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 ( in_array( $name, array( 'width', 'height', 'page' ) ) ) {
  77                         if ( $value <= 0 ) {
  78                                 return false;
  79                         } else {
  80                                 return true;
  81                         }
  82                 } else {
  83                         return false;
  84                 }
  85         }
  86
  87         /**
  88          * @param array $params
  89          * @return bool|string
  90          */
  91         function makeParamString( $params ) {
  92                 $page = isset( $params['page'] ) ? $params['page'] : 1;
  93                 if ( !isset( $params['width'] ) ) {
  94                         return false;
  95                 }
  96
  97                 return "page{$page}-{$params['width']}px";
  98         }
  99
 100         /**
 101          * @param string $str
 102          * @return array|bool
 103          */
 104         function parseParamString( $str ) {
 105                 $m = false;
 106                 if ( preg_match( '/^page(\d+)-(\d+)px$/', $str, $m ) ) {
 107                         return array( 'width' => $m[2], 'page' => $m[1] );
 108                 } else {
 109                         return false;
 110                 }
 111         }
 112
 113         /**
 114          * @param array $params
 115          * @return array
 116          */
 117         function getScriptParams( $params ) {
 118                 return array(
 119                         'width' => $params['width'],
 120                         'page' => $params['page'],
 121                 );
 122         }
 123
 124         /**
 125          * @param File $image
 126          * @param string $dstPath
 127          * @param string $dstUrl
 128          * @param array $params
 129          * @param int $flags
 130          * @return MediaTransformError|ThumbnailImage|TransformParameterError
 131          */
 132         function doTransform( $image, $dstPath, $dstUrl, $params, $flags = 0 ) {
 133                 global $wgDjvuRenderer, $wgDjvuPostProcessor;
 134
 135                 // Fetch XML and check it, to give a more informative error message than the one which
 136                 // normaliseParams will inevitably give.
 137                 $xml = $image->getMetadata();
 138                 if ( !$xml ) {
 139                         $width = isset( $params['width'] ) ? $params['width'] : 0;
 140                         $height = isset( $params['height'] ) ? $params['height'] : 0;
 141
 142                         return new MediaTransformError( 'thumbnail_error', $width, $height,
 143                                 wfMessage( 'djvu_no_xml' )->text() );
 144                 }
 145
 146                 if ( !$this->normaliseParams( $image, $params ) ) {
 147                         return new TransformParameterError( $params );
 148                 }
 149                 $width = $params['width'];
 150                 $height = $params['height'];
 151                 $page = $params['page'];
 152                 if ( $page > $this->pageCount( $image ) ) {
 153                         return new MediaTransformError(
 154                                 'thumbnail_error',
 155                                 $width,
 156                                 $height,
 157                                 wfMessage( 'djvu_page_error' )->text()
 158                         );
 159                 }
 160
 161                 if ( $flags & self::TRANSFORM_LATER ) {
 162                         $params = array(
 163                                 'width' => $width,
 164                                 'height' => $height,
 165                                 'page' => $page
 166                         );
 167
 168                         return new ThumbnailImage( $image, $dstUrl, $dstPath, $params );
 169                 }
 170
 171                 if ( !wfMkdirParents( dirname( $dstPath ), null, __METHOD__ ) ) {
 172                         return new MediaTransformError(
 173                                 'thumbnail_error',
 174                                 $width,
 175                                 $height,
 176                                 wfMessage( 'thumbnail_dest_directory' )->text()
 177                         );
 178                 }
 179
 180                 $srcPath = $image->getLocalRefPath();
 181                 # Use a subshell (brackets) to aggregate stderr from both pipeline commands
 182                 # before redirecting it to the overall stdout. This works in both Linux and Windows XP.
 183                 $cmd = '(' . wfEscapeShellArg(
 184                         $wgDjvuRenderer,
 185                         "-format=ppm",
 186                         "-page={$page}",
 187                         "-size={$params['physicalWidth']}x{$params['physicalHeight']}",
 188                         $srcPath );
 189                 if ( $wgDjvuPostProcessor ) {
 190                         $cmd .= " | {$wgDjvuPostProcessor}";
 191                 }
 192                 $cmd .= ' > ' . wfEscapeShellArg( $dstPath ) . ') 2>&1';
 193                 wfProfileIn( 'ddjvu' );
 194                 wfDebug( __METHOD__ . ": $cmd\n" );
 195                 $retval = '';
 196                 $err = wfShellExec( $cmd, $retval );
 197                 wfProfileOut( 'ddjvu' );
 198
 199                 $removed = $this->removeBadFile( $dstPath, $retval );
 200                 if ( $retval != 0 || $removed ) {
 201                         $this->logErrorForExternalProcess( $retval, $err, $cmd );
 202                         return new MediaTransformError( 'thumbnail_error', $width, $height, $err );
 203                 } else {
 204                         $params = array(
 205                                 'width' => $width,
 206                                 'height' => $height,
 207                                 'page' => $page
 208                         );
 209
 210                         return new ThumbnailImage( $image, $dstUrl, $dstPath, $params );
 211                 }
 212         }
 213
 214         /**
 215          * Cache an instance of DjVuImage in an Image object, return that instance
 216          *
 217          * @param File $image
 218          * @param string $path
 219          * @return DjVuImage
 220          */
 221         function getDjVuImage( $image, $path ) {
 222                 if ( !$image ) {
 223                         $deja = new DjVuImage( $path );
 224                 } elseif ( !isset( $image->dejaImage ) ) {
 225                         $deja = $image->dejaImage = new DjVuImage( $path );
 226                 } else {
 227                         $deja = $image->dejaImage;
 228                 }
 229
 230                 return $deja;
 231         }
 232
 233         /**
 234          * Get metadata, unserializing it if neccessary.
 235          *
 236          * @param File $file The DjVu file in question
 237          * @return String XML metadata as a string.
 238          */
 239         private function getUnserializedMetadata( File $file ) {
 240                 $metadata = $file->getMetadata();
 241                 if ( substr( $metadata, 0, 3 ) === '<?xml' ) {
 242                         // Old style. Not serialized but instead just a raw string of XML.
 243                         return $metadata;
 244                 }
 245
 246                 wfSuppressWarnings();
 247                 $unser = unserialize( $metadata );
 248                 wfRestoreWarnings();
 249                 if ( is_array( $unser ) ) {
 250                         return $unser['xml'];
 251                 }
 252
 253                 // unserialize failed. Guess it wasn't really serialized after all,
 254                 return $metadata;
 255         }
 256
 257         /**
 258          * Cache a document tree for the DjVu XML metadata
 259          * @param File $image
 260          * @param bool $gettext DOCUMENT (Default: false)
 261          * @return bool|SimpleXMLElement
 262          */
 263         function getMetaTree( $image, $gettext = false ) {
 264                 if ( $gettext && isset( $image->djvuTextTree ) ) {
 265                         return $image->djvuTextTree;
 266                 }
 267                 if ( !$gettext && isset( $image->dejaMetaTree ) ) {
 268                         return $image->dejaMetaTree;
 269                 }
 270
 271                 $metadata = $this->getUnserializedMetadata( $image );
 272                 if ( !$this->isMetadataValid( $image, $metadata ) ) {
 273                         wfDebug( "DjVu XML metadata is invalid or missing, should have been fixed in upgradeRow\n" );
 274
 275                         return false;
 276                 }
 277                 wfProfileIn( __METHOD__ );
 278
 279                 wfSuppressWarnings();
 280                 try {
 281                         // Set to false rather than null to avoid further attempts
 282                         $image->dejaMetaTree = false;
 283                         $image->djvuTextTree = false;
 284                         $tree = new SimpleXMLElement( $metadata );
 285                         if ( $tree->getName() == 'mw-djvu' ) {
 286                                 /** @var SimpleXMLElement $b */
 287                                 foreach ( $tree->children() as $b ) {
 288                                         if ( $b->getName() == 'DjVuTxt' ) {
 289                                                 // @todo File::djvuTextTree and File::dejaMetaTree are declared
 290                                                 // dynamically. Add a public File::$data to facilitate this?
