MediaWiki  master
deleteAutoPatrolLogs.php
Go to the documentation of this file.
1 <?php
19 require_once __DIR__ . '/Maintenance.php';
20 
27 
28  public function __construct() {
29  parent::__construct();
30  $this->addDescription( 'Remove autopatrol logs in the logging table' );
31  $this->addOption( 'dry-run', 'Print debug info instead of actually deleting' );
32  $this->addOption(
33  'check-old',
34  'Check old patrol logs (for deleting old format autopatrols).'
35  );
36  $this->addOption(
37  'before',
38  'Timestamp to delete only before that time, all MediaWiki timestamp formats are accepted',
39  false,
40  true
41  );
42  $this->addOption(
43  'from-id',
44  'First row (log id) to start updating from',
45  false,
46  true
47  );
48  $this->addOption(
49  'sleep',
50  'Sleep time (in seconds) between every batch',
51  false,
52  true
53  );
54  $this->setBatchSize( 1000 );
55  }
56 
57  public function execute() {
58  $this->setBatchSize( $this->getOption( 'batch-size', $this->getBatchSize() ) );
59 
60  $sleep = (int)$this->getOption( 'sleep', 10 );
61  $fromId = $this->getOption( 'from-id', null );
62  $this->countDown( 5 );
63  while ( true ) {
64  if ( $this->hasOption( 'check-old' ) ) {
65  $rowsData = $this->getRowsOld( $fromId );
66  // We reached end of the table
67  if ( !$rowsData ) {
68  break;
69  }
70  $rows = $rowsData['rows'];
71  $fromId = $rowsData['lastId'];
72 
73  // There is nothing to delete in this batch
74  if ( !$rows ) {
75  continue;
76  }
77  } else {
78  $rows = $this->getRows( $fromId );
79  if ( !$rows ) {
80  break;
81  }
82  $fromId = end( $rows );
83  }
84 
85  if ( $this->hasOption( 'dry-run' ) ) {
86  $this->output( 'These rows will get deleted: ' . implode( ', ', $rows ) . "\n" );
87  } else {
88  $this->deleteRows( $rows );
89  $this->output( 'Processed up to row id ' . end( $rows ) . "\n" );
90  }
91 
92  if ( $sleep > 0 ) {
93  sleep( $sleep );
94  }
95  }
96  }
97 
98  private function getRows( $fromId ) {
99  $lb = MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancer();
100  $dbr = $lb->getConnectionRef( DB_REPLICA );
101  $before = $this->getOption( 'before', false );
102 
103  $conds = [
104  'log_type' => 'patrol',
105  'log_action' => 'autopatrol',
106  ];
107 
108  if ( $fromId ) {
109  $conds[] = 'log_id > ' . $dbr->addQuotes( $fromId );
110  }
111 
112  if ( $before ) {
113  $conds[] = 'log_timestamp < ' . $dbr->addQuotes( $dbr->timestamp( $before ) );
114  }
115 
116  return $dbr->newSelectQueryBuilder()
117  ->select( 'log_id' )
118  ->from( 'logging' )
119  ->where( $conds )
120  ->orderBy( 'log_id' )
121  ->limit( $this->getBatchSize() )
122  ->caller( __METHOD__ )
123  ->fetchFieldValues();
124  }
125 
126  private function getRowsOld( $fromId ) {
127  $lb = MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancer();
128  $dbr = $lb->getConnectionRef( DB_REPLICA );
129  $batchSize = $this->getBatchSize();
130  $before = $this->getOption( 'before', false );
131 
132  $conds = [
133  'log_type' => 'patrol',
134  'log_action' => 'patrol',
135  ];
136 
137  if ( $fromId ) {
138  $conds[] = 'log_id > ' . $dbr->addQuotes( $fromId );
139  }
140 
141  if ( $before ) {
142  $conds[] = 'log_timestamp < ' . $dbr->addQuotes( $dbr->timestamp( $before ) );
143  }
144 
145  $result = $dbr->newSelectQueryBuilder()
146  ->select( [ 'log_id', 'log_params' ] )
147  ->from( 'logging' )
148  ->where( $conds )
149  ->orderBy( 'log_id' )
150  ->limit( $batchSize )
151  ->caller( __METHOD__ )
152  ->fetchResultSet();
153 
154  $last = null;
155  $autopatrols = [];
156  foreach ( $result as $row ) {
157  $last = $row->log_id;
158  $logEntry = DatabaseLogEntry::newFromRow( $row );
159  $params = $logEntry->getParameters();
160  if ( !is_array( $params ) ) {
161  continue;
162  }
163 
164  // This logic belongs to PatrolLogFormatter::getMessageKey
165  // and LogFormatter::extractParameters the 'auto' value is logically presented as key [5].
166  // For legacy case the logical key is index + 3, meaning [2].
167  // For the modern case, the logical key is index - 1 meaning [6].
168  if ( array_key_exists( '6::auto', $params ) ) {
169  // Between 2011-2016 autopatrol logs
170  $auto = $params['6::auto'] === true;
171  } elseif ( $logEntry->isLegacy() === true && array_key_exists( 2, $params ) ) {
172  // Pre-2011 autopatrol logs
173  $auto = $params[2] === '1';
174  } else {
175  continue;
176  }
177 
178  if ( $auto ) {
179  $autopatrols[] = $row->log_id;
180  }
181  }
182 
183  if ( $last === null ) {
184  return null;
185  }
186 
187  return [ 'rows' => $autopatrols, 'lastId' => $last ];
188  }
189 
190  private function deleteRows( array $rows ) {
191  $lb = MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancer();
192  $dbw = $lb->getConnectionRef( DB_PRIMARY );
193 
194  $dbw->delete(
195  'logging',
196  [ 'log_id' => $rows ],
197  __METHOD__
198  );
199 
200  $lbFactory = MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
201  $lbFactory->waitForReplication();
202  }
203 
204 }
205 
206 $maintClass = DeleteAutoPatrolLogs::class;
207 require_once RUN_MAINTENANCE_IF_MAIN;
static newFromRow( $row)
Constructs new LogEntry from database result row.
This program is free software; you can redistribute it and/or modify it under the terms of the GNU Ge...
execute()
Do the actual work.
__construct()
Default constructor.
Abstract maintenance class for quickly writing and churning out maintenance scripts with minimal effo...
Definition: Maintenance.php:66
output( $out, $channel=null)
Throw some output to the user.
hasOption( $name)
Checks to see if a particular option was set.
countDown( $seconds)
Count down from $seconds to zero on the terminal, with a one-second pause between showing each number...
getBatchSize()
Returns batch size.
addDescription( $text)
Set the description text.
addOption( $name, $description, $required=false, $withArg=false, $shortName=false, $multiOccurrence=false)
Add a parameter to the script.
getOption( $name, $default=null)
Get an option, or return the default.
setBatchSize( $s=0)
static getInstance()
Returns the global default instance of the top level service locator.
const DB_REPLICA
Definition: defines.php:26
const DB_PRIMARY
Definition: defines.php:28