maintenance/rebuildrecentchanges.php

   1 <?php
   2 /**
   3  * Rebuild recent changes from scratch.  This takes several hours,
   4  * depending on the database size and server configuration.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License along
  17  * with this program; if not, write to the Free Software Foundation, Inc.,
  18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  19  * http://www.gnu.org/copyleft/gpl.html
  20  *
  21  * @file
  22  * @ingroup Maintenance
  23  * @todo Document
  24  */
  25
  26 require_once __DIR__ . '/Maintenance.php';
  27
  28 use MediaWiki\MediaWikiServices;
  29 use Wikimedia\Rdbms\ILBFactory;
  30
  31 /**
  32  * Maintenance script that rebuilds recent changes from scratch.
  33  *
  34  * @ingroup Maintenance
  35  */
  36 class RebuildRecentchanges extends Maintenance {
  37         /** @var int UNIX timestamp */
  38         private $cutoffFrom;
  39         /** @var int UNIX timestamp */
  40         private $cutoffTo;
  41
  42         public function __construct() {
  43                 parent::__construct();
  44                 $this->addDescription( 'Rebuild recent changes' );
  45
  46                 $this->addOption(
  47                         'from',
  48                         "Only rebuild rows in requested time range (in YYYYMMDDHHMMSS format)",
  49                         false,
  50                         true
  51                 );
  52                 $this->addOption(
  53                         'to',
  54                         "Only rebuild rows in requested time range (in YYYYMMDDHHMMSS format)",
  55                         false,
  56                         true
  57                 );
  58                 $this->setBatchSize( 200 );
  59         }
  60
  61         public function execute() {
  62                 if (
  63                         ( $this->hasOption( 'from' ) && !$this->hasOption( 'to' ) ) ||
  64                         ( !$this->hasOption( 'from' ) && $this->hasOption( 'to' ) )
  65                 ) {
  66                         $this->fatalError( "Both 'from' and 'to' must be given, or neither" );
  67                 }
  68
  69                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
  70                 $this->rebuildRecentChangesTablePass1( $lbFactory );
  71                 $this->rebuildRecentChangesTablePass2( $lbFactory );
  72                 $this->rebuildRecentChangesTablePass3( $lbFactory );
  73                 $this->rebuildRecentChangesTablePass4( $lbFactory );
  74                 $this->rebuildRecentChangesTablePass5( $lbFactory );
  75                 if ( !( $this->hasOption( 'from' ) && $this->hasOption( 'to' ) ) ) {
  76                         $this->purgeFeeds();
  77                 }
  78                 $this->output( "Done.\n" );
  79         }
  80
  81         /**
  82          * Rebuild pass 1: Insert `recentchanges` entries for page revisions.
  83          */
  84         private function rebuildRecentChangesTablePass1( ILBFactory $lbFactory ) {
  85                 $dbw = $this->getDB( DB_MASTER );
  86                 $commentStore = CommentStore::getStore();
  87
  88                 if ( $this->hasOption( 'from' ) && $this->hasOption( 'to' ) ) {
  89                         $this->cutoffFrom = wfTimestamp( TS_UNIX, $this->getOption( 'from' ) );
  90                         $this->cutoffTo = wfTimestamp( TS_UNIX, $this->getOption( 'to' ) );
  91
  92                         $sec = $this->cutoffTo - $this->cutoffFrom;
  93                         $days = $sec / 24 / 3600;
  94                         $this->output( "Rebuilding range of $sec seconds ($days days)\n" );
  95                 } else {
  96                         global $wgRCMaxAge;
  97
  98                         $days = $wgRCMaxAge / 24 / 3600;
  99                         $this->output( "Rebuilding \$wgRCMaxAge=$wgRCMaxAge seconds ($days days)\n" );
 100
 101                         $this->cutoffFrom = time() - $wgRCMaxAge;
 102                         $this->cutoffTo = time();
 103                 }
 104
 105                 $this->output( "Clearing recentchanges table for time range...\n" );
 106                 $rcids = $dbw->selectFieldValues(
 107                         'recentchanges',
 108                         'rc_id',
 109                         [
 110                                 'rc_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 111                                 'rc_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) )
 112                         ]
 113                 );
 114                 foreach ( array_chunk( $rcids, $this->getBatchSize() ) as $rcidBatch ) {
 115                         $dbw->delete( 'recentchanges', [ 'rc_id' => $rcidBatch ], __METHOD__ );
 116                         $lbFactory->waitForReplication();
 117                 }
 118
 119                 $this->output( "Loading from page and revision tables...\n" );
 120
 121                 $commentQuery = $commentStore->getJoin( 'rev_comment' );
 122                 $actorQuery = ActorMigration::newMigration()->getJoin( 'rev_user' );
 123                 $res = $dbw->select(
