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\IDatabase;
  30 use Wikimedia\Rdbms\ILBFactory;
  31
  32 /**
  33  * Maintenance script that rebuilds recent changes from scratch.
  34  *
  35  * @ingroup Maintenance
  36  */
  37 class RebuildRecentchanges extends Maintenance {
  38         /** @var int UNIX timestamp */
  39         private $cutoffFrom;
  40         /** @var int UNIX timestamp */
  41         private $cutoffTo;
  42
  43         public function __construct() {
  44                 parent::__construct();
  45                 $this->addDescription( 'Rebuild recent changes' );
  46
  47                 $this->addOption(
  48                         'from',
  49                         "Only rebuild rows in requested time range (in YYYYMMDDHHMMSS format)",
  50                         false,
  51                         true
  52                 );
  53                 $this->addOption(
  54                         'to',
  55                         "Only rebuild rows in requested time range (in YYYYMMDDHHMMSS format)",
  56                         false,
  57                         true
  58                 );
  59                 $this->setBatchSize( 200 );
  60         }
  61
  62         public function execute() {
  63                 if (
  64                         ( $this->hasOption( 'from' ) && !$this->hasOption( 'to' ) ) ||
  65                         ( !$this->hasOption( 'from' ) && $this->hasOption( 'to' ) )
  66                 ) {
  67                         $this->fatalError( "Both 'from' and 'to' must be given, or neither" );
  68                 }
  69
  70                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
  71                 $this->rebuildRecentChangesTablePass1( $lbFactory );
  72                 $this->rebuildRecentChangesTablePass2( $lbFactory );
  73                 $this->rebuildRecentChangesTablePass3( $lbFactory );
  74                 $this->rebuildRecentChangesTablePass4( $lbFactory );
  75                 $this->rebuildRecentChangesTablePass5( $lbFactory );
  76                 if ( !( $this->hasOption( 'from' ) && $this->hasOption( 'to' ) ) ) {
  77                         $this->purgeFeeds();
  78                 }
  79                 $this->output( "Done.\n" );
  80         }
  81
  82         /**
  83          * Rebuild pass 1: Insert `recentchanges` entries for page revisions.
  84          *
  85          * @param ILBFactory $lbFactory
  86          */
  87         private function rebuildRecentChangesTablePass1( ILBFactory $lbFactory ) {
  88                 $dbw = $this->getDB( DB_PRIMARY );
  89                 $commentStore = CommentStore::getStore();
  90
  91                 if ( $this->hasOption( 'from' ) && $this->hasOption( 'to' ) ) {
  92                         $this->cutoffFrom = (int)wfTimestamp( TS_UNIX, $this->getOption( 'from' ) );
  93                         $this->cutoffTo = (int)wfTimestamp( TS_UNIX, $this->getOption( 'to' ) );
  94
  95                         $sec = $this->cutoffTo - $this->cutoffFrom;
  96                         $days = $sec / 24 / 3600;
  97                         $this->output( "Rebuilding range of $sec seconds ($days days)\n" );
  98                 } else {
  99                         global $wgRCMaxAge;
 100
 101                         $days = $wgRCMaxAge / 24 / 3600;
 102                         $this->output( "Rebuilding \$wgRCMaxAge=$wgRCMaxAge seconds ($days days)\n" );
 103
 104                         $this->cutoffFrom = time() - $wgRCMaxAge;
 105                         $this->cutoffTo = time();
 106                 }
 107
 108                 $this->output( "Clearing recentchanges table for time range...\n" );
 109                 $rcids = $dbw->selectFieldValues(
 110                         'recentchanges',
 111                         'rc_id',
 112                         [
 113                                 'rc_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 114                                 'rc_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) )
 115                         ],
 116                         __METHOD__
 117                 );
 118                 foreach ( array_chunk( $rcids, $this->getBatchSize() ) as $rcidBatch ) {
 119                         $dbw->delete( 'recentchanges', [ 'rc_id' => $rcidBatch ], __METHOD__ );
