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 // @codeCoverageIgnoreStart
  27 require_once __DIR__ . '/Maintenance.php';
  28 // @codeCoverageIgnoreEnd
  29
  30 use MediaWiki\Maintenance\Maintenance;
  31 use Wikimedia\Rdbms\IDatabase;
  32 use Wikimedia\Rdbms\IReadableDatabase;
  33 use Wikimedia\Rdbms\SelectQueryBuilder;
  34
  35 /**
  36  * Maintenance script that rebuilds recent changes from scratch.
  37  *
  38  * @ingroup Maintenance
  39  */
  40 class RebuildRecentchanges extends Maintenance {
  41         /** @var int UNIX timestamp */
  42         private $cutoffFrom;
  43         /** @var int UNIX timestamp */
  44         private $cutoffTo;
  45
  46         public function __construct() {
  47                 parent::__construct();
  48                 $this->addDescription( 'Rebuild recent changes' );
  49
  50                 $this->addOption(
  51                         'from',
  52                         "Only rebuild rows in requested time range (in YYYYMMDDHHMMSS format)",
  53                         false,
  54                         true
  55                 );
  56                 $this->addOption(
  57                         'to',
  58                         "Only rebuild rows in requested time range (in YYYYMMDDHHMMSS format)",
  59                         false,
  60                         true
  61                 );
  62                 $this->setBatchSize( 200 );
  63         }
  64
  65         public function execute() {
  66                 if (
  67                         ( $this->hasOption( 'from' ) && !$this->hasOption( 'to' ) ) ||
  68                         ( !$this->hasOption( 'from' ) && $this->hasOption( 'to' ) )
  69                 ) {
  70                         $this->fatalError( "Both 'from' and 'to' must be given, or neither" );
  71                 }
  72
  73                 $this->rebuildRecentChangesTablePass1();
  74                 $this->rebuildRecentChangesTablePass2();
  75                 $this->rebuildRecentChangesTablePass3();
  76                 $this->rebuildRecentChangesTablePass4();
  77                 $this->rebuildRecentChangesTablePass5();
  78                 if ( !( $this->hasOption( 'from' ) && $this->hasOption( 'to' ) ) ) {
  79                         $this->purgeFeeds();
  80                 }
  81                 $this->output( "Done.\n" );
  82         }
  83
  84         /**
  85          * Rebuild pass 1: Insert `recentchanges` entries for page revisions.
  86          */
  87         private function rebuildRecentChangesTablePass1() {
  88                 $dbw = $this->getPrimaryDB();
  89                 $commentStore = $this->getServiceContainer()->getCommentStore();
  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->newSelectQueryBuilder()
 110                         ->select( 'rc_id' )
 111                         ->from( 'recentchanges' )
 112                         ->where( $dbw->expr( 'rc_timestamp', '>', $dbw->timestamp( $this->cutoffFrom ) ) )
 113                         ->andWhere( $dbw->expr( 'rc_timestamp', '<', $dbw->timestamp( $this->cutoffTo ) ) )
 114                         ->caller( __METHOD__ )->fetchFieldValues();
 115                 foreach ( array_chunk( $rcids, $this->getBatchSize() ) as $rcidBatch ) {
 116                         $dbw->newDeleteQueryBuilder()
 117                                 ->deleteFrom( 'recentchanges' )
 118                                 ->where( [ 'rc_id' => $rcidBatch ] )
 119                                 ->caller( __METHOD__ )->execute();
 120                         $this->waitForReplication();
 121                 }
 122
 123                 $this->output( "Loading from page and revision tables...\n" );
 124
 125                 $res = $dbw->newSelectQueryBuilder()
 126                         ->select(
 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                                         'rev_comment_text' => 'comment_rev_comment.comment_text',
 137                                         'rev_comment_data' => 'comment_rev_comment.comment_data',
 138                                         'rev_comment_cid' => 'comment_rev_comment.comment_id',
 139                                         'rev_user' => 'actor_rev_user.actor_user',
 140                                         'rev_user_text' => 'actor_rev_user.actor_name',
 141                                         'rev_actor' => 'rev_actor',
 142                                 ]
 143                         )
 144                         ->from( 'revision' )
 145                         ->join( 'page', null, 'rev_page=page_id' )
 146                         ->join( 'comment', 'comment_rev_comment', 'comment_rev_comment.comment_id = rev_comment_id' )
 147                         ->join( 'actor', 'actor_rev_user', 'actor_rev_user.actor_id = rev_actor' )
 148                         ->where(
 149                                 [
