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
  58                 return $info !== false;
  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
  76                 return false;
  77         }
  78
  79         // ---------
  80
  81         /**
  82          * For debugging; dump the IFF chunk structure
  83          */
  84         function dump() {
  85                 $file = fopen( $this->mFilename, 'rb' );
  86                 $header = fread( $file, 12 );
  87                 // @todo FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
  88                 extract( unpack( 'a4magic/a4chunk/NchunkLength', $header ) );
  89                 echo "$chunk $chunkLength\n";
  90                 $this->dumpForm( $file, $chunkLength, 1 );
  91                 fclose( $file );
  92         }
  93
  94         private function dumpForm( $file, $length, $indent ) {
  95                 $start = ftell( $file );
  96                 $secondary = fread( $file, 4 );
  97                 echo str_repeat( ' ', $indent * 4 ) . "($secondary)\n";
  98                 while ( ftell( $file ) - $start < $length ) {
  99                         $chunkHeader = fread( $file, 8 );
 100                         if ( $chunkHeader == '' ) {
 101                                 break;
 102                         }
 103                         // @todo FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 104                         extract( unpack( 'a4chunk/NchunkLength', $chunkHeader ) );
 105                         echo str_repeat( ' ', $indent * 4 ) . "$chunk $chunkLength\n";
 106
 107                         if ( $chunk == 'FORM' ) {
 108                                 $this->dumpForm( $file, $chunkLength, $indent + 1 );
 109                         } else {
 110                                 fseek( $file, $chunkLength, SEEK_CUR );
 111                                 if ( $chunkLength & 1 == 1 ) {
 112                                         // Padding byte between chunks
 113                                         fseek( $file, 1, SEEK_CUR );
 114                                 }
 115                         }
 116                 }
 117         }
 118
 119         function getInfo() {
 120                 wfSuppressWarnings();
 121                 $file = fopen( $this->mFilename, 'rb' );
 122                 wfRestoreWarnings();
 123                 if ( $file === false ) {
 124                         wfDebug( __METHOD__ . ": missing or failed file read\n" );
 125
 126                         return false;
 127                 }
 128
 129                 $header = fread( $file, 16 );
 130                 $info = false;
 131
 132                 if ( strlen( $header ) < 16 ) {
 133                         wfDebug( __METHOD__ . ": too short file header\n" );
 134                 } else {
 135                         // @todo FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 136                         extract( unpack( 'a4magic/a4form/NformLength/a4subtype', $header ) );
 137
 138                         if ( $magic != 'AT&T' ) {
 139                                 wfDebug( __METHOD__ . ": not a DjVu file\n" );
 140                         } elseif ( $subtype == 'DJVU' ) {
 141                                 // Single-page document
 142                                 $info = $this->getPageInfo( $file, $formLength );
 143                         } elseif ( $subtype == 'DJVM' ) {
 144                                 // Multi-page document
 145                                 $info = $this->getMultiPageInfo( $file, $formLength );
 146                         } else {
 147                                 wfDebug( __METHOD__ . ": unrecognized DJVU file type '$formType'\n" );
 148                         }
 149                 }
 150                 fclose( $file );
 151
 152                 return $info;
 153         }
 154
 155         private function readChunk( $file ) {
 156                 $header = fread( $file, 8 );
 157                 if ( strlen( $header ) < 8 ) {
 158                         return array( false, 0 );
 159                 } else {
 160                         // @todo FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 161                         extract( unpack( 'a4chunk/Nlength', $header ) );
 162
 163                         return array( $chunk, $length );
 164                 }
 165         }
 166
 167         private function skipChunk( $file, $chunkLength ) {
 168                 fseek( $file, $chunkLength, SEEK_CUR );
 169
 170                 if ( $chunkLength & 0x01 == 1 && !feof( $file ) ) {
 171                         // padding byte
 172                         fseek( $file, 1, SEEK_CUR );
 173                 }
 174         }
 175
 176         private function getMultiPageInfo( $file, $formLength ) {
 177                 // For now, we'll just look for the first page in the file
 178                 // and report its information, hoping others are the same size.
