includes/media/DjVuImage.php

   1 <?php
   2 /**
   3  * DjVu image handler.
   4  *
   5  * Copyright © 2006 Brion Vibber <brion@pobox.com>
   6  * http://www.mediawiki.org/
   7  *
   8  * This program is free software; you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation; either version 2 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * This program is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License along
  19  * with this program; if not, write to the Free Software Foundation, Inc.,
  20  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  21  * http://www.gnu.org/copyleft/gpl.html
  22  *
  23  * @file
  24  * @ingroup Media
  25  */
  26
  27 /**
  28  * Support for detecting/validating DjVu image files and getting
  29  * some basic file metadata (resolution etc)
  30  *
  31  * File format docs are available in source package for DjVuLibre:
  32  * http://djvulibre.djvuzone.org/
  33  *
  34  * @ingroup Media
  35  */
  36 class DjVuImage {
  37         /**
  38          * Constructor
  39          *
  40          * @param string $filename The DjVu file name.
  41          */
  42         function __construct( $filename ) {
  43                 $this->mFilename = $filename;
  44         }
  45
  46         /**
  47          * @const DJVUTXT_MEMORY_LIMIT Memory limit for the DjVu description software
  48          */
  49         const DJVUTXT_MEMORY_LIMIT = 300000;
  50
  51         /**
  52          * Check if the given file is indeed a valid DjVu image file
  53          * @return bool
  54          */
  55         public function isValid() {
  56                 $info = $this->getInfo();
  57                 return $info !== false;
  58         }
  59
  60         /**
  61          * Return data in the style of getimagesize()
  62          * @return array or false on failure
  63          */
  64         public function getImageSize() {
  65                 $data = $this->getInfo();
  66
  67                 if( $data !== false ) {
  68                         $width = $data['width'];
  69                         $height = $data['height'];
  70
  71                         return array( $width, $height, 'DjVu',
  72                                 "width=\"$width\" height=\"$height\"" );
  73                 }
  74                 return false;
  75         }
  76
  77         // ---------
  78
  79         /**
  80          * For debugging; dump the IFF chunk structure
  81          */
  82         function dump() {
  83                 $file = fopen( $this->mFilename, 'rb' );
  84                 $header = fread( $file, 12 );
  85                 // @todo FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
  86                 extract( unpack( 'a4magic/a4chunk/NchunkLength', $header ) );
  87                 echo "$chunk $chunkLength\n";
  88                 $this->dumpForm( $file, $chunkLength, 1 );
  89                 fclose( $file );
  90         }
  91
  92         private function dumpForm( $file, $length, $indent ) {
  93                 $start = ftell( $file );
  94                 $secondary = fread( $file, 4 );
  95                 echo str_repeat( ' ', $indent * 4 ) . "($secondary)\n";
  96                 while( ftell( $file ) - $start < $length ) {
  97                         $chunkHeader = fread( $file, 8 );
  98                         if( $chunkHeader == '' ) {
  99                                 break;
 100                         }
 101                         // @todo FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 102                         extract( unpack( 'a4chunk/NchunkLength', $chunkHeader ) );
 103                         echo str_repeat( ' ', $indent * 4 ) . "$chunk $chunkLength\n";
 104
 105                         if( $chunk == 'FORM' ) {
 106                                 $this->dumpForm( $file, $chunkLength, $indent + 1 );
 107                         } else {
 108                                 fseek( $file, $chunkLength, SEEK_CUR );
 109                                 if( $chunkLength & 1 == 1 ) {
 110                                         // Padding byte between chunks
 111                                         fseek( $file, 1, SEEK_CUR );
 112                                 }
 113                         }
 114                 }
 115         }
 116
 117         function getInfo() {
 118                 wfSuppressWarnings();
 119                 $file = fopen( $this->mFilename, 'rb' );
 120                 wfRestoreWarnings();
 121                 if( $file === false ) {
 122                         wfDebug( __METHOD__ . ": missing or failed file read\n" );
 123                         return false;
 124                 }
 125