 124                         [ 'revision', 'page' ] + $commentQuery['tables'] + $actorQuery['tables'],
 125                         [
 126                                 'rev_timestamp',
 127                                 'rev_minor_edit',
 128                                 'rev_id',
 129                                 'rev_deleted',
 130                                 'page_namespace',
 131                                 'page_title',
 132                                 'page_is_new',
 133                                 'page_id'
 134                         ] + $commentQuery['fields'] + $actorQuery['fields'],
 135                         [
 136                                 'rev_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 137                                 'rev_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) )
 138                         ],
 139                         __METHOD__,
 140                         [ 'ORDER BY' => 'rev_timestamp DESC' ],
 141                         [
 142                                 'page' => [ 'JOIN', 'rev_page=page_id' ],
 143                         ] + $commentQuery['joins'] + $actorQuery['joins']
 144                 );
 145
 146                 $this->output( "Inserting from page and revision tables...\n" );
 147                 $inserted = 0;
 148                 $actorMigration = ActorMigration::newMigration();
 149                 foreach ( $res as $row ) {
 150                         $comment = $commentStore->getComment( 'rev_comment', $row );
 151                         $user = User::newFromAnyId( $row->rev_user, $row->rev_user_text, $row->rev_actor );
 152                         $dbw->insert(
 153                                 'recentchanges',
 154                                 [
 155                                         'rc_timestamp' => $row->rev_timestamp,
 156                                         'rc_namespace' => $row->page_namespace,
 157                                         'rc_title' => $row->page_title,
 158                                         'rc_minor' => $row->rev_minor_edit,
 159                                         'rc_bot' => 0,
 160                                         'rc_new' => $row->page_is_new,
 161                                         'rc_cur_id' => $row->page_id,
 162                                         'rc_this_oldid' => $row->rev_id,
 163                                         'rc_last_oldid' => 0, // is this ok?
 164                                         'rc_type' => $row->page_is_new ? RC_NEW : RC_EDIT,
 165                                         'rc_source' => $row->page_is_new ? RecentChange::SRC_NEW : RecentChange::SRC_EDIT,
 166                                         'rc_deleted' => $row->rev_deleted
 167                                 ] + $commentStore->insert( $dbw, 'rc_comment', $comment )
 168                                         + $actorMigration->getInsertValues( $dbw, 'rc_user', $user ),
 169                                 __METHOD__
 170                         );
 171                         if ( ( ++$inserted % $this->getBatchSize() ) == 0 ) {
 172                                 $lbFactory->waitForReplication();
 173                         }
 174                 }
 175         }
 176
 177         /**
 178          * Rebuild pass 2: Enhance entries for page revisions with references to the previous revision
 179          * (rc_last_oldid, rc_new etc.) and size differences (rc_old_len, rc_new_len).
 180          */
 181         private function rebuildRecentChangesTablePass2( ILBFactory $lbFactory ) {
 182                 $dbw = $this->getDB( DB_MASTER );
 183
 184                 $this->output( "Updating links and size differences...\n" );
 185
 186                 # Fill in the rc_last_oldid field, which points to the previous edit
 187                 $res = $dbw->select(
 188                         'recentchanges',
 189                         [ 'rc_cur_id', 'rc_this_oldid', 'rc_timestamp' ],
 190                         [
 191                                 "rc_timestamp > " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 192                                 "rc_timestamp < " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) )
 193                         ],
 194                         __METHOD__,
 195                         [ 'ORDER BY' => 'rc_cur_id,rc_timestamp' ]
 196                 );
 197
 198                 $lastCurId = 0;
 199                 $lastOldId = 0;
 200                 $lastSize = null;
 201                 $updated = 0;
 202                 foreach ( $res as $obj ) {
 203                         $new = 0;
 204
 205                         if ( $obj->rc_cur_id != $lastCurId ) {
 206                                 # Switch! Look up the previous last edit, if any
 207                                 $lastCurId = intval( $obj->rc_cur_id );
 208                                 $emit = $obj->rc_timestamp;
 209
 210                                 $row = $dbw->selectRow(
 211                                         'revision',
 212                                         [ 'rev_id', 'rev_len' ],
