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