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
  88                 // something that explicitly initializes local variables.
  89                 extract( unpack( 'a4magic/a4chunk/NchunkLength', $header ) );
  90                 echo "$chunk $chunkLength\n";
  91                 $this->dumpForm( $file, $chunkLength, 1 );
  92                 fclose( $file );
  93         }
  94
  95         private function dumpForm( $file, $length, $indent ) {
  96                 $start = ftell( $file );
  97                 $secondary = fread( $file, 4 );
  98                 echo str_repeat( ' ', $indent * 4 ) . "($secondary)\n";
  99                 while ( ftell( $file ) - $start < $length ) {
 100                         $chunkHeader = fread( $file, 8 );
 101                         if ( $chunkHeader == '' ) {
 102                                 break;
 103                         }
 104                         // @todo FIXME: Would be good to replace this extract() call with
 105                         // something that explicitly initializes local variables.
 106                         extract( unpack( 'a4chunk/NchunkLength', $chunkHeader ) );
 107                         echo str_repeat( ' ', $indent * 4 ) . "$chunk $chunkLength\n";
 108
 109                         if ( $chunk == 'FORM' ) {
 110                                 $this->dumpForm( $file, $chunkLength, $indent + 1 );
 111                         } else {
 112                                 fseek( $file, $chunkLength, SEEK_CUR );
 113                                 if ( $chunkLength & 1 == 1 ) {
 114                                         // Padding byte between chunks
 115                                         fseek( $file, 1, SEEK_CUR );
 116                                 }
 117                         }
 118                 }
 119         }
 120
 121         function getInfo() {
 122                 wfSuppressWarnings();
 123                 $file = fopen( $this->mFilename, 'rb' );
 124                 wfRestoreWarnings();
 125                 if ( $file === false ) {
 126                         wfDebug( __METHOD__ . ": missing or failed file read\n" );
 127
 128                         return false;
 129                 }
 130
 131                 $header = fread( $file, 16 );
 132                 $info = false;
 133
 134                 if ( strlen( $header ) < 16 ) {
 135                         wfDebug( __METHOD__ . ": too short file header\n" );
 136                 } else {
 137                         // @todo FIXME: Would be good to replace this extract() call with
 138                         // something that explicitly initializes local variables.
 139                         extract( unpack( 'a4magic/a4form/NformLength/a4subtype', $header ) );
 140
 141                         if ( $magic != 'AT&T' ) {
 142                                 wfDebug( __METHOD__ . ": not a DjVu file\n" );
 143                         } elseif ( $subtype == 'DJVU' ) {
 144                                 // Single-page document
 145                                 $info = $this->getPageInfo( $file, $formLength );
 146                         } elseif ( $subtype == 'DJVM' ) {
 147                                 // Multi-page document
 148                                 $info = $this->getMultiPageInfo( $file, $formLength );
 149                         } else {
 150                                 wfDebug( __METHOD__ . ": unrecognized DJVU file type '$formType'\n" );
 151                         }
 152                 }
 153                 fclose( $file );
 154
 155                 return $info;
 156         }
 157
 158         private function readChunk( $file ) {
 159                 $header = fread( $file, 8 );
 160                 if ( strlen( $header ) < 8 ) {
 161                         return array( false, 0 );
 162                 } else {
 163                         // @todo FIXME: Would be good to replace this extract() call with
 164                         // something that explicitly initializes local variables.
 165                         extract( unpack( 'a4chunk/Nlength', $header ) );
 166
 167                         return array( $chunk, $length );
 168                 }
 169         }
 170
 171         private function skipChunk( $file, $chunkLength ) {
 172                 fseek( $file, $chunkLength, SEEK_CUR );
 173
 174                 if ( $chunkLength & 0x01 == 1 && !feof( $file ) ) {
 175                         // padding byte
 176                         fseek( $file, 1, SEEK_CUR );
 177                 }
 178         }
 179
 180         private function getMultiPageInfo( $file, $formLength ) {
 181                 // For now, we'll just look for the first page in the file
 182                 // and report its information, hoping others are the same size.
