maintenance/importTextFiles.php

   1 <?php
   2 /**
   3  * Import pages from text files
   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\MediaWikiServices;
  25
  26 require_once __DIR__ . '/Maintenance.php';
  27
  28 /**
  29  * Maintenance script which reads in text files
  30  * and imports their content to a page of the wiki.
  31  *
  32  * @ingroup Maintenance
  33  */
  34 class ImportTextFiles extends Maintenance {
  35         public function __construct() {
  36                 parent::__construct();
  37                 $this->addDescription( 'Reads in text files and imports their content to pages of the wiki' );
  38                 $this->addOption( 'user', 'Username to which edits should be attributed. ' .
  39                         'Default: "Maintenance script"', false, true, 'u' );
  40                 $this->addOption( 'summary', 'Specify edit summary for the edits', false, true, 's' );
  41                 $this->addOption( 'use-timestamp', 'Use the modification date of the text file ' .
  42                         'as the timestamp for the edit' );
  43                 $this->addOption( 'overwrite', 'Overwrite existing pages. If --use-timestamp is passed, this ' .
  44                         'will only overwrite pages if the file has been modified since the page was last modified.' );
  45                 $this->addOption( 'prefix', 'A string to place in front of the file name', false, true, 'p' );
  46                 $this->addOption( 'bot', 'Mark edits as bot edits in the recent changes list.' );
  47                 $this->addOption( 'rc', 'Place revisions in RecentChanges.' );
  48                 $this->addArg( 'files', 'Files to import' );
  49         }
  50
  51         public function execute() {
  52                 $userName = $this->getOption( 'user', false );
  53                 $summary = $this->getOption( 'summary', 'Imported from text file' );
  54                 $useTimestamp = $this->hasOption( 'use-timestamp' );
  55                 $rc = $this->hasOption( 'rc' );
  56                 $bot = $this->hasOption( 'bot' );
  57                 $overwrite = $this->hasOption( 'overwrite' );
  58                 $prefix = $this->getOption( 'prefix', '' );
  59
  60                 // Get all the arguments. A loop is required since Maintenance doesn't
  61                 // support an arbitrary number of arguments.
  62                 $files = [];
  63                 $i = 0;
  64                 while ( $arg = $this->getArg( $i++ ) ) {
  65                         if ( file_exists( $arg ) ) {
  66                                 $files[$arg] = file_get_contents( $arg );
  67                         } else {
  68                                 // use glob to support the Windows shell, which doesn't automatically
  69                                 // expand wildcards
  70                                 $found = false;
  71                                 foreach ( glob( $arg ) as $filename ) {
  72                                         $found = true;
  73                                         $files[$filename] = file_get_contents( $filename );
  74                                 }
  75                                 if ( !$found ) {
  76                                         $this->fatalError( "Fatal error: The file '$arg' does not exist!" );
  77                                 }
  78                         }
  79                 };
  80
  81                 $count = count( $files );
  82                 $this->output( "Importing $count pages...\n" );
  83
  84                 if ( $userName === false ) {
  85                         $user = User::newSystemUser( 'Maintenance script', [ 'steal' => true ] );
  86                 } else {
  87                         $user = User::newFromName( $userName );
  88                 }
  89
  90                 if ( !$user ) {
  91                         $this->fatalError( "Invalid username\n" );
  92                 }
  93                 if ( $user->isAnon() ) {
  94                         $user->addToDatabase();
  95                 }
  96
  97                 $exit = 0;
  98
  99                 $successCount = 0;
 100                 $failCount = 0;
 101                 $skipCount = 0;
 102
 103                 foreach ( $files as $file => $text ) {
 104                         $pageName = $prefix . pathinfo( $file, PATHINFO_FILENAME );
 105                         $timestamp = $useTimestamp ? wfTimestamp( TS_UNIX, filemtime( $file ) ) : wfTimestampNow();
 106
 107                         $title = Title::newFromText( $pageName );
 108                         // Have to check for # manually, since it gets interpreted as a fragment
 109                         if ( !$title || $title->hasFragment() ) {
 110                                 $this->error( "Invalid title $pageName. Skipping.\n" );
 111                                 $skipCount++;
 112                                 continue;
 113                         }
 114
 115                         $exists = $title->exists();
 116                         $oldRevID = $title->getLatestRevID();
 117                         $oldRev = $oldRevID ? Revision::newFromId( $oldRevID ) : null;
 118                         $actualTitle = $title->getPrefixedText();
 119
 120                         if ( $exists ) {
 121                                 $touched = wfTimestamp( TS_UNIX, $title->getTouched() );
 122                                 if ( !$overwrite ) {
 123                                         $this->output( "Title $actualTitle already exists. Skipping.\n" );
 124                                         $skipCount++;
 125                                         continue;
 126                                 } elseif ( $useTimestamp && intval( $touched ) >= intval( $timestamp ) ) {
 127                                         $this->output( "File for title $actualTitle has not been modified since the " .
