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  * @ingroup Media
  33  */
  34 class DjVuImage {
  35         function __construct( $filename ) {
  36                 $this->mFilename = $filename;
  37         }
  38
  39         /**
  40          * Check if the given file is indeed a valid DjVu image file
  41          * @return bool
  42          */
  43         public function isValid() {
  44                 $info = $this->getInfo();
  45                 return $info !== false;
  46         }
  47
  48
  49         /**
  50          * Return data in the style of getimagesize()
  51          * @return array or false on failure
  52          */
  53         public function getImageSize() {
  54                 $data = $this->getInfo();
  55
  56                 if( $data !== false ) {
  57                         $width  = $data['width'];
  58                         $height = $data['height'];
  59
  60                         return array( $width, $height, 'DjVu',
  61                                 "width=\"$width\" height=\"$height\"" );
  62                 }
  63                 return false;
  64         }
  65
  66         // ---------
  67
  68         /**
  69          * For debugging; dump the IFF chunk structure
  70          */
  71         function dump() {
  72                 $file = fopen( $this->mFilename, 'rb' );
  73                 $header = fread( $file, 12 );
  74                 // FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
  75                 extract( unpack( 'a4magic/a4chunk/NchunkLength', $header ) );
  76                 echo "$chunk $chunkLength\n";
  77                 $this->dumpForm( $file, $chunkLength, 1 );
  78                 fclose( $file );
  79         }
  80
  81         private function dumpForm( $file, $length, $indent ) {
  82                 $start = ftell( $file );
  83                 $secondary = fread( $file, 4 );
  84                 echo str_repeat( ' ', $indent * 4 ) . "($secondary)\n";
  85                 while( ftell( $file ) - $start < $length ) {
  86                         $chunkHeader = fread( $file, 8 );
  87                         if( $chunkHeader == '' ) {
  88                                 break;
  89                         }
  90                         // FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
  91                         extract( unpack( 'a4chunk/NchunkLength', $chunkHeader ) );
  92                         echo str_repeat( ' ', $indent * 4 ) . "$chunk $chunkLength\n";
  93
  94                         if( $chunk == 'FORM' ) {
  95                                 $this->dumpForm( $file, $chunkLength, $indent + 1 );
  96                         } else {
  97                                 fseek( $file, $chunkLength, SEEK_CUR );
  98                                 if( $chunkLength & 1 == 1 ) {
  99                                         // Padding byte between chunks
 100                                         fseek( $file, 1, SEEK_CUR );
 101                                 }
 102                         }
 103                 }
 104         }
 105
 106         function getInfo() {
 107                 wfSuppressWarnings();
 108                 $file = fopen( $this->mFilename, 'rb' );
 109                 wfRestoreWarnings();
 110                 if( $file === false ) {
 111                         wfDebug( __METHOD__ . ": missing or failed file read\n" );
 112                         return false;
 113                 }
 114
 115                 $header = fread( $file, 16 );
 116                 $info = false;
 117
 118                 if( strlen( $header ) < 16 ) {
 119                         wfDebug( __METHOD__ . ": too short file header\n" );
 120                 } else {
 121                         // FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 122                         extract( unpack( 'a4magic/a4form/NformLength/a4subtype', $header ) );
 123
 124                         if( $magic != 'AT&T' ) {
 125                                 wfDebug( __METHOD__ . ": not a DjVu file\n" );
 126                         } elseif( $subtype == 'DJVU' ) {
 127                                 // Single-page document
 128                                 $info = $this->getPageInfo( $file, $formLength );
 129                         } elseif( $subtype == 'DJVM' ) {
 130                                 // Multi-page document
 131                                 $info = $this->getMultiPageInfo( $file, $formLength );
 132                         } else  {
 133                                 wfDebug( __METHOD__ . ": unrecognized DJVU file type '$formType'\n" );
 134                         }
 135                 }
 136                 fclose( $file );
 137                 return $info;
 138         }
 139
 140         private function readChunk( $file ) {
 141                 $header = fread( $file, 8 );
 142                 if( strlen( $header ) < 8 ) {
 143                         return array( false, 0 );
 144                 } else {
 145                         // FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 146                         extract( unpack( 'a4chunk/Nlength', $header ) );
 147                         return array( $chunk, $length );
 148                 }
 149         }
 150
 151         private function skipChunk( $file, $chunkLength ) {
 152                 fseek( $file, $chunkLength, SEEK_CUR );
 153
 154                 if( $chunkLength & 0x01 == 1 && !feof( $file ) ) {
 155                         // padding byte
 156                         fseek( $file, 1, SEEK_CUR );
 157                 }
 158         }
 159
 160         private function getMultiPageInfo( $file, $formLength ) {
 161                 // For now, we'll just look for the first page in the file
 162                 // and report its information, hoping others are the same size.