 179                 $start = ftell( $file );
 180                 do {
 181                         list( $chunk, $length ) = $this->readChunk( $file );
 182                         if ( !$chunk ) {
 183                                 break;
 184                         }
 185
 186                         if ( $chunk == 'FORM' ) {
 187                                 $subtype = fread( $file, 4 );
 188                                 if ( $subtype == 'DJVU' ) {
 189                                         wfDebug( __METHOD__ . ": found first subpage\n" );
 190
 191                                         return $this->getPageInfo( $file, $length );
 192                                 }
 193                                 $this->skipChunk( $file, $length - 4 );
 194                         } else {
 195                                 wfDebug( __METHOD__ . ": skipping '$chunk' chunk\n" );
 196                                 $this->skipChunk( $file, $length );
 197                         }
 198                 } while ( $length != 0 && !feof( $file ) && ftell( $file ) - $start < $formLength );
 199
 200                 wfDebug( __METHOD__ . ": multi-page DJVU file contained no pages\n" );
 201
 202                 return false;
 203         }
 204
 205         private function getPageInfo( $file, $formLength ) {
 206                 list( $chunk, $length ) = $this->readChunk( $file );
 207                 if ( $chunk != 'INFO' ) {
 208                         wfDebug( __METHOD__ . ": expected INFO chunk, got '$chunk'\n" );
 209
 210                         return false;
 211                 }
 212
 213                 if ( $length < 9 ) {
 214                         wfDebug( __METHOD__ . ": INFO should be 9 or 10 bytes, found $length\n" );
 215
 216                         return false;
 217                 }
 218                 $data = fread( $file, $length );
 219                 if ( strlen( $data ) < $length ) {
 220                         wfDebug( __METHOD__ . ": INFO chunk cut off\n" );
 221
 222                         return false;
 223                 }
 224
 225                 // @todo FIXME: Would be good to replace this extract() call with something that explicitly initializes local variables.
 226                 extract( unpack(
 227                         'nwidth/' .
 228                         'nheight/' .
 229                         'Cminor/' .
 230                         'Cmajor/' .
 231                         'vresolution/' .
 232                         'Cgamma', $data ) );
 233
 234                 # Newer files have rotation info in byte 10, but we don't use it yet.
 235
 236                 return array(
 237                         'width' => $width,
 238                         'height' => $height,
 239                         'version' => "$major.$minor",
 240                         'resolution' => $resolution,
 241                         'gamma' => $gamma / 10.0 );
 242         }
 243
 244         /**
 245          * Return an XML string describing the DjVu image
 246          * @return string
 247          */
 248         function retrieveMetaData() {
 249                 global $wgDjvuToXML, $wgDjvuDump, $wgDjvuTxt;
 250                 wfProfileIn( __METHOD__ );
 251
 252                 if ( isset( $wgDjvuDump ) ) {
 253                         # djvudump is faster as of version 3.5
 254                         # http://sourceforge.net/tracker/index.php?func=detail&aid=1704049&group_id=32953&atid=406583
 255                         wfProfileIn( 'djvudump' );
 256                         $cmd = wfEscapeShellArg( $wgDjvuDump ) . ' ' . wfEscapeShellArg( $this->mFilename );
 257                         $dump = wfShellExec( $cmd );
 258                         $xml = $this->convertDumpToXML( $dump );
 259                         wfProfileOut( 'djvudump' );
 260                 } elseif ( isset( $wgDjvuToXML ) ) {
 261                         wfProfileIn( 'djvutoxml' );
 262                         $cmd = wfEscapeShellArg( $wgDjvuToXML ) . ' --without-anno --without-text ' .
 263                                 wfEscapeShellArg( $this->mFilename );
 264                         $xml = wfShellExec( $cmd );
 265                         wfProfileOut( 'djvutoxml' );
 266                 } else {
 267                         $xml = null;
 268                 }
 269                 # Text layer
 270                 if ( isset( $wgDjvuTxt ) ) {
 271                         wfProfileIn( 'djvutxt' );
 272                         $cmd = wfEscapeShellArg( $wgDjvuTxt ) . ' --detail=page ' . wfEscapeShellArg( $this->mFilename );
 273                         wfDebug( __METHOD__ . ": $cmd\n" );
 274                         $retval = '';
 275                         $txt = wfShellExec( $cmd, $retval, array(), array( 'memory' => self::DJVUTXT_MEMORY_LIMIT ) );
 276                         wfProfileOut( 'djvutxt' );
 277                         if ( $retval == 0 ) {
 278                                 # Strip some control characters
 279                                 $txt = preg_replace( "/[\013\035\037]/", "", $txt );
 280                                 $reg = <<<EOR
 281                                         /\(page\s[\d-]*\s[\d-]*\s[\d-]*\s[\d-]*\s*"
 282                                         ((?>    # Text to match is composed of atoms of either:
 283                                           \\\\. # - any escaped character
 284                                           |     # - any character different from " and \
 285                                           [^"\\\\]+
 286                                         )*?)
