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 use MediaWiki\MediaWikiServices;
  36
  37 require_once __DIR__ . '/Maintenance.php';
  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                 $processed = $added = $ignored = $skipped = $overwritten = $failed = 0;
 131
 132                 $this->output( "Importing Files\n\n" );
 133
 134                 $dir = $this->getArg( 0 );
 135
 136                 # Check Protection
 137                 if ( $this->hasOption( 'protect' ) && $this->hasOption( 'unprotect' ) ) {
 138                         $this->fatalError( "Cannot specify both protect and unprotect.  Only 1 is allowed.\n" );
 139                 }
 140
 141                 if ( $this->hasOption( 'protect' ) && trim( $this->getOption( 'protect' ) ) ) {
 142                         $this->fatalError( "You must specify a protection option.\n" );
 143                 }
 144
 145                 # Prepare the list of allowed extensions
 146                 $extensions = $this->hasOption( 'extensions' )
 147                         ? explode( ',', strtolower( $this->getOption( 'extensions' ) ) )
 148                         : $wgFileExtensions;
 149
 150                 # Search the path provided for candidates for import
 151                 $files = $this->findFiles( $dir, $extensions, $this->hasOption( 'search-recursively' ) );
 152
 153                 # Initialise the user for this operation
 154                 $user = $this->hasOption( 'user' )
 155                         ? User::newFromName( $this->getOption( 'user' ) )
 156                         : User::newSystemUser( 'Maintenance script', [ 'steal' => true ] );
 157                 if ( !$user instanceof User ) {
 158                         $user = User::newSystemUser( 'Maintenance script', [ 'steal' => true ] );
 159                 }
 160                 $wgUser = $user;
 161
 162                 # Get block check. If a value is given, this specified how often the check is performed
 163                 $checkUserBlock = (int)$this->getOption( 'check-userblock' );
 164
 165                 $from = $this->getOption( 'from' );
 166                 $sleep = (int)$this->getOption( 'sleep' );
 167                 $limit = (int)$this->getOption( 'limit' );
 168                 $timestamp = $this->getOption( 'timestamp', false );
 169
 170                 # Get the upload comment. Provide a default one in case there's no comment given.
 171                 $commentFile = $this->getOption( 'comment-file' );
 172                 if ( $commentFile !== null ) {
 173                         $comment = file_get_contents( $commentFile );
 174                         if ( $comment === false || $comment === null ) {
 175                                 $this->fatalError( "failed to read comment file: {$commentFile}\n" );
 176                         }
 177                 } else {
 178                         $comment = $this->getOption( 'comment', 'Importing file' );
 179                 }
 180                 $commentExt = $this->getOption( 'comment-ext' );
 181                 $summary = $this->getOption( 'summary', '' );
 182
 183                 $license = $this->getOption( 'license', '' );
 184
 185                 $sourceWikiUrl = $this->getOption( 'source-wiki-url' );
 186
 187                 # Batch "upload" operation
 188                 $count = count( $files );
 189                 if ( $count > 0 ) {
 190                         foreach ( $files as $file ) {
 191                                 if ( $sleep && ( $processed > 0 ) ) {
 192                                         sleep( $sleep );
 193                                 }
 194
 195                                 $base = UtfNormal\Validator::cleanUp( wfBaseName( $file ) );
 196
 197                                 # Validate a title
 198                                 $title = Title::makeTitleSafe( NS_FILE, $base );
 199                                 if ( !is_object( $title ) ) {
 200                                         $this->output(
 201                                                 "{$base} could not be imported; a valid title cannot be produced\n" );
 202                                         continue;
 203                                 }
 204
 205                                 if ( $from ) {
 206                                         if ( $from == $title->getDBkey() ) {
 207                                                 $from = null;
 208                                         } else {
 209                                                 $ignored++;
 210                                                 continue;
 211                                         }
 212                                 }
 213
 214                                 if ( $checkUserBlock && ( ( $processed % $checkUserBlock ) == 0 ) ) {
 215                                         $user->clearInstanceCache( 'name' ); // reload from DB!