 213                                         [ 'rev_page' => $lastCurId, "rev_timestamp < " . $dbw->addQuotes( $emit ) ],
 214                                         __METHOD__,
 215                                         [ 'ORDER BY' => 'rev_timestamp DESC' ]
 216                                 );
 217                                 if ( $row ) {
 218                                         $lastOldId = intval( $row->rev_id );
 219                                         # Grab the last text size if available
 220                                         $lastSize = !is_null( $row->rev_len ) ? intval( $row->rev_len ) : null;
 221                                 } else {
 222                                         # No previous edit
 223                                         $lastOldId = 0;
 224                                         $lastSize = 0;
 225                                         $new = 1; // probably true
 226                                 }
 227                         }
 228
 229                         if ( $lastCurId == 0 ) {
 230                                 $this->output( "Uhhh, something wrong? No curid\n" );
 231                         } else {
 232                                 # Grab the entry's text size
 233                                 $size = (int)$dbw->selectField(
 234                                         'revision',
 235                                         'rev_len',
 236                                         [ 'rev_id' => $obj->rc_this_oldid ],
 237                                         __METHOD__
 238                                 );
 239
 240                                 $dbw->update(
 241                                         'recentchanges',
 242                                         [
 243                                                 'rc_last_oldid' => $lastOldId,
 244                                                 'rc_new' => $new,
 245                                                 'rc_type' => $new ? RC_NEW : RC_EDIT,
 246                                                 'rc_source' => $new === 1 ? RecentChange::SRC_NEW : RecentChange::SRC_EDIT,
 247                                                 'rc_old_len' => $lastSize,
 248                                                 'rc_new_len' => $size,
 249                                         ],
 250                                         [
 251                                                 'rc_cur_id' => $lastCurId,
 252                                                 'rc_this_oldid' => $obj->rc_this_oldid,
 253                                                 'rc_timestamp' => $obj->rc_timestamp // index usage
 254                                         ],
 255                                         __METHOD__
 256                                 );
 257
 258                                 $lastOldId = intval( $obj->rc_this_oldid );
 259                                 $lastSize = $size;
 260
 261                                 if ( ( ++$updated % $this->getBatchSize() ) == 0 ) {
 262                                         $lbFactory->waitForReplication();
 263                                 }
 264                         }
 265                 }
 266         }
 267
 268         /**
 269          * Rebuild pass 3: Insert `recentchanges` entries for action logs.
 270          */
 271         private function rebuildRecentChangesTablePass3( ILBFactory $lbFactory ) {
 272                 global $wgLogRestrictions, $wgFilterLogTypes;
 273
 274                 $dbw = $this->getDB( DB_MASTER );
 275                 $commentStore = CommentStore::getStore();
 276                 $nonRCLogs = array_merge( array_keys( $wgLogRestrictions ),
 277                         array_keys( $wgFilterLogTypes ),
 278                         [ 'create' ] );
 279
 280                 $this->output( "Loading from user, page, and logging tables...\n" );
 281
 282                 $commentQuery = $commentStore->getJoin( 'log_comment' );
 283                 $actorQuery = ActorMigration::newMigration()->getJoin( 'log_user' );
 284                 $res = $dbw->select(
 285                         [ 'logging', 'page' ] + $commentQuery['tables'] + $actorQuery['tables'],
 286                         [
 287                                 'log_timestamp',
 288                                 'log_namespace',
 289                                 'log_title',
 290                                 'page_id',
 291                                 'log_type',
 292                                 'log_action',
 293                                 'log_id',
 294                                 'log_params',
 295                                 'log_deleted'
 296                         ] + $commentQuery['fields'] + $actorQuery['fields'],
 297                         [
 298                                 'log_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 299                                 'log_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) ),
 300                                 // Some logs don't go in RC since they are private, or are included in the filterable log types.
