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         /**
  62          * Return data in the style of getimagesize()
  63          * @return array or false on failure
  64          */
  65         public function getImageSize() {
  66                 $data = $this->getInfo();
  67
  68                 if( $data !== false ) {
  69                         $width  = $data['width'];
  70                         $height = $data['height'];
  71
  72                         return array( $width, $height, 'DjVu',
  73                                 "width=\"$width\" height=\"$height\"" );
  74                 }
  75                 return false;
  76         }
  77
  78         // ---------
  79
  80         /**
  81          * For debugging; dump the IFF chunk structure
  82          */
  83         function dump() {
  84                 $file = fopen( $this->mFilename, 'rb' );
  85                 $header = fread( $file, 12 );
  86                 // @todo FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
  87                 extract( unpack( 'a4magic/a4chunk/NchunkLength', $header ) );
  88                 echo "$chunk $chunkLength\n";
  89                 $this->dumpForm( $file, $chunkLength, 1 );
  90                 fclose( $file );
  91         }
  92
  93         private function dumpForm( $file, $length, $indent ) {
  94                 $start = ftell( $file );
  95                 $secondary = fread( $file, 4 );
  96                 echo str_repeat( ' ', $indent * 4 ) . "($secondary)\n";
  97                 while( ftell( $file ) - $start < $length ) {
  98                         $chunkHeader = fread( $file, 8 );
  99                         if( $chunkHeader == '' ) {
 100                                 break;
 101                         }
 102                         // @todo FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 103                         extract( unpack( 'a4chunk/NchunkLength', $chunkHeader ) );
 104                         echo str_repeat( ' ', $indent * 4 ) . "$chunk $chunkLength\n";
 105
 106                         if( $chunk == 'FORM' ) {
 107                                 $this->dumpForm( $file, $chunkLength, $indent + 1 );
 108                         } else {
 109                                 fseek( $file, $chunkLength, SEEK_CUR );
 110                                 if( $chunkLength & 1 == 1 ) {
 111                                         // Padding byte between chunks
 112                                         fseek( $file, 1, SEEK_CUR );
 113                                 }
 114                         }
 115                 }
 116         }
 117
 118         function getInfo() {
 119                 wfSuppressWarnings();
 120                 $file = fopen( $this->mFilename, 'rb' );
 121                 wfRestoreWarnings();
 122                 if( $file === false ) {
 123                         wfDebug( __METHOD__ . ": missing or failed file read\n" );
 124                         return false;
 125                 }
 126
 127                 $header = fread( $file, 16 );
 128                 $info = false;
 129
 130                 if( strlen( $header ) < 16 ) {
 131                         wfDebug( __METHOD__ . ": too short file header\n" );
 132                 } else {
 133                         // @todo FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 134                         extract( unpack( 'a4magic/a4form/NformLength/a4subtype', $header ) );
 135
 136                         if( $magic != 'AT&T' ) {
 137                                 wfDebug( __METHOD__ . ": not a DjVu file\n" );
 138                         } elseif( $subtype == 'DJVU' ) {
 139                                 // Single-page document
 140                                 $info = $this->getPageInfo( $file, $formLength );
 141                         } elseif( $subtype == 'DJVM' ) {
 142                                 // Multi-page document
 143                                 $info = $this->getMultiPageInfo( $file, $formLength );
 144                         } else  {
 145                                 wfDebug( __METHOD__ . ": unrecognized DJVU file type '$formType'\n" );
 146                         }
 147                 }
 148                 fclose( $file );
 149                 return $info;
 150         }
 151
 152         private function readChunk( $file ) {
 153                 $header = fread( $file, 8 );
 154                 if( strlen( $header ) < 8 ) {
 155                         return array( false, 0 );
 156                 } else {
 157                         // @todo FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 158                         extract( unpack( 'a4chunk/Nlength', $header ) );
 159                         return array( $chunk, $length );
 160                 }
 161         }
 162
 163         private function skipChunk( $file, $chunkLength ) {
 164                 fseek( $file, $chunkLength, SEEK_CUR );
 165
 166                 if( $chunkLength & 0x01 == 1 && !feof( $file ) ) {
 167                         // padding byte
 168                         fseek( $file, 1, SEEK_CUR );
 169                 }
 170         }
 171
 172         private function getMultiPageInfo( $file, $formLength ) {
 173                 // For now, we'll just look for the first page in the file
 174                 // and report its information, hoping others are the same size.
