maintenance/importImages.php

   1 <?php
   2 /**
   3  * Import one or more images from the local file system into the wiki without
   4  * using the web-based interface.
   5  *
   6  * "Smart import" additions:
   7  * - aim: preserve the essential metadata (user, description) when importing media
   8  *   files from an existing wiki.
   9  * - process:
  10  *      - interface with the source wiki, don't use bare files only (see --source-wiki-url).
  11  *      - fetch metadata from source wiki for each file to import.
  12  *      - commit the fetched metadata to the destination wiki while submitting.
  13  *
  14  * This program is free software; you can redistribute it and/or modify
  15  * it under the terms of the GNU General Public License as published by
  16  * the Free Software Foundation; either version 2 of the License, or
  17  * (at your option) any later version.
  18  *
  19  * This program is distributed in the hope that it will be useful,
  20  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  21  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  22  * GNU General Public License for more details.
  23  *
  24  * You should have received a copy of the GNU General Public License along
  25  * with this program; if not, write to the Free Software Foundation, Inc.,
  26  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  27  * http://www.gnu.org/copyleft/gpl.html
  28  *
  29  * @file
  30  * @ingroup Maintenance
  31  * @author Rob Church <robchur@gmail.com>
  32  * @author Mij <mij@bitchx.it>
  33  */
  34
  35 require_once __DIR__ . '/Maintenance.php';
  36
  37 use MediaWiki\MediaWikiServices;
  38
  39 class ImportImages extends Maintenance {
  40
  41         public function __construct() {
  42                 parent::__construct();
  43
  44                 $this->addDescription( 'Imports images and other media files into the wiki' );
  45                 $this->addArg( 'dir', 'Path to the directory containing images to be imported' );
  46
  47                 $this->addOption( 'extensions',
  48                         'Comma-separated list of allowable extensions, defaults to $wgFileExtensions',
  49                         false,
  50                         true
  51                 );
  52                 $this->addOption( 'overwrite',
  53                         'Overwrite existing images with the same name (default is to skip them)' );
  54                 $this->addOption( 'limit',
  55                         'Limit the number of images to process. Ignored or skipped images are not counted',
  56                         false,
  57                         true
  58                 );
  59                 $this->addOption( 'from',
  60                         "Ignore all files until the one with the given name. Useful for resuming aborted "
  61                                 . "imports. The name should be the file's canonical database form.",
  62                         false,
  63                         true
  64                 );
  65                 $this->addOption( 'skip-dupes',
  66                         'Skip images that were already uploaded under a different name (check SHA1)' );
  67                 $this->addOption( 'search-recursively', 'Search recursively for files in subdirectories' );
  68                 $this->addOption( 'sleep',
  69                         'Sleep between files. Useful mostly for debugging',
  70                         false,
  71                         true
  72                 );
  73                 $this->addOption( 'user',
  74                         "Set username of uploader, default 'Maintenance script'",
  75                         false,
  76                         true
  77                 );
  78                 // This parameter can optionally have an argument. If none specified, getOption()
  79                 // returns 1 which is precisely what we need.
