|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
declare( strict_types = 1 ); |
|
4
|
|
|
namespace Wikibase\Repo\Maintenance; |
|
5
|
|
|
|
|
6
|
|
|
use Maintenance; |
|
7
|
|
|
use MediaWiki\MediaWikiServices; |
|
8
|
|
|
use Wikibase\DataModel\Entity\Item; |
|
9
|
|
|
use Wikibase\Lib\WikibaseSettings; |
|
10
|
|
|
use Wikibase\Repo\WikibaseRepo; |
|
11
|
|
|
use Wikimedia\Rdbms\IDatabase; |
|
12
|
|
|
use Wikimedia\Rdbms\ILBFactory; |
|
13
|
|
|
|
|
14
|
|
|
$basePath = getenv( 'MW_INSTALL_PATH' ) !== false ? getenv( 'MW_INSTALL_PATH' ) : __DIR__ . '/../../../..'; |
|
15
|
|
|
|
|
16
|
|
|
require_once $basePath . '/maintenance/Maintenance.php'; |
|
17
|
|
|
|
|
18
|
|
|
/** |
|
19
|
|
|
* Maintenance script for pruning rows belonging to deleted or redirected items |
|
20
|
|
|
* from the wb_items_per_site table. |
|
21
|
|
|
* |
|
22
|
|
|
* @license GPL-2.0-or-later |
|
23
|
|
|
* @author Marius Hoch < [email protected] > |
|
24
|
|
|
*/ |
|
25
|
|
|
class PruneItemsPerSite extends Maintenance { |
|
26
|
|
|
|
|
27
|
|
|
public function __construct() { |
|
28
|
|
|
parent::__construct(); |
|
29
|
|
|
|
|
30
|
|
|
$this->addDescription( 'Pune rows belonging to deleted or redirected Items from the wb_items_per_site table' ); |
|
31
|
|
|
|
|
32
|
|
|
$this->addOption( 'select-batch-size', "Number of table rows to scan per select (100000 by default)", false, true ); |
|
33
|
|
|
} |
|
34
|
|
|
|
|
35
|
|
|
/** |
|
36
|
|
|
* @inheritDoc |
|
37
|
|
|
*/ |
|
38
|
|
|
public function execute() { |
|
39
|
|
|
if ( !WikibaseSettings::isRepoEnabled() ) { |
|
40
|
|
|
$this->fatalError( "You need to have Wikibase enabled in order to use this maintenance script!\n\n" ); |
|
41
|
|
|
} |
|
42
|
|
|
|
|
43
|
|
|
$wikibaseRepo = WikibaseRepo::getDefaultInstance(); |
|
44
|
|
|
if ( !in_array( Item::ENTITY_TYPE, $wikibaseRepo->getLocalEntitySource()->getEntityTypes() ) ) { |
|
45
|
|
|
$this->fatalError( |
|
46
|
|
|
"This script assumes Items to be part of the local entity source." |
|
47
|
|
|
); |
|
48
|
|
|
} |
|
49
|
|
|
|
|
50
|
|
|
$itemNamespace = $wikibaseRepo->getEntityNamespaceLookup()->getEntityNamespace( Item::ENTITY_TYPE ); |
|
51
|
|
|
|
|
52
|
|
|
$loadBalancerFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory(); |
|
53
|
|
|
$selectBatchSize = (int)$this->getOption( 'select-batch-size', 100000 ); |
|
54
|
|
|
|
|
55
|
|
|
$this->prune( $loadBalancerFactory, $itemNamespace, $selectBatchSize ); |
|
56
|
|
|
} |
|
57
|
|
|
|
|
58
|
|
|
private function prune( |
|
59
|
|
|
ILBFactory $loadBalancerFactory, |
|
60
|
|
|
int $itemNamespace, |
|
61
|
|
|
int $selectBatchSize |
|
62
|
|
|
) { |
|
63
|
|
|
$dbr = $loadBalancerFactory->getMainLB()->getConnection( DB_REPLICA, [ 'vslow' ] ); |
|
64
|
|
|