 175                 $start = ftell( $file );
 176                 do {
 177                         list( $chunk, $length ) = $this->readChunk( $file );
 178                         if( !$chunk ) {
 179                                 break;
 180                         }
 181
 182                         if( $chunk == 'FORM' ) {
 183                                 $subtype = fread( $file, 4 );
 184                                 if( $subtype == 'DJVU' ) {
 185                                         wfDebug( __METHOD__ . ": found first subpage\n" );
 186                                         return $this->getPageInfo( $file, $length );
 187                                 }
 188                                 $this->skipChunk( $file, $length - 4 );
 189                         } else {
 190                                 wfDebug( __METHOD__ . ": skipping '$chunk' chunk\n" );
 191                                 $this->skipChunk( $file, $length );
 192                         }
 193                 } while( $length != 0 && !feof( $file ) && ftell( $file ) - $start < $formLength );
 194
 195                 wfDebug( __METHOD__ . ": multi-page DJVU file contained no pages\n" );
 196                 return false;
 197         }
 198
 199         private function getPageInfo( $file, $formLength ) {
 200                 list( $chunk, $length ) = $this->readChunk( $file );
 201                 if( $chunk != 'INFO' ) {
 202                         wfDebug( __METHOD__ . ": expected INFO chunk, got '$chunk'\n" );
 203                         return false;
 204                 }
 205
 206                 if( $length < 9 ) {
 207                         wfDebug( __METHOD__ . ": INFO should be 9 or 10 bytes, found $length\n" );
 208                         return false;
 209                 }
 210                 $data = fread( $file, $length );
 211                 if( strlen( $data ) < $length ) {
 212                         wfDebug( __METHOD__ . ": INFO chunk cut off\n" );
 213                         return false;
 214                 }
 215
 216                 // @todo FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 217                 extract( unpack(
 218                         'nwidth/' .
 219                         'nheight/' .
 220                         'Cminor/' .
 221                         'Cmajor/' .
 222                         'vresolution/' .
 223                         'Cgamma', $data ) );
 224                 # Newer files have rotation info in byte 10, but we don't use it yet.
 225
 226                 return array(
 227                         'width' => $width,
 228                         'height' => $height,
 229                         'version' => "$major.$minor",
 230                         'resolution' => $resolution,
 231                         'gamma' => $gamma / 10.0 );
 232         }
 233
 234         /**
 235          * Return an XML string describing the DjVu image
 236          * @return string
 237          */
 238         function retrieveMetaData() {
 239                 global $wgDjvuToXML, $wgDjvuDump, $wgDjvuTxt;
 240                 wfProfileIn( __METHOD__ );
 241
 242                 if ( isset( $wgDjvuDump ) ) {
 243                         # djvudump is faster as of version 3.5
 244                         # http://sourceforge.net/tracker/index.php?func=detail&aid=1704049&group_id=32953&atid=406583
 245                         wfProfileIn( 'djvudump' );
 246                         $cmd = wfEscapeShellArg( $wgDjvuDump ) . ' ' . wfEscapeShellArg( $this->mFilename );
 247                         $dump = wfShellExec( $cmd );
 248                         $xml = $this->convertDumpToXML( $dump );
 249                         wfProfileOut( 'djvudump' );
 250                 } elseif ( isset( $wgDjvuToXML ) ) {
 251                         wfProfileIn( 'djvutoxml' );
 252                         $cmd = wfEscapeShellArg( $wgDjvuToXML ) . ' --without-anno --without-text ' .
 253                                 wfEscapeShellArg( $this->mFilename );
 254                         $xml = wfShellExec( $cmd );
 255                         wfProfileOut( 'djvutoxml' );
 256                 } else {
 257                         $xml = null;
 258                 }
 259                 # Text layer
 260                 if ( isset( $wgDjvuTxt ) ) {
 261                         wfProfileIn( 'djvutxt' );
 262                         $cmd = wfEscapeShellArg( $wgDjvuTxt ) . ' --detail=page ' . wfEscapeShellArg( $this->mFilename ) ;
 263                         wfDebug( __METHOD__.": $cmd\n" );
 264                         $retval = '';
 265                         $txt = wfShellExec( $cmd, $retval, array(), array( 'memory' => self::DJVUTXT_MEMORY_LIMIT ) );
 266                         wfProfileOut( 'djvutxt' );
 267                         if( $retval == 0) {
 268                                 # Strip some control characters
 269                                 $txt = preg_replace( "/[\013\035\037]/", "", $txt );
 270                                 $reg = <<<EOR
 271                                         /\(page\s[\d-]*\s[\d-]*\s[\d-]*\s[\d-]*\s*"
 272                                         ((?>    # Text to match is composed of atoms of either:
 273                                           \\\\. # - any escaped character
 274                                           |     # - any character different from " and \
 275                                           [^"\\\\]+
 276                                         )*?)
