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