includes/DjVuImage.php

   1 <?php
   2
   3 /**
   4  *
   5  * Copyright (C) 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  */
  24
  25 /**
  26  * Support for detecting/validating DjVu image files and getting
  27  * some basic file metadata (resolution etc)
  28  *
  29  * File format docs are available in source package for DjVuLibre:
  30  * http://djvulibre.djvuzone.org/
  31  */
  32 class DjVuImage {
  33         function __construct( $filename ) {
  34                 $this->mFilename = $filename;
  35         }
  36
  37         /**
  38          * Check if the given file is indeed a valid DjVu image file
  39          * @return bool
  40          */
  41         public function isValid() {
  42                 $info = $this->getInfo();
  43                 return $info !== false;
  44         }
  45
  46
  47         /**
  48          * Return data in the style of getimagesize()
  49          * @return array or false on failure
  50          */
  51         public function getImageSize() {
  52                 $data = $this->getInfo();
  53
  54                 if( $data !== false ) {
  55                         $width  = $data['width'];
  56                         $height = $data['height'];
  57
  58                         return array( $width, $height, 'DjVu',
  59                                 "width=\"$width\" height=\"$height\"" );
  60                 }
  61                 return false;
  62         }
  63
  64         // ---------
  65
  66         /**
  67          * For debugging; dump the IFF chunk structure
  68          */
  69         function dump() {
  70                 $file = fopen( $this->mFilename, 'rb' );
  71                 $header = fread( $file, 12 );
  72                 // FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
  73                 extract( unpack( 'a4magic/a4chunk/NchunkLength', $header ) );
  74                 echo "$chunk $chunkLength\n";
  75                 $this->dumpForm( $file, $chunkLength, 1 );
  76                 fclose( $file );
  77         }
  78
  79         private function dumpForm( $file, $length, $indent ) {
  80                 $start = ftell( $file );
  81                 $secondary = fread( $file, 4 );
  82                 echo str_repeat( ' ', $indent * 4 ) . "($secondary)\n";
  83                 while( ftell( $file ) - $start < $length ) {
  84                         $chunkHeader = fread( $file, 8 );
  85                         if( $chunkHeader == '' ) {
  86                                 break;
  87                         }
  88                         // FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
  89                         extract( unpack( 'a4chunk/NchunkLength', $chunkHeader ) );
  90                         echo str_repeat( ' ', $indent * 4 ) . "$chunk $chunkLength\n";
  91
  92                         if( $chunk == 'FORM' ) {
  93                                 $this->dumpForm( $file, $chunkLength, $indent + 1 );
  94                         } else {
  95                                 fseek( $file, $chunkLength, SEEK_CUR );
  96                                 if( $chunkLength & 1 == 1 ) {
  97                                         // Padding byte between chunks
  98                                         fseek( $file, 1, SEEK_CUR );
  99                                 }
 100                         }
 101                 }
 102         }
 103
 104         function getInfo() {
 105                 $file = fopen( $this->mFilename, 'rb' );
 106                 if( $file === false ) {
 107                         wfDebug( __METHOD__ . ": missing or failed file read\n" );
 108                         return false;
 109                 }
 110
 111                 $header = fread( $file, 16 );
 112                 $info = false;
 113
 114                 if( strlen( $header ) < 16 ) {
 115                         wfDebug( __METHOD__ . ": too short file header\n" );
 116                 } else {
 117                         // FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 118                         extract( unpack( 'a4magic/a4form/NformLength/a4subtype', $header ) );
 119
 120                         if( $magic != 'AT&T' ) {
 121                                 wfDebug( __METHOD__ . ": not a DjVu file\n" );
 122                         } elseif( $subtype == 'DJVU' ) {
 123                                 // Single-page document
 124                                 $info = $this->getPageInfo( $file, $formLength );
 125                         } elseif( $subtype == 'DJVM' ) {
 126                                 // Multi-page document
 127                                 $info = $this->getMultiPageInfo( $file, $formLength );
 128                         } else  {
 129                                 wfDebug( __METHOD__ . ": unrecognized DJVU file type '$formType'\n" );
 130                         }
 131                 }
 132                 fclose( $file );
 133                 return $info;
 134         }
 135
 136         private function readChunk( $file ) {
 137                 $header = fread( $file, 8 );
 138                 if( strlen( $header ) < 8 ) {
 139                         return array( false, 0 );
 140                 } else {
 141                         // FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 142                         extract( unpack( 'a4chunk/Nlength', $header ) );
 143                         return array( $chunk, $length );
 144                 }
 145         }
 146
 147         private function skipChunk( $file, $chunkLength ) {
 148                 fseek( $file, $chunkLength, SEEK_CUR );
 149
 150                 if( $chunkLength & 0x01 == 1 && !feof( $file ) ) {
 151                         // padding byte
 152                         fseek( $file, 1, SEEK_CUR );
 153                 }
 154         }
 155
 156         private function getMultiPageInfo( $file, $formLength ) {
 157                 // For now, we'll just look for the first page in the file
 158                 // and report its information, hoping others are the same size.
 159                 $start = ftell( $file );
 160                 do {
 161                         list( $chunk, $length ) = $this->readChunk( $file );
 162                         if( !$chunk ) {
 163                                 break;
 164                         }
 165
 166                         if( $chunk == 'FORM' ) {
 167                                 $subtype = fread( $file, 4 );
 168                                 if( $subtype == 'DJVU' ) {
 169                                         wfDebug( __METHOD__ . ": found first subpage\n" );
 170                                         return $this->getPageInfo( $file, $length );
 171                                 }
 172                                 $this->skipChunk( $file, $length - 4 );
 173                         } else {
 174                                 wfDebug( __METHOD__ . ": skipping '$chunk' chunk\n" );
 175                                 $this->skipChunk( $file, $length );
 176                         }
 177                 } while( $length != 0 && !feof( $file ) && ftell( $file ) - $start < $formLength );
 178
 179                 wfDebug( __METHOD__ . ": multi-page DJVU file contained no pages\n" );
 180                 return false;
 181         }
 182
 183         private function getPageInfo( $file, $formLength ) {
 184                 list( $chunk, $length ) = $this->readChunk( $file );
 185                 if( $chunk != 'INFO' ) {
 186                         wfDebug( __METHOD__ . ": expected INFO chunk, got '$chunk'\n" );
 187                         return false;
 188                 }
 189
 190                 if( $length < 9 ) {
 191                         wfDebug( __METHOD__ . ": INFO should be 9 or 10 bytes, found $length\n" );
 192                         return false;
 193                 }
 194                 $data = fread( $file, $length );
 195                 if( strlen( $data ) < $length ) {
 196                         wfDebug( __METHOD__ . ": INFO chunk cut off\n" );
 197                         return false;
 198                 }
 199
 200                 // FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 201                 extract( unpack(
 202                         'nwidth/' .
 203                         'nheight/' .
 204                         'Cminor/' .
 205                         'Cmajor/' .
 206                         'vresolution/' .
 207                         'Cgamma', $data ) );
 208                 # Newer files have rotation info in byte 10, but we don't use it yet.
 209
 210                 return array(
 211                         'width' => $width,
 212                         'height' => $height,
 213                         'version' => "$major.$minor",
 214                         'resolution' => $resolution,
 215                         'gamma' => $gamma / 10.0 );
 216         }
 217
 218         /**
 219          * Return an XML string describing the DjVu image
 220          * @return string
 221          */
 222         function retrieveMetaData() {
 223                 global $wgDjvuToXML, $wgDjvuDump;
 224                 if ( isset( $wgDjvuDump ) ) {
 225                         # djvudump is faster as of version 3.5
 226                         # http://sourceforge.net/tracker/index.php?func=detail&aid=1704049&group_id=32953&atid=406583
 227                         wfProfileIn( 'djvudump' );
 228                         $cmd = wfEscapeShellArg( $wgDjvuDump ) . ' ' . wfEscapeShellArg( $this->mFilename );
 229                         $dump = wfShellExec( $cmd );
 230                         $xml = $this->convertDumpToXML( $dump );
 231                         wfProfileOut( 'djvudump' );
 232                 } elseif ( isset( $wgDjvuToXML ) ) {
 233                         wfProfileIn( 'djvutoxml' );
 234                         $cmd = wfEscapeShellArg( $wgDjvuToXML ) . ' --without-anno --without-text ' .
