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          * @param ILBFactory $lbFactory
  85          */
  86         private function rebuildRecentChangesTablePass1( ILBFactory $lbFactory ) {
  87                 $dbw = $this->getDB( DB_MASTER );
  88                 $commentStore = CommentStore::getStore();
  89
  90                 if ( $this->hasOption( 'from' ) && $this->hasOption( 'to' ) ) {
  91                         $this->cutoffFrom = wfTimestamp( TS_UNIX, $this->getOption( 'from' ) );
  92                         $this->cutoffTo = wfTimestamp( TS_UNIX, $this->getOption( 'to' ) );
  93
  94                         $sec = $this->cutoffTo - $this->cutoffFrom;
  95                         $days = $sec / 24 / 3600;
  96                         $this->output( "Rebuilding range of $sec seconds ($days days)\n" );
  97                 } else {
  98                         global $wgRCMaxAge;
  99
 100                         $days = $wgRCMaxAge / 24 / 3600;
 101                         $this->output( "Rebuilding \$wgRCMaxAge=$wgRCMaxAge seconds ($days days)\n" );
 102
 103                         $this->cutoffFrom = time() - $wgRCMaxAge;
 104                         $this->cutoffTo = time();
 105                 }
 106
 107                 $this->output( "Clearing recentchanges table for time range...\n" );
 108                 $rcids = $dbw->selectFieldValues(
 109                         'recentchanges',
 110                         'rc_id',
 111                         [
 112                                 'rc_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 113                                 'rc_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) )
 114                         ]
 115                 );
 116                 foreach ( array_chunk( $rcids, $this->getBatchSize() ) as $rcidBatch ) {
 117                         $dbw->delete( 'recentchanges', [ 'rc_id' => $rcidBatch ], __METHOD__ );
 118                         $lbFactory->waitForReplication();
 119                 }
 120
 121                 $this->output( "Loading from page and revision tables...\n" );
 122
 123                 $commentQuery = $commentStore->getJoin( 'rev_comment' );
 124                 $actorQuery = ActorMigration::newMigration()->getJoin( 'rev_user' );
 125                 $res = $dbw->select(
 126                         [ 'revision', 'page' ] + $commentQuery['tables'] + $actorQuery['tables'],
 127                         [
 128                                 'rev_timestamp',
 129                                 'rev_minor_edit',
 130                                 'rev_id',
 131                                 'rev_deleted',
 132                                 'page_namespace',
 133                                 'page_title',
 134                                 'page_is_new',
 135                                 'page_id'
 136                         ] + $commentQuery['fields'] + $actorQuery['fields'],
 137                         [
 138                                 'rev_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 139                                 'rev_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) )
 140                         ],
 141                         __METHOD__,
 142                         [ 'ORDER BY' => 'rev_timestamp DESC' ],
 143                         [
 144                                 'page' => [ 'JOIN', 'rev_page=page_id' ],
 145                         ] + $commentQuery['joins'] + $actorQuery['joins']
 146                 );
 147
 148                 $this->output( "Inserting from page and revision tables...\n" );
 149                 $inserted = 0;
 150                 $actorMigration = ActorMigration::newMigration();
 151                 foreach ( $res as $row ) {
 152                         $comment = $commentStore->getComment( 'rev_comment', $row );
 153                         $user = User::newFromAnyId( $row->rev_user, $row->rev_user_text, $row->rev_actor );
 154                         $dbw->insert(
 155                                 'recentchanges',
 156                                 [
 157                                         'rc_timestamp' => $row->rev_timestamp,
 158                                         'rc_namespace' => $row->page_namespace,
 159                                         'rc_title' => $row->page_title,
 160                                         'rc_minor' => $row->rev_minor_edit,
 161                                         'rc_bot' => 0,
 162                                         'rc_new' => $row->page_is_new,
 163                                         'rc_cur_id' => $row->page_id,
 164                                         'rc_this_oldid' => $row->rev_id,
 165                                         'rc_last_oldid' => 0, // is this ok?
