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 use MediaWiki\MediaWikiServices;
  28
  29 /**
  30  * Maintenance script that rebuilds recent changes from scratch.
  31  *
  32  * @ingroup Maintenance
  33  */
  34 class RebuildRecentchanges extends Maintenance {
  35         /** @var integer UNIX timestamp */
  36         private $cutoffFrom;
  37         /** @var integer UNIX timestamp */
  38         private $cutoffTo;
  39
  40         public function __construct() {
  41                 parent::__construct();
  42                 $this->addDescription( 'Rebuild recent changes' );
  43
  44                 $this->addOption(
  45                         'from',
  46                         "Only rebuild rows in requested time range (in YYYYMMDDHHMMSS format)",
  47                         false,
  48                         true
  49                 );
  50                 $this->addOption(
  51                         'to',
  52                         "Only rebuild rows in requested time range (in YYYYMMDDHHMMSS format)",
  53                         false,
  54                         true
  55                 );
  56                 $this->setBatchSize( 200 );
  57         }
  58
  59         public function execute() {
  60                 if (
  61                         ( $this->hasOption( 'from' ) && !$this->hasOption( 'to' ) ) ||
  62                         ( !$this->hasOption( 'from' ) && $this->hasOption( 'to' ) )
  63                 ) {
  64                         $this->error( "Both 'from' and 'to' must be given, or neither", 1 );
  65                 }
  66
  67                 $this->rebuildRecentChangesTablePass1();
  68                 $this->rebuildRecentChangesTablePass2();
  69                 $this->rebuildRecentChangesTablePass3();
  70                 $this->rebuildRecentChangesTablePass4();
  71                 $this->rebuildRecentChangesTablePass5();
  72                 if ( !( $this->hasOption( 'from' ) && $this->hasOption( 'to' ) ) ) {
  73                         $this->purgeFeeds();
  74                 }
  75                 $this->output( "Done.\n" );
  76         }
  77
  78         /**
  79          * Rebuild pass 1: Insert `recentchanges` entries for page revisions.
  80          */
  81         private function rebuildRecentChangesTablePass1() {
  82                 $dbw = $this->getDB( DB_MASTER );
  83
  84                 if ( $this->hasOption( 'from' ) && $this->hasOption( 'to' ) ) {
  85                         $this->cutoffFrom = wfTimestamp( TS_UNIX, $this->getOption( 'from' ) );
  86                         $this->cutoffTo = wfTimestamp( TS_UNIX, $this->getOption( 'to' ) );
  87
  88                         $sec = $this->cutoffTo - $this->cutoffFrom;
  89                         $days = $sec / 24 / 3600;
  90                         $this->output( "Rebuilding range of $sec seconds ($days days)\n" );
  91                 } else {
  92                         global $wgRCMaxAge;
  93
  94                         $days = $wgRCMaxAge / 24 / 3600;
  95                         $this->output( "Rebuilding \$wgRCMaxAge=$wgRCMaxAge seconds ($days days)\n" );
  96
  97                         $this->cutoffFrom = time() - $wgRCMaxAge;
  98                         $this->cutoffTo = time();
  99                 }
 100
 101                 $this->output( "Clearing recentchanges table for time range...\n" );
 102                 $rcids = $dbw->selectFieldValues(
 103                         'recentchanges',
 104                         'rc_id',
 105                         [
 106                                 'rc_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 107                                 'rc_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) )
 108                         ]
 109                 );
 110                 foreach ( array_chunk( $rcids, $this->mBatchSize ) as $rcidBatch ) {
 111                         $dbw->delete( 'recentchanges', [ 'rc_id' => $rcidBatch ], __METHOD__ );
 112                         wfGetLBFactory()->waitForReplication();
 113                 }
 114
 115                 $this->output( "Loading from page and revision tables...\n" );
 116                 $res = $dbw->select(
 117                         [ 'page', 'revision' ],
 118                         [
 119                                 'rev_timestamp',
 120                                 'rev_user',
 121                                 'rev_user_text',
 122                                 'rev_comment',
 123                                 'rev_minor_edit',
 124                                 'rev_id',
 125                                 'rev_deleted',