 150                                         $dbw->expr( 'rev_timestamp', '>', $dbw->timestamp( $this->cutoffFrom ) ),
 151                                         $dbw->expr( 'rev_timestamp', '<', $dbw->timestamp( $this->cutoffTo ) )
 152                                 ]
 153                         )
 154                         ->orderBy( 'rev_timestamp', SelectQueryBuilder::SORT_DESC )
 155                         ->caller( __METHOD__ )->fetchResultSet();
 156
 157                 $this->output( "Inserting from page and revision tables...\n" );
 158                 $inserted = 0;
 159                 foreach ( $res as $row ) {
 160                         $comment = $commentStore->getComment( 'rev_comment', $row );
 161                         $dbw->newInsertQueryBuilder()
 162                                 ->insertInto( 'recentchanges' )
 163                                 ->row( [
 164                                         'rc_timestamp' => $row->rev_timestamp,
 165                                         'rc_actor' => $row->rev_actor,
 166                                         'rc_namespace' => $row->page_namespace,
 167                                         'rc_title' => $row->page_title,
 168                                         'rc_minor' => $row->rev_minor_edit,
 169                                         'rc_bot' => 0,
 170                                         'rc_new' => $row->page_is_new,
 171                                         'rc_cur_id' => $row->page_id,
 172                                         'rc_this_oldid' => $row->rev_id,
 173                                         'rc_last_oldid' => 0, // is this ok?
 174                                         'rc_type' => $row->page_is_new ? RC_NEW : RC_EDIT,
 175                                         'rc_source' => $row->page_is_new ? RecentChange::SRC_NEW : RecentChange::SRC_EDIT,
 176                                         'rc_deleted' => $row->rev_deleted
 177                                         ] + $commentStore->insert( $dbw, 'rc_comment', $comment ) )
 178                                 ->caller( __METHOD__ )->execute();
 179
 180                         $rcid = $dbw->insertId();
 181                         $dbw->newUpdateQueryBuilder()
 182                                 ->update( 'change_tag' )
 183                                 ->set( [ 'ct_rc_id' => $rcid ] )
 184                                 ->where( [ 'ct_rev_id' => $row->rev_id ] )
 185                                 ->caller( __METHOD__ )->execute();
 186
 187                         if ( ( ++$inserted % $this->getBatchSize() ) == 0 ) {
 188                                 $this->waitForReplication();
 189                         }
 190                 }
 191         }
 192
 193         /**
 194          * Rebuild pass 2: Enhance entries for page revisions with references to the previous revision
 195          * (rc_last_oldid, rc_new etc.) and size differences (rc_old_len, rc_new_len).
 196          */
 197         private function rebuildRecentChangesTablePass2() {
 198                 $dbw = $this->getPrimaryDB();
 199
 200                 $this->output( "Updating links and size differences...\n" );
 201
 202                 # Fill in the rc_last_oldid field, which points to the previous edit
 203                 $res = $dbw->newSelectQueryBuilder()
 204                         ->select( [ 'rc_cur_id', 'rc_this_oldid', 'rc_timestamp' ] )
 205                         ->from( 'recentchanges' )
 206                         ->where( $dbw->expr( 'rc_timestamp', '>', $dbw->timestamp( $this->cutoffFrom ) ) )
 207                         ->andWhere( $dbw->expr( 'rc_timestamp', '<', $dbw->timestamp( $this->cutoffTo ) ) )
 208                         ->orderBy( [ 'rc_cur_id', 'rc_timestamp' ] )
 209                         ->caller( __METHOD__ )->fetchResultSet();
 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->newSelectQueryBuilder()
 224                                         ->select( [ 'rev_id', 'rev_len' ] )
 225                                         ->from( 'revision' )
 226                                         ->where( [ 'rev_page' => $lastCurId, $dbw->expr( 'rev_timestamp', '<', $emit ) ] )
 227                                         ->orderBy( 'rev_timestamp DESC' )
 228                                         ->caller( __METHOD__ )->fetchRow();
 229                                 if ( $revRow ) {
 230                                         $lastOldId = intval( $revRow->rev_id );
 231                                         # Grab the last text size if available
 232                                         $lastSize = $revRow->rev_len !== null ? intval( $revRow->rev_len ) : null;
 233                                 } else {
 234                                         # No previous edit
 235                                         $lastOldId = 0;
 236                                         $lastSize = 0;