  80                 $this->addOption( 'check-userblock', 'Check if the user got blocked during import' );
  81                 $this->addOption( 'comment',
  82                         "Set file description, default 'Importing file'",
  83                         false,
  84                         true
  85                 );
  86                 $this->addOption( 'comment-file',
  87                         'Set description to the content of this file',
  88                         false,
  89                         true
  90                 );
  91                 $this->addOption( 'comment-ext',
  92                         'Causes the description for each file to be loaded from a file with the same name, but '
  93                                 . 'the extension provided. If a global description is also given, it is appended.',
  94                         false,
  95                         true
  96                 );
  97                 $this->addOption( 'summary',
  98                         'Upload summary, description will be used if not provided',
  99                         false,
 100                         true
 101                 );
 102                 $this->addOption( 'license',
 103                         'Use an optional license template',
 104                         false,
 105                         true
 106                 );
 107                 $this->addOption( 'timestamp',
 108                         'Override upload time/date, all MediaWiki timestamp formats are accepted',
 109                         false,
 110                         true
 111                 );
 112                 $this->addOption( 'protect',
 113                         'Specify the protect value (autoconfirmed,sysop)',
 114                         false,
 115                         true
 116                 );
 117                 $this->addOption( 'unprotect', 'Unprotects all uploaded images' );
 118                 $this->addOption( 'source-wiki-url',
 119                         'If specified, take User and Comment data for each imported file from this URL. '
 120                                 . 'For example, --source-wiki-url="http://en.wikipedia.org/',
 121                         false,
 122                         true
 123                 );
 124                 $this->addOption( 'dry', "Dry run, don't import anything" );
 125         }
 126
 127         public function execute() {
 128                 global $wgFileExtensions, $wgUser, $wgRestrictionLevels;
 129
 130                 $permissionManager = MediaWikiServices::getInstance()->getPermissionManager();
 131
 132                 $processed = $added = $ignored = $skipped = $overwritten = $failed = 0;
 133
 134                 $this->output( "Importing Files\n\n" );
 135
 136                 $dir = $this->getArg( 0 );
 137
 138                 # Check Protection
 139                 if ( $this->hasOption( 'protect' ) && $this->hasOption( 'unprotect' ) ) {
 140                         $this->fatalError( "Cannot specify both protect and unprotect.  Only 1 is allowed.\n" );
 141                 }
 142
 143                 if ( $this->hasOption( 'protect' ) && trim( $this->getOption( 'protect' ) ) ) {
 144                         $this->fatalError( "You must specify a protection option.\n" );
 145                 }
 146
 147                 # Prepare the list of allowed extensions
 148                 $extensions = $this->hasOption( 'extensions' )
 149                         ? explode( ',', strtolower( $this->getOption( 'extensions' ) ) )
 150                         : $wgFileExtensions;
 151
 152                 # Search the path provided for candidates for import
 153                 $files = $this->findFiles( $dir, $extensions, $this->hasOption( 'search-recursively' ) );
 154
 155                 # Initialise the user for this operation
 156                 $user = $this->hasOption( 'user' )
 157                         ? User::newFromName( $this->getOption( 'user' ) )
 158                         : User::newSystemUser( 'Maintenance script', [ 'steal' => true ] );
 159                 if ( !$user instanceof User ) {
 160                         $user = User::newSystemUser( 'Maintenance script', [ 'steal' => true ] );
 161                 }
 162                 $wgUser = $user;
 163
 164                 # Get block check. If a value is given, this specified how often the check is performed
 165                 $checkUserBlock = (int)$this->getOption( 'check-userblock' );
 166
 167                 $from = $this->getOption( 'from' );
 168                 $sleep = (int)$this->getOption( 'sleep' );
 169                 $limit = (int)$this->getOption( 'limit' );
 170                 $timestamp = $this->getOption( 'timestamp', false );
 171
 172                 # Get the upload comment. Provide a default one in case there's no comment given.
 173                 $commentFile = $this->getOption( 'comment-file' );
 174                 if ( $commentFile !== null ) {
 175                         $comment = file_get_contents( $commentFile );
 176                         if ( $comment === false || $comment === null ) {
 177                                 $this->fatalError( "failed to read comment file: {$commentFile}\n" );
 178                         }
 179                 } else {
 180                         $comment = $this->getOption( 'comment', 'Importing file' );
 181                 }
 182                 $commentExt = $this->getOption( 'comment-ext' );
 183                 $summary = $this->getOption( 'summary', '' );
 184
 185                 $license = $this->getOption( 'license', '' );
 186
 187                 $sourceWikiUrl = $this->getOption( 'source-wiki-url' );
 188
 189                 # Batch "upload" operation
 190                 $count = count( $files );
 191                 if ( $count > 0 ) {
 192                         foreach ( $files as $file ) {
 193                                 if ( $sleep && ( $processed > 0 ) ) {
 194                                         sleep( $sleep );
 195                                 }
 196
 197                                 $base = UtfNormal\Validator::cleanUp( wfBaseName( $file ) );
 198
 199                                 # Validate a title
 200                                 $title = Title::makeTitleSafe( NS_FILE, $base );
 201                                 if ( !is_object( $title ) ) {
 202                                         $this->output(
 203                                                 "{$base} could not be imported; a valid title cannot be produced\n" );
 204                                         continue;
 205                                 }
 206
 207                                 if ( $from ) {
 208                                         if ( $from == $title->getDBkey() ) {
 209                                                 $from = null;
 210                                         } else {
 211                                                 $ignored++;
 212                                                 continue;
 213                                         }
 214                                 }
 215
 216                                 if ( $checkUserBlock && ( ( $processed % $checkUserBlock ) == 0 ) ) {
 217                                         $user->clearInstanceCache( 'name' ); // reload from DB!