 163                 $start = ftell( $file );
 164                 do {
 165                         list( $chunk, $length ) = $this->readChunk( $file );
 166                         if( !$chunk ) {
 167                                 break;
 168                         }
 169
 170                         if( $chunk == 'FORM' ) {
 171                                 $subtype = fread( $file, 4 );
 172                                 if( $subtype == 'DJVU' ) {
 173                                         wfDebug( __METHOD__ . ": found first subpage\n" );
 174                                         return $this->getPageInfo( $file, $length );
 175                                 }
 176                                 $this->skipChunk( $file, $length - 4 );
 177                         } else {
 178                                 wfDebug( __METHOD__ . ": skipping '$chunk' chunk\n" );
 179                                 $this->skipChunk( $file, $length );
 180                         }
 181                 } while( $length != 0 && !feof( $file ) && ftell( $file ) - $start < $formLength );
 182
 183                 wfDebug( __METHOD__ . ": multi-page DJVU file contained no pages\n" );
 184                 return false;
 185         }
 186
 187         private function getPageInfo( $file, $formLength ) {
 188                 list( $chunk, $length ) = $this->readChunk( $file );
 189                 if( $chunk != 'INFO' ) {
 190                         wfDebug( __METHOD__ . ": expected INFO chunk, got '$chunk'\n" );
 191                         return false;
 192                 }
 193
 194                 if( $length < 9 ) {
 195                         wfDebug( __METHOD__ . ": INFO should be 9 or 10 bytes, found $length\n" );
 196                         return false;
 197                 }
 198                 $data = fread( $file, $length );
 199                 if( strlen( $data ) < $length ) {
 200                         wfDebug( __METHOD__ . ": INFO chunk cut off\n" );
 201                         return false;
 202                 }
 203
 204                 // FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 205                 extract( unpack(
 206                         'nwidth/' .
 207                         'nheight/' .
 208                         'Cminor/' .
 209                         'Cmajor/' .
 210                         'vresolution/' .
 211                         'Cgamma', $data ) );
 212                 # Newer files have rotation info in byte 10, but we don't use it yet.
 213
 214                 return array(
 215                         'width' => $width,
 216                         'height' => $height,
 217                         'version' => "$major.$minor",
 218                         'resolution' => $resolution,
 219                         'gamma' => $gamma / 10.0 );
 220         }
 221
 222         /**
 223          * Return an XML string describing the DjVu image
 224          * @return string
 225          */
 226         function retrieveMetaData() {
 227                 global $wgDjvuToXML, $wgDjvuDump, $wgDjvuTxt;
 228                 if ( isset( $wgDjvuDump ) ) {
 229                         # djvudump is faster as of version 3.5
 230                         # http://sourceforge.net/tracker/index.php?func=detail&aid=1704049&group_id=32953&atid=406583
 231                         wfProfileIn( 'djvudump' );
 232                         $cmd = wfEscapeShellArg( $wgDjvuDump ) . ' ' . wfEscapeShellArg( $this->mFilename );
 233                         $dump = wfShellExec( $cmd );
 234                         $xml = $this->convertDumpToXML( $dump );
 235                         wfProfileOut( 'djvudump' );
 236                 } elseif ( isset( $wgDjvuToXML ) ) {
 237                         wfProfileIn( 'djvutoxml' );
 238                         $cmd = wfEscapeShellArg( $wgDjvuToXML ) . ' --without-anno --without-text ' .