 237                                         $new = 1; // probably true
 238                                 }
 239                         }
 240
 241                         if ( $lastCurId == 0 ) {
 242                                 $this->output( "Uhhh, something wrong? No curid\n" );
 243                         } else {
 244                                 # Grab the entry's text size
 245                                 $size = (int)$dbw->newSelectQueryBuilder()
 246                                         ->select( 'rev_len' )
 247                                         ->from( 'revision' )
 248                                         ->where( [ 'rev_id' => $row->rc_this_oldid ] )
 249                                         ->caller( __METHOD__ )->fetchField();
 250
 251                                 $dbw->newUpdateQueryBuilder()
 252                                         ->update( 'recentchanges' )
 253                                         ->set( [
 254                                                 'rc_last_oldid' => $lastOldId,
 255                                                 'rc_new' => $new,
 256                                                 'rc_type' => $new ? RC_NEW : RC_EDIT,
 257                                                 'rc_source' => $new === 1 ? RecentChange::SRC_NEW : RecentChange::SRC_EDIT,
 258                                                 'rc_old_len' => $lastSize,
 259                                                 'rc_new_len' => $size,
 260                                         ] )
 261                                         ->where( [
 262                                                 'rc_cur_id' => $lastCurId,
 263                                                 'rc_this_oldid' => $row->rc_this_oldid,
 264                                                 // index usage
 265                                                 'rc_timestamp' => $row->rc_timestamp,
 266                                         ] )
 267                                         ->caller( __METHOD__ )->execute();
 268
 269                                 $lastOldId = intval( $row->rc_this_oldid );
 270                                 $lastSize = $size;
 271
 272                                 if ( ( ++$updated % $this->getBatchSize() ) == 0 ) {
 273                                         $this->waitForReplication();
 274                                 }
 275                         }
 276                 }
 277         }
 278
 279         /**
 280          * Rebuild pass 3: Insert `recentchanges` entries for action logs.
 281          */
 282         private function rebuildRecentChangesTablePass3() {
 283                 global $wgLogRestrictions, $wgFilterLogTypes;
 284
 285                 $dbw = $this->getDB( DB_PRIMARY );
 286                 $commentStore = $this->getServiceContainer()->getCommentStore();
 287                 $nonRCLogs = array_merge(
 288                         array_keys( $wgLogRestrictions ),
 289                         array_keys( $wgFilterLogTypes ),
 290                         [ 'create' ]
 291                 );
 292
 293                 $this->output( "Loading from user and logging tables...\n" );
 294
 295                 $res = $dbw->newSelectQueryBuilder()
 296                         ->select(
 297                                 [
 298                                         'log_timestamp',
 299                                         'log_actor',
 300                                         'log_namespace',
 301                                         'log_title',
 302                                         'log_page',
 303                                         'log_type',
 304                                         'log_action',
 305                                         'log_id',
 306                                         'log_params',
 307                                         'log_deleted',
 308                                         'log_comment_text' => 'comment_log_comment.comment_text',
 309                                         'log_comment_data' => 'comment_log_comment.comment_data',
 310                                         'log_comment_cid' => 'comment_log_comment.comment_id',
 311                                 ]
 312                         )
 313                         ->from( 'logging' )
 314                         ->join( 'comment', 'comment_log_comment', 'comment_log_comment.comment_id = log_comment_id' )
 315                         ->where(
 316                                 [
 317                                         $dbw->expr( 'log_timestamp', '>', $dbw->timestamp( $this->cutoffFrom ) ),
 318                                         $dbw->expr( 'log_timestamp', '<', $dbw->timestamp( $this->cutoffTo ) ),
 319                                         // Some logs don't go in RC since they are private, or are included in the filterable log types.
