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