 216                                         // @TODO Use PermissionManager::isBlockedFrom() instead.
 217                                         if ( $user->getBlock() ) {
 218                                                 $this->output( $user->getName() . " was blocked! Aborting.\n" );
 219                                                 break;
 220                                         }
 221                                 }
 222
 223                                 # Check existence
 224                                 $image = MediaWikiServices::getInstance()->getRepoGroup()->getLocalRepo()
 225                                         ->newFile( $title );
 226                                 if ( $image->exists() ) {
 227                                         if ( $this->hasOption( 'overwrite' ) ) {
 228                                                 $this->output( "{$base} exists, overwriting..." );
 229                                                 $svar = 'overwritten';
 230                                         } else {
 231                                                 $this->output( "{$base} exists, skipping\n" );
 232                                                 $skipped++;
 233                                                 continue;
 234                                         }
 235                                 } else {
 236                                         if ( $this->hasOption( 'skip-dupes' ) ) {
 237                                                 $repo = $image->getRepo();
 238                                                 # XXX: we end up calculating this again when actually uploading. that sucks.
 239                                                 $sha1 = FSFile::getSha1Base36FromPath( $file );
 240
 241                                                 $dupes = $repo->findBySha1( $sha1 );
 242
 243                                                 if ( $dupes ) {
 244                                                         $this->output(
 245                                                                 "{$base} already exists as {$dupes[0]->getName()}, skipping\n" );
 246                                                         $skipped++;
 247                                                         continue;
 248                                                 }
 249                                         }
 250
 251                                         $this->output( "Importing {$base}..." );
 252                                         $svar = 'added';
 253                                 }
 254
 255                                 if ( $sourceWikiUrl ) {
 256                                         /* find comment text directly from source wiki, through MW's API */
 257                                         $real_comment = $this->getFileCommentFromSourceWiki( $sourceWikiUrl, $base );
 258                                         if ( $real_comment === false ) {
 259                                                 $commentText = $comment;
 260                                         } else {
 261                                                 $commentText = $real_comment;
 262                                         }
 263
 264                                         /* find user directly from source wiki, through MW's API */
 265                                         $real_user = $this->getFileUserFromSourceWiki( $sourceWikiUrl, $base );
 266                                         if ( $real_user === false ) {
 267                                                 $wgUser = $user;
 268                                         } else {
 269                                                 $wgUser = User::newFromName( $real_user );
 270                                                 if ( $wgUser === false ) {
 271                                                         # user does not exist in target wiki
 272                                                         $this->output(
 273                                                                 "failed: user '$real_user' does not exist in target wiki." );
 274                                                         continue;
 275                                                 }
 276                                         }
 277                                 } else {
 278                                         # Find comment text
 279                                         $commentText = false;
 280
 281                                         if ( $commentExt ) {
 282                                                 $f = $this->findAuxFile( $file, $commentExt );
 283                                                 if ( !$f ) {
 284                                                         $this->output( " No comment file with extension {$commentExt} found "
 285                                                                  . "for {$file}, using default comment. " );
 286                                                 } else {
 287                                                         $commentText = file_get_contents( $f );
 288                                                         if ( !$commentText ) {
 289                                                                 $this->output(
 290                                                                         " Failed to load comment file {$f}, using default comment. " );
 291                                                         }
 292                                                 }
 293                                         }
 294
 295                                         if ( !$commentText ) {
 296                                                 $commentText = $comment;
 297                                         }
 298                                 }
 299
 300                                 # Import the file
 301                                 if ( $this->hasOption( 'dry' ) ) {
 302                                         $this->output(
 303                                                 " publishing {$file} by '{$wgUser->getName()}', comment '$commentText'... "
 304                                         );
 305                                 } else {
 306                                         $mwProps = new MWFileProps( MediaWiki\MediaWikiServices::getInstance()->getMimeAnalyzer() );
 307                                         $props = $mwProps->getPropsFromPath( $file, true );
 308                                         $flags = 0;
 309                                         $publishOptions = [];
 310                                         $handler = MediaHandler::getHandler( $props['mime'] );
 311                                         if ( $handler ) {
 312                                                 $metadata = \Wikimedia\AtEase\AtEase::quietCall( 'unserialize', $props['metadata'] );
 313
 314                                                 $publishOptions['headers'] = $handler->getContentHeaders( $metadata );
 315                                         } else {
 316                                                 $publishOptions['headers'] = [];
 317                                         }
 318                                         $archive = $image->publish( $file, $flags, $publishOptions );
 319                                         if ( !$archive->isGood() ) {
 320                                                 $this->output( "failed. (" .
 321                                                          $archive->getWikiText( false, false, 'en' ) .
 322                                                          ")\n" );
 323                                                 $failed++;
 324                                                 continue;
 325                                         }
 326                                 }
 327
 328                                 $commentText = SpecialUpload::getInitialPageText( $commentText, $license );
 329                                 if ( !$this->hasOption( 'summary' ) ) {
 330                                         $summary = $commentText;
 331                                 }
 332
 333                                 if ( $this->hasOption( 'dry' ) ) {
 334                                         $this->output( "done.\n" );
 335                                         // @phan-suppress-next-line PhanUndeclaredMethod
 336                                 } elseif ( $image->recordUpload2(
 337                                         $archive->value,
 338                                         $summary,
 339                                         $commentText,
 340                                         $props,
 341                                         $timestamp
 342                                 )->isOK() ) {
 343                                         $this->output( "done.\n" );
 344
 345                                         $doProtect = false;
 346
 347                                         $protectLevel = $this->getOption( 'protect' );
 348
 349                                         if ( $protectLevel && in_array( $protectLevel, $wgRestrictionLevels ) ) {
 350                                                 $doProtect = true;
 351                                         }
 352                                         if ( $this->hasOption( 'unprotect' ) ) {
 353                                                 $protectLevel = '';
 354                                                 $doProtect = true;
 355                                         }
 356
 357                                         if ( $doProtect ) {
 358                                                 # Protect the file
 359                                                 $this->output( "\nWaiting for replica DBs...\n" );
 360                                                 // Wait for replica DBs.
 361                                                 sleep( 2 ); # Why this sleep?
 362                                                 wfWaitForSlaves();
 363
 364                                                 $this->output( "\nSetting image restrictions ... " );
 365
 366                                                 $cascade = false;
 367                                                 $restrictions = [];
 368                                                 foreach ( $title->getRestrictionTypes() as $type ) {
 369                                                         $restrictions[$type] = $protectLevel;
 370                                                 }
 371
 372                                                 $page = WikiPage::factory( $title );
 373                                                 $status = $page->doUpdateRestrictions( $restrictions, [], $cascade, '', $user );
 374                                                 $this->output( ( $status->isOK() ? 'done' : 'failed' ) . "\n" );
 375                                         }
 376                                 } else {
 377                                         $this->output( "failed. (at recordUpload stage)\n" );
 378                                         $svar = 'failed';
 379                                 }
 380
 381                                 $$svar++;
 382                                 $processed++;
 383
 384                                 if ( $limit && $processed >= $limit ) {
 385                                         break;
 386                                 }
 387                         }
 388
 389                         # Print out some statistics
 390                         $this->output( "\n" );
 391                         foreach (
 392                                 [
 393                                         'count' => 'Found',
 394                                         'limit' => 'Limit',
 395                                         'ignored' => 'Ignored',
 396                                         'added' => 'Added',
 397                                         'skipped' => 'Skipped',
 398                                         'overwritten' => 'Overwritten',
 399                                         'failed' => 'Failed'
 400                                 ] as $var => $desc
 401                         ) {
 402                                 if ( $$var > 0 ) {
 403                                         $this->output( "{$desc}: {$$var}\n" );
 404                                 }
 405                         }
 406                 } else {
 407                         $this->output( "No suitable files could be found for import.\n" );