 126                 $header = fread( $file, 16 );
 127                 $info = false;
 128
 129                 if( strlen( $header ) < 16 ) {
 130                         wfDebug( __METHOD__ . ": too short file header\n" );
 131                 } else {
 132                         // @todo FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 133                         extract( unpack( 'a4magic/a4form/NformLength/a4subtype', $header ) );
 134
 135                         if( $magic != 'AT&T' ) {
 136                                 wfDebug( __METHOD__ . ": not a DjVu file\n" );
 137                         } elseif( $subtype == 'DJVU' ) {
 138                                 // Single-page document
 139                                 $info = $this->getPageInfo( $file, $formLength );
 140                         } elseif( $subtype == 'DJVM' ) {
 141                                 // Multi-page document
 142                                 $info = $this->getMultiPageInfo( $file, $formLength );
 143                         } else  {
 144                                 wfDebug( __METHOD__ . ": unrecognized DJVU file type '$formType'\n" );
 145                         }
 146                 }
 147                 fclose( $file );
 148                 return $info;
 149         }
 150
 151         private function readChunk( $file ) {
 152                 $header = fread( $file, 8 );
 153                 if( strlen( $header ) < 8 ) {
 154                         return array( false, 0 );
 155                 } else {
 156                         // @todo FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 157                         extract( unpack( 'a4chunk/Nlength', $header ) );
 158                         return array( $chunk, $length );
 159                 }
 160         }
 161
 162         private function skipChunk( $file, $chunkLength ) {
 163                 fseek( $file, $chunkLength, SEEK_CUR );
 164
 165                 if( $chunkLength & 0x01 == 1 && !feof( $file ) ) {
 166                         // padding byte
 167                         fseek( $file, 1, SEEK_CUR );
 168                 }
 169         }
 170
 171         private function getMultiPageInfo( $file, $formLength ) {
 172                 // For now, we'll just look for the first page in the file
 173                 // and report its information, hoping others are the same size.
 174                 $start = ftell( $file );
 175                 do {
 176                         list( $chunk, $length ) = $this->readChunk( $file );
 177                         if( !$chunk ) {
 178                                 break;
 179                         }
 180
 181                         if( $chunk == 'FORM' ) {
 182                                 $subtype = fread( $file, 4 );
 183                                 if( $subtype == 'DJVU' ) {
 184                                         wfDebug( __METHOD__ . ": found first subpage\n" );
 185                                         return $this->getPageInfo( $file, $length );
 186                                 }
 187                                 $this->skipChunk( $file, $length - 4 );
 188                         } else {
 189                                 wfDebug( __METHOD__ . ": skipping '$chunk' chunk\n" );
 190                                 $this->skipChunk( $file, $length );
 191                         }
 192                 } while( $length != 0 && !feof( $file ) && ftell( $file ) - $start < $formLength );
 193
 194                 wfDebug( __METHOD__ . ": multi-page DJVU file contained no pages\n" );
 195                 return false;
 196         }
 197
 198         private function getPageInfo( $file, $formLength ) {
 199                 list( $chunk, $length ) = $this->readChunk( $file );
 200                 if( $chunk != 'INFO' ) {
 201                         wfDebug( __METHOD__ . ": expected INFO chunk, got '$chunk'\n" );
 202                         return false;
 203                 }
 204
 205                 if( $length < 9 ) {
 206                         wfDebug( __METHOD__ . ": INFO should be 9 or 10 bytes, found $length\n" );
 207                         return false;
 208                 }
 209                 $data = fread( $file, $length );
 210                 if( strlen( $data ) < $length ) {
 211                         wfDebug( __METHOD__ . ": INFO chunk cut off\n" );
 212                         return false;
 213                 }
 214
 215                 // @todo FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 216                 extract( unpack(
 217                         'nwidth/' .
 218                         'nheight/' .
 219                         'Cminor/' .