 183                 $start = ftell( $file );
 184                 do {
 185                         list( $chunk, $length ) = $this->readChunk( $file );
 186                         if ( !$chunk ) {
 187                                 break;
 188                         }
 189
 190                         if ( $chunk == 'FORM' ) {
 191                                 $subtype = fread( $file, 4 );
 192                                 if ( $subtype == 'DJVU' ) {
 193                                         wfDebug( __METHOD__ . ": found first subpage\n" );
 194
 195                                         return $this->getPageInfo( $file, $length );
 196                                 }
 197                                 $this->skipChunk( $file, $length - 4 );
 198                         } else {
 199                                 wfDebug( __METHOD__ . ": skipping '$chunk' chunk\n" );
 200                                 $this->skipChunk( $file, $length );
 201                         }
 202                 } while ( $length != 0 && !feof( $file ) && ftell( $file ) - $start < $formLength );
 203
 204                 wfDebug( __METHOD__ . ": multi-page DJVU file contained no pages\n" );
 205
 206                 return false;
 207         }
 208
 209         private function getPageInfo( $file, $formLength ) {
 210                 list( $chunk, $length ) = $this->readChunk( $file );
 211                 if ( $chunk != 'INFO' ) {
 212                         wfDebug( __METHOD__ . ": expected INFO chunk, got '$chunk'\n" );
 213
 214                         return false;
 215                 }
 216
 217                 if ( $length < 9 ) {
 218                         wfDebug( __METHOD__ . ": INFO should be 9 or 10 bytes, found $length\n" );
 219
 220                         return false;
 221                 }
 222                 $data = fread( $file, $length );
 223                 if ( strlen( $data ) < $length ) {
 224                         wfDebug( __METHOD__ . ": INFO chunk cut off\n" );
 225
 226                         return false;
 227                 }
 228
 229                 // @todo FIXME: Would be good to replace this extract() call with
 230                 // something that explicitly initializes local variables.
 231                 extract( unpack(
 232                         'nwidth/' .
 233                         'nheight/' .
 234                         'Cminor/' .
 235                         'Cmajor/' .
 236                         'vresolution/' .
 237                         'Cgamma', $data ) );
 238
 239                 # Newer files have rotation info in byte 10, but we don't use it yet.
 240
 241                 return array(
 242                         'width' => $width,
 243                         'height' => $height,
 244                         'version' => "$major.$minor",
 245                         'resolution' => $resolution,
 246                         'gamma' => $gamma / 10.0 );
 247         }
 248
 249         /**
 250          * Return an XML string describing the DjVu image
 251          * @return string
 252          */
 253         function retrieveMetaData() {
 254                 global $wgDjvuToXML, $wgDjvuDump, $wgDjvuTxt;
 255                 wfProfileIn( __METHOD__ );
 256
 257                 if ( isset( $wgDjvuDump ) ) {
 258                         # djvudump is faster as of version 3.5
 259                         # http://sourceforge.net/tracker/index.php?func=detail&aid=1704049&group_id=32953&atid=406583
 260                         wfProfileIn( 'djvudump' );
 261                         $cmd = wfEscapeShellArg( $wgDjvuDump ) . ' ' . wfEscapeShellArg( $this->mFilename );
 262                         $dump = wfShellExec( $cmd );
 263                         $xml = $this->convertDumpToXML( $dump );
 264                         wfProfileOut( 'djvudump' );
 265                 } elseif ( isset( $wgDjvuToXML ) ) {
 266                         wfProfileIn( 'djvutoxml' );
 267                         $cmd = wfEscapeShellArg( $wgDjvuToXML ) . ' --without-anno --without-text ' .
 268                                 wfEscapeShellArg( $this->mFilename );
 269                         $xml = wfShellExec( $cmd );
 270                         wfProfileOut( 'djvutoxml' );
 271                 } else {
 272                         $xml = null;
 273                 }
 274                 # Text layer
 275                 if ( isset( $wgDjvuTxt ) ) {
 276                         wfProfileIn( 'djvutxt' );
 277                         $cmd = wfEscapeShellArg( $wgDjvuTxt ) . ' --detail=page ' . wfEscapeShellArg( $this->mFilename );
 278                         wfDebug( __METHOD__ . ": $cmd\n" );
 279                         $retval = '';
 280                         $txt = wfShellExec( $cmd, $retval, array(), array( 'memory' => self::DJVUTXT_MEMORY_LIMIT ) );
 281                         wfProfileOut( 'djvutxt' );
 282                         if ( $retval == 0 ) {
 283                                 # Strip some control characters
 284                                 $txt = preg_replace( "/[\013\035\037]/", "", $txt );
 285                                 $reg = <<<EOR
 286                                         /\(page\s[\d-]*\s[\d-]*\s[\d-]*\s[\d-]*\s*"
 287                                         ((?>    # Text to match is composed of atoms of either:
 288                                           \\\\. # - any escaped character
 289                                           |     # - any character different from " and \
 290                                           [^"\\\\]+
 291                                         )*?)