 218                                         if ( $permissionManager->isBlockedFrom( $user, $title ) ) {
 219                                                 $this->output(
 220                                                         "{$user->getName()} is blocked from {$title->getPrefixedText()}! skipping.\n"
 221                                                 );
 222                                                 $skipped++;
 223                                                 continue;
 224                                         }
 225                                 }
 226
 227                                 # Check existence
 228                                 $image = MediaWikiServices::getInstance()->getRepoGroup()->getLocalRepo()
 229                                         ->newFile( $title );
 230                                 if ( $image->exists() ) {
 231                                         if ( $this->hasOption( 'overwrite' ) ) {
 232                                                 $this->output( "{$base} exists, overwriting..." );
 233                                                 $svar = 'overwritten';
 234                                         } else {
 235                                                 $this->output( "{$base} exists, skipping\n" );
 236                                                 $skipped++;
 237                                                 continue;
 238                                         }
 239                                 } else {
 240                                         if ( $this->hasOption( 'skip-dupes' ) ) {
 241                                                 $repo = $image->getRepo();
 242                                                 # XXX: we end up calculating this again when actually uploading. that sucks.
 243                                                 $sha1 = FSFile::getSha1Base36FromPath( $file );
 244
 245                                                 $dupes = $repo->findBySha1( $sha1 );
 246
 247                                                 if ( $dupes ) {
 248                                                         $this->output(
 249                                                                 "{$base} already exists as {$dupes[0]->getName()}, skipping\n" );
 250                                                         $skipped++;
 251                                                         continue;
 252                                                 }
 253                                         }
 254
 255                                         $this->output( "Importing {$base}..." );
 256                                         $svar = 'added';
 257                                 }
 258
 259                                 if ( $sourceWikiUrl ) {
 260                                         /* find comment text directly from source wiki, through MW's API */
 261                                         $real_comment = $this->getFileCommentFromSourceWiki( $sourceWikiUrl, $base );
 262                                         if ( $real_comment === false ) {
 263                                                 $commentText = $comment;
 264                                         } else {
 265                                                 $commentText = $real_comment;
 266                                         }
 267
 268                                         /* find user directly from source wiki, through MW's API */
 269                                         $real_user = $this->getFileUserFromSourceWiki( $sourceWikiUrl, $base );
 270                                         if ( $real_user === false ) {
 271                                                 $wgUser = $user;
 272                                         } else {
 273                                                 $wgUser = User::newFromName( $real_user );
 274                                                 if ( $wgUser === false ) {
 275                                                         # user does not exist in target wiki
 276                                                         $this->output(
 277                                                                 "failed: user '$real_user' does not exist in target wiki." );
 278                                                         continue;
 279                                                 }
 280                                         }
 281                                 } else {
 282                                         # Find comment text
 283                                         $commentText = false;
 284
 285                                         if ( $commentExt ) {
 286                                                 $f = $this->findAuxFile( $file, $commentExt );
 287                                                 if ( !$f ) {
 288                                                         $this->output( " No comment file with extension {$commentExt} found "
 289                                                                  . "for {$file}, using default comment. " );
 290                                                 } else {
 291                                                         $commentText = file_get_contents( $f );
 292                                                         if ( !$commentText ) {
 293                                                                 $this->output(
 294                                                                         " Failed to load comment file {$f}, using default comment. " );
 295                                                         }
 296                                                 }
 297                                         }
 298
 299                                         if ( !$commentText ) {
 300                                                 $commentText = $comment;
 301                                         }
 302                                 }
 303
 304                                 # Import the file
 305                                 if ( $this->hasOption( 'dry' ) ) {
 306                                         $this->output(
 307                                                 " publishing {$file} by '{$wgUser->getName()}', comment '$commentText'... "
 308                                         );
 309                                 } else {
 310                                         $mwProps = new MWFileProps( MediaWiki\MediaWikiServices::getInstance()->getMimeAnalyzer() );
 311                                         $props = $mwProps->getPropsFromPath( $file, true );
 312                                         $flags = 0;
 313                                         $publishOptions = [];
 314                                         $handler = MediaHandler::getHandler( $props['mime'] );
 315                                         if ( $handler ) {
 316                                                 $metadata = \Wikimedia\AtEase\AtEase::quietCall( 'unserialize', $props['metadata'] );
 317
 318                                                 $publishOptions['headers'] = $handler->getContentHeaders( $metadata );
 319                                         } else {
 320                                                 $publishOptions['headers'] = [];
 321                                         }
 322                                         $archive = $image->publish( $file, $flags, $publishOptions );
 323                                         if ( !$archive->isGood() ) {
 324                                                 $this->output( "failed. (" .
