includes/media/GIFMetadataExtractor.php

   1 <?php
   2 /**
   3  * GIF frame counter.
   4  *
   5  * Originally written in Perl by Steve Sanbeg.
   6  * Ported to PHP by Andrew Garrett
   7  * Deliberately not using MWExceptions to avoid external dependencies, encouraging
   8  * redistribution.
   9  *
  10  * This program is free software; you can redistribute it and/or modify
  11  * it under the terms of the GNU General Public License as published by
  12  * the Free Software Foundation; either version 2 of the License, or
  13  * (at your option) any later version.
  14  *
  15  * This program is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  18  * GNU General Public License for more details.
  19  *
  20  * You should have received a copy of the GNU General Public License along
  21  * with this program; if not, write to the Free Software Foundation, Inc.,
  22  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  23  * http://www.gnu.org/copyleft/gpl.html
  24  *
  25  * @file
  26  * @ingroup Media
  27  */
  28
  29 /**
  30  * GIF frame counter.
  31  *
  32  * @ingroup Media
  33  */
  34 class GIFMetadataExtractor {
  35         static $gif_frame_sep;
  36
  37         static $gif_extension_sep;
  38
  39         static $gif_term;
  40
  41         const VERSION = 1;
  42
  43         // Each sub-block is less than or equal to 255 bytes.
  44         // Most of the time its 255 bytes, except for in XMP
  45         // blocks, where it's usually between 32-127 bytes each.
  46         const MAX_SUBBLOCKS = 262144; // 5mb divided by 20.
  47
  48         /**
  49          * @throws Exception
  50          * @param $filename string
  51          * @return array
  52          */
  53         static function getMetadata( $filename ) {
  54                 self::$gif_frame_sep = pack( "C", ord( "," ) );
  55                 self::$gif_extension_sep = pack( "C", ord( "!" ) );
  56                 self::$gif_term = pack( "C", ord( ";" ) );
  57
  58                 $frameCount = 0;
  59                 $duration = 0.0;
  60                 $isLooped = false;
  61                 $xmp = "";
  62                 $comment = array();
  63
  64                 if ( !$filename ) {
  65                         throw new Exception( "No file name specified" );
  66                 } elseif ( !file_exists( $filename ) || is_dir( $filename ) ) {
  67                         throw new Exception( "File $filename does not exist" );
  68                 }
  69
  70                 $fh = fopen( $filename, 'rb' );
  71
  72                 if ( !$fh ) {
  73                         throw new Exception( "Unable to open file $filename" );
  74                 }
  75
  76                 // Check for the GIF header
  77                 $buf = fread( $fh, 6 );
  78                 if ( !( $buf == 'GIF87a' || $buf == 'GIF89a' ) ) {
  79                         throw new Exception( "Not a valid GIF file; header: $buf" );
  80                 }
  81
  82                 // Skip over width and height.
  83                 fread( $fh, 4 );
  84
  85                 // Read BPP
  86                 $buf = fread( $fh, 1 );
  87                 $bpp = self::decodeBPP( $buf );
  88
  89                 // Skip over background and aspect ratio
  90                 fread( $fh, 2 );
  91
  92                 // Skip over the GCT
  93                 self::readGCT( $fh, $bpp );
  94
  95                 while ( !feof( $fh ) ) {
  96                         $buf = fread( $fh, 1 );
  97
  98                         if ( $buf == self::$gif_frame_sep ) {
  99                                 // Found a frame
 100                                 $frameCount++;
 101
 102                                 ## Skip bounding box
 103                                 fread( $fh, 8 );
 104
 105                                 ## Read BPP
 106                                 $buf = fread( $fh, 1 );
 107                                 $bpp = self::decodeBPP( $buf );
 108
 109                                 ## Read GCT
 110                                 self::readGCT( $fh, $bpp );
 111                                 fread( $fh, 1 );
 112                                 self::skipBlock( $fh );
 113                         } elseif ( $buf == self::$gif_extension_sep ) {
 114                                 $buf = fread( $fh, 1 );
 115                                 if ( strlen( $buf ) < 1 ) {
 116                                         throw new Exception( "Ran out of input" );
 117                                 }
 118                                 $extension_code = unpack( 'C', $buf );
 119                                 $extension_code = $extension_code[1];
 120
 121                                 if ( $extension_code == 0xF9 ) {
 122                                         // Graphics Control Extension.
 123                                         fread( $fh, 1 ); // Block size
 124
 125                                         fread( $fh, 1 ); // Transparency, disposal method, user input
 126
 127                                         $buf = fread( $fh, 2 ); // Delay, in hundredths of seconds.
 128                                         if ( strlen( $buf ) < 2 ) {
 129                                                 throw new Exception( "Ran out of input" );
 130                                         }
 131                                         $delay = unpack( 'v', $buf );
 132                                         $delay = $delay[1];
 133                                         $duration += $delay * 0.01;
 134
 135                                         fread( $fh, 1 ); // Transparent colour index
 136
 137                                         $term = fread( $fh, 1 ); // Should be a terminator
 138                                         if ( strlen( $term ) < 1 ) {
 139                                                 throw new Exception( "Ran out of input" );
 140                                         }
 141                                         $term = unpack( 'C', $term );
 142                                         $term = $term[1];
 143                                         if ( $term != 0 ) {
 144                                                 throw new Exception( "Malformed Graphics Control Extension block" );
 145                                         }
 146                                 } elseif ( $extension_code == 0xFE ) {
 147                                         // Comment block(s).