 291                                                 $image->djvuTextTree = $b;
 292                                         } elseif ( $b->getName() == 'DjVuXML' ) {
 293                                                 $image->dejaMetaTree = $b;
 294                                         }
 295                                 }
 296                         } else {
 297                                 $image->dejaMetaTree = $tree;
 298                         }
 299                 } catch ( Exception $e ) {
 300                         wfDebug( "Bogus multipage XML metadata on '{$image->getName()}'\n" );
 301                 }
 302                 wfRestoreWarnings();
 303                 wfProfileOut( __METHOD__ );
 304                 if ( $gettext ) {
 305                         return $image->djvuTextTree;
 306                 } else {
 307                         return $image->dejaMetaTree;
 308                 }
 309         }
 310
 311         /**
 312          * @param File $image
 313          * @param string $path
 314          * @return bool|array False on failure
 315          */
 316         function getImageSize( $image, $path ) {
 317                 return $this->getDjVuImage( $image, $path )->getImageSize();
 318         }
 319
 320         function getThumbType( $ext, $mime, $params = null ) {
 321                 global $wgDjvuOutputExtension;
 322                 static $mime;
 323                 if ( !isset( $mime ) ) {
 324                         $magic = MimeMagic::singleton();
 325                         $mime = $magic->guessTypesForExtension( $wgDjvuOutputExtension );
 326                 }
 327
 328                 return array( $wgDjvuOutputExtension, $mime );
 329         }
 330
 331         function getMetadata( $image, $path ) {
 332                 wfDebug( "Getting DjVu metadata for $path\n" );
 333
 334                 $xml = $this->getDjVuImage( $image, $path )->retrieveMetaData();
 335                 if ( $xml === false ) {
 336                         return false;
 337                 } else {
 338                         return serialize( array( 'xml' => $xml ) );
 339                 }
 340         }
 341
 342         function getMetadataType( $image ) {
 343                 return 'djvuxml';
 344         }
 345
 346         function isMetadataValid( $image, $metadata ) {
 347                 return !empty( $metadata ) && $metadata != serialize( array() );
 348         }
 349
 350         function pageCount( $image ) {
 351                 $tree = $this->getMetaTree( $image );
 352                 if ( !$tree ) {
 353                         return false;
 354                 }
 355
 356                 return count( $tree->xpath( '//OBJECT' ) );
 357         }
 358
 359         function getPageDimensions( $image, $page ) {
 360                 $tree = $this->getMetaTree( $image );
 361                 if ( !$tree ) {
 362                         return false;
 363                 }
 364
 365                 $o = $tree->BODY[0]->OBJECT[$page - 1];
 366                 if ( $o ) {
 367                         return array(
 368                                 'width' => intval( $o['width'] ),
 369                                 'height' => intval( $o['height'] )
 370                         );
 371                 } else {
 372                         return false;
 373                 }
 374         }
 375
 376         /**
 377          * @param File $image
 378          * @param int $page Page number to get information for
 379          * @return bool|string Page text or false when no text found.
 380          */
 381         function getPageText( $image, $page ) {
 382                 $tree = $this->getMetaTree( $image, true );
 383                 if ( !$tree ) {
 384                         return false;
 385                 }
 386
 387                 $o = $tree->BODY[0]->PAGE[$page - 1];
 388                 if ( $o ) {
 389                         $txt = $o['value'];
 390
 391                         return $txt;
 392                 } else {
 393                         return false;
 394                 }
 395         }
 396 }