|
154
|
|
|
*/ |
|
155
|
|
|
public function removeByProcessId($processId): void |
|
156
|
|
|
{ |
|
157
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable(self::TABLE_NAME); |
|
158
|
|
|
|
|
159
|
|
|
$queryBuilder |
|
160
|
|
|
->delete(self::TABLE_NAME) |
|
161
|
|
|
->where( |
|
162
|
|
|
$queryBuilder->expr()->eq('process_id', $queryBuilder->createNamedParameter($processId, PDO::PARAM_STR)) |
|
163
|
|
|
)->execute(); |
|
164
|
|
|
} |
|
165
|
|
|
|
|
166
|
|
|
/** |
|
167
|
|
|
* Returns the number of active processes. |
|
168
|
|
|
* |
|
169
|
|
|
* @return int |
|
170
|
|
|
* @deprecated Using ProcessRepository->countActive() is deprecated since 9.1.1 and will be removed in v11.x, please use ProcessRepository->findAllActive->count() instead |
|
171
|
|
|
* @codeCoverageIgnore |
|
172
|
|
|
*/ |
|
173
|
|
|
public function countActive() |
|
174
|
|
|
{ |
|
175
|
|
|
return $this->findAllActive()->count(); |
|
176
|
|
|
} |
|
177
|
|
|
|
|
178
|
|
|
/** |
|
179
|
|
|
* @return array|null |
|
180
|
|
|
* |
|
181
|
|
|
* Function is moved from ProcessCleanUpHook |
|
182
|
|
|
* TODO: Check why we need both getActiveProcessesOlderThanOneHour and getActiveOrphanProcesses, the get getActiveOrphanProcesses does not really check for Orphan in this implementation. |
|
183
|
|
|
*/ |
|
184
|
|
|
public function getActiveProcessesOlderThanOneHour() |
|
185
|
|
|
{ |
|
186
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable(self::TABLE_NAME); |
|
187
|
|
|
$activeProcesses = []; |
|
188
|
|
|
$statement = $queryBuilder |
|
189
|
|
|
->select('process_id', 'system_process_id') |
|
190
|
|
|
->from(self::TABLE_NAME) |
|
191
|
|
|
->where( |
|
192
|
|
|
$queryBuilder->expr()->lte('ttl', intval(time() - $this->extensionSettings['processMaxRunTime'] - 3600)), |
|
193
|
|
|
$queryBuilder->expr()->eq('active', 1) |
|
194
|
|
|
) |
|
195
|
|
|
->execute(); |
|
196
|
|
|
|
|
197
|
|
|
while ($row = $statement->fetch()) { |
|
198
|
|
|
$activeProcesses[] = $row; |
|
199
|
|
|
} |
|
200
|
|
|
|
|
201
|
|
|
return $activeProcesses; |
|
202
|
|
|
} |
|
203
|
|
|
|
|
204
|
|
|
/** |
|
205
|
|
|
* Function is moved from ProcessCleanUpHook |
|
206
|
|
|
* |
|
207
|
|
|
* @return array |
|
208
|
|
|
* @see getActiveProcessesOlderThanOneHour |
|
209
|
|
|
*/ |
|
210
|
|
|
public function getActiveOrphanProcesses() |
|
211
|
|
|
{ |
|
212
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable(self::TABLE_NAME); |
|
213
|
|
|
|
|
214
|
|
|
return $queryBuilder |
|
215
|
|
|
->select('process_id', 'system_process_id') |
|
216
|
|
|
->from(self::TABLE_NAME) |
|
217
|
|
|
->where( |
|
218
|
|
|
$queryBuilder->expr()->lte('ttl', intval(time() - $this->extensionSettings['processMaxRunTime'])), |
|
219
|
|
|
$queryBuilder->expr()->eq('active', 1) |
|
220
|
|
|
) |
|
221
|
|
|
->execute()->fetchAll(); |
|
222
|
|
|
} |
|
223
|
|
|
|
|
224
|
|
|
/** |
|
225
|
|
|
* Returns the number of processes that live longer than the given timestamp. |
|
226
|
|
|
*/ |
|
227
|
|
|
public function countNotTimeouted(int $ttl): int |
|
228
|
|
|
{ |
|
229
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable(self::TABLE_NAME); |
|
230
|
|
|
|
|
231
|
|
|
return $queryBuilder |
|
232
|
|
|
->count('*') |
|
233
|
|
|
->from(self::TABLE_NAME) |
|
234
|
|
|
->where( |
|
235
|
|
|
$queryBuilder->expr()->eq('deleted', 0), |
|
236
|
|
|
$queryBuilder->expr()->gt('ttl', intval($ttl)) |
|
237
|
|
|
) |
|
238
|
|
|
->execute() |
|
239
|
|
|
->fetchColumn(0); |
|
240
|
|
|
} |
|
241
|
|
|
|
|
242
|
|
|
/** |
|
243
|
|
|
* Get limit clause |
|
244
|
|
|
* @deprecated Using ProcessRepository::getLimitFromItemCountAndOffset() is deprecated since 9.1.1 and will be removed in v11.x, was not used, so will be removed |
|
245
|
|
|
*/ |
|
246
|
|
|
public static function getLimitFromItemCountAndOffset(int $itemCount, int $offset): string |
