maintenance/userDupes.inc

   1 <?php
   2 # Copyright (C) 2005 Brion Vibber <brion@pobox.com>
   3 # http://www.mediawiki.org/
   4 #
   5 # This program is free software; you can redistribute it and/or modify
   6 # it under the terms of the GNU General Public License as published by
   7 # the Free Software Foundation; either version 2 of the License, or
   8 # (at your option) any later version.
   9 #
  10 # This program is distributed in the hope that it will be useful,
  11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13 # GNU General Public License for more details.
  14 #
  15 # You should have received a copy of the GNU General Public License along
  16 # with this program; if not, write to the Free Software Foundation, Inc.,
  17 # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18 # http://www.gnu.org/copyleft/gpl.html
  19
  20 /**
  21  * Look for duplicate user table entries and optionally prune them.
  22  */
  23 class UserDupes {
  24         var $db;
  25         var $reassigned;
  26         var $trimmed;
  27         var $failed;
  28
  29         function UserDupes( &$database ) {
  30                 $this->db =& $database;
  31         }
  32
  33         /**
  34          * Check if this database's user table has already had a unique
  35          * user_name index applied.
  36          * @return bool
  37          */
  38         function hasUniqueIndex() {
  39                 $fname = 'UserDupes::hasUniqueIndex';
  40                 $info = $this->db->indexInfo( 'user', 'user_name', $fname );
  41                 if( !$info ) {
  42                         echo "WARNING: doesn't seem to have user_name index at all!\n";
  43                         return false;
  44                 }
  45
  46                 # Confusingly, 'Non_unique' is 0 for *unique* indexes,
  47                 # and 1 for *non-unique* indexes. Pass the crack, MySQL,
  48                 # it's obviously some good stuff!
  49                 return ( $info->Non_unique == 0 );
  50         }
  51
  52         /**
  53          * Checks the database for duplicate user account records
  54          * and remove them in preparation for application of a unique
  55          * index on the user_name field. Returns true if the table is
  56          * clean or if duplicates have been resolved automatically.
  57          *
  58          * May return false if there are unresolvable problems.
  59          * Status information will be echo'd to stdout.
  60          *
  61          * @return bool
  62          */
  63         function clearDupes() {
  64                 return $this->checkDupes( true );
  65         }
  66
  67         /**
  68          * Checks the database for duplicate user account records
  69          * in preparation for application of a unique index on the
  70          * user_name field. Returns true if the table is clean or
  71          * if duplicates can be resolved automatically.
  72          *
  73          * Returns false if there are duplicates and resolution was
  74          * not requested. (If doing resolution, edits may be reassigned.)
  75          * Status information will be echo'd to stdout.
  76          *
  77          * @param bool $doDelete pass true to actually remove things
  78          *                       from the database; false to just check.
  79          * @return bool
  80          */
  81         function checkDupes( $doDelete = false ) {
  82                 global $wgDBname;
  83
  84                 if( $this->hasUniqueIndex() ) {
  85                         echo "$wgDBname already has a unique index on its user table.\n";
  86                         return true;
  87                 }
  88
  89                 $this->lock();
  90
  91                 echo "Checking for duplicate accounts...\n";
  92                 $dupes = $this->getDupes();
  93                 $count = count( $dupes );
  94
  95                 echo "Found $count accounts with duplicate records on $wgDBname.\n";
  96                 $this->trimmed    = 0;
  97                 $this->reassigned = 0;
  98                 $this->failed     = 0;
  99                 foreach( $dupes as $name ) {
 100                         $this->examine( $name, $doDelete );
 101                 }
 102
 103                 $this->unlock();
 104
 105                 echo "\n";
 106
 107                 if( $this->reassigned > 0 ) {
 108                         if( $doDelete ) {
 109                                 echo "$this->reassigned duplicate accounts had edits reassigned to a canonical record id.\n";
 110                         } else {
 111                                 echo "$this->reassigned duplicate accounts need to have edits reassigned.\n";
 112                         }
 113                 }
 114
 115                 if( $this->trimmed > 0 ) {
 116                         if( $doDelete ) {
 117                                 echo "$this->trimmed duplicate user records were deleted from $wgDBname.\n";
 118                         } else {
 119                                 echo "$this->trimmed duplicate user accounts were found on $wgDBname which can be removed safely.\n";
 120                         }
 121                 }
 122
 123                 if( $this->failed > 0 ) {
 124                         echo "Something terribly awry; $this->failed duplicate accounts were not removed.\n";
 125                         return false;
 126                 }
 127
 128                 if( $this->trimmed == 0 || $doDelete ) {
 129                         echo "It is now safe to apply the unique index on user_name.\n";
 130                         return true;
 131                 } else {
 132                         echo "Run this script again with the --fix option to automatically delete them.\n";
 133                         return false;
 134                 }
 135         }
 136
 137         /**
 138          * We don't want anybody to mess with our stuff...
 139          * @access private
 140          */
 141         function lock() {
 142                 $fname = 'UserDupes::lock';
 143                 if( $this->newSchema() ) {
 144                         $set = array( 'user', 'revision' );
 145                 } else {
 146                         $set = array( 'user', 'cur', 'old' );
 147                 }
 148                 $names = array_map( array( $this, 'lockTable' ), $set );
 149                 $tables = implode( ',', $names );
 150
 151                 $result = $this->db->query( "LOCK TABLES $tables", $fname );
 152         }
 153
 154         function lockTable( $table ) {
 155                 return $this->db->tableName( $table ) . ' WRITE';
 156         }
 157
 158         /**
 159          * @return bool
 160          * @access private
 161          */
 162         function newSchema() {
 163                 return class_exists( 'Revision' );
 164         }
 165
 166         /**
 167          * @access private
 168          */
 169         function unlock() {
 170                 $fname = 'UserDupes::unlock';
 171                 $result = $this->db->query( "UNLOCK TABLES", $fname );
 172         }
 173
 174         /**
 175          * Grab usernames for which multiple records are present in the database.