 126                                 'page_namespace',
 127                                 'page_title',
 128                                 'page_is_new',
 129                                 'page_id'
 130                         ],
 131                         [
 132                                 'rev_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 133                                 'rev_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) ),
 134                                 'rev_page=page_id'
 135                         ],
 136                         __METHOD__,
 137                         [ 'ORDER BY' => 'rev_timestamp DESC' ]
 138                 );
 139
 140                 $this->output( "Inserting from page and revision tables...\n" );
 141                 $inserted = 0;
 142                 foreach ( $res as $row ) {
 143                         $dbw->insert(
 144                                 'recentchanges',
 145                                 [
 146                                         'rc_timestamp' => $row->rev_timestamp,
 147                                         'rc_user' => $row->rev_user,
 148                                         'rc_user_text' => $row->rev_user_text,
 149                                         'rc_namespace' => $row->page_namespace,
 150                                         'rc_title' => $row->page_title,
 151                                         'rc_comment' => $row->rev_comment,
 152                                         'rc_minor' => $row->rev_minor_edit,
 153                                         'rc_bot' => 0,
 154                                         'rc_new' => $row->page_is_new,
 155                                         'rc_cur_id' => $row->page_id,
 156                                         'rc_this_oldid' => $row->rev_id,
 157                                         'rc_last_oldid' => 0, // is this ok?
 158                                         'rc_type' => $row->page_is_new ? RC_NEW : RC_EDIT,
 159                                         'rc_source' => $row->page_is_new ? RecentChange::SRC_NEW : RecentChange::SRC_EDIT
 160                                         ,
 161                                         'rc_deleted' => $row->rev_deleted
 162                                 ],
 163                                 __METHOD__
 164                         );
 165                         if ( ( ++$inserted % $this->mBatchSize ) == 0 ) {
 166                                 wfGetLBFactory()->waitForReplication();
 167                         }
 168                 }
 169         }
 170
 171         /**
 172          * Rebuild pass 2: Enhance entries for page revisions with references to the previous revision
 173          * (rc_last_oldid, rc_new etc.) and size differences (rc_old_len, rc_new_len).
 174          */
 175         private function rebuildRecentChangesTablePass2() {
 176                 $dbw = $this->getDB( DB_MASTER );
 177
 178                 $this->output( "Updating links and size differences...\n" );
 179
 180                 # Fill in the rc_last_oldid field, which points to the previous edit
 181                 $res = $dbw->select(
 182                         'recentchanges',
 183                         [ 'rc_cur_id', 'rc_this_oldid', 'rc_timestamp' ],
 184                         [
 185                                 "rc_timestamp > " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 186                                 "rc_timestamp < " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) )
 187                         ],
 188                         __METHOD__,
 189                         [ 'ORDER BY' => 'rc_cur_id,rc_timestamp' ]
 190                 );
 191
 192                 $lastCurId = 0;
 193                 $lastOldId = 0;
 194                 $lastSize = null;
 195                 $updated = 0;
 196                 foreach ( $res as $obj ) {
 197                         $new = 0;
 198
 199                         if ( $obj->rc_cur_id != $lastCurId ) {
 200                                 # Switch! Look up the previous last edit, if any
 201                                 $lastCurId = intval( $obj->rc_cur_id );
 202                                 $emit = $obj->rc_timestamp;
 203
 204                                 $row = $dbw->selectRow(
 205                                         'revision',
 206                                         [ 'rev_id', 'rev_len' ],
 207                                         [ 'rev_page' => $lastCurId, "rev_timestamp < " . $dbw->addQuotes( $emit ) ],
 208                                         __METHOD__,
 209                                         [ 'ORDER BY' => 'rev_timestamp DESC' ]
