MediaWiki  master
removeUnusedAccounts.php
Go to the documentation of this file.
1 <?php
28 
29 require_once __DIR__ . '/Maintenance.php';
30 
37  public function __construct() {
38  parent::__construct();
39  $this->addOption( 'delete', 'Actually delete the account' );
40  $this->addOption( 'ignore-groups', 'List of comma-separated groups to exclude', false, true );
41  $this->addOption( 'ignore-touched', 'Skip accounts touched in last N days', false, true );
42  }
43 
44  public function execute() {
45  $services = $this->getServiceContainer();
46  $userFactory = $services->getUserFactory();
47  $userGroupManager = $services->getUserGroupManager();
48  $this->output( "Remove unused accounts\n\n" );
49 
50  # Do an initial scan for inactive accounts and report the result
51  $this->output( "Checking for unused user accounts...\n" );
52  $delUser = [];
53  $delActor = [];
54  $dbr = $this->getDB( DB_REPLICA );
55  $res = $dbr->newSelectQueryBuilder()
56  ->select( [ 'user_id', 'user_name', 'user_touched', 'actor_id' ] )
57  ->from( 'user' )
58  ->leftJoin( 'actor', null, 'user_id = actor_user' )
59  ->caller( __METHOD__ )->fetchResultSet();
60  if ( $this->hasOption( 'ignore-groups' ) ) {
61  $excludedGroups = explode( ',', $this->getOption( 'ignore-groups' ) );
62  } else {
63  $excludedGroups = [];
64  }
65  $touched = $this->getOption( 'ignore-touched', "1" );
66  if ( !ctype_digit( $touched ) ) {
67  $this->fatalError( "Please put a valid positive integer on the --ignore-touched parameter." );
68  }
69  $touchedSeconds = 86400 * $touched;
70  foreach ( $res as $row ) {
71  # Check the account, but ignore it if it's within a $excludedGroups
72  # group or if it's touched within the $touchedSeconds seconds.
73  $instance = $userFactory->newFromId( $row->user_id );
74  if ( count(
75  array_intersect( $userGroupManager->getUserEffectiveGroups( $instance ), $excludedGroups ) ) == 0
76  && $this->isInactiveAccount( $instance, $row->actor_id ?? null, true )
77  && wfTimestamp( TS_UNIX, $row->user_touched ) < wfTimestamp( TS_UNIX, time() - $touchedSeconds
78  )
79  ) {
80  # Inactive; print out the name and flag it
81  $delUser[] = $row->user_id;
82  if ( isset( $row->actor_id ) && $row->actor_id ) {
83  $delActor[] = $row->actor_id;
84  }
85  $this->output( $row->user_name . "\n" );
86  }
87  }
88  $count = count( $delUser );
89  $this->output( "...found {$count}.\n" );
90 
91  # If required, go back and delete each marked account
92  if ( $count > 0 && $this->hasOption( 'delete' ) ) {
93  $this->output( "\nDeleting unused accounts..." );
94  $dbw = $this->getDB( DB_PRIMARY );
95  $dbw->delete( 'user', [ 'user_id' => $delUser ], __METHOD__ );
96  # Keep actor rows referenced from ipblocks
97  $keep = $dbw->newSelectQueryBuilder()
98  ->select( 'ipb_by_actor' )
99  ->from( 'ipblocks' )
100  ->where( [ 'ipb_by_actor' => $delActor ] )
101  ->caller( __METHOD__ )->fetchFieldValues();
102  $del = array_diff( $delActor, $keep );
103  if ( $del ) {
104  $dbw->delete( 'actor', [ 'actor_id' => $del ], __METHOD__ );
105  }
106  if ( $keep ) {
107  $dbw->update( 'actor', [ 'actor_user' => null ], [ 'actor_id' => $keep ], __METHOD__ );
108  }
109  $dbw->delete( 'user_groups', [ 'ug_user' => $delUser ], __METHOD__ );
110  $dbw->delete( 'user_former_groups', [ 'ufg_user' => $delUser ], __METHOD__ );
111  $dbw->delete( 'user_properties', [ 'up_user' => $delUser ], __METHOD__ );
112  $dbw->delete( 'logging', [ 'log_actor' => $delActor ], __METHOD__ );
113  $dbw->delete( 'recentchanges', [ 'rc_actor' => $delActor ], __METHOD__ );
114  $this->output( "done.\n" );
115  # Update the site_stats.ss_users field
116  $users = $dbw->newSelectQueryBuilder()
117  ->select( 'COUNT(*)' )
118  ->from( 'user' )
119  ->caller( __METHOD__ )->fetchField();
120  $dbw->update(
121  'site_stats',
122  [ 'ss_users' => $users ],
123  [ 'ss_row_id' => 1 ],
124  __METHOD__
125  );
126  } elseif ( $count > 0 ) {
127  $this->output( "\nRun the script again with --delete to remove them from the database.\n" );
128  }
129  $this->output( "\n" );
130  }
131 
141  private function isInactiveAccount( $user, $actor, $primary = false ) {
142  if ( $actor === null ) {
143  // There's no longer a way for a user to be active in any of
144  // these tables without having an actor ID. The only way to link
145  // to a user row is via an actor row.