|
247
|
|
|
{ |
|
248
|
|
|
$itemCount = filter_var($itemCount, FILTER_VALIDATE_INT, ['options' => ['min_range' => 1, 'default' => 20]]); |
|
249
|
|
|
$offset = filter_var($offset, FILTER_VALIDATE_INT, ['options' => ['min_range' => 0, 'default' => 0]]); |
|
250
|
|
|
|
|
251
|
|
|
return $offset . ', ' . $itemCount; |
|
252
|
|
|
} |
|
253
|
|
|
|
|
254
|
|
|
public function deleteProcessesWithoutItemsAssigned(): void |
|
255
|
|
|
{ |
|
256
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable(self::TABLE_NAME); |
|
257
|
|
|
$queryBuilder |
|
258
|
|
|
->delete(self::TABLE_NAME) |
|
259
|
|
|
->where( |
|
260
|
|
|
$queryBuilder->expr()->eq('assigned_items_count', 0) |
|
261
|
|
|
) |
|
262
|
|
|
->execute(); |
|
263
|
|
|
} |
|
264
|
|
|
|
|
265
|
|
|
public function deleteProcessesMarkedAsDeleted(): void |
|
266
|
|
|
{ |
|
267
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable(self::TABLE_NAME); |
|
268
|
|
|
$queryBuilder |
|
269
|
|
|
->delete(self::TABLE_NAME) |
|
270
|
|
|
->where( |
|
271
|
|
|
$queryBuilder->expr()->eq('deleted', 1) |
|
272
|
|
|
) |
|
273
|
|
|
->execute(); |
|
274
|
|
|
} |
|
275
|
|
|
|
|
276
|
|
|
public function isProcessActive(string $processId): bool |
|
277
|
|
|
{ |
|
278
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable(self::TABLE_NAME); |
|
279
|
|
|
$isActive = $queryBuilder |
|
280
|
|
|
->select('active') |
|
281
|
|
|
->from(self::TABLE_NAME) |
|
282
|
|
|
->where( |
|
283
|
|
|
$queryBuilder->expr()->eq('process_id', $queryBuilder->createNamedParameter($processId)) |
|
284
|
|
|
) |
|
285
|
|
|
->orderBy('ttl') |
|
286
|
|
|
->execute() |
|
287
|
|
|
->fetchColumn(0); |
|
288
|
|
|
|
|
289
|
|
|
return (bool) $isActive; |
|
290
|
|
|
} |
|
291
|
|
|
|
|
292
|
|
|
/** |
|
293
|
|
|
* @param $numberOfAffectedRows |
|
294
|
|
|
*/ |
|
295
|
|
|
public function updateProcessAssignItemsCount($numberOfAffectedRows, string $processId): void |
|
296
|
|
|
{ |
|
297
|
|
|
GeneralUtility::makeInstance(ConnectionPool::class)->getConnectionForTable(self::TABLE_NAME) |
|
298
|
|
|
->update( |
|
299
|
|
|
self::TABLE_NAME, |
|
300
|
|
|
['assigned_items_count' => (int) $numberOfAffectedRows], |
|
301
|
|
|
['process_id' => $processId] |
|
302
|
|
|
); |
|
303
|
|
|
} |
|
304
|
|
|
|
|
305
|
|
|
public function markRequestedProcessesAsNotActive(array $processIds): void |
|
306
|
|
|
{ |
|
307
|
|
|
$queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class)->getQueryBuilderForTable(self::TABLE_NAME); |
|
308
|
|
|
$queryBuilder->update(self::TABLE_NAME) |
|
309
|
|
|
->where( |
|
310
|
|
|
'NOT EXISTS ( |
|
311
|
|
|
SELECT * FROM tx_crawler_queue |
|
312
|
|
|
WHERE tx_crawler_queue.process_id = tx_crawler_process.process_id |
|
313
|
|
|
AND tx_crawler_queue.exec_time = 0 |
|
314
|
|
|
)', |
|
315
|
|
|
$queryBuilder->expr()->in('process_id', $queryBuilder->createNamedParameter($processIds, Connection::PARAM_STR_ARRAY)), |
|
316
|
|
|
$queryBuilder->expr()->eq('deleted', 0) |
|
317
|
|
|
) |
|
318
|
|
|
->set('active', 0) |
|
319
|
|
|
->execute(); |
|
320
|
|
|
} |
|
321
|
|
|
|
|
322
|
|
|
public function addProcess(string $processId, int $systemProcessId): void |
|
323
|
|
|
{ |
|
324
|
|
|
GeneralUtility::makeInstance(ConnectionPool::class)->getConnectionForTable(self::TABLE_NAME)->insert( |
|
325
|
|
|
self::TABLE_NAME, |
|
326
|
|
|
[ |
|
327
|
|
|
'process_id' => $processId, |
|
328
|
|
|
'active' => 1, |
|
329
|
|
|
'ttl' => time() + (int) $this->extensionSettings['processMaxRunTime'], |
|
330
|
|
|
'system_process_id' => $systemProcessId, |
|
331
|
|
|
] |
|
332
|
|
|
); |
|
333
|
|
|
} |
|
334
|
|
|
} |
|
335
|
|
|
|
In the issue above, the returned value is violating the contract defined by the mentioned interface.
Let's take a look at an example: