includes/media/DjVuHandler.php

   1 <?php
   2 /**
   3  * Handler for DjVu images.
   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 Media
  22  */
  23 use MediaWiki\MediaWikiServices;
  24 use MediaWiki\Shell\Shell;
  25
  26 /**
  27  * Handler for DjVu images
  28  *
  29  * @ingroup Media
  30  */
  31 class DjVuHandler extends ImageHandler {
  32         const EXPENSIVE_SIZE_LIMIT = 10485760; // 10MiB
  33
  34         /**
  35          * @return bool
  36          */
  37         public function isEnabled() {
  38                 global $wgDjvuRenderer, $wgDjvuDump, $wgDjvuToXML;
  39                 if ( !$wgDjvuRenderer || ( !$wgDjvuDump && !$wgDjvuToXML ) ) {
  40                         wfDebug( "DjVu is disabled, please set \$wgDjvuRenderer and \$wgDjvuDump\n" );
  41
  42                         return false;
  43                 } else {
  44                         return true;
  45                 }
  46         }
  47
  48         /**
  49          * @param File $file
  50          * @return bool
  51          */
  52         public function mustRender( $file ) {
  53                 return true;
  54         }
  55
  56         /**
  57          * True if creating thumbnails from the file is large or otherwise resource-intensive.
  58          * @param File $file
  59          * @return bool
  60          */
  61         public function isExpensiveToThumbnail( $file ) {
  62                 return $file->getSize() > static::EXPENSIVE_SIZE_LIMIT;
  63         }
  64
  65         /**
  66          * @param File $file
  67          * @return bool
  68          */
  69         public function isMultiPage( $file ) {
  70                 return true;
  71         }
  72
  73         /**
  74          * @return array
  75          */
  76         public function getParamMap() {
  77                 return [
  78                         'img_width' => 'width',
  79                         'img_page' => 'page',
  80                 ];
  81         }
  82
  83         /**
  84          * @param string $name
  85          * @param mixed $value
  86          * @return bool
  87          */
  88         public function validateParam( $name, $value ) {
  89                 if ( $name === 'page' && trim( $value ) !== (string)intval( $value ) ) {
  90                         // Extra junk on the end of page, probably actually a caption
  91                         // e.g. [[File:Foo.djvu|thumb|Page 3 of the document shows foo]]
  92                         return false;
  93                 }
  94                 if ( in_array( $name, [ 'width', 'height', 'page' ] ) ) {
  95                         if ( $value <= 0 ) {
  96                                 return false;
  97                         } else {
  98                                 return true;
  99                         }
 100                 } else {
 101                         return false;
 102                 }
 103         }
 104
 105         /**
 106          * @param array $params
 107          * @return bool|string
 108          */
 109         public function makeParamString( $params ) {
 110                 $page = $params['page'] ?? 1;
 111                 if ( !isset( $params['width'] ) ) {
 112                         return false;
 113                 }
 114
 115                 return "page{$page}-{$params['width']}px";
 116         }
 117
 118         /**
 119          * @param string $str
 120          * @return array|bool
 121          */
 122         public function parseParamString( $str ) {
 123                 $m = false;
 124                 if ( preg_match( '/^page(\d+)-(\d+)px$/', $str, $m ) ) {
 125                         return [ 'width' => $m[2], 'page' => $m[1] ];
 126                 } else {
 127                         return false;
 128                 }
 129         }
 130
 131         /**
 132          * @param array $params
 133          * @return array
 134          */
 135         protected function getScriptParams( $params ) {
 136                 return [
 137                         'width' => $params['width'],
 138                         'page' => $params['page'],
 139                 ];
 140         }
 141
 142         /**
 143          * @param File $image
 144          * @param string $dstPath
 145          * @param string $dstUrl
 146          * @param array $params
 147          * @param int $flags
 148          * @return MediaTransformError|ThumbnailImage|TransformParameterError
 149          */
 150         function doTransform( $image, $dstPath, $dstUrl, $params, $flags = 0 ) {
 151                 global $wgDjvuRenderer, $wgDjvuPostProcessor;
 152
 153                 if ( !$this->normaliseParams( $image, $params ) ) {
 154                         return new TransformParameterError( $params );
 155                 }
 156                 $width = $params['width'];
 157                 $height = $params['height'];
 158                 $page = $params['page'];
 159
 160                 if ( $flags & self::TRANSFORM_LATER ) {
 161                         $params = [
 162                                 'width' => $width,
 163                                 'height' => $height,
 164                                 'page' => $page
 165                         ];
 166
 167                         return new ThumbnailImage( $image, $dstUrl, $dstPath, $params );
 168                 }
 169
 170                 if ( !wfMkdirParents( dirname( $dstPath ), null, __METHOD__ ) ) {
 171                         return new MediaTransformError(
 172                                 'thumbnail_error',
 173                                 $width,
 174                                 $height,
 175                                 wfMessage( 'thumbnail_dest_directory' )
 176                         );
 177                 }
 178
 179                 // Get local copy source for shell scripts
 180                 // Thumbnail extraction is very inefficient for large files.
 181                 // Provide a way to pool count limit the number of downloaders.
 182                 if ( $image->getSize() >= 1e7 ) { // 10MB
 183                         $work = new PoolCounterWorkViaCallback( 'GetLocalFileCopy', sha1( $image->getName() ),
 184                                 [
 185                                         'doWork' => function () use ( $image ) {
 186                                                 return $image->getLocalRefPath();
 187                                         }
 188                                 ]
 189                         );
 190                         $srcPath = $work->execute();
 191                 } else {
 192                         $srcPath = $image->getLocalRefPath();
 193                 }
 194
 195                 if ( $srcPath === false ) { // Failed to get local copy
 196                         wfDebugLog( 'thumbnail',
 197                                 sprintf( 'Thumbnail failed on %s: could not get local copy of "%s"',
 198                                         wfHostname(), $image->getName() ) );
 199
 200                         return new MediaTransformError( 'thumbnail_error',
 201                                 $params['width'], $params['height'],
 202                                 wfMessage( 'filemissing' )
 203                         );
 204                 }
 205
 206                 # Use a subshell (brackets) to aggregate stderr from both pipeline commands
 207                 # before redirecting it to the overall stdout. This works in both Linux and Windows XP.
 208                 $cmd = '(' . Shell::escape(
 209                         $wgDjvuRenderer,
 210                         "-format=ppm",
 211                         "-page={$page}",
 212                         "-size={$params['physicalWidth']}x{$params['physicalHeight']}",
 213                         $srcPath );
 214                 if ( $wgDjvuPostProcessor ) {
 215                         $cmd .= " | {$wgDjvuPostProcessor}";
 216                 }
 217                 $cmd .= ' > ' . Shell::escape( $dstPath ) . ') 2>&1';
 218                 wfDebug( __METHOD__ . ": $cmd\n" );
 219                 $retval = '';
 220                 $err = wfShellExec( $cmd, $retval );
 221
 222                 $removed = $this->removeBadFile( $dstPath, $retval );
 223                 if ( $retval != 0 || $removed ) {
 224                         $this->logErrorForExternalProcess( $retval, $err, $cmd );
 225                         return new MediaTransformError( 'thumbnail_error', $width, $height, $err );
 226                 } else {
 227                         $params = [
 228                                 'width' => $width,
 229                                 'height' => $height,
 230                                 'page' => $page
 231                         ];
 232
 233                         return new ThumbnailImage( $image, $dstUrl, $dstPath, $params );
 234                 }
 235         }
 236
 237         /**
 238          * Cache an instance of DjVuImage in an Image object, return that instance
 239          *
 240          * @param File|FSFile $image
 241          * @param string $path
 242          * @return DjVuImage
 243          * @suppress PhanUndeclaredProperty Custom property
 244          */
 245         function getDjVuImage( $image, $path ) {
 246                 if ( !$image ) {
 247                         $deja = new DjVuImage( $path );
 248                 } elseif ( !isset( $image->dejaImage ) ) {
 249                         $deja = $image->dejaImage = new DjVuImage( $path );
 250                 } else {
 251                         $deja = $image->dejaImage;
 252                 }
 253
 254                 return $deja;
 255         }
 256
 257         /**
 258          * Get metadata, unserializing it if necessary.