 210                                 );
 211                                 if ( $row ) {
 212                                         $lastOldId = intval( $row->rev_id );
 213                                         # Grab the last text size if available
 214                                         $lastSize = !is_null( $row->rev_len ) ? intval( $row->rev_len ) : null;
 215                                 } else {
 216                                         # No previous edit
 217                                         $lastOldId = 0;
 218                                         $lastSize = null;
 219                                         $new = 1; // probably true
 220                                 }
 221                         }
 222
 223                         if ( $lastCurId == 0 ) {
 224                                 $this->output( "Uhhh, something wrong? No curid\n" );
 225                         } else {
 226                                 # Grab the entry's text size
 227                                 $size = (int)$dbw->selectField(
 228                                         'revision',
 229                                         'rev_len',
 230                                         [ 'rev_id' => $obj->rc_this_oldid ],
 231                                         __METHOD__
 232                                 );
 233
 234                                 $dbw->update(
 235                                         'recentchanges',
 236                                         [
 237                                                 'rc_last_oldid' => $lastOldId,
 238                                                 'rc_new' => $new,
 239                                                 'rc_type' => $new ? RC_NEW : RC_EDIT,
 240                                                 'rc_source' => $new === 1 ? RecentChange::SRC_NEW : RecentChange::SRC_EDIT,
 241                                                 'rc_old_len' => $lastSize,
 242                                                 'rc_new_len' => $size,
 243                                         ],
 244                                         [
 245                                                 'rc_cur_id' => $lastCurId,
 246                                                 'rc_this_oldid' => $obj->rc_this_oldid,
 247                                                 'rc_timestamp' => $obj->rc_timestamp // index usage
 248                                         ],
 249                                         __METHOD__
 250                                 );
 251
 252                                 $lastOldId = intval( $obj->rc_this_oldid );
 253                                 $lastSize = $size;
 254
 255                                 if ( ( ++$updated % $this->mBatchSize ) == 0 ) {
 256                                         wfGetLBFactory()->waitForReplication();
 257                                 }
 258                         }
 259                 }
 260         }
 261
 262         /**
 263          * Rebuild pass 3: Insert `recentchanges` entries for action logs.
 264          */
 265         private function rebuildRecentChangesTablePass3() {
 266                 global $wgLogTypes, $wgLogRestrictions;
 267
 268                 $dbw = $this->getDB( DB_MASTER );
 269
 270                 $this->output( "Loading from user, page, and logging tables...\n" );
 271
 272                 $res = $dbw->select(
 273                         [ 'user', 'logging', 'page' ],
 274                         [
 275                                 'log_timestamp',
 276                                 'log_user',
 277                                 'user_name',
 278                                 'log_namespace',
 279                                 'log_title',
 280                                 'log_comment',
 281                                 'page_id',
 282                                 'log_type',
 283                                 'log_action',
 284                                 'log_id',
 285                                 'log_params',
 286                                 'log_deleted'
 287                         ],
 288                         [
 289                                 'log_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 290                                 'log_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) ),
 291                                 'log_user=user_id',
 292                                 // Some logs don't go in RC since they are private.
 293                                 // @FIXME: core/extensions also have spammy logs that don't go in RC.
 294                                 'log_type' => array_diff( $wgLogTypes, array_keys( $wgLogRestrictions ) ),
 295                         ],
 296                         __METHOD__,
 297                         [ 'ORDER BY' => 'log_timestamp DESC' ],
 298                         [
 299                                 'page' =>
 300                                         [ 'LEFT JOIN', [ 'log_namespace=page_namespace', 'log_title=page_title' ] ]
 301                         ]
 302                 );
 303
 304                 $field = $dbw->fieldInfo( 'recentchanges', 'rc_cur_id' );
 305
 306                 $inserted = 0;
 307                 foreach ( $res as $row ) {
 308                         $dbw->insert(
 309                                 'recentchanges',
 310                                 [
 311                                         'rc_timestamp' => $row->log_timestamp,
 312                                         'rc_user' => $row->log_user,
 313                                         'rc_user_text' => $row->user_name,
 314                                         'rc_namespace' => $row->log_namespace,
 315                                         'rc_title' => $row->log_title,
 316                                         'rc_comment' => $row->log_comment,
 317                                         'rc_minor' => 0,
 318                                         'rc_bot' => 0,
 319                                         'rc_patrolled' => 1,
 320                                         'rc_new' => 0,
 321                                         'rc_this_oldid' => 0,
 322                                         'rc_last_oldid' => 0,
 323                                         'rc_type' => RC_LOG,
 324                                         'rc_source' => RecentChange::SRC_LOG,
 325                                         'rc_cur_id' => $field->isNullable()
 326                                                 ? $row->page_id
 327                                                 : (int)$row->page_id, // NULL => 0,
 328                                         'rc_log_type' => $row->log_type,
 329                                         'rc_log_action' => $row->log_action,
 330                                         'rc_logid' => $row->log_id,
 331                                         'rc_params' => $row->log_params,
 332                                         'rc_deleted' => $row->log_deleted
 333                                 ],
 334                                 __METHOD__
 335                         );
 336
 337                         if ( ( ++$inserted % $this->mBatchSize ) == 0 ) {
 338                                 wfGetLBFactory()->waitForReplication();
 339                         }
 340                 }
 341         }
 342
 343         /**
 344          * Rebuild pass 4: Mark bot and autopatrolled entries.
 345          */
 346         private function rebuildRecentChangesTablePass4() {
 347                 global $wgUseRCPatrol, $wgMiserMode;
 348
 349                 $dbw = $this->getDB( DB_MASTER );
 350
 351                 list( $recentchanges, $usergroups, $user ) =
 352                         $dbw->tableNamesN( 'recentchanges', 'user_groups', 'user' );
 353
 354                 # @FIXME: recognize other bot account groups (not the same as users with 'bot' rights)
 355                 # @NOTE: users with 'bot' rights choose when edits are bot edits or not. That information
 356                 # may be lost at this point (aside from joining on the patrol log table entries).
 357                 $botgroups = [ 'bot' ];
 358                 $autopatrolgroups = $wgUseRCPatrol ? User::getGroupsWithPermission( 'autopatrol' ) : [];
 359
 360                 # Flag our recent bot edits
 361                 if ( $botgroups ) {
 362                         $botwhere = $dbw->makeList( $botgroups );
 363
 364                         $this->output( "Flagging bot account edits...\n" );
 365
 366                         # Find all users that are bots
 367                         $sql = "SELECT DISTINCT user_name FROM $usergroups, $user " .
 368                                 "WHERE ug_group IN($botwhere) AND user_id = ug_user";
 369                         $res = $dbw->query( $sql, __METHOD__ );
 370
 371                         $botusers = [];
 372                         foreach ( $res as $obj ) {
 373                                 $botusers[] = $obj->user_name;
 374                         }
 375
 376                         # Fill in the rc_bot field
 377                         if ( $botusers ) {
 378                                 $rcids = $dbw->selectFieldValues(
 379                                         'recentchanges',
 380                                         'rc_id',
 381                                         [
 382                                                 'rc_user_text' => $botusers,
 383                                                 "rc_timestamp > " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 384                                                 "rc_timestamp < " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) )
 385                                         ],
 386                                         __METHOD__
 387                                 );
 388
 389                                 foreach ( array_chunk( $rcids, $this->mBatchSize ) as $rcidBatch ) {
 390                                         $dbw->update(
 391                                                 'recentchanges',
 392                                                 [ 'rc_bot' => 1 ],
 393                                                 [ 'rc_id' => $rcidBatch ],
 394                                                 __METHOD__
 395                                         );
 396                                         wfGetLBFactory()->waitForReplication();
 397                                 }
 398                         }
 399                 }
 400
 401                 # Flag our recent autopatrolled edits
 402                 if ( !$wgMiserMode && $autopatrolgroups ) {
 403                         $patrolwhere = $dbw->makeList( $autopatrolgroups );
 404                         $patrolusers = [];
 405
 406                         $this->output( "Flagging auto-patrolled edits...\n" );
 407
 408                         # Find all users in RC with autopatrol rights
 409                         $sql = "SELECT DISTINCT user_name FROM $usergroups, $user " .
 410                                 "WHERE ug_group IN($patrolwhere) AND user_id = ug_user";
 411                         $res = $dbw->query( $sql, __METHOD__ );
 412
 413                         foreach ( $res as $obj ) {
 414                                 $patrolusers[] = $dbw->addQuotes( $obj->user_name );
 415                         }
 416
 417                         # Fill in the rc_patrolled field
 418                         if ( $patrolusers ) {
 419                                 $patrolwhere = implode( ',', $patrolusers );
 420                                 $sql2 = "UPDATE $recentchanges SET rc_patrolled=1 " .
 421                                         "WHERE rc_user_text IN($patrolwhere) " .
 422                                         "AND rc_timestamp > " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ) . ' ' .
 423                                         "AND rc_timestamp < " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) );
 424                                 $dbw->query( $sql2 );
 425                         }
 426                 }
 427         }
 428
 429         /**
 430          * Rebuild pass 5: Delete duplicate entries where we generate both a page revision and a log entry
 431          * for a single action (upload only, at the moment, but potentially also move, protect, ...).
 432          */
 433         private function rebuildRecentChangesTablePass5() {
 434                 $dbw = wfGetDB( DB_MASTER );
 435
 436                 $this->output( "Removing duplicate revision and logging entries...\n" );
 437
 438                 $res = $dbw->select(
 439                         [ 'logging', 'log_search' ],
 440                         [ 'ls_value', 'ls_log_id' ],
 441                         [
 442                                 'ls_log_id = log_id',
 443                                 'ls_field' => 'associated_rev_id',
 444                                 'log_type' => 'upload',
 445                                 'log_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom ) ),
 446                                 'log_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo ) ),
 447                         ],
 448                         __METHOD__
 449                 );
 450
 451                 $updates = 0;
 452                 foreach ( $res as $obj ) {
 453                         $rev_id = $obj->ls_value;
 454                         $log_id = $obj->ls_log_id;
 455
 456                         // Mark the logging row as having an associated rev id
 457                         $dbw->update(
 458                                 'recentchanges',
 459                                 /*SET*/ [ 'rc_this_oldid' => $rev_id ],
 460                                 /*WHERE*/ [ 'rc_logid' => $log_id ],
 461                                 __METHOD__
 462                         );
 463
 464                         // Delete the revision row
 465                         $dbw->delete(
 466                                 'recentchanges',
 467                                 /*WHERE*/ [ 'rc_this_oldid' => $rev_id, 'rc_logid' => 0 ],
 468                                 __METHOD__
 469                         );
 470
 471                         if ( ( ++$updates % $this->mBatchSize ) == 0 ) {
 472                                 wfGetLBFactory()->waitForReplication();
 473                         }
 474                 }
 475         }
 476
 477         /**
 478          * Purge cached feeds in $wanCache
 479          */
 480         private function purgeFeeds() {
 481                 global $wgFeedClasses;
 482
 483                 $this->output( "Deleting feed timestamps.\n" );
 484
 485                 $wanCache = MediaWikiServices::getInstance()->getMainWANObjectCache();
 486                 foreach ( $wgFeedClasses as $feed => $className ) {
 487                         $wanCache->delete( $wanCache->makeKey( 'rcfeed', $feed, 'timestamp' ) ); # Good enough for now.
 488                 }
 489         }
 490 }
 491
 492 $maintClass = "RebuildRecentchanges";
 493 require_once RUN_MAINTENANCE_IF_MAIN;