 166                                         'rc_type' => $row->page_is_new ? RC_NEW : RC_EDIT,
 167                                         'rc_source' => $row->page_is_new ? RecentChange::SRC_NEW : RecentChange::SRC_EDIT,
 168                                         'rc_deleted' => $row->rev_deleted
 169                                 ] + $commentStore->insert( $dbw, 'rc_comment', $comment )
 170                                         + $actorMigration->getInsertValues( $dbw, 'rc_user', $user ),
 171                                 __METHOD__
 172                         );
 173
 174                         $rcid = $dbw->insertId();
 175                         $dbw->update(
 176                                 'change_tag',
 177                                 [ 'ct_rc_id' => $rcid ],
 178                                 [ 'ct_rev_id' => $row->rev_id ],
 179                                 __METHOD__
 180                         );
 181
 182                         if ( ( ++$inserted % $this->getBatchSize() ) == 0 ) {
 183                                 $lbFactory->waitForReplication();
 184                         }
 185                 }
 186         }
 187
 188         /**
 189          * Rebuild pass 2: Enhance entries for page revisions with references to the previous revision
 190          * (rc_last_oldid, rc_new etc.) and size differences (rc_old_len, rc_new_len).
 191          *
 192          * @param ILBFactory $lbFactory
 193          */
 194         private function rebuildRecentChangesTablePass2( ILBFactory $lbFactory ) {
 195                 $dbw = $this->getDB( DB_MASTER );
 196
 197                 $this->output( "Updating links and size differences...\n" );
 198
 199                 # Fill in the rc_last_oldid field, which points to the previous edit
 200                 $res = $dbw->select(
 201                         'recentchanges',
 202                         [ 'rc_cur_id', 'rc_this_oldid', 'rc_timestamp' ],
 203                         [
 204                                 "rc_timestamp > " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 205                                 "rc_timestamp < " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) )
 206                         ],
 207                         __METHOD__,
 208                         [ 'ORDER BY' => 'rc_cur_id,rc_timestamp' ]
 209                 );
 210
 211                 $lastCurId = 0;
 212                 $lastOldId = 0;
 213                 $lastSize = null;
 214                 $updated = 0;
 215                 foreach ( $res as $row ) {
 216                         $new = 0;
 217
 218                         if ( $row->rc_cur_id != $lastCurId ) {
 219                                 # Switch! Look up the previous last edit, if any
 220                                 $lastCurId = intval( $row->rc_cur_id );
 221                                 $emit = $row->rc_timestamp;
 222
 223                                 $revRow = $dbw->selectRow(
 224                                         'revision',
 225                                         [ 'rev_id', 'rev_len' ],
 226                                         [ 'rev_page' => $lastCurId, "rev_timestamp < " . $dbw->addQuotes( $emit ) ],
 227                                         __METHOD__,
 228                                         [ 'ORDER BY' => 'rev_timestamp DESC' ]
 229                                 );
 230                                 if ( $revRow ) {
 231                                         $lastOldId = intval( $revRow->rev_id );
 232                                         # Grab the last text size if available
 233                                         $lastSize = !is_null( $revRow->rev_len ) ? intval( $revRow->rev_len ) : null;
 234                                 } else {
 235                                         # No previous edit
 236                                         $lastOldId = 0;
 237                                         $lastSize = 0;
 238                                         $new = 1; // probably true
 239                                 }
 240                         }
 241
 242                         if ( $lastCurId == 0 ) {
 243                                 $this->output( "Uhhh, something wrong? No curid\n" );
 244                         } else {
 245                                 # Grab the entry's text size
 246                                 $size = (int)$dbw->selectField(
 247                                         'revision',
 248                                         'rev_len',
 249                                         [ 'rev_id' => $row->rc_this_oldid ],
 250                                         __METHOD__
 251                                 );
 252
 253                                 $dbw->update(
 254                                         'recentchanges',
 255                                         [
 256                                                 'rc_last_oldid' => $lastOldId,
 257                                                 'rc_new' => $new,
 258                                                 'rc_type' => $new ? RC_NEW : RC_EDIT,
 259                                                 'rc_source' => $new === 1 ? RecentChange::SRC_NEW : RecentChange::SRC_EDIT,
 260                                                 'rc_old_len' => $lastSize,
 261                                                 'rc_new_len' => $size,
 262                                         ],
 263                                         [
 264                                                 'rc_cur_id' => $lastCurId,
 265                                                 'rc_this_oldid' => $row->rc_this_oldid,
 266                                                 'rc_timestamp' => $row->rc_timestamp // index usage
 267                                         ],
 268                                         __METHOD__
 269                                 );
 270
 271                                 $lastOldId = intval( $row->rc_this_oldid );
 272                                 $lastSize = $size;
 273
 274                                 if ( ( ++$updated % $this->getBatchSize() ) == 0 ) {
 275                                         $lbFactory->waitForReplication();
 276                                 }
 277                         }
 278                 }
 279         }
 280
 281         /**
 282          * Rebuild pass 3: Insert `recentchanges` entries for action logs.