 259          *
 260          * @param File $file The DjVu file in question
 261          * @return string XML metadata as a string.
 262          * @throws MWException
 263          */
 264         private function getUnserializedMetadata( File $file ) {
 265                 $metadata = $file->getMetadata();
 266                 if ( substr( $metadata, 0, 3 ) === '<?xml' ) {
 267                         // Old style. Not serialized but instead just a raw string of XML.
 268                         return $metadata;
 269                 }
 270
 271                 Wikimedia\suppressWarnings();
 272                 $unser = unserialize( $metadata );
 273                 Wikimedia\restoreWarnings();
 274                 if ( is_array( $unser ) ) {
 275                         if ( isset( $unser['error'] ) ) {
 276                                 return false;
 277                         } elseif ( isset( $unser['xml'] ) ) {
 278                                 return $unser['xml'];
 279                         } else {
 280                                 // Should never ever reach here.
 281                                 throw new MWException( "Error unserializing DjVu metadata." );
 282                         }
 283                 }
 284
 285                 // unserialize failed. Guess it wasn't really serialized after all,
 286                 return $metadata;
 287         }
 288
 289         /**
 290          * Cache a document tree for the DjVu XML metadata
 291          * @param File $image
 292          * @param bool $gettext DOCUMENT (Default: false)
 293          * @return bool|SimpleXMLElement
 294          * @suppress PhanUndeclaredProperty Custom property
 295          */
 296         public function getMetaTree( $image, $gettext = false ) {
 297                 if ( $gettext && isset( $image->djvuTextTree ) ) {
 298                         return $image->djvuTextTree;
 299                 }
 300                 if ( !$gettext && isset( $image->dejaMetaTree ) ) {
 301                         return $image->dejaMetaTree;
 302                 }
 303
 304                 $metadata = $this->getUnserializedMetadata( $image );
 305                 if ( !$this->isMetadataValid( $image, $metadata ) ) {
 306                         wfDebug( "DjVu XML metadata is invalid or missing, should have been fixed in upgradeRow\n" );
 307
 308                         return false;
 309                 }
 310
 311                 $trees = $this->extractTreesFromMetadata( $metadata );
 312                 $image->djvuTextTree = $trees['TextTree'];
 313                 $image->dejaMetaTree = $trees['MetaTree'];
 314
 315                 if ( $gettext ) {
 316                         return $image->djvuTextTree;
 317                 } else {
 318                         return $image->dejaMetaTree;
 319                 }
 320         }
 321
 322         /**
 323          * Extracts metadata and text trees from metadata XML in string form
 324          * @param string $metadata XML metadata as a string
 325          * @return array
 326          */
 327         protected function extractTreesFromMetadata( $metadata ) {
 328                 Wikimedia\suppressWarnings();
 329                 try {
 330                         // Set to false rather than null to avoid further attempts
 331                         $metaTree = false;
 332                         $textTree = false;
 333                         $tree = new SimpleXMLElement( $metadata, LIBXML_PARSEHUGE );
 334                         if ( $tree->getName() == 'mw-djvu' ) {
 335                                 /** @var SimpleXMLElement $b */
 336                                 foreach ( $tree->children() as $b ) {
 337                                         if ( $b->getName() == 'DjVuTxt' ) {
 338                                                 // @todo File::djvuTextTree and File::dejaMetaTree are declared
 339                                                 // dynamically. Add a public File::$data to facilitate this?