 128                                                 "destination page was touched. Skipping.\n" );
 129                                         $skipCount++;
 130                                         continue;
 131                                 }
 132                         }
 133
 134                         $rev = new WikiRevision( MediaWikiServices::getInstance()->getMainConfig() );
 135                         $rev->setText( rtrim( $text ) );
 136                         $rev->setTitle( $title );
 137                         $rev->setUserObj( $user );
 138                         $rev->setComment( $summary );
 139                         $rev->setTimestamp( $timestamp );
 140
 141                         if ( $exists && $overwrite && $rev->getContent()->equals( $oldRev->getContent() ) ) {
 142                                 $this->output( "File for title $actualTitle contains no changes from the current " .
 143                                         "revision. Skipping.\n" );
 144                                 $skipCount++;
 145                                 continue;
 146                         }
 147
 148                         $status = $rev->importOldRevision();
 149                         $newId = $title->getLatestRevID();
 150
 151                         if ( $status ) {
 152                                 $action = $exists ? 'updated' : 'created';
 153                                 $this->output( "Successfully $action $actualTitle\n" );
 154                                 $successCount++;
 155                         } else {
 156                                 $action = $exists ? 'update' : 'create';
 157                                 $this->output( "Failed to $action $actualTitle\n" );
 158                                 $failCount++;
 159                                 $exit = 1;
 160                         }
 161
 162                         // Create the RecentChanges entry if necessary
 163                         if ( $rc && $status ) {
 164                                 if ( $exists ) {
 165                                         if ( is_object( $oldRev ) ) {
 166                                                 $oldContent = $oldRev->getContent();
 167                                                 RecentChange::notifyEdit(
 168                                                         $timestamp,
 169                                                         $title,
 170                                                         $rev->getMinor(),
 171                                                         $user,
 172                                                         $summary,
 173                                                         $oldRevID,
 174                                                         $oldRev->getTimestamp(),
 175                                                         $bot,
 176                                                         '',
 177                                                         $oldContent ? $oldContent->getSize() : 0,
 178                                                         $rev->getContent()->getSize(),
 179                                                         $newId,
 180                                                         1 /* the pages don't need to be patrolled */
 181                                                 );
 182                                         }
 183                                 } else {
 184                                         RecentChange::notifyNew(
 185                                                 $timestamp,
 186                                                 $title,
 187                                                 $rev->getMinor(),
 188                                                 $user,
 189                                                 $summary,
 190                                                 $bot,
 191                                                 '',
 192                                                 $rev->getContent()->getSize(),
 193                                                 $newId,
 194                                                 1
 195                                         );
 196                                 }
 197                         }
 198                 }
 199
 200                 $this->output( "Done! $successCount succeeded, $skipCount skipped.\n" );
 201                 if ( $exit ) {
 202                         $this->fatalError( "Import failed with $failCount failed pages.\n", $exit );
 203                 }
 204         }
 205 }
 206
 207 $maintClass = ImportTextFiles::class;
 208 require_once RUN_MAINTENANCE_IF_MAIN;