maintenance/cleanupSpam.php

   1 <?php
   2 /**
   3  * Cleanup all spam from a given hostname.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup Maintenance
  22  */
  23
  24 use MediaWiki\ExternalLinks\LinkFilter;
  25 use MediaWiki\Permissions\Authority;
  26 use MediaWiki\Revision\RevisionRecord;
  27 use MediaWiki\Revision\SlotRecord;
  28 use MediaWiki\StubObject\StubGlobalUser;
  29 use MediaWiki\Title\Title;
  30 use MediaWiki\User\User;
  31
  32 require_once __DIR__ . '/Maintenance.php';
  33
  34 /**
  35  * Maintenance script to cleanup all spam from a given hostname.
  36  *
  37  * @ingroup Maintenance
  38  */
  39 class CleanupSpam extends Maintenance {
  40
  41         public function __construct() {
  42                 parent::__construct();
  43                 $this->addDescription( 'Cleanup all spam from a given hostname' );
  44                 $this->addOption( 'all', 'Check all wikis in $wgLocalDatabases' );
  45                 $this->addOption( 'delete', 'Delete pages containing only spam instead of blanking them' );
  46                 $this->addArg(
  47                         'hostname',
  48                         'Hostname that was spamming, single * wildcard in the beginning allowed'
  49                 );
  50         }
  51
  52         public function execute() {
  53                 global $IP, $wgLocalDatabases;
  54
  55                 $username = wfMessage( 'spambot_username' )->text();
  56                 $user = User::newSystemUser( $username );
  57                 if ( !$user ) {
  58                         $this->fatalError( "Invalid username specified in 'spambot_username' message: $username" );
  59                 }
  60                 // Hack: Grant bot rights so we don't flood RecentChanges
  61                 $this->getServiceContainer()->getUserGroupManager()->addUserToGroup( $user, 'bot' );
  62                 StubGlobalUser::setUser( $user );
  63
  64                 $spec = $this->getArg( 0 );
  65
  66                 $protConds = [];
  67                 foreach ( [ 'http://', 'https://' ] as $prot ) {
  68                         $conds = LinkFilter::getQueryConditions( $spec, [ 'protocol' => $prot ] );
  69                         if ( !$conds ) {
  70                                 $this->fatalError( "Not a valid hostname specification: $spec" );
  71                         }
  72                         $protConds[$prot] = $conds;
  73                 }
  74
  75                 if ( $this->hasOption( 'all' ) ) {
  76                         // Clean up spam on all wikis
  77                         $this->output( "Finding spam on " . count( $wgLocalDatabases ) . " wikis\n" );
  78                         $found = false;
  79                         foreach ( $wgLocalDatabases as $wikiId ) {
  80                                 /** @var Database $dbr */
  81                                 $dbr = $this->getDB( DB_REPLICA, [], $wikiId );
  82
  83                                 foreach ( $protConds as $conds ) {
  84                                         $count = $dbr->newSelectQueryBuilder()
  85                                                 ->select( 'COUNT(*)' )
  86                                                 ->from( 'externallinks' )
  87                                                 ->where( $conds )
  88                                                 ->caller( __METHOD__ )
  89                                                 ->fetchField();
  90                                         if ( $count ) {
  91                                                 $found = true;
  92                                                 $cmd = wfShellWikiCmd(
  93                                                         "$IP/maintenance/cleanupSpam.php",
  94                                                         [ '--wiki', $wikiId, $spec ]
  95                                                 );
  96                                                 // phpcs:ignore MediaWiki.Usage.ForbiddenFunctions.passthru
  97                                                 passthru( "$cmd | sed 's/^/$wikiId:  /'" );
  98                                         }
  99                                 }
 100                         }
 101                         if ( $found ) {
 102                                 $this->output( "All done\n" );
 103                         } else {
 104                                 $this->output( "None found\n" );
 105                         }
 106                 } else {
 107                         // Clean up spam on this wiki
 108
 109                         $count = 0;
 110                         /** @var Database $dbr */
 111                         $dbr = $this->getReplicaDB();
 112                         foreach ( $protConds as $prot => $conds ) {
 113                                 $res = $dbr->newSelectQueryBuilder()
 114                                         ->select( 'el_from' )
 115                                         ->distinct()
 116                                         ->from( 'externallinks' )
 117                                         ->where( $conds )
 118                                         ->caller( __METHOD__ )
 119                                         ->fetchResultSet();