 283          *
 284          * @param ILBFactory $lbFactory
 285          */
 286         private function rebuildRecentChangesTablePass3( ILBFactory $lbFactory ) {
 287                 global $wgLogRestrictions, $wgFilterLogTypes;
 288
 289                 $dbw = $this->getDB( DB_MASTER );
 290                 $commentStore = CommentStore::getStore();
 291                 $nonRCLogs = array_merge( array_keys( $wgLogRestrictions ),
 292                         array_keys( $wgFilterLogTypes ),
 293                         [ 'create' ] );
 294
 295                 $this->output( "Loading from user and logging tables...\n" );
 296
 297                 $commentQuery = $commentStore->getJoin( 'log_comment' );
 298                 $actorQuery = ActorMigration::newMigration()->getJoin( 'log_user' );
 299                 $res = $dbw->select(
 300                         [ 'logging' ] + $commentQuery['tables'] + $actorQuery['tables'],
 301                         [
 302                                 'log_timestamp',
 303                                 'log_namespace',
 304                                 'log_title',
 305                                 'log_page',
 306                                 'log_type',
 307                                 'log_action',
 308                                 'log_id',
 309                                 'log_params',
 310                                 'log_deleted'
 311                         ] + $commentQuery['fields'] + $actorQuery['fields'],
 312                         [
 313                                 'log_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 314                                 'log_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) ),
 315                                 // Some logs don't go in RC since they are private, or are included in the filterable log types.
 316                                 'log_type' => array_diff( LogPage::validTypes(), $nonRCLogs ),
 317                         ],
 318                         __METHOD__,
 319                         [ 'ORDER BY' => 'log_timestamp DESC' ],
 320                         $commentQuery['joins'] + $actorQuery['joins']
 321                 );
 322
 323                 $field = $dbw->fieldInfo( 'recentchanges', 'rc_cur_id' );
 324
 325                 $inserted = 0;
 326                 $actorMigration = ActorMigration::newMigration();
 327                 foreach ( $res as $row ) {
 328                         $comment = $commentStore->getComment( 'log_comment', $row );
 329                         $user = User::newFromAnyId( $row->log_user, $row->log_user_text, $row->log_actor );
 330                         $dbw->insert(
 331                                 'recentchanges',
 332                                 [
 333                                         'rc_timestamp' => $row->log_timestamp,
 334                                         'rc_namespace' => $row->log_namespace,
 335                                         'rc_title' => $row->log_title,
 336                                         'rc_minor' => 0,
 337                                         'rc_bot' => 0,
 338                                         'rc_patrolled' => $row->log_type == 'upload' ? 0 : 2,
 339                                         'rc_new' => 0,
 340                                         'rc_this_oldid' => 0,
 341                                         'rc_last_oldid' => 0,
 342                                         'rc_type' => RC_LOG,
 343                                         'rc_source' => RecentChange::SRC_LOG,
 344                                         'rc_cur_id' => $field->isNullable()
 345                                                 ? $row->log_page
 346                                                 : (int)$row->log_page, // NULL => 0,
 347                                         'rc_log_type' => $row->log_type,
 348                                         'rc_log_action' => $row->log_action,
 349                                         'rc_logid' => $row->log_id,
 350                                         'rc_params' => $row->log_params,
 351                                         'rc_deleted' => $row->log_deleted
 352                                 ] + $commentStore->insert( $dbw, 'rc_comment', $comment )
 353                                         + $actorMigration->getInsertValues( $dbw, 'rc_user', $user ),
 354                                 __METHOD__
 355                         );
 356
 357                         $rcid = $dbw->insertId();
 358                         $dbw->update(
 359                                 'change_tag',
 360                                 [ 'ct_rc_id' => $rcid ],
 361                                 [ 'ct_log_id' => $row->log_id ],
 362                                 __METHOD__
 363                         );
 364
 365                         if ( ( ++$inserted % $this->getBatchSize() ) == 0 ) {
 366                                 $lbFactory->waitForReplication();
 367                         }
 368                 }
 369         }
 370
 371         /**
 372          * Rebuild pass 4: Mark bot and autopatrolled entries.