146  return true;
147  }
148 
149  $dbo = $this->getDB( $primary ? DB_PRIMARY : DB_REPLICA );
150  $checks = [
151  'archive' => 'ar',
152  'image' => 'img',
153  'oldimage' => 'oi',
154  'filearchive' => 'fa'
155  // re-add when actor migration is complete
156  // 'revision' => 'rev'
157  ];
158  $count = 0;
159 
160  $this->beginTransaction( $dbo, __METHOD__ );
161  foreach ( $checks as $table => $prefix ) {
162  $count += (int)$dbo->selectField(
163  $table,
164  'COUNT(*)',
165  [ "{$prefix}_actor" => $actor ],
166  __METHOD__
167  );
168  }
169 
170  // Delete this special case when the actor migration is complete
171  $actorQuery = ActorMigration::newMigration()->getWhere( $dbo, 'rev_user', $user );
172  $count += (int)$dbo->selectField(
173  [ 'revision' ] + $actorQuery['tables'],
174  'COUNT(*)',
175  $actorQuery['conds'],
176  __METHOD__,
177  [],
178  $actorQuery['joins']
179  );
180 
181  $count += (int)$dbo->newSelectQueryBuilder()
182  ->select( 'COUNT(*)' )
183  ->from( 'logging' )
184  ->where( [ 'log_actor' => $actor, 'log_type != ' . $dbo->addQuotes( 'newusers' ) ] )
185  ->caller( __METHOD__ )->fetchField();
186 
187  $this->commitTransaction( $dbo, __METHOD__ );
188 
189  return $count == 0;
190  }
191 }
192 
193 $maintClass = RemoveUnusedAccounts::class;
194 require_once RUN_MAINTENANCE_IF_MAIN;
wfTimestamp( $outputtype=TS_UNIX, $ts=0)
Get a timestamp string in one of various formats.
Abstract maintenance class for quickly writing and churning out maintenance scripts with minimal effo...
Definition: Maintenance.php:66
getDB( $db, $groups=[], $dbDomain=false)
Returns a database to be used by current maintenance script.
beginTransaction(IDatabase $dbw, $fname)
Begin a transaction on a DB.
commitTransaction(IDatabase $dbw, $fname)
Commit the transaction on a DB handle and wait for replica DBs to catch up.
output( $out, $channel=null)
Throw some output to the user.
hasOption( $name)
Checks to see if a particular option was set.
getServiceContainer()
Returns the main service container.
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.
fatalError( $msg, $exitCode=1)
Output a message and terminate the current script.
This is not intended to be a long-term part of MediaWiki; it will be deprecated and removed once acto...
Maintenance script that removes unused user accounts from the database.
execute()
Do the actual work.
__construct()
Default constructor.
Interface for objects representing user identity.
const DB_REPLICA
Definition: defines.php:26
const DB_PRIMARY
Definition: defines.php:28