 292                                         "\s*\)
 293                                         | # Or page can be empty ; in this case, djvutxt dumps ()
 294                                         \(\s*()\)/sx
 295 EOR;
 296                                 $txt = preg_replace_callback( $reg, array( $this, 'pageTextCallback' ), $txt );
 297                                 $txt = "<DjVuTxt>\n<HEAD></HEAD>\n<BODY>\n" . $txt . "</BODY>\n</DjVuTxt>\n";
 298                                 $xml = preg_replace( "/<DjVuXML>/", "<mw-djvu><DjVuXML>", $xml, 1 );
 299                                 $xml = $xml . $txt . '</mw-djvu>';
 300                         }
 301                 }
 302                 wfProfileOut( __METHOD__ );
 303
 304                 return $xml;
 305         }
 306
 307         function pageTextCallback( $matches ) {
 308                 # Get rid of invalid UTF-8, strip control characters
 309                 return '<PAGE value="' . htmlspecialchars( UtfNormal::cleanUp( $matches[1] ) ) . '" />';
 310         }
 311
 312         /**
 313          * Hack to temporarily work around djvutoxml bug
 314          * @return bool|string
 315          */
 316         function convertDumpToXML( $dump ) {
 317                 if ( strval( $dump ) == '' ) {
 318                         return false;
 319                 }
 320
 321                 $xml = <<<EOT
 322 <?xml version="1.0" ?>
 323 <!DOCTYPE DjVuXML PUBLIC "-//W3C//DTD DjVuXML 1.1//EN" "pubtext/DjVuXML-s.dtd">
 324 <DjVuXML>
 325 <HEAD></HEAD>
 326 <BODY>
 327 EOT;
 328
 329                 $dump = str_replace( "\r", '', $dump );
 330                 $line = strtok( $dump, "\n" );
 331                 $m = false;
 332                 $good = false;
 333                 if ( preg_match( '/^( *)FORM:DJVU/', $line, $m ) ) {
 334                         # Single-page
 335                         if ( $this->parseFormDjvu( $line, $xml ) ) {
 336                                 $good = true;
 337                         } else {
 338                                 return false;
 339                         }
 340                 } elseif ( preg_match( '/^( *)FORM:DJVM/', $line, $m ) ) {
 341                         # Multi-page
 342                         $parentLevel = strlen( $m[1] );
 343                         # Find DIRM
 344                         $line = strtok( "\n" );
 345                         while ( $line !== false ) {
 346                                 $childLevel = strspn( $line, ' ' );
 347                                 if ( $childLevel <= $parentLevel ) {
 348                                         # End of chunk
 349                                         break;
 350                                 }
 351
 352                                 if ( preg_match( '/^ *DIRM.*indirect/', $line ) ) {
 353                                         wfDebug( "Indirect multi-page DjVu document, bad for server!\n" );
 354
 355                                         return false;
 356                                 }
 357                                 if ( preg_match( '/^ *FORM:DJVU/', $line ) ) {
 358                                         # Found page
 359                                         if ( $this->parseFormDjvu( $line, $xml ) ) {
 360                                                 $good = true;
 361                                         } else {
 362                                                 return false;
 363                                         }
 364                                 }
 365                                 $line = strtok( "\n" );
 366                         }
 367                 }
 368                 if ( !$good ) {
 369                         return false;
 370                 }
 371
 372                 $xml .= "</BODY>\n</DjVuXML>\n";
 373
 374                 return $xml;
 375         }
 376
 377         function parseFormDjvu( $line, &$xml ) {
 378                 $parentLevel = strspn( $line, ' ' );
 379                 $line = strtok( "\n" );
 380
 381                 # Find INFO
 382                 while ( $line !== false ) {
 383                         $childLevel = strspn( $line, ' ' );
 384                         if ( $childLevel <= $parentLevel ) {
 385                                 # End of chunk
 386                                 break;
 387                         }
 388
 389                         if ( preg_match(
 390                                 '/^ *INFO *\[\d*\] *DjVu *(\d+)x(\d+), *\w*, *(\d+) *dpi, *gamma=([0-9.-]+)/',
 391                                 $line,
 392                                 $m
 393                         ) ) {
 394                                 $xml .= Xml::tags(
 395                                         'OBJECT',
 396                                         array(
 397                                                 #'data' => '',
 398                                                 #'type' => 'image/x.djvu',
 399                                                 'height' => $m[2],
 400                                                 'width' => $m[1],
 401                                                 #'usemap' => '',
 402                                         ),
 403                                         "\n" .
 404                                                 Xml::element( 'PARAM', array( 'name' => 'DPI', 'value' => $m[3] ) ) . "\n" .
 405                                                 Xml::element( 'PARAM', array( 'name' => 'GAMMA', 'value' => $m[4] ) ) . "\n"
 406                                 ) . "\n";
 407
 408                                 return true;
 409                         }
 410                         $line = strtok( "\n" );
 411                 }
 412
 413                 # Not found
 414                 return false;
 415         }
 416 }