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 = MediaWikiServices::getInstance();
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->select(
56  [ 'user', 'actor' ],
57  [ 'user_id', 'user_name', 'user_touched', 'actor_id' ],
58  '',
59  __METHOD__,
60  [],
61  [ 'actor' => [ 'LEFT JOIN', 'user_id = actor_user' ] ]
62  );
63  if ( $this->hasOption( 'ignore-groups' ) ) {
64  $excludedGroups = explode( ',', $this->getOption( 'ignore-groups' ) );
65  } else {
66  $excludedGroups = [];
67  }
68  $touched = $this->getOption( 'ignore-touched', "1" );
69  if ( !ctype_digit( $touched ) ) {
70  $this->fatalError( "Please put a valid positive integer on the --ignore-touched parameter." );
71  }
72  $touchedSeconds = 86400 * $touched;
73  foreach ( $res as $row ) {
74  # Check the account, but ignore it if it's within a $excludedGroups
75  # group or if it's touched within the $touchedSeconds seconds.
76  $instance = $userFactory->newFromId( $row->user_id );
77  if ( count(
78  array_intersect( $userGroupManager->getUserEffectiveGroups( $instance ), $excludedGroups ) ) == 0
79  && $this->isInactiveAccount( $instance, $row->actor_id ?? null, true )
80  && wfTimestamp( TS_UNIX, $row->user_touched ) < wfTimestamp( TS_UNIX, time() - $touchedSeconds
81  )
82  ) {
83  # Inactive; print out the name and flag it
84  $delUser[] = $row->user_id;
85  if ( isset( $row->actor_id ) && $row->actor_id ) {
86  $delActor[] = $row->actor_id;
87  }
88  $this->output( $row->user_name . "\n" );
89  }
90  }
91  $count = count( $delUser );
92  $this->output( "...found {$count}.\n" );
93 
94  # If required, go back and delete each marked account
95  if ( $count > 0 && $this->hasOption( 'delete' ) ) {
96  $this->output( "\nDeleting unused accounts..." );
97  $dbw = $this->getDB( DB_PRIMARY );
98  $dbw->delete( 'user', [ 'user_id' => $delUser ], __METHOD__ );
99  # Keep actor rows referenced from ipblocks
100  $keep = $dbw->selectFieldValues(
101  'ipblocks', 'ipb_by_actor', [ 'ipb_by_actor' => $delActor ], __METHOD__
102  );
103  $del = array_diff( $delActor, $keep );
104  if ( $del ) {
105  $dbw->delete( 'actor', [ 'actor_id' => $del ], __METHOD__ );
106  }
107  if ( $keep ) {
108  $dbw->update( 'actor', [ 'actor_user' => null ], [ 'actor_id' => $keep ], __METHOD__ );
109  }
110  $dbw->delete( 'user_groups', [ 'ug_user' => $delUser ], __METHOD__ );
111  $dbw->delete( 'user_former_groups', [ 'ufg_user' => $delUser ], __METHOD__ );
112  $dbw->delete( 'user_properties', [ 'up_user' => $delUser ], __METHOD__ );
113  $dbw->delete( 'logging', [ 'log_actor' => $delActor ], __METHOD__ );
114  $dbw->delete( 'recentchanges', [ 'rc_actor' => $delActor ], __METHOD__ );
115  $this->output( "done.\n" );
116  # Update the site_stats.ss_users field
117  $users = $dbw->selectField( 'user', 'COUNT(*)', [], __METHOD__ );
118  $dbw->update(
119  'site_stats',
120  [ 'ss_users' => $users ],
121  [ 'ss_row_id' => 1 ],
122  __METHOD__
123  );
124  } elseif ( $count > 0 ) {
125  $this->output( "\nRun the script again with --delete to remove them from the database.\n" );
126  }
127  $this->output( "\n" );
128  }
129 
139  private function isInactiveAccount( $user, $actor, $primary = false ) {
140  if ( $actor === null ) {
141  // There's no longer a way for a user to be active in any of
142  // these tables without having an actor ID. The only way to link
143  // to a user row is via an actor row.
144  return true;
145  }
146 
147  $dbo = $this->getDB( $primary ? DB_PRIMARY : DB_REPLICA );
148  $checks = [
149  'archive' => 'ar',
150  'image' => 'img',
151  'oldimage' => 'oi',
152  'filearchive' => 'fa'
153  // re-add when actor migration is complete
154  // 'revision' => 'rev'
155  ];
156  $count = 0;
157 
158  $this->beginTransaction( $dbo, __METHOD__ );
159  foreach ( $checks as $table => $prefix ) {
160  $count += (int)$dbo->selectField(
161  $table,
162  'COUNT(*)',
163  [ "{$prefix}_actor" => $actor ],
164  __METHOD__
165  );
166  }
167 
168  // Delete this special case when the actor migration is complete
169  $actorQuery = ActorMigration::newMigration()->getWhere( $dbo, 'rev_user', $user );
170  $count += (int)$dbo->selectField(
171  [ 'revision' ] + $actorQuery['tables'],
172  'COUNT(*)',
173  $actorQuery['conds'],
174  __METHOD__,
175  [],
176  $actorQuery['joins']
177  );
178 
179  $count += (int)$dbo->selectField(
180  [ 'logging' ],
181  'COUNT(*)',
182  [
183  'log_actor' => $actor,
184  'log_type != ' . $dbo->addQuotes( 'newusers' )
185  ],
186  __METHOD__
187  );
188 
189  $this->commitTransaction( $dbo, __METHOD__ );
190 
191  return $count == 0;
192  }
193 }
194 
195 $maintClass = RemoveUnusedAccounts::class;
196 require_once RUN_MAINTENANCE_IF_MAIN;
wfTimestamp( $outputtype=TS_UNIX, $ts=0)
Get a timestamp string in one of various formats.
static newMigration()
Static constructor.
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.
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.
Service locator for MediaWiki core services.
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