 120                         $lbFactory->waitForReplication();
 121                 }
 122
 123                 $this->output( "Loading from page and revision tables...\n" );
 124
 125                 $commentQuery = $commentStore->getJoin( 'rev_comment' );
 126                 $actorQuery = ActorMigration::newMigration()->getJoin( 'rev_user' );
 127                 $res = $dbw->select(
 128                         [ 'revision', 'page' ] + $commentQuery['tables'] + $actorQuery['tables'],
 129                         [
 130                                 'rev_timestamp',
 131                                 'rev_minor_edit',
 132                                 'rev_id',
 133                                 'rev_deleted',
 134                                 'page_namespace',
 135                                 'page_title',
 136                                 'page_is_new',
 137                                 'page_id'
 138                         ] + $commentQuery['fields'] + $actorQuery['fields'],
 139                         [
 140                                 'rev_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 141                                 'rev_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) )
 142                         ],
 143                         __METHOD__,
 144                         [ 'ORDER BY' => 'rev_timestamp DESC' ],
 145                         [
 146                                 'page' => [ 'JOIN', 'rev_page=page_id' ],
 147                         ] + $commentQuery['joins'] + $actorQuery['joins']
 148                 );
 149
 150                 $this->output( "Inserting from page and revision tables...\n" );
 151                 $inserted = 0;
 152                 foreach ( $res as $row ) {
 153                         $comment = $commentStore->getComment( 'rev_comment', $row );
 154                         $dbw->insert(
 155                                 'recentchanges',
 156                                 [
 157                                         'rc_timestamp' => $row->rev_timestamp,
 158                                         'rc_actor' => $row->rev_actor,
 159                                         'rc_namespace' => $row->page_namespace,
 160                                         'rc_title' => $row->page_title,
 161                                         'rc_minor' => $row->rev_minor_edit,
 162                                         'rc_bot' => 0,
 163                                         'rc_new' => $row->page_is_new,
 164                                         'rc_cur_id' => $row->page_id,
 165                                         'rc_this_oldid' => $row->rev_id,
 166                                         'rc_last_oldid' => 0, // is this ok?
 167                                         'rc_type' => $row->page_is_new ? RC_NEW : RC_EDIT,
 168                                         'rc_source' => $row->page_is_new ? RecentChange::SRC_NEW : RecentChange::SRC_EDIT,
 169                                         'rc_deleted' => $row->rev_deleted
 170                                 ] + $commentStore->insert( $dbw, 'rc_comment', $comment ),
 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_PRIMARY );
 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 = $revRow->rev_len !== null ? 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_PRIMARY );
 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                 $res = $dbw->select(
 299                         [ 'logging' ] + $commentQuery['tables'],
 300                         [
 301                                 'log_timestamp',
 302                                 'log_actor',
 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'],
 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']
 321                 );
 322
 323                 $field = $dbw->fieldInfo( 'recentchanges', 'rc_cur_id' );
 324
 325                 $inserted = 0;
 326                 foreach ( $res as $row ) {
 327                         $comment = $commentStore->getComment( 'log_comment', $row );
 328                         $dbw->insert(
 329                                 'recentchanges',
 330                                 [
 331                                         'rc_timestamp' => $row->log_timestamp,
 332                                         'rc_actor' => $row->log_actor,