 239                                 wfEscapeShellArg( $this->mFilename );
 240                         $xml = wfShellExec( $cmd );
 241                         wfProfileOut( 'djvutoxml' );
 242                 } else {
 243                         $xml = null;
 244                 }
 245                 # Text layer
 246                 if ( isset( $wgDjvuTxt ) ) {
 247                         wfProfileIn( 'djvutxt' );
 248                         $cmd = wfEscapeShellArg( $wgDjvuTxt ) . ' --detail=page ' . wfEscapeShellArg( $this->mFilename ) ;
 249                         wfDebug( __METHOD__.": $cmd\n" );
 250                         $txt = wfShellExec( $cmd, $retval );
 251                         wfProfileOut( 'djvutxt' );
 252                         if( $retval == 0) {
 253                                 # Get rid of invalid UTF-8, strip control characters
 254                                 if( is_callable( 'iconv' ) ) {
 255                                         wfSuppressWarnings();
 256                                         $txt = iconv( "UTF-8","UTF-8//IGNORE", $txt );
 257                                         wfRestoreWarnings();
 258                                 } else {
 259                                         $txt = UtfNormal::cleanUp( $txt );
 260                                 }
 261                                 $txt = preg_replace( "/[\013\035\037]/", "", $txt );
 262                                 $txt = htmlspecialchars($txt);
 263                                 $txt = preg_replace( "/\((page\s[\d-]*\s[\d-]*\s[\d-]*\s[\d-]*\s*\&quot;([^<]*?)\&quot;\s*|)\)/s", "<PAGE value=\"$2\" />", $txt  );
 264                                 $txt = "<DjVuTxt>\n<HEAD></HEAD>\n<BODY>\n" . $txt . "</BODY>\n</DjVuTxt>\n";
 265                                 $xml = preg_replace( "/<DjVuXML>/", "<mw-djvu><DjVuXML>", $xml );
 266                                 $xml = $xml . $txt. '</mw-djvu>' ;
 267                         }
 268                 }
 269                 return $xml;
 270         }
 271
 272         /**
 273          * Hack to temporarily work around djvutoxml bug
 274          */
 275         function convertDumpToXML( $dump ) {
 276                 if ( strval( $dump ) == '' ) {
 277                         return false;
 278                 }
 279
 280                 $xml = <<<EOT
 281 <?xml version="1.0" ?>
 282 <!DOCTYPE DjVuXML PUBLIC "-//W3C//DTD DjVuXML 1.1//EN" "pubtext/DjVuXML-s.dtd">
 283 <DjVuXML>
 284 <HEAD></HEAD>
 285 <BODY>
 286 EOT;
 287
 288                 $dump = str_replace( "\r", '', $dump );
 289                 $line = strtok( $dump, "\n" );
 290                 $m = false;
 291                 $good = false;
 292                 if ( preg_match( '/^( *)FORM:DJVU/', $line, $m ) ) {
 293                         # Single-page
 294                         if ( $this->parseFormDjvu( $line, $xml ) ) {
 295                                 $good = true;
 296                         } else {
 297                                 return false;
 298                         }
 299                 } elseif ( preg_match( '/^( *)FORM:DJVM/', $line, $m ) ) {
 300                         # Multi-page
 301                         $parentLevel = strlen( $m[1] );
 302                         # Find DIRM
 303                         $line = strtok( "\n" );
 304                         while ( $line !== false ) {
 305                                 $childLevel = strspn( $line, ' ' );
 306                                 if ( $childLevel <= $parentLevel ) {
 307                                         # End of chunk
 308                                         break;
 309                                 }
 310
 311                                 if ( preg_match( '/^ *DIRM.*indirect/', $line ) ) {
 312                                         wfDebug( "Indirect multi-page DjVu document, bad for server!\n" );
 313                                         return false;
 314                                 }
 315                                 if ( preg_match( '/^ *FORM:DJVU/', $line ) ) {
 316                                         # Found page
 317                                         if ( $this->parseFormDjvu( $line, $xml ) ) {
 318                                                 $good = true;
 319                                         } else {
 320                                                 return false;
 321                                         }
 322                                 }
 323                                 $line = strtok( "\n" );
 324                         }
 325                 }
 326                 if ( !$good ) {
 327                         return false;
 328                 }
 329
 330                 $xml .= "</BODY>\n</DjVuXML>\n";
 331                 return $xml;
 332         }
 333
 334         function parseFormDjvu( $line, &$xml ) {
 335                 $parentLevel = strspn( $line, ' ' );
 336                 $line = strtok( "\n" );
 337
 338                 # Find INFO
 339                 while ( $line !== false ) {
 340                         $childLevel = strspn( $line, ' ' );
 341                         if ( $childLevel <= $parentLevel ) {
 342                                 # End of chunk
 343                                 break;
 344                         }
 345
 346                         if ( preg_match( '/^ *INFO *\[\d*\] *DjVu *(\d+)x(\d+), *\w*, *(\d+) *dpi, *gamma=([0-9.-]+)/', $line, $m ) ) {
 347                                 $xml .= Xml::tags( 'OBJECT',
 348                                         array(
 349                                                 #'data' => '',
 350                                                 #'type' => 'image/x.djvu',
 351                                                 'height' => $m[2],
 352                                                 'width' => $m[1],
 353                                                 #'usemap' => '',
 354                                         ),
 355                                         "\n" .
 356                                         Xml::element( 'PARAM', array( 'name' => 'DPI', 'value' => $m[3] ) ) . "\n" .
 357                                         Xml::element( 'PARAM', array( 'name' => 'GAMMA', 'value' => $m[4] ) ) . "\n"
 358                                 ) . "\n";
 359                                 return true;
 360                         }
 361                         $line = strtok( "\n" );
 362                 }
 363                 # Not found
 364                 return false;
 365         }
 366 }