 320                                         'log_type' => array_diff( LogPage::validTypes(), $nonRCLogs ),
 321                                 ]
 322                         )
 323                         ->orderBy( [ 'log_timestamp DESC', 'log_id DESC' ] )
 324                         ->caller( __METHOD__ )->fetchResultSet();
 325
 326                 $field = $dbw->fieldInfo( 'recentchanges', 'rc_cur_id' );
 327
 328                 $inserted = 0;
 329                 foreach ( $res as $row ) {
 330                         $comment = $commentStore->getComment( 'log_comment', $row );
 331                         $dbw->newInsertQueryBuilder()
 332                                 ->insertInto( 'recentchanges' )
 333                                 ->row( [
 334                                         'rc_timestamp' => $row->log_timestamp,
 335                                         'rc_actor' => $row->log_actor,
 336                                         'rc_namespace' => $row->log_namespace,
 337                                         'rc_title' => $row->log_title,
 338                                         'rc_minor' => 0,
 339                                         'rc_bot' => 0,
 340                                         'rc_patrolled' => $row->log_type == 'upload' ? 0 : 2,
 341                                         'rc_new' => 0,
 342                                         'rc_this_oldid' => 0,
 343                                         'rc_last_oldid' => 0,
 344                                         'rc_type' => RC_LOG,
 345                                         'rc_source' => RecentChange::SRC_LOG,
 346                                         'rc_cur_id' => $field->isNullable()
 347                                                 ? $row->log_page
 348                                                 : (int)$row->log_page, // NULL => 0,
 349                                         'rc_log_type' => $row->log_type,
 350                                         'rc_log_action' => $row->log_action,
 351                                         'rc_logid' => $row->log_id,
 352                                         'rc_params' => $row->log_params,
 353                                         'rc_deleted' => $row->log_deleted
 354                                         ] + $commentStore->insert( $dbw, 'rc_comment', $comment ) )
 355                                 ->caller( __METHOD__ )->execute();
 356
 357                         $rcid = $dbw->insertId();
 358                         $dbw->newUpdateQueryBuilder()
 359                                 ->update( 'change_tag' )
 360                                 ->set( [ 'ct_rc_id' => $rcid ] )
 361                                 ->where( [ 'ct_log_id' => $row->log_id ] )
 362                                 ->caller( __METHOD__ )->execute();
 363
 364                         if ( ( ++$inserted % $this->getBatchSize() ) == 0 ) {
 365                                 $this->waitForReplication();
 366                         }
 367                 }
 368         }
 369
 370         /**
 371          * Find rc_id values that have a user with one of the specified groups
 372          *
 373          * @param IReadableDatabase $db
 374          * @param string[] $groups
 375          * @param array $conds Extra query conditions
 376          * @return int[]
 377          */
 378         private function findRcIdsWithGroups( $db, $groups, $conds = [] ) {
 379                 if ( !count( $groups ) ) {
 380                         return [];
 381                 }
 382                 return $db->newSelectQueryBuilder()
 383                         ->select( 'rc_id' )
 384                         ->distinct()
 385                         ->from( 'recentchanges' )
 386                         ->join( 'actor', null, 'actor_id=rc_actor' )
 387                         ->join( 'user_groups', null, 'ug_user=actor_user' )
 388                         ->where( $conds )
 389                         ->andWhere( [
 390                                 $db->expr( 'rc_timestamp', '>', $db->timestamp( $this->cutoffFrom ) ),
 391                                 $db->expr( 'rc_timestamp', '<', $db->timestamp( $this->cutoffTo ) ),
 392                                 'ug_group' => $groups
 393                         ] )
 394                         ->caller( __METHOD__ )->fetchFieldValues();
 395         }
 396
 397         /**
 398          * Rebuild pass 4: Mark bot and autopatrolled entries.
 399          */
 400         private function rebuildRecentChangesTablePass4() {
 401                 global $wgUseRCPatrol, $wgUseNPPatrol, $wgUseFilePatrol, $wgMiserMode;
 402
 403                 $dbw = $this->getPrimaryDB();
 404
 405                 # @FIXME: recognize other bot account groups (not the same as users with 'bot' rights)
 406                 # @NOTE: users with 'bot' rights choose when edits are bot edits or not. That information
 407                 # may be lost at this point (aside from joining on the patrol log table entries).
 408                 $botgroups = [ 'bot' ];
 409                 $autopatrolgroups = ( $wgUseRCPatrol || $wgUseNPPatrol || $wgUseFilePatrol ) ?