 235                                 wfEscapeShellArg( $this->mFilename );
 236                         $xml = wfShellExec( $cmd );
 237                         wfProfileOut( 'djvutoxml' );
 238                 } else {
 239                         $xml = null;
 240                 }
 241                 return $xml;
 242         }
 243
 244         /**
 245          * Hack to temporarily work around djvutoxml bug
 246          */
 247         function convertDumpToXML( $dump ) {
 248                 if ( strval( $dump ) == '' ) {
 249                         return false;
 250                 }
 251
 252                 $xml = <<<EOT
 253 <?xml version="1.0" ?>
 254 <!DOCTYPE DjVuXML PUBLIC "-//W3C//DTD DjVuXML 1.1//EN" "pubtext/DjVuXML-s.dtd">
 255 <DjVuXML>
 256 <HEAD></HEAD>
 257 <BODY>
 258 EOT;
 259
 260                 $dump = str_replace( "\r", '', $dump );
 261                 $line = strtok( $dump, "\n" );
 262                 $m = false;
 263                 $good = false;
 264                 if ( preg_match( '/^( *)FORM:DJVU/', $line, $m ) ) {
 265                         # Single-page
 266                         if ( $this->parseFormDjvu( $line, $xml ) ) {
 267                                 $good = true;
 268                         } else {
 269                                 return false;
 270                         }
 271                 } elseif ( preg_match( '/^( *)FORM:DJVM/', $line, $m ) ) {
 272                         # Multi-page
 273                         $parentLevel = strlen( $m[1] );
 274                         # Find DIRM
 275                         $line = strtok( "\n" );
 276                         while ( $line !== false ) {
 277                                 $childLevel = strspn( $line, ' ' );
 278                                 if ( $childLevel <= $parentLevel ) {
 279                                         # End of chunk
 280                                         break;
 281                                 }
 282
 283                                 if ( preg_match( '/^ *DIRM.*indirect/', $line ) ) {
 284                                         wfDebug( "Indirect multi-page DjVu document, bad for server!\n" );
 285                                         return false;
 286                                 }
 287                                 if ( preg_match( '/^ *FORM:DJVU/', $line ) ) {
 288                                         # Found page
 289                                         if ( $this->parseFormDjvu( $line, $xml ) ) {
 290                                                 $good = true;
 291                                         } else {
 292                                                 return false;
 293                                         }
 294                                 }
 295                                 $line = strtok( "\n" );
 296                         }
 297                 }
 298                 if ( !$good ) {
 299                         return false;
 300                 }
 301
 302                 $xml .= "</BODY>\n</DjVuXML>\n";
 303                 return $xml;
 304         }
 305
 306         function parseFormDjvu( $line, &$xml ) {
 307                 $parentLevel = strspn( $line, ' ' );
 308                 $line = strtok( "\n" );
 309
 310                 # Find INFO
 311                 while ( $line !== false ) {
 312                         $childLevel = strspn( $line, ' ' );
 313                         if ( $childLevel <= $parentLevel ) {
 314                                 # End of chunk
 315                                 break;
 316                         }
 317
 318                         if ( preg_match( '/^ *INFO *\[\d*\] *DjVu *(\d+)x(\d+), *\w*, *(\d+) *dpi, *gamma=([0-9.-]+)/', $line, $m ) ) {
 319                                 $xml .= Xml::tags( 'OBJECT',
 320                                         array(
 321                                                 #'data' => '',
 322                                                 #'type' => 'image/x.djvu',
 323                                                 'height' => $m[2],
 324                                                 'width' => $m[1],
 325                                                 #'usemap' => '',
 326                                         ),
 327                                         "\n" .
 328                                         Xml::element( 'PARAM', array( 'name' => 'DPI', 'value' => $m[3] ) ) . "\n" .
 329                                         Xml::element( 'PARAM', array( 'name' => 'GAMMA', 'value' => $m[4] ) ) . "\n"
 330                                 ) . "\n";
 331                                 return true;
 332                         }
 333                         $line = strtok( "\n" );
 334                 }
 335                 # Not found
 336                 return false;
 337         }
 338 }
 339
 340
 341 ?>