 325                                                          $archive->getWikiText( false, false, 'en' ) .
 326                                                          ")\n" );
 327                                                 $failed++;
 328                                                 continue;
 329                                         }
 330                                 }
 331
 332                                 $commentText = SpecialUpload::getInitialPageText( $commentText, $license );
 333                                 if ( !$this->hasOption( 'summary' ) ) {
 334                                         $summary = $commentText;
 335                                 }
 336
 337                                 if ( $this->hasOption( 'dry' ) ) {
 338                                         $this->output( "done.\n" );
 339                                 } elseif ( $image->recordUpload2(
 340                                         $archive->value,
 341                                         $summary,
 342                                         $commentText,
 343                                         $props,
 344                                         $timestamp
 345                                 )->isOK() ) {
 346                                         $this->output( "done.\n" );
 347
 348                                         $doProtect = false;
 349
 350                                         $protectLevel = $this->getOption( 'protect' );
 351
 352                                         if ( $protectLevel && in_array( $protectLevel, $wgRestrictionLevels ) ) {
 353                                                 $doProtect = true;
 354                                         }
 355                                         if ( $this->hasOption( 'unprotect' ) ) {
 356                                                 $protectLevel = '';
 357                                                 $doProtect = true;
 358                                         }
 359
 360                                         if ( $doProtect ) {
 361                                                 # Protect the file
 362                                                 $this->output( "\nWaiting for replica DBs...\n" );
 363                                                 // Wait for replica DBs.
 364                                                 sleep( 2 ); # Why this sleep?
 365                                                 wfWaitForSlaves();
 366
 367                                                 $this->output( "\nSetting image restrictions ... " );
 368
 369                                                 $cascade = false;
 370                                                 $restrictions = [];
 371                                                 foreach ( $title->getRestrictionTypes() as $type ) {
 372                                                         $restrictions[$type] = $protectLevel;
 373                                                 }
 374
 375                                                 $page = WikiPage::factory( $title );
 376                                                 $status = $page->doUpdateRestrictions( $restrictions, [], $cascade, '', $user );
 377                                                 $this->output( ( $status->isOK() ? 'done' : 'failed' ) . "\n" );
 378                                         }
 379                                 } else {
 380                                         $this->output( "failed. (at recordUpload stage)\n" );
 381                                         $svar = 'failed';
 382                                 }
 383
 384                                 $$svar++;
 385                                 $processed++;
 386
 387                                 if ( $limit && $processed >= $limit ) {
 388                                         break;
 389                                 }
 390                         }
 391
 392                         # Print out some statistics
 393                         $this->output( "\n" );
 394                         foreach (
 395                                 [
 396                                         'count' => 'Found',
 397                                         'limit' => 'Limit',
 398                                         'ignored' => 'Ignored',
 399                                         'added' => 'Added',
 400                                         'skipped' => 'Skipped',
 401                                         'overwritten' => 'Overwritten',
 402                                         'failed' => 'Failed'
 403                                 ] as $var => $desc
 404                         ) {
 405                                 if ( $$var > 0 ) {
 406                                         $this->output( "{$desc}: {$$var}\n" );
 407                                 }
 408                         }
 409                 } else {
 410                         $this->output( "No suitable files could be found for import.\n" );
 411                 }
