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