 277                                         "\s*\)
 278                                         | # Or page can be empty ; in this case, djvutxt dumps ()
 279                                         \(\s*()\)/sx
 280 EOR;
 281                                 $txt = preg_replace_callback( $reg, array( $this, 'pageTextCallback' ), $txt );
 282                                 $txt = "<DjVuTxt>\n<HEAD></HEAD>\n<BODY>\n" . $txt . "</BODY>\n</DjVuTxt>\n";
 283                                 $xml = preg_replace( "/<DjVuXML>/", "<mw-djvu><DjVuXML>", $xml, 1 );
 284                                 $xml = $xml . $txt. '</mw-djvu>' ;
 285                         }
 286                 }
 287                 wfProfileOut( __METHOD__ );
 288                 return $xml;
 289         }
 290
 291         function pageTextCallback( $matches ) {
 292                 # Get rid of invalid UTF-8, strip control characters
 293                 return '<PAGE value="' . htmlspecialchars( UtfNormal::cleanUp( $matches[1] ) ) . '" />';
 294         }
 295
 296         /**
 297          * Hack to temporarily work around djvutoxml bug
 298          * @return bool|string
 299          */
 300         function convertDumpToXML( $dump ) {
 301                 if ( strval( $dump ) == '' ) {
 302                         return false;
 303                 }
 304
 305                 $xml = <<<EOT
 306 <?xml version="1.0" ?>
 307 <!DOCTYPE DjVuXML PUBLIC "-//W3C//DTD DjVuXML 1.1//EN" "pubtext/DjVuXML-s.dtd">
 308 <DjVuXML>
 309 <HEAD></HEAD>
 310 <BODY>
 311 EOT;
 312
 313                 $dump = str_replace( "\r", '', $dump );
 314                 $line = strtok( $dump, "\n" );
 315                 $m = false;
 316                 $good = false;
 317                 if ( preg_match( '/^( *)FORM:DJVU/', $line, $m ) ) {
 318                         # Single-page
 319                         if ( $this->parseFormDjvu( $line, $xml ) ) {
 320                                 $good = true;
 321                         } else {
 322                                 return false;
 323                         }
 324                 } elseif ( preg_match( '/^( *)FORM:DJVM/', $line, $m ) ) {
 325                         # Multi-page
 326                         $parentLevel = strlen( $m[1] );
 327                         # Find DIRM
 328                         $line = strtok( "\n" );
 329                         while ( $line !== false ) {
 330                                 $childLevel = strspn( $line, ' ' );
 331                                 if ( $childLevel <= $parentLevel ) {
 332                                         # End of chunk
 333                                         break;
 334                                 }
 335
 336                                 if ( preg_match( '/^ *DIRM.*indirect/', $line ) ) {
 337                                         wfDebug( "Indirect multi-page DjVu document, bad for server!\n" );
 338                                         return false;
 339                                 }
 340                                 if ( preg_match( '/^ *FORM:DJVU/', $line ) ) {
 341                                         # Found page
 342                                         if ( $this->parseFormDjvu( $line, $xml ) ) {
 343                                                 $good = true;
 344                                         } else {
 345                                                 return false;
 346                                         }
 347                                 }
 348                                 $line = strtok( "\n" );
 349                         }
 350                 }
 351                 if ( !$good ) {
 352                         return false;
 353                 }
 354
 355                 $xml .= "</BODY>\n</DjVuXML>\n";
 356                 return $xml;
 357         }
 358
 359         function parseFormDjvu( $line, &$xml ) {
 360                 $parentLevel = strspn( $line, ' ' );
 361                 $line = strtok( "\n" );
 362
 363                 # Find INFO
 364                 while ( $line !== false ) {
 365                         $childLevel = strspn( $line, ' ' );
 366                         if ( $childLevel <= $parentLevel ) {
 367                                 # End of chunk
 368                                 break;
 369                         }
 370
 371                         if ( preg_match( '/^ *INFO *\[\d*\] *DjVu *(\d+)x(\d+), *\w*, *(\d+) *dpi, *gamma=([0-9.-]+)/', $line, $m ) ) {
 372                                 $xml .= Xml::tags( 'OBJECT',
 373                                         array(
 374                                                 #'data' => '',
 375                                                 #'type' => 'image/x.djvu',
 376                                                 'height' => $m[2],
 377                                                 'width' => $m[1],
 378                                                 #'usemap' => '',
 379                                         ),
 380                                         "\n" .
 381                                         Xml::element( 'PARAM', array( 'name' => 'DPI', 'value' => $m[3] ) ) . "\n" .
 382                                         Xml::element( 'PARAM', array( 'name' => 'GAMMA', 'value' => $m[4] ) ) . "\n"
 383                                 ) . "\n";
 384                                 return true;
 385                         }
 386                         $line = strtok( "\n" );
 387                 }
 388                 # Not found
 389                 return false;
 390         }
 391 }