 301                                 'log_type' => array_diff( LogPage::validTypes(), $nonRCLogs ),
 302                         ],
 303                         __METHOD__,
 304                         [ 'ORDER BY' => 'log_timestamp DESC' ],
 305                         [
 306                                 'page' =>
 307                                         [ 'LEFT JOIN', [ 'log_namespace=page_namespace', 'log_title=page_title' ] ]
 308                         ] + $commentQuery['joins'] + $actorQuery['joins']
 309                 );
 310
 311                 $field = $dbw->fieldInfo( 'recentchanges', 'rc_cur_id' );
 312
 313                 $inserted = 0;
 314                 $actorMigration = ActorMigration::newMigration();
 315                 foreach ( $res as $row ) {
 316                         $comment = $commentStore->getComment( 'log_comment', $row );
 317                         $user = User::newFromAnyId( $row->log_user, $row->log_user_text, $row->log_actor );
 318                         $dbw->insert(
 319                                 'recentchanges',
 320                                 [
 321                                         'rc_timestamp' => $row->log_timestamp,
 322                                         'rc_namespace' => $row->log_namespace,
 323                                         'rc_title' => $row->log_title,
 324                                         'rc_minor' => 0,
 325                                         'rc_bot' => 0,
 326                                         'rc_patrolled' => 1,
 327                                         'rc_new' => 0,
 328                                         'rc_this_oldid' => 0,
 329                                         'rc_last_oldid' => 0,
 330                                         'rc_type' => RC_LOG,
 331                                         'rc_source' => RecentChange::SRC_LOG,
 332                                         'rc_cur_id' => $field->isNullable()
 333                                                 ? $row->page_id
 334                                                 : (int)$row->page_id, // NULL => 0,
 335                                         'rc_log_type' => $row->log_type,
 336                                         'rc_log_action' => $row->log_action,
 337                                         'rc_logid' => $row->log_id,
 338                                         'rc_params' => $row->log_params,
 339                                         'rc_deleted' => $row->log_deleted
 340                                 ] + $commentStore->insert( $dbw, 'rc_comment', $comment )
 341                                         + $actorMigration->getInsertValues( $dbw, 'rc_user', $user ),
 342                                 __METHOD__
 343                         );
 344
 345                         if ( ( ++$inserted % $this->getBatchSize() ) == 0 ) {
 346                                 $lbFactory->waitForReplication();
 347                         }
 348                 }
 349         }
 350
 351         /**
 352          * Rebuild pass 4: Mark bot and autopatrolled entries.
 353          */
 354         private function rebuildRecentChangesTablePass4( ILBFactory $lbFactory ) {
 355                 global $wgUseRCPatrol, $wgMiserMode;
 356
 357                 $dbw = $this->getDB( DB_MASTER );
 358
 359                 $userQuery = User::getQueryInfo();
 360
 361                 # @FIXME: recognize other bot account groups (not the same as users with 'bot' rights)
 362                 # @NOTE: users with 'bot' rights choose when edits are bot edits or not. That information
 363                 # may be lost at this point (aside from joining on the patrol log table entries).
 364                 $botgroups = [ 'bot' ];
 365                 $autopatrolgroups = $wgUseRCPatrol ? User::getGroupsWithPermission( 'autopatrol' ) : [];
 366
 367                 # Flag our recent bot edits
 368                 if ( $botgroups ) {
 369                         $this->output( "Flagging bot account edits...\n" );
 370
 371                         # Find all users that are bots
 372                         $res = $dbw->select(
 373                                 array_merge( [ 'user_groups' ], $userQuery['tables'] ),
 374                                 $userQuery['fields'],
 375                                 [ 'ug_group' => $botgroups ],
 376                                 __METHOD__,
 377                                 [ 'DISTINCT' ],
 378                                 [ 'user_group' => [ 'JOIN', 'user_id = ug_user' ] ] + $userQuery['joins']
 379                         );
 380
 381                         $botusers = [];
 382                         foreach ( $res as $obj ) {
 383                                 $botusers[] = User::newFromRow( $obj );
 384                         }
 385
 386                         # Fill in the rc_bot field
 387                         if ( $botusers ) {
 388                                 $actorQuery = ActorMigration::newMigration()->getWhere( $dbw, 'rc_user', $botusers, false );
 389                                 $rcids = [];
 390                                 foreach ( $actorQuery['orconds'] as $cond ) {
 391                                         $rcids = array_merge( $rcids, $dbw->selectFieldValues(
 392                                                 [ 'recentchanges' ] + $actorQuery['tables'],
 393                                                 'rc_id',
 394                                                 [
 395                                                         "rc_timestamp > " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 396                                                         "rc_timestamp < " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) ),
 397                                                         $cond,
 398                                                 ],
 399                                                 __METHOD__,
 400                                                 [],
 401                                                 $actorQuery['joins']