 373          *
 374          * @param ILBFactory $lbFactory
 375          */
 376         private function rebuildRecentChangesTablePass4( ILBFactory $lbFactory ) {
 377                 global $wgUseRCPatrol, $wgMiserMode;
 378
 379                 $dbw = $this->getDB( DB_MASTER );
 380
 381                 $userQuery = User::getQueryInfo();
 382
 383                 # @FIXME: recognize other bot account groups (not the same as users with 'bot' rights)
 384                 # @NOTE: users with 'bot' rights choose when edits are bot edits or not. That information
 385                 # may be lost at this point (aside from joining on the patrol log table entries).
 386                 $botgroups = [ 'bot' ];
 387                 $autopatrolgroups = $wgUseRCPatrol ? MediaWikiServices::getInstance()
 388                         ->getPermissionManager()
 389                         ->getGroupsWithPermission( 'autopatrol' ) : [];
 390
 391                 # Flag our recent bot edits
 392                 if ( $botgroups ) {
 393                         $this->output( "Flagging bot account edits...\n" );
 394
 395                         # Find all users that are bots
 396                         $res = $dbw->select(
 397                                 array_merge( [ 'user_groups' ], $userQuery['tables'] ),
 398                                 $userQuery['fields'],
 399                                 [ 'ug_group' => $botgroups ],
 400                                 __METHOD__,
 401                                 [ 'DISTINCT' ],
 402                                 [ 'user_groups' => [ 'JOIN', 'user_id = ug_user' ] ] + $userQuery['joins']
 403                         );
 404
 405                         $botusers = [];
 406                         foreach ( $res as $row ) {
 407                                 $botusers[] = User::newFromRow( $row );
 408                         }
 409
 410                         # Fill in the rc_bot field
 411                         if ( $botusers ) {
 412                                 $actorQuery = ActorMigration::newMigration()->getWhere( $dbw, 'rc_user', $botusers, false );
 413                                 $rcids = [];
 414                                 foreach ( $actorQuery['orconds'] as $cond ) {
 415                                         $rcids = array_merge( $rcids, $dbw->selectFieldValues(
 416                                                 [ 'recentchanges' ] + $actorQuery['tables'],
 417                                                 'rc_id',
 418                                                 [
 419                                                         "rc_timestamp > " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 420                                                         "rc_timestamp < " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) ),
 421                                                         $cond,
 422                                                 ],
 423                                                 __METHOD__,
 424                                                 [],
 425                                                 $actorQuery['joins']
 426                                         ) );
 427                                 }
 428                                 $rcids = array_values( array_unique( $rcids ) );
 429
 430                                 foreach ( array_chunk( $rcids, $this->getBatchSize() ) as $rcidBatch ) {
 431                                         $dbw->update(
 432                                                 'recentchanges',
 433                                                 [ 'rc_bot' => 1 ],
 434                                                 [ 'rc_id' => $rcidBatch ],
 435                                                 __METHOD__
 436                                         );
 437                                         $lbFactory->waitForReplication();
 438                                 }
 439                         }
 440                 }
 441
 442                 # Flag our recent autopatrolled edits
 443                 if ( !$wgMiserMode && $autopatrolgroups ) {
 444                         $patrolusers = [];
 445
 446                         $this->output( "Flagging auto-patrolled edits...\n" );
 447
 448                         # Find all users in RC with autopatrol rights
 449                         $res = $dbw->select(
 450                                 array_merge( [ 'user_groups' ], $userQuery['tables'] ),
 451                                 $userQuery['fields'],
 452                                 [ 'ug_group' => $autopatrolgroups ],
 453                                 __METHOD__,
 454                                 [ 'DISTINCT' ],