 408                 }
 409         }
 410
 411         /**
 412          * Search a directory for files with one of a set of extensions
 413          *
 414          * @param string $dir Path to directory to search
 415          * @param array $exts Array of extensions to search for
 416          * @param bool $recurse Search subdirectories recursively
 417          * @return array|bool Array of filenames on success, or false on failure
 418          */
 419         private function findFiles( $dir, $exts, $recurse = false ) {
 420                 if ( is_dir( $dir ) ) {
 421                         $dhl = opendir( $dir );
 422                         if ( $dhl ) {
 423                                 $files = [];
 424                                 while ( ( $file = readdir( $dhl ) ) !== false ) {
 425                                         if ( is_file( $dir . '/' . $file ) ) {
 426                                                 $ext = pathinfo( $file, PATHINFO_EXTENSION );
 427                                                 if ( array_search( strtolower( $ext ), $exts ) !== false ) {
 428                                                         $files[] = $dir . '/' . $file;
 429                                                 }
 430                                         } elseif ( $recurse && is_dir( $dir . '/' . $file ) && $file !== '..' && $file !== '.' ) {
 431                                                 $files = array_merge( $files, $this->findFiles( $dir . '/' . $file, $exts, true ) );
 432                                         }
 433                                 }
 434
 435                                 return $files;
 436                         } else {
 437                                 return [];
 438                         }
 439                 } else {
 440                         return [];
 441                 }
 442         }
 443
 444         /**
 445          * Find an auxilliary file with the given extension, matching
 446          * the give base file path. $maxStrip determines how many extensions
 447          * may be stripped from the original file name before appending the
 448          * new extension. For example, with $maxStrip = 1 (the default),
 449          * file files acme.foo.bar.txt and acme.foo.txt would be auxilliary
 450          * files for acme.foo.bar and the extension ".txt". With $maxStrip = 2,
 451          * acme.txt would also be acceptable.
 452          *
 453          * @param string $file Base path
 454          * @param string $auxExtension The extension to be appended to the base path
 455          * @param int $maxStrip The maximum number of extensions to strip from the base path (default: 1)
 456          * @return string|bool
 457          */
 458         private function findAuxFile( $file, $auxExtension, $maxStrip = 1 ) {
 459                 if ( strpos( $auxExtension, '.' ) !== 0 ) {
 460                         $auxExtension = '.' . $auxExtension;
 461                 }
 462
 463                 $d = dirname( $file );
 464                 $n = basename( $file );
 465
 466                 while ( $maxStrip >= 0 ) {
 467                         $f = $d . '/' . $n . $auxExtension;
 468
 469                         if ( file_exists( $f ) ) {
 470                                 return $f;
 471                         }
 472
 473                         $idx = strrpos( $n, '.' );
 474                         if ( !$idx ) {
 475                                 break;
 476                         }
 477
 478                         $n = substr( $n, 0, $idx );
 479                         $maxStrip -= 1;
 480                 }
 481
 482                 return false;
 483         }
 484
 485         # @todo FIXME: Access the api in a saner way and performing just one query
 486         # (preferably batching files too).
 487         private function getFileCommentFromSourceWiki( $wiki_host, $file ) {
 488                 $url = $wiki_host . '/api.php?action=query&format=xml&titles=File:'
 489                         . rawurlencode( $file ) . '&prop=imageinfo&&iiprop=comment';
 490                 $body = Http::get( $url, [], __METHOD__ );
 491                 if ( preg_match( '#<ii comment="([^"]*)" />#', $body, $matches ) == 0 ) {
 492                         return false;
 493                 }
 494
 495                 return html_entity_decode( $matches[1] );
 496         }
 497
 498         private function getFileUserFromSourceWiki( $wiki_host, $file ) {
 499                 $url = $wiki_host . '/api.php?action=query&format=xml&titles=File:'
 500                         . rawurlencode( $file ) . '&prop=imageinfo&&iiprop=user';
 501                 $body = Http::get( $url, [], __METHOD__ );
 502                 if ( preg_match( '#<ii user="([^"]*)" />#', $body, $matches ) == 0 ) {
 503                         return false;
 504                 }
 505
 506                 return html_entity_decode( $matches[1] );
 507         }
 508
 509 }
 510
 511 $maintClass = ImportImages::class;
 512 require_once RUN_MAINTENANCE_IF_MAIN;