 412         }
 413
 414         /**
 415          * Search a directory for files with one of a set of extensions
 416          *
 417          * @param string $dir Path to directory to search
 418          * @param array $exts Array of extensions to search for
 419          * @param bool $recurse Search subdirectories recursively
 420          * @return array|bool Array of filenames on success, or false on failure
 421          */
 422         private function findFiles( $dir, $exts, $recurse = false ) {
 423                 if ( is_dir( $dir ) ) {
 424                         $dhl = opendir( $dir );
 425                         if ( $dhl ) {
 426                                 $files = [];
 427                                 while ( ( $file = readdir( $dhl ) ) !== false ) {
 428                                         if ( is_file( $dir . '/' . $file ) ) {
 429                                                 $ext = pathinfo( $file, PATHINFO_EXTENSION );
 430                                                 if ( array_search( strtolower( $ext ), $exts ) !== false ) {
 431                                                         $files[] = $dir . '/' . $file;
 432                                                 }
 433                                         } elseif ( $recurse && is_dir( $dir . '/' . $file ) && $file !== '..' && $file !== '.' ) {
 434                                                 $files = array_merge( $files, $this->findFiles( $dir . '/' . $file, $exts, true ) );
 435                                         }
 436                                 }
 437
 438                                 return $files;
 439                         } else {
 440                                 return [];
 441                         }
 442                 } else {
 443                         return [];
 444                 }
 445         }
 446
 447         /**
 448          * Find an auxilliary file with the given extension, matching
 449          * the give base file path. $maxStrip determines how many extensions
 450          * may be stripped from the original file name before appending the
 451          * new extension. For example, with $maxStrip = 1 (the default),
 452          * file files acme.foo.bar.txt and acme.foo.txt would be auxilliary
 453          * files for acme.foo.bar and the extension ".txt". With $maxStrip = 2,
 454          * acme.txt would also be acceptable.
 455          *
 456          * @param string $file Base path
 457          * @param string $auxExtension The extension to be appended to the base path
 458          * @param int $maxStrip The maximum number of extensions to strip from the base path (default: 1)
 459          * @return string|bool
 460          */
 461         private function findAuxFile( $file, $auxExtension, $maxStrip = 1 ) {
 462                 if ( strpos( $auxExtension, '.' ) !== 0 ) {
 463                         $auxExtension = '.' . $auxExtension;
 464                 }
 465
 466                 $d = dirname( $file );
 467                 $n = basename( $file );
 468
 469                 while ( $maxStrip >= 0 ) {
 470                         $f = $d . '/' . $n . $auxExtension;
 471
 472                         if ( file_exists( $f ) ) {
 473                                 return $f;
 474                         }
 475
 476                         $idx = strrpos( $n, '.' );
 477                         if ( !$idx ) {
 478                                 break;
 479                         }
 480
 481                         $n = substr( $n, 0, $idx );
 482                         $maxStrip -= 1;
 483                 }
 484
 485                 return false;
 486         }
 487
 488         # @todo FIXME: Access the api in a saner way and performing just one query
 489         # (preferably batching files too).
 490         private function getFileCommentFromSourceWiki( $wiki_host, $file ) {
 491                 $url = $wiki_host . '/api.php?action=query&format=xml&titles=File:'
 492                         . rawurlencode( $file ) . '&prop=imageinfo&&iiprop=comment';
 493                 $body = Http::get( $url, [], __METHOD__ );
 494                 if ( preg_match( '#<ii comment="([^"]*)" />#', $body, $matches ) == 0 ) {
 495                         return false;
 496                 }
 497
 498                 return html_entity_decode( $matches[1] );
 499         }
 500
 501         private function getFileUserFromSourceWiki( $wiki_host, $file ) {
 502                 $url = $wiki_host . '/api.php?action=query&format=xml&titles=File:'
 503                         . rawurlencode( $file ) . '&prop=imageinfo&&iiprop=user';
 504                 $body = Http::get( $url, [], __METHOD__ );
 505                 if ( preg_match( '#<ii user="([^"]*)" />#', $body, $matches ) == 0 ) {
 506                         return false;
 507                 }
 508
 509                 return html_entity_decode( $matches[1] );
 510         }
 511
 512 }
 513
 514 $maintClass = ImportImages::class;
 515 require_once RUN_MAINTENANCE_IF_MAIN;