 410                         $this->getServiceContainer()->getGroupPermissionsLookup()
 411                         ->getGroupsWithPermission( 'autopatrol' ) : [];
 412
 413                 # Flag our recent bot edits
 414                 // @phan-suppress-next-line PhanRedundantCondition
 415                 if ( $botgroups ) {
 416                         $this->output( "Flagging bot account edits...\n" );
 417
 418                         # Fill in the rc_bot field
 419                         $rcids = $this->findRcIdsWithGroups( $dbw, $botgroups );
 420
 421                         foreach ( array_chunk( $rcids, $this->getBatchSize() ) as $rcidBatch ) {
 422                                 $dbw->newUpdateQueryBuilder()
 423                                         ->update( 'recentchanges' )
 424                                         ->set( [ 'rc_bot' => 1 ] )
 425                                         ->where( [ 'rc_id' => $rcidBatch ] )
 426                                         ->caller( __METHOD__ )->execute();
 427                                 $this->waitForReplication();
 428                         }
 429                 }
 430
 431                 # Flag our recent autopatrolled edits
 432                 if ( !$wgMiserMode && $autopatrolgroups ) {
 433                         $this->output( "Flagging auto-patrolled edits...\n" );
 434
 435                         $conds = [ 'rc_patrolled' => 0 ];
 436                         if ( !$wgUseRCPatrol ) {
 437                                 $subConds = [];
 438                                 if ( $wgUseNPPatrol ) {
 439                                         $subConds[] = $dbw->expr( 'rc_source', '=', RecentChange::SRC_NEW );
 440                                 }
 441                                 if ( $wgUseFilePatrol ) {
 442                                         $subConds[] = $dbw->expr( 'rc_log_type', '=', 'upload' );
 443                                 }
 444                                 $conds[] = $dbw->makeList( $subConds, IDatabase::LIST_OR );
 445                         }
 446
 447                         $rcids = $this->findRcIdsWithGroups( $dbw, $autopatrolgroups, $conds );
 448                         foreach ( array_chunk( $rcids, $this->getBatchSize() ) as $rcidBatch ) {
 449                                 $dbw->newUpdateQueryBuilder()
 450                                         ->update( 'recentchanges' )
 451                                         ->set( [ 'rc_patrolled' => 2 ] )
 452                                         ->where( [ 'rc_id' => $rcidBatch ] )
 453                                         ->caller( __METHOD__ )->execute();
 454                                 $this->waitForReplication();
 455                         }
 456                 }
 457         }
 458
 459         /**
 460          * Rebuild pass 5: Delete duplicate entries where we generate both a page revision and a log
 461          * entry for a single action (upload, move, protect, import, etc.).
 462          */
 463         private function rebuildRecentChangesTablePass5() {
 464                 $dbw = $this->getPrimaryDB();
 465
 466                 $this->output( "Removing duplicate revision and logging entries...\n" );
 467
 468                 $res = $dbw->newSelectQueryBuilder()
 469                         ->select( [ 'ls_value', 'ls_log_id' ] )
 470                         ->from( 'logging' )
 471                         ->join( 'log_search', null, 'ls_log_id = log_id' )
 472                         ->where( [
 473                                 'ls_field' => 'associated_rev_id',
 474                                 $dbw->expr( 'log_type', '!=', 'create' ),
 475                                 $dbw->expr( 'log_timestamp', '>', $dbw->timestamp( $this->cutoffFrom ) ),
 476                                 $dbw->expr( 'log_timestamp', '<', $dbw->timestamp( $this->cutoffTo ) ),
 477                         ] )
 478                         ->caller( __METHOD__ )->fetchResultSet();
 479
 480                 $updates = 0;
 481                 foreach ( $res as $row ) {
 482                         $rev_id = $row->ls_value;
 483                         $log_id = $row->ls_log_id;
 484
 485                         // Mark the logging row as having an associated rev id
 486                         $dbw->newUpdateQueryBuilder()
 487                                 ->update( 'recentchanges' )
 488                                 ->set( [ 'rc_this_oldid' => $rev_id ] )
 489                                 ->where( [ 'rc_logid' => $log_id ] )
 490                                 ->caller( __METHOD__ )->execute();
 491
 492                         // Delete the revision row
 493                         $dbw->newDeleteQueryBuilder()
 494                                 ->deleteFrom( 'recentchanges' )
 495                                 ->where( [ 'rc_this_oldid' => $rev_id, 'rc_logid' => 0 ] )
 496                                 ->caller( __METHOD__ )->execute();
 497
 498                         if ( ( ++$updates % $this->getBatchSize() ) == 0 ) {
 499                                 $this->waitForReplication();
 500                         }
 501                 }
 502         }
 503
 504         /**
 505          * Purge cached feeds in $wanCache
 506          */
 507         private function purgeFeeds() {
 508                 global $wgFeedClasses;
 509
 510                 $this->output( "Deleting feed timestamps.\n" );
 511
 512                 $wanCache = $this->getServiceContainer()->getMainWANObjectCache();
 513                 foreach ( $wgFeedClasses as $feed => $className ) {
 514                         $wanCache->delete( $wanCache->makeKey( 'rcfeed', $feed, 'timestamp' ) ); # Good enough for now.
 515                 }
 516         }
 517 }
 518
 519 // @codeCoverageIgnoreStart
 520 $maintClass = RebuildRecentchanges::class;
 521 require_once RUN_MAINTENANCE_IF_MAIN;
 522 // @codeCoverageIgnoreEnd