 287                                         "\s*\)
 288                                         | # Or page can be empty ; in this case, djvutxt dumps ()
 289                                         \(\s*()\)/sx
 290 EOR;
 291                                 $txt = preg_replace_callback( $reg, array( $this, 'pageTextCallback' ), $txt );
 292                                 $txt = "<DjVuTxt>\n<HEAD></HEAD>\n<BODY>\n" . $txt . "</BODY>\n</DjVuTxt>\n";
 293                                 $xml = preg_replace( "/<DjVuXML>/", "<mw-djvu><DjVuXML>", $xml, 1 );
 294                                 $xml = $xml . $txt . '</mw-djvu>';
 295                         }
 296                 }
 297                 wfProfileOut( __METHOD__ );
 298
 299                 return $xml;
 300         }
 301
 302         function pageTextCallback( $matches ) {
 303                 # Get rid of invalid UTF-8, strip control characters
 304                 return '<PAGE value="' . htmlspecialchars( UtfNormal::cleanUp( $matches[1] ) ) . '" />';
 305         }
 306
 307         /**
 308          * Hack to temporarily work around djvutoxml bug
 309          * @return bool|string
 310          */
 311         function convertDumpToXML( $dump ) {
 312                 if ( strval( $dump ) == '' ) {
 313                         return false;
 314                 }
 315
 316                 $xml = <<<EOT
 317 <?xml version="1.0" ?>
 318 <!DOCTYPE DjVuXML PUBLIC "-//W3C//DTD DjVuXML 1.1//EN" "pubtext/DjVuXML-s.dtd">
 319 <DjVuXML>
 320 <HEAD></HEAD>
 321 <BODY>
 322 EOT;
 323
 324                 $dump = str_replace( "\r", '', $dump );
 325                 $line = strtok( $dump, "\n" );
 326                 $m = false;
 327                 $good = false;
 328                 if ( preg_match( '/^( *)FORM:DJVU/', $line, $m ) ) {
 329                         # Single-page
 330                         if ( $this->parseFormDjvu( $line, $xml ) ) {
 331                                 $good = true;
 332                         } else {
 333                                 return false;
 334                         }
 335                 } elseif ( preg_match( '/^( *)FORM:DJVM/', $line, $m ) ) {
 336                         # Multi-page
 337                         $parentLevel = strlen( $m[1] );
 338                         # Find DIRM
 339                         $line = strtok( "\n" );
 340                         while ( $line !== false ) {
 341                                 $childLevel = strspn( $line, ' ' );
 342                                 if ( $childLevel <= $parentLevel ) {
 343                                         # End of chunk
 344                                         break;
 345                                 }
 346
 347                                 if ( preg_match( '/^ *DIRM.*indirect/', $line ) ) {
 348                                         wfDebug( "Indirect multi-page DjVu document, bad for server!\n" );
 349
 350                                         return false;
 351                                 }
 352                                 if ( preg_match( '/^ *FORM:DJVU/', $line ) ) {
 353                                         # Found page
 354                                         if ( $this->parseFormDjvu( $line, $xml ) ) {
 355                                                 $good = true;
 356                                         } else {
 357                                                 return false;
 358                                         }
 359                                 }
 360                                 $line = strtok( "\n" );
 361                         }
 362                 }
 363                 if ( !$good ) {
 364                         return false;
 365                 }
 366
 367                 $xml .= "</BODY>\n</DjVuXML>\n";
 368
 369                 return $xml;
 370         }
 371
 372         function parseFormDjvu( $line, &$xml ) {
 373                 $parentLevel = strspn( $line, ' ' );
 374                 $line = strtok( "\n" );
 375
 376                 # Find INFO
 377                 while ( $line !== false ) {
 378                         $childLevel = strspn( $line, ' ' );
 379                         if ( $childLevel <= $parentLevel ) {
 380                                 # End of chunk
 381                                 break;
 382                         }
 383
 384                         if ( preg_match( '/^ *INFO *\[\d*\] *DjVu *(\d+)x(\d+), *\w*, *(\d+) *dpi, *gamma=([0-9.-]+)/', $line, $m ) ) {
 385                                 $xml .= Xml::tags(
 386                                         'OBJECT',
 387                                         array(
 388                                                 #'data' => '',
 389                                                 #'type' => 'image/x.djvu',
 390                                                 'height' => $m[2],
 391                                                 'width' => $m[1],
 392                                                 #'usemap' => '',
 393                                         ),
 394                                         "\n" .
 395                                                 Xml::element( 'PARAM', array( 'name' => 'DPI', 'value' => $m[3] ) ) . "\n" .
 396                                                 Xml::element( 'PARAM', array( 'name' => 'GAMMA', 'value' => $m[4] ) ) . "\n"
 397                                 ) . "\n";
 398
 399                                 return true;
 400                         }
 401                         $line = strtok( "\n" );
 402                 }
 403
 404                 # Not found
 405                 return false;
 406         }
 407 }