 220                         'Cmajor/' .
 221                         'vresolution/' .
 222                         'Cgamma', $data ) );
 223                 # Newer files have rotation info in byte 10, but we don't use it yet.
 224
 225                 return array(
 226                         'width' => $width,
 227                         'height' => $height,
 228                         'version' => "$major.$minor",
 229                         'resolution' => $resolution,
 230                         'gamma' => $gamma / 10.0 );
 231         }
 232
 233         /**
 234          * Return an XML string describing the DjVu image
 235          * @return string
 236          */
 237         function retrieveMetaData() {
 238                 global $wgDjvuToXML, $wgDjvuDump, $wgDjvuTxt;
 239                 wfProfileIn( __METHOD__ );
 240
 241                 if ( isset( $wgDjvuDump ) ) {
 242                         # djvudump is faster as of version 3.5
 243                         # http://sourceforge.net/tracker/index.php?func=detail&aid=1704049&group_id=32953&atid=406583
 244                         wfProfileIn( 'djvudump' );
 245                         $cmd = wfEscapeShellArg( $wgDjvuDump ) . ' ' . wfEscapeShellArg( $this->mFilename );
 246                         $dump = wfShellExec( $cmd );
 247                         $xml = $this->convertDumpToXML( $dump );
 248                         wfProfileOut( 'djvudump' );
 249                 } elseif ( isset( $wgDjvuToXML ) ) {
 250                         wfProfileIn( 'djvutoxml' );
 251                         $cmd = wfEscapeShellArg( $wgDjvuToXML ) . ' --without-anno --without-text ' .
 252                                 wfEscapeShellArg( $this->mFilename );
 253                         $xml = wfShellExec( $cmd );
 254                         wfProfileOut( 'djvutoxml' );
 255                 } else {
 256                         $xml = null;
 257                 }
 258                 # Text layer
 259                 if ( isset( $wgDjvuTxt ) ) {
 260                         wfProfileIn( 'djvutxt' );
 261                         $cmd = wfEscapeShellArg( $wgDjvuTxt ) . ' --detail=page ' . wfEscapeShellArg( $this->mFilename );
 262                         wfDebug( __METHOD__ . ": $cmd\n" );
 263                         $retval = '';
 264                         $txt = wfShellExec( $cmd, $retval, array(), array( 'memory' => self::DJVUTXT_MEMORY_LIMIT ) );
 265                         wfProfileOut( 'djvutxt' );
 266                         if( $retval == 0 ) {
 267                                 # Strip some control characters
 268                                 $txt = preg_replace( "/[\013\035\037]/", "", $txt );
 269                                 $reg = <<<EOR
 270                                         /\(page\s[\d-]*\s[\d-]*\s[\d-]*\s[\d-]*\s*"
 271                                         ((?>    # Text to match is composed of atoms of either:
 272                                           \\\\. # - any escaped character
 273                                           |     # - any character different from " and \
 274                                           [^"\\\\]+
 275                                         )*?)