 402                                         ) );
 403                                 }
 404                                 $rcids = array_values( array_unique( $rcids ) );
 405
 406                                 foreach ( array_chunk( $rcids, $this->getBatchSize() ) as $rcidBatch ) {
 407                                         $dbw->update(
 408                                                 'recentchanges',
 409                                                 [ 'rc_bot' => 1 ],
 410                                                 [ 'rc_id' => $rcidBatch ],
 411                                                 __METHOD__
 412                                         );
 413                                         $lbFactory->waitForReplication();
 414                                 }
 415                         }
 416                 }
 417
 418                 # Flag our recent autopatrolled edits
 419                 if ( !$wgMiserMode && $autopatrolgroups ) {
 420                         $patrolusers = [];
 421
 422                         $this->output( "Flagging auto-patrolled edits...\n" );
 423
 424                         # Find all users in RC with autopatrol rights
 425                         $res = $dbw->select(
 426                                 array_merge( [ 'user_groups' ], $userQuery['tables'] ),
 427                                 $userQuery['fields'],
 428                                 [ 'ug_group' => $autopatrolgroups ],
 429                                 __METHOD__,
 430                                 [ 'DISTINCT' ],
 431                                 [ 'user_group' => [ 'JOIN', 'user_id = ug_user' ] ] + $userQuery['joins']
 432                         );
 433
 434                         foreach ( $res as $obj ) {
 435                                 $patrolusers[] = User::newFromRow( $obj );
 436                         }
 437
 438                         # Fill in the rc_patrolled field
 439                         if ( $patrolusers ) {
 440                                 $actorQuery = ActorMigration::newMigration()->getWhere( $dbw, 'rc_user', $patrolusers, false );
 441                                 foreach ( $actorQuery['orconds'] as $cond ) {
 442                                         $dbw->update(
 443                                                 'recentchanges',
 444                                                 [ 'rc_patrolled' => 1 ],
 445                                                 [
 446                                                         $cond,
 447                                                         'rc_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 448                                                         'rc_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) ),
 449                                                 ],
 450                                                 __METHOD__
 451                                         );
 452                                         $lbFactory->waitForReplication();
 453                                 }
 454                         }
 455                 }
 456         }
 457
 458         /**
 459          * Rebuild pass 5: Delete duplicate entries where we generate both a page revision and a log entry
 460          * for a single action (upload only, at the moment, but potentially also move, protect, ...).
 461          */
 462         private function rebuildRecentChangesTablePass5( ILBFactory $lbFactory ) {
 463                 $dbw = wfGetDB( DB_MASTER );
 464
 465                 $this->output( "Removing duplicate revision and logging entries...\n" );
 466
 467                 $res = $dbw->select(
 468                         [ 'logging', 'log_search' ],
 469                         [ 'ls_value', 'ls_log_id' ],
 470                         [
 471                                 'ls_log_id = log_id',
 472                                 'ls_field' => 'associated_rev_id',
 473                                 'log_type' => 'upload',
 474                                 'log_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 475                                 'log_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) ),
 476                         ],
 477                         __METHOD__
 478                 );
 479
 480                 $updates = 0;
 481                 foreach ( $res as $obj ) {
 482                         $rev_id = $obj->ls_value;
 483                         $log_id = $obj->ls_log_id;
 484
 485                         // Mark the logging row as having an associated rev id
 486                         $dbw->update(
 487                                 'recentchanges',
 488                                 /*SET*/ [ 'rc_this_oldid' => $rev_id ],
 489                                 /*WHERE*/ [ 'rc_logid' => $log_id ],
 490                                 __METHOD__
 491                         );
 492
 493                         // Delete the revision row
 494                         $dbw->delete(
 495                                 'recentchanges',
 496                                 /*WHERE*/ [ 'rc_this_oldid' => $rev_id, 'rc_logid' => 0 ],
 497                                 __METHOD__
 498                         );
 499
 500                         if ( ( ++$updates % $this->getBatchSize() ) == 0 ) {
 501                                 $lbFactory->waitForReplication();
 502                         }
 503                 }
 504         }
 505
 506         /**
 507          * Purge cached feeds in $wanCache
 508          */
 509         private function purgeFeeds() {
 510                 global $wgFeedClasses;
 511
 512                 $this->output( "Deleting feed timestamps.\n" );
 513
 514                 $wanCache = MediaWikiServices::getInstance()->getMainWANObjectCache();
 515                 foreach ( $wgFeedClasses as $feed => $className ) {
 516                         $wanCache->delete( $wanCache->makeKey( 'rcfeed', $feed, 'timestamp' ) ); # Good enough for now.
 517                 }
 518         }
 519 }
 520
 521 $maintClass = RebuildRecentchanges::class;
 522 require_once RUN_MAINTENANCE_IF_MAIN;