 455                                 [ 'user_groups' => [ 'JOIN', 'user_id = ug_user' ] ] + $userQuery['joins']
 456                         );
 457
 458                         foreach ( $res as $row ) {
 459                                 $patrolusers[] = User::newFromRow( $row );
 460                         }
 461
 462                         # Fill in the rc_patrolled field
 463                         if ( $patrolusers ) {
 464                                 $actorQuery = ActorMigration::newMigration()->getWhere( $dbw, 'rc_user', $patrolusers, false );
 465                                 foreach ( $actorQuery['orconds'] as $cond ) {
 466                                         $dbw->update(
 467                                                 'recentchanges',
 468                                                 [ 'rc_patrolled' => 2 ],
 469                                                 [
 470                                                         $cond,
 471                                                         'rc_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 472                                                         'rc_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) ),
 473                                                         'rc_patrolled' => 0
 474                                                 ],
 475                                                 __METHOD__
 476                                         );
 477                                         $lbFactory->waitForReplication();
 478                                 }
 479                         }
 480                 }
 481         }
 482
 483         /**
 484          * Rebuild pass 5: Delete duplicate entries where we generate both a page revision and a log
 485          * entry for a single action (upload only, at the moment, but potentially move, protect, ...).
 486          *
 487          * @param ILBFactory $lbFactory
 488          */
 489         private function rebuildRecentChangesTablePass5( ILBFactory $lbFactory ) {
 490                 $dbw = wfGetDB( DB_MASTER );
 491
 492                 $this->output( "Removing duplicate revision and logging entries...\n" );
 493
 494                 $res = $dbw->select(
 495                         [ 'logging', 'log_search' ],
 496                         [ 'ls_value', 'ls_log_id' ],
 497                         [
 498                                 'ls_log_id = log_id',
 499                                 'ls_field' => 'associated_rev_id',
 500                                 'log_type' => 'upload',
 501                                 'log_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 502                                 'log_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) ),
 503                         ],
 504                         __METHOD__
 505                 );
 506
 507                 $updates = 0;
 508                 foreach ( $res as $row ) {
 509                         $rev_id = $row->ls_value;
 510                         $log_id = $row->ls_log_id;
 511
 512                         // Mark the logging row as having an associated rev id
 513                         $dbw->update(
 514                                 'recentchanges',
 515                                 /*SET*/ [ 'rc_this_oldid' => $rev_id ],
 516                                 /*WHERE*/ [ 'rc_logid' => $log_id ],
 517                                 __METHOD__
 518                         );
 519
 520                         // Delete the revision row
 521                         $dbw->delete(
 522                                 'recentchanges',
 523                                 /*WHERE*/ [ 'rc_this_oldid' => $rev_id, 'rc_logid' => 0 ],
 524                                 __METHOD__
 525                         );
 526
 527                         if ( ( ++$updates % $this->getBatchSize() ) == 0 ) {
 528                                 $lbFactory->waitForReplication();
 529                         }
 530                 }
 531         }
 532
 533         /**
 534          * Purge cached feeds in $wanCache
 535          */
 536         private function purgeFeeds() {
 537                 global $wgFeedClasses;
 538
 539                 $this->output( "Deleting feed timestamps.\n" );
 540
 541                 $wanCache = MediaWikiServices::getInstance()->getMainWANObjectCache();
 542                 foreach ( $wgFeedClasses as $feed => $className ) {
 543                         $wanCache->delete( $wanCache->makeKey( 'rcfeed', $feed, 'timestamp' ) ); # Good enough for now.
 544                 }
 545         }
 546 }
 547
 548 $maintClass = RebuildRecentchanges::class;
 549 require_once RUN_MAINTENANCE_IF_MAIN;