$dbw = $loadBalancerFactory->getMainLB()->getConnection( DB_MASTER ); |
|
65
|
|
|
|
|
66
|
|
|
$maxIpsRowId = (int)$dbr->selectField( 'wb_items_per_site', 'MAX(ips_row_id)', '', __METHOD__ ); |
|
67
|
|
|
// Add 1%, but at least 50, to the maxIpsRowId to use, for items created during the script run |
|
68
|
|
|
$maxIpsRowId = max( $maxIpsRowId * 1.01, $maxIpsRowId + 50 ); |
|
69
|
|
|
|
|
70
|
|
|
$startRowId = (int)$dbr->selectField( 'wb_items_per_site', 'MIN(ips_row_id)', '', __METHOD__ ); |
|
71
|
|
|
while ( $startRowId < $maxIpsRowId ) { |
|
72
|
|
|
$endRowId = $startRowId + $selectBatchSize; |
|
73
|
|
|
$rowsToDelete = $this->selectInRange( $dbr, $itemNamespace, $startRowId, $endRowId ); |
|
74
|
|
|
$this->output( "Read up to ips_row_id $endRowId.\n" ); |
|
75
|
|
|
|
|
76
|
|
|
if ( $rowsToDelete ) { |
|
|
|
|
|
|
77
|
|
|
$affectedRows = $this->deleteRows( $dbw, $rowsToDelete ); |
|
78
|
|
|
$this->output( "Deleted $affectedRows rows.\n" ); |
|
79
|
|
|
$loadBalancerFactory->waitForReplication(); |
|
80
|
|
|
} |
|
81
|
|
|
|
|
82
|
|
|
$startRowId = $endRowId; |
|
83
|
|
|
} |
|
84
|
|
|
} |
|
85
|
|
|
|
|
86
|
|
|
private function selectInRange( IDatabase $dbr, int $itemNamespace, int $startRowId, int $endRowId ): array { |
|
87
|
|
|
return $dbr->selectFieldValues( |
|
88
|
|
|
[ 'wb_items_per_site', 'page' ], |
|
89
|
|
|
'ips_row_id', |
|
90
|
|
|
[ |
|
91
|
|
|
'ips_row_id >= ' . $startRowId, |
|
92
|
|
|
'ips_row_id < ' . $endRowId, |
|
93
|
|
|
'page_id IS NULL', |
|
94
|
|
|
], |
|
95
|
|
|
__METHOD__, |
|
96
|
|
|
[], |
|
97
|
|
|
[ |
|
98
|
|
|
'page' => [ |
|
99
|
|
|
'LEFT JOIN', |
|
100
|
|
|
[ |
|
101
|
|
|
'page_title = ' . $dbr->buildConcat( [ |
|
102
|
|
|
$dbr->addQuotes( "Q" ), |
|
103
|
|
|
"ips_item_id", |
|
104
|
|
|
] ), |
|
105
|
|
|
'page_namespace' => $itemNamespace, |
|
106
|
|
|
'page_is_redirect' => 0, |
|
107
|
|
|
], |
|
108
|
|
|
], |
|
109
|
|
|
] |
|
110
|
|
|
); |
|
111
|
|
|
} |
|
112
|
|
|
|
|
113
|
|
|
private function deleteRows( IDatabase $dbw, array $rowsToDelete ): int { |
|
114
|
|
|
$dbw->delete( |
|
115
|
|
|
'wb_items_per_site', |
|
116
|
|
|
[ |
|
117
|
|
|
'ips_row_id' => $rowsToDelete |
|
118
|
|
|
], |
|
119
|
|
|
__METHOD__ |
|
120
|
|
|
); |
|
121
|
|
|
|
|
122
|
|
|
return $dbw->affectedRows(); |
|
123
|
|
|
} |
|
124
|
|
|
|
|
125
|
|
|
} |
|
126
|
|
|
|
|
127
|
|
|
$maintClass = PruneItemsPerSite::class; |
|
128
|
|
|
require_once RUN_MAINTENANCE_IF_MAIN; |
|
129
|
|
|
|
wikimedia /
mediawiki-extensions-Wikibase
This check marks implicit conversions of arrays to boolean values in a comparison. While in PHP an empty array is considered to be equal (but not identical) to false, this is not always apparent.
Consider making the comparison explicit by using
empty(..)or! empty(...)instead.