 120                                 $count += $res->numRows();
 121                                 $this->output( "Found $count articles containing $spec so far...\n" );
 122                                 foreach ( $res as $row ) {
 123                                         $this->cleanupArticle(
 124                                                 $row->el_from,
 125                                                 $spec,
 126                                                 $prot,
 127                                                 $user
 128                                         );
 129                                 }
 130                         }
 131                         if ( $count ) {
 132                                 $this->output( "Done\n" );
 133                         }
 134                 }
 135         }
 136
 137         /**
 138          * @param int $id
 139          * @param string $domain
 140          * @param string $protocol
 141          * @param Authority $performer
 142          */
 143         private function cleanupArticle( $id, $domain, $protocol, Authority $performer ) {
 144                 $title = Title::newFromID( $id );
 145                 if ( !$title ) {
 146                         $this->error( "Internal error: no page for ID $id" );
 147
 148                         return;
 149                 }
 150
 151                 $this->output( $title->getPrefixedDBkey() . " ..." );
 152
 153                 $services = $this->getServiceContainer();
 154                 $revLookup = $services->getRevisionLookup();
 155                 $rev = $revLookup->getRevisionByTitle( $title );
 156                 $currentRevId = $rev->getId();
 157
 158                 while ( $rev && ( $rev->isDeleted( RevisionRecord::DELETED_TEXT ) ||
 159                         LinkFilter::matchEntry(
 160                                 // @phan-suppress-next-line PhanTypeMismatchArgumentNullable RAW never returns null
 161                                 $rev->getContent( SlotRecord::MAIN, RevisionRecord::RAW ),
 162                                 $domain,
 163                                 $protocol
 164                         ) )
 165                 ) {
 166                         $rev = $revLookup->getPreviousRevision( $rev );
 167                 }
 168
 169                 if ( $rev && $rev->getId() == $currentRevId ) {
 170                         // The regex didn't match the current article text
 171                         // This happens e.g. when a link comes from a template rather than the page itself
 172                         $this->output( "False match\n" );
 173                 } else {
 174                         $dbw = $this->getPrimaryDB();
 175                         $this->beginTransaction( $dbw, __METHOD__ );
 176                         $page = $services->getWikiPageFactory()->newFromTitle( $title );
 177                         if ( $rev ) {
 178                                 // Revert to this revision
 179                                 $content = $rev->getContent( SlotRecord::MAIN, RevisionRecord::RAW );
 180
 181                                 $this->output( "reverting\n" );
 182                                 $page->doUserEditContent(
 183                                         // @phan-suppress-next-line PhanTypeMismatchArgumentNullable RAW never returns null
 184                                         $content,
 185                                         $performer,
 186                                         wfMessage( 'spam_reverting', $domain )->inContentLanguage()->text(),
 187                                         EDIT_UPDATE | EDIT_FORCE_BOT,
 188                                         $rev->getId()
 189                                 );
 190                         } elseif ( $this->hasOption( 'delete' ) ) {
 191                                 // Didn't find a non-spammy revision, blank the page
 192                                 $this->output( "deleting\n" );
 193                                 $deletePage = $services->getDeletePageFactory()->newDeletePage( $page, $performer );
 194                                 $deletePage->deleteUnsafe( wfMessage( 'spam_deleting', $domain )->inContentLanguage()->text() );
 195                         } else {
 196                                 // Didn't find a non-spammy revision, blank the page
 197                                 $handler = $services->getContentHandlerFactory()
 198                                         ->getContentHandler( $title->getContentModel() );
 199                                 $content = $handler->makeEmptyContent();
 200
 201                                 $this->output( "blanking\n" );
 202                                 $page->doUserEditContent(
 203                                         $content,
 204                                         $performer,
 205                                         wfMessage( 'spam_blanking', $domain )->inContentLanguage()->text(),
 206                                         EDIT_UPDATE | EDIT_FORCE_BOT
 207                                 );
 208                         }
 209                         $this->commitTransaction( $dbw, __METHOD__ );
 210                 }
 211         }
 212 }
 213
 214 $maintClass = CleanupSpam::class;
 215 require_once RUN_MAINTENANCE_IF_MAIN;