 276                                         "\s*\)
 277                                         | # Or page can be empty ; in this case, djvutxt dumps ()
 278                                         \(\s*()\)/sx
 279 EOR;
 280                                 $txt = preg_replace_callback( $reg, array( $this, 'pageTextCallback' ), $txt );
 281                                 $txt = "<DjVuTxt>\n<HEAD></HEAD>\n<BODY>\n" . $txt . "</BODY>\n</DjVuTxt>\n";
 282                                 $xml = preg_replace( "/<DjVuXML>/", "<mw-djvu><DjVuXML>", $xml, 1 );
 283                                 $xml = $xml . $txt . '</mw-djvu>';
 284                         }
 285                 }
 286                 wfProfileOut( __METHOD__ );
 287                 return $xml;
 288         }
 289
 290         function pageTextCallback( $matches ) {
 291                 # Get rid of invalid UTF-8, strip control characters
 292                 return '<PAGE value="' . htmlspecialchars( UtfNormal::cleanUp( $matches[1] ) ) . '" />';
 293         }
 294
 295         /**
 296          * Hack to temporarily work around djvutoxml bug
 297          * @return bool|string
 298          */
 299         function convertDumpToXML( $dump ) {
 300                 if ( strval( $dump ) == '' ) {
 301                         return false;
 302                 }
 303
 304                 $xml = <<<EOT
 305 <?xml version="1.0" ?>
 306 <!DOCTYPE DjVuXML PUBLIC "-//W3C//DTD DjVuXML 1.1//EN" "pubtext/DjVuXML-s.dtd">
 307 <DjVuXML>
 308 <HEAD></HEAD>
 309 <BODY>
 310 EOT;
 311
 312                 $dump = str_replace( "\r", '', $dump );
 313                 $line = strtok( $dump, "\n" );
 314                 $m = false;
 315                 $good = false;
 316                 if ( preg_match( '/^( *)FORM:DJVU/', $line, $m ) ) {
 317                         # Single-page
 318                         if ( $this->parseFormDjvu( $line, $xml ) ) {
 319                                 $good = true;
 320                         } else {
 321                                 return false;
 322                         }
 323                 } elseif ( preg_match( '/^( *)FORM:DJVM/', $line, $m ) ) {
 324                         # Multi-page
 325                         $parentLevel = strlen( $m[1] );
 326                         # Find DIRM
 327                         $line = strtok( "\n" );
 328                         while ( $line !== false ) {
 329                                 $childLevel = strspn( $line, ' ' );
 330                                 if ( $childLevel <= $parentLevel ) {
 331                                         # End of chunk
 332                                         break;
 333                                 }
 334
 335                                 if ( preg_match( '/^ *DIRM.*indirect/', $line ) ) {
 336                                         wfDebug( "Indirect multi-page DjVu document, bad for server!\n" );
 337                                         return false;
 338                                 }
 339                                 if ( preg_match( '/^ *FORM:DJVU/', $line ) ) {
 340                                         # Found page
 341                                         if ( $this->parseFormDjvu( $line, $xml ) ) {
 342                                                 $good = true;
 343                                         } else {
 344                                                 return false;
 345                                         }
 346                                 }
 347                                 $line = strtok( "\n" );
 348                         }
 349                 }
 350                 if ( !$good ) {
 351                         return false;
 352                 }
 353
 354                 $xml .= "</BODY>\n</DjVuXML>\n";
 355                 return $xml;
 356         }
 357
 358         function parseFormDjvu( $line, &$xml ) {
 359                 $parentLevel = strspn( $line, ' ' );
 360                 $line = strtok( "\n" );
 361
 362                 # Find INFO
 363                 while ( $line !== false ) {
 364                         $childLevel = strspn( $line, ' ' );
 365                         if ( $childLevel <= $parentLevel ) {
 366                                 # End of chunk
 367                                 break;
 368                         }
 369
 370                         if ( preg_match( '/^ *INFO *\[\d*\] *DjVu *(\d+)x(\d+), *\w*, *(\d+) *dpi, *gamma=([0-9.-]+)/', $line, $m ) ) {
 371                                 $xml .= Xml::tags( 'OBJECT',
 372                                         array(
 373                                                 #'data' => '',
 374                                                 #'type' => 'image/x.djvu',
 375                                                 'height' => $m[2],
 376                                                 'width' => $m[1],
 377                                                 #'usemap' => '',
 378                                         ),
 379                                         "\n" .
 380                                         Xml::element( 'PARAM', array( 'name' => 'DPI', 'value' => $m[3] ) ) . "\n" .
 381                                         Xml::element( 'PARAM', array( 'name' => 'GAMMA', 'value' => $m[4] ) ) . "\n"
 382                                 ) . "\n";
 383                                 return true;
 384                         }
 385                         $line = strtok( "\n" );
 386                 }
 387                 # Not found
 388                 return false;
 389         }
 390 }