 333                                         'rc_namespace' => $row->log_namespace,
 334                                         'rc_title' => $row->log_title,
 335                                         'rc_minor' => 0,
 336                                         'rc_bot' => 0,
 337                                         'rc_patrolled' => $row->log_type == 'upload' ? 0 : 2,
 338                                         'rc_new' => 0,
 339                                         'rc_this_oldid' => 0,
 340                                         'rc_last_oldid' => 0,
 341                                         'rc_type' => RC_LOG,
 342                                         'rc_source' => RecentChange::SRC_LOG,
 343                                         'rc_cur_id' => $field->isNullable()
 344                                                 ? $row->log_page
 345                                                 : (int)$row->log_page, // NULL => 0,
 346                                         'rc_log_type' => $row->log_type,
 347                                         'rc_log_action' => $row->log_action,
 348                                         'rc_logid' => $row->log_id,
 349                                         'rc_params' => $row->log_params,
 350                                         'rc_deleted' => $row->log_deleted
 351                                 ] + $commentStore->insert( $dbw, 'rc_comment', $comment ),
 352                                 __METHOD__
 353                         );
 354
 355                         $rcid = $dbw->insertId();
 356                         $dbw->update(
 357                                 'change_tag',
 358                                 [ 'ct_rc_id' => $rcid ],
 359                                 [ 'ct_log_id' => $row->log_id ],
 360                                 __METHOD__
 361                         );
 362
 363                         if ( ( ++$inserted % $this->getBatchSize() ) == 0 ) {
 364                                 $lbFactory->waitForReplication();
 365                         }
 366                 }
 367         }
 368
 369         /**
 370          * Find rc_id values that have a user with one of the specified groups
 371          *
 372          * @param IDatabase $db
 373          * @param string[] $groups
 374          * @param array $conds Extra query conditions
 375          * @return int[]
 376          */
 377         private function findRcIdsWithGroups( $db, $groups, $conds = [] ) {
 378                 if ( !count( $groups ) ) {
 379                         return [];
 380                 }
 381                 return $db->selectFieldValues(
 382                         [ 'recentchanges', 'actor', 'user_groups' ],
 383                         'rc_id',
 384                         $conds + [
 385                                 "rc_timestamp > " . $db->addQuotes( $db->timestamp( $this->cutoffFrom ) ),
 386                                 "rc_timestamp < " . $db->addQuotes( $db->timestamp( $this->cutoffTo ) ),
 387                                 'ug_group' => $groups
 388                         ],
 389                         __METHOD__,
 390                         [ 'DISTINCT' ],
 391                         [
 392                                 'actor' => [ 'JOIN', 'actor_id=rc_actor' ],
 393                                 'user_groups' => [ 'JOIN', 'ug_user=actor_user' ]
 394                         ]
 395                 );
 396         }
 397
 398         /**
 399          * Rebuild pass 4: Mark bot and autopatrolled entries.
 400          *
 401          * @param ILBFactory $lbFactory
 402          */
 403         private function rebuildRecentChangesTablePass4( ILBFactory $lbFactory ) {
 404                 global $wgUseRCPatrol, $wgUseNPPatrol, $wgUseFilePatrol, $wgMiserMode;
 405
 406                 $dbw = $this->getDB( DB_PRIMARY );
 407
 408                 # @FIXME: recognize other bot account groups (not the same as users with 'bot' rights)
 409                 # @NOTE: users with 'bot' rights choose when edits are bot edits or not. That information
 410                 # may be lost at this point (aside from joining on the patrol log table entries).
 411                 $botgroups = [ 'bot' ];
 412                 $autopatrolgroups = ( $wgUseRCPatrol || $wgUseNPPatrol || $wgUseFilePatrol ) ?