 176          * @return array
 177          * @access private
 178          */
 179         function getDupes() {
 180                 $fname = 'UserDupes::listDupes';
 181                 $user = $this->db->tableName( 'user' );
 182                 $result = $this->db->query(
 183                          "SELECT user_name,COUNT(*) AS n
 184                             FROM $user
 185                         GROUP BY user_name
 186                           HAVING n > 1", $fname );
 187
 188                 $list = array();
 189                 while( $row = $this->db->fetchObject( $result ) ) {
 190                         $list[] = $row->user_name;
 191                 }
 192                 $this->db->freeResult( $result );
 193
 194                 return $list;
 195         }
 196
 197         /**
 198          * Examine user records for the given name. Try to see which record
 199          * will be the one that actually gets used, then check remaining records
 200          * for edits. If the dupes have no edits, we can safely remove them.
 201          * @param string $name
 202          * @param bool $doDelete
 203          * @access private
 204          */
 205         function examine( $name, $doDelete ) {
 206                 $fname = 'UserDupes::listDupes';
 207                 $result = $this->db->select( 'user',
 208                         array( 'user_id' ),
 209                         array( 'user_name' => $name ),
 210                         $fname );
 211
 212                 $firstRow = $this->db->fetchObject( $result );
 213                 $firstId  = $firstRow->user_id;
 214                 echo "Record that will be used for '$name' is user_id=$firstId\n";
 215
 216                 while( $row = $this->db->fetchObject( $result ) ) {
 217                         $dupeId = $row->user_id;
 218                         echo "... dupe id $dupeId: ";
 219                         $edits = $this->editCount( $dupeId );
 220                         if( $edits > 0 ) {
 221                                 $this->reassigned++;
 222                                 echo "has $edits edits! ";
 223                                 if( $doDelete ) {
 224                                         $this->reassignEdits( $dupeId, $firstId );
 225                                         $newEdits = $this->editCount( $dupeId );
 226                                         if( $newEdits == 0 ) {
 227                                                 echo "confirmed cleaned. ";
 228                                         } else {
 229                                                 $this->failed++;
 230                                                 echo "WARNING! $newEdits remaining edits for $dupeId; NOT deleting user.\n";
 231                                                 continue;
 232                                         }
 233                                 } else {
 234                                         echo "(will need to reassign edits on fix)";
 235                                 }
 236                         } else {
 237                                 echo "ok, no edits. ";
 238                         }
 239                         $this->trimmed++;
 240                         if( $doDelete ) {
 241                                 $this->trimAccount( $dupeId );
 242                         }
 243                         echo "\n";
 244                 }
 245                 $this->db->freeResult( $result );
 246         }
 247
 248         /**
 249          * Count the number of edits attributed to this user.
 250          * Does not currently check log table or other things
 251          * where it might show up...
 252          * @param int $userid
 253          * @return int
 254          * @access private
 255          */
 256         function editCount( $userid ) {
 257                 if( $this->newSchema() ) {
 258                         return $this->editCountOn( 'revision', 'rev_user', $userid );
 259                 } else {
 260                         return $this->editCountOn( 'cur', 'cur_user', $userid ) +
 261                                 $this->editCountOn( 'old', 'old_user', $userid );
 262                 }
 263         }
 264
 265         /**
 266          * Count the number of hits on a given table for this account.
 267          * @param string $table
 268          * @param string $field
 269          * @param int $userid
 270          * @return int
 271          * @access private
 272          */
 273         function editCountOn( $table, $field, $userid ) {
 274                 $fname = 'UserDupes::editCountOn';
 275                 return intval( $this->db->selectField(
 276                         $table,
 277                         'COUNT(*)',
 278                         array( $field => $userid ),
 279                         $fname ) );
 280         }
 281
 282         /**
 283          * @param int $from
 284          * @param int $to
 285          * @access private
 286          */
 287         function reassignEdits( $from, $to ) {
 288                 $set = $this->newSchema()
 289                         ? array( 'revision' => 'rev_user' )
 290                         : array( 'cur' => 'cur_user', 'old' => 'old_user' );
 291                 foreach( $set as $table => $field ) {
 292                         $this->reassignEditsOn( $table, $field, $from, $to );
 293                 }
 294         }
 295
 296         /**
 297          * @param string $table
 298          * @param string $field
 299          * @param int $from
 300          * @param int $to
 301          * @access private
 302          */
 303         function reassignEditsOn( $table, $field, $from, $to ) {
 304                 $fname = 'UserDupes::reassignEditsOn';
 305                 echo "reassigning on $table... ";
 306                 $result = $this->db->update( $table,
 307                         array( $field => $to ),
 308                         array( $field => $from ),
 309                         $fname );
 310                 echo "ok. ";
 311         }
 312
 313         /**
 314          * Remove a user account line.
 315          * @param int $userid
 316          * @access private
 317          */
 318         function trimAccount( $userid ) {
 319                 $fname = 'UserDupes::trimAccount';
 320                 echo "deleting...";
 321                 $this->db->delete( 'user', array( 'user_id' => $userid ), $fname );
 322                 echo " ok";
 323         }
 324
 325 }
 326
 327
 328 ?>