 340                                                 $textTree = $b;
 341                                         } elseif ( $b->getName() == 'DjVuXML' ) {
 342                                                 $metaTree = $b;
 343                                         }
 344                                 }
 345                         } else {
 346                                 $metaTree = $tree;
 347                         }
 348                 } catch ( Exception $e ) {
 349                         wfDebug( "Bogus multipage XML metadata\n" );
 350                 }
 351                 Wikimedia\restoreWarnings();
 352
 353                 return [ 'MetaTree' => $metaTree, 'TextTree' => $textTree ];
 354         }
 355
 356         function getImageSize( $image, $path ) {
 357                 return $this->getDjVuImage( $image, $path )->getImageSize();
 358         }
 359
 360         public function getThumbType( $ext, $mime, $params = null ) {
 361                 global $wgDjvuOutputExtension;
 362                 static $mime;
 363                 if ( !isset( $mime ) ) {
 364                         $magic = MediaWiki\MediaWikiServices::getInstance()->getMimeAnalyzer();
 365                         $mime = $magic->guessTypesForExtension( $wgDjvuOutputExtension );
 366                 }
 367
 368                 return [ $wgDjvuOutputExtension, $mime ];
 369         }
 370
 371         public function getMetadata( $image, $path ) {
 372                 wfDebug( "Getting DjVu metadata for $path\n" );
 373
 374                 $xml = $this->getDjVuImage( $image, $path )->retrieveMetaData();
 375                 if ( $xml === false ) {
 376                         // Special value so that we don't repetitively try and decode a broken file.
 377                         return serialize( [ 'error' => 'Error extracting metadata' ] );
 378                 } else {
 379                         return serialize( [ 'xml' => $xml ] );
 380                 }
 381         }
 382
 383         function getMetadataType( $image ) {
 384                 return 'djvuxml';
 385         }
 386
 387         public function isMetadataValid( $image, $metadata ) {
 388                 return !empty( $metadata ) && $metadata != serialize( [] );
 389         }
 390
 391         public function pageCount( File $image ) {
 392                 $info = $this->getDimensionInfo( $image );
 393
 394                 return $info ? $info['pageCount'] : false;
 395         }
 396
 397         public function getPageDimensions( File $image, $page ) {
 398                 $index = $page - 1; // MW starts pages at 1
 399
 400                 $info = $this->getDimensionInfo( $image );
 401                 if ( $info && isset( $info['dimensionsByPage'][$index] ) ) {
 402                         return $info['dimensionsByPage'][$index];
 403                 }
 404
 405                 return false;
 406         }
 407
 408         protected function getDimensionInfo( File $file ) {
 409                 $cache = MediaWikiServices::getInstance()->getMainWANObjectCache();
 410                 return $cache->getWithSetCallback(
 411                         $cache->makeKey( 'file-djvu', 'dimensions', $file->getSha1() ),
 412                         $cache::TTL_INDEFINITE,
 413                         function () use ( $file ) {
 414                                 $tree = $this->getMetaTree( $file );
 415                                 return $this->getDimensionInfoFromMetaTree( $tree );
 416                         },
 417                         [ 'pcTTL' => $cache::TTL_INDEFINITE ]
 418                 );
 419         }
 420
 421         /**
 422          * Given an XML metadata tree, returns dimension information about the document
 423          * @param bool|SimpleXMLElement $metatree The file's XML metadata tree
 424          * @return bool|array
 425          */
 426         protected function getDimensionInfoFromMetaTree( $metatree ) {
 427                 if ( !$metatree ) {
 428                         return false;
 429                 }
 430
 431                 $dimsByPage = [];
 432                 $count = count( $metatree->xpath( '//OBJECT' ) );
 433                 for ( $i = 0; $i < $count; $i++ ) {
 434                         $o = $metatree->BODY[0]->OBJECT[$i];
 435                         if ( $o ) {
 436                                 $dimsByPage[$i] = [
 437                                         'width' => (int)$o['width'],
 438                                         'height' => (int)$o['height'],
 439                                 ];
 440                         } else {
 441                                 $dimsByPage[$i] = false;
 442                         }
 443                 }
 444
 445                 return [ 'pageCount' => $count, 'dimensionsByPage' => $dimsByPage ];
 446         }
 447
 448         /**
 449          * @param File $image
 450          * @param int $page Page number to get information for
 451          * @return bool|string Page text or false when no text found.
 452          */
 453         function getPageText( File $image, $page ) {
 454                 $tree = $this->getMetaTree( $image, true );
 455                 if ( !$tree ) {
 456                         return false;
 457                 }
 458
 459                 $o = $tree->BODY[0]->PAGE[$page - 1];
 460                 if ( $o ) {
 461                         $txt = $o['value'];
 462
 463                         return $txt;
 464                 } else {
 465                         return false;
 466                 }
 467         }
 468 }