 413                         MediaWikiServices::getInstance()->getGroupPermissionsLookup()
 414                         ->getGroupsWithPermission( 'autopatrol' ) : [];
 415
 416                 # Flag our recent bot edits
 417                 // @phan-suppress-next-line PhanRedundantCondition
 418                 if ( $botgroups ) {
 419                         $this->output( "Flagging bot account edits...\n" );
 420
 421                         # Fill in the rc_bot field
 422                         $rcids = $this->findRcIdsWithGroups( $dbw, $botgroups );
 423
 424                         foreach ( array_chunk( $rcids, $this->getBatchSize() ) as $rcidBatch ) {
 425                                 $dbw->update(
 426                                         'recentchanges',
 427                                         [ 'rc_bot' => 1 ],
 428                                         [ 'rc_id' => $rcidBatch ],
 429                                         __METHOD__
 430                                 );
 431                                 $lbFactory->waitForReplication();
 432                         }
 433                 }
 434
 435                 # Flag our recent autopatrolled edits
 436                 if ( !$wgMiserMode && $autopatrolgroups ) {
 437                         $this->output( "Flagging auto-patrolled edits...\n" );
 438
 439                         $conds = [ 'rc_patrolled' => 0 ];
 440                         if ( !$wgUseRCPatrol ) {
 441                                 $subConds = [];
 442                                 if ( $wgUseNPPatrol ) {
 443                                         $subConds[] = 'rc_source = ' . $dbw->addQuotes( RecentChange::SRC_NEW );
 444                                 }
 445                                 if ( $wgUseFilePatrol ) {
 446                                         $subConds[] = 'rc_log_type = ' . $dbw->addQuotes( 'upload' );
 447                                 }
 448                                 $conds[] = $dbw->makeList( $subConds, IDatabase::LIST_OR );
 449                         }
 450
 451                         $rcids = $this->findRcIdsWithGroups( $dbw, $autopatrolgroups, $conds );
 452                         foreach ( array_chunk( $rcids, $this->getBatchSize() ) as $rcidBatch ) {
 453                                 $dbw->update(
 454                                         'recentchanges',
 455                                         [ 'rc_patrolled' => 2 ],
 456                                         [ 'rc_id' => $rcidBatch ],
 457                                         __METHOD__
 458                                 );
 459                                 $lbFactory->waitForReplication();
 460                         }
 461                 }
 462         }
 463
 464         /**
 465          * Rebuild pass 5: Delete duplicate entries where we generate both a page revision and a log
 466          * entry for a single action (upload, move, protect, import, etc.).
 467          *
 468          * @param ILBFactory $lbFactory
 469          */
 470         private function rebuildRecentChangesTablePass5( ILBFactory $lbFactory ) {
 471                 $dbw = $this->getDB( DB_PRIMARY );
 472
 473                 $this->output( "Removing duplicate revision and logging entries...\n" );
 474
 475                 $res = $dbw->select(
 476                         [ 'logging', 'log_search' ],
 477                         [ 'ls_value', 'ls_log_id' ],
 478                         [
 479                                 'ls_log_id = log_id',
 480                                 'ls_field' => 'associated_rev_id',
 481                                 'log_type != ' . $dbw->addQuotes( 'create' ),
 482                                 'log_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 483                                 'log_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) ),
 484                         ],
 485                         __METHOD__
 486                 );
 487
 488                 $updates = 0;
 489                 foreach ( $res as $row ) {
 490                         $rev_id = $row->ls_value;
 491                         $log_id = $row->ls_log_id;
 492
 493                         // Mark the logging row as having an associated rev id
 494                         $dbw->update(
 495                                 'recentchanges',
 496                                 /*SET*/ [ 'rc_this_oldid' => $rev_id ],
 497                                 /*WHERE*/ [ 'rc_logid' => $log_id ],
 498                                 __METHOD__
 499                         );
 500
 501                         // Delete the revision row
 502                         $dbw->delete(
 503                                 'recentchanges',
 504                                 /*WHERE*/ [ 'rc_this_oldid' => $rev_id, 'rc_logid' => 0 ],
 505                                 __METHOD__
 506                         );
 507
 508                         if ( ( ++$updates % $this->getBatchSize() ) == 0 ) {
 509                                 $lbFactory->waitForReplication();
 510                         }
 511                 }
 512         }
 513
 514         /**
 515          * Purge cached feeds in $wanCache
 516          */
 517         private function purgeFeeds() {
 518                 global $wgFeedClasses;
 519
 520                 $this->output( "Deleting feed timestamps.\n" );
 521
 522                 $wanCache = MediaWikiServices::getInstance()->getMainWANObjectCache();
 523                 foreach ( $wgFeedClasses as $feed => $className ) {
 524                         $wanCache->delete( $wanCache->makeKey( 'rcfeed', $feed, 'timestamp' ) ); # Good enough for now.
 525                 }
 526         }
 527 }
 528
 529 $maintClass = RebuildRecentchanges::class;
 530 require_once RUN_MAINTENANCE_IF_MAIN;