 148                                         $data = self::readBlock( $fh );
 149                                         if ( $data === "" ) {
 150                                                 throw new Exception( 'Read error, zero-length comment block' );
 151                                         }
 152
 153                                         // The standard says this should be ASCII, however its unclear if
 154                                         // thats true in practise. Check to see if its valid utf-8, if so
 155                                         // assume its that, otherwise assume its windows-1252 (iso-8859-1)
 156                                         $dataCopy = $data;
 157                                         // quickIsNFCVerify has the side effect of replacing any invalid characters
 158                                         UtfNormal::quickIsNFCVerify( $dataCopy );
 159
 160                                         if ( $dataCopy !== $data ) {
 161                                                 wfSuppressWarnings();
 162                                                 $data = iconv( 'windows-1252', 'UTF-8', $data );
 163                                                 wfRestoreWarnings();
 164                                         }
 165
 166                                         $commentCount = count( $comment );
 167                                         if ( $commentCount === 0
 168                                                 || $comment[$commentCount - 1] !== $data
 169                                         ) {
 170                                                 // Some applications repeat the same comment on each
 171                                                 // frame of an animated GIF image, so if this comment
 172                                                 // is identical to the last, only extract once.
 173                                                 $comment[] = $data;
 174                                         }
 175                                 } elseif ( $extension_code == 0xFF ) {
 176                                         // Application extension (Netscape info about the animated gif)
 177                                         // or XMP (or theoretically any other type of extension block)
 178                                         $blockLength = fread( $fh, 1 );
 179                                         if ( strlen( $blockLength ) < 1 ) {
 180                                                 throw new Exception( "Ran out of input" );
 181                                         }
 182                                         $blockLength = unpack( 'C', $blockLength );
 183                                         $blockLength = $blockLength[1];
 184                                         $data = fread( $fh, $blockLength );
 185
 186                                         if ( $blockLength != 11 ) {
 187                                                 wfDebug( __METHOD__ . " GIF application block with wrong length\n" );
 188                                                 fseek( $fh, -( $blockLength + 1 ), SEEK_CUR );
 189                                                 self::skipBlock( $fh );
 190                                                 continue;
 191                                         }
 192
 193                                         // NETSCAPE2.0 (application name for animated gif)
 194                                         if ( $data == 'NETSCAPE2.0' ) {
 195                                                 $data = fread( $fh, 2 ); // Block length and introduction, should be 03 01
 196
 197                                                 if ( $data != "\x03\x01" ) {
 198                                                         throw new Exception( "Expected \x03\x01, got $data" );
 199                                                 }
 200
 201                                                 // Unsigned little-endian integer, loop count or zero for "forever"
 202                                                 $loopData = fread( $fh, 2 );
 203                                                 if ( strlen( $loopData ) < 2 ) {
 204                                                         throw new Exception( "Ran out of input" );
 205                                                 }
 206                                                 $loopData = unpack( 'v', $loopData );
 207                                                 $loopCount = $loopData[1];
 208
 209                                                 if ( $loopCount != 1 ) {
 210                                                         $isLooped = true;
 211                                                 }
 212
 213                                                 // Read out terminator byte
 214                                                 fread( $fh, 1 );
 215                                         } elseif ( $data == 'XMP DataXMP' ) {
 216                                                 // application name for XMP data.
 217                                                 // see pg 18 of XMP spec part 3.
 218
 219                                                 $xmp = self::readBlock( $fh, true );
 220
 221                                                 if ( substr( $xmp, -257, 3 ) !== "\x01\xFF\xFE"
 222                                                         || substr( $xmp, -4 ) !== "\x03\x02\x01\x00"
 223                                                 ) {
 224                                                         // this is just a sanity check.
 225                                                         throw new Exception( "XMP does not have magic trailer!" );
 226                                                 }
 227
 228                                                 // strip out trailer.
 229                                                 $xmp = substr( $xmp, 0, -257 );
 230                                         } else {
 231                                                 // unrecognized extension block
 232                                                 fseek( $fh, -( $blockLength + 1 ), SEEK_CUR );
 233                                                 self::skipBlock( $fh );
 234                                                 continue;
 235                                         }
 236                                 } else {
 237                                         self::skipBlock( $fh );
 238                                 }
 239                         } elseif ( $buf == self::$gif_term ) {
 240                                 break;
 241                         } else {
 242                                 if ( strlen( $buf ) < 1 ) {
 243                                         throw new Exception( "Ran out of input" );
 244                                 }
 245                                 $byte = unpack( 'C', $buf );
 246                                 $byte = $byte[1];
 247                                 throw new Exception( "At position: " . ftell( $fh ) . ", Unknown byte " . $byte );
 248                         }
 249                 }
 250
 251                 return array(
 252                         'frameCount' => $frameCount,
 253                         'looped' => $isLooped,
 254                         'duration' => $duration,
 255                         'xmp' => $xmp,
 256                         'comment' => $comment,
 257                 );
 258         }
 259
 260         /**
 261          * @param $fh
 262          * @param $bpp
 263          * @return void
 264          */
 265         static function readGCT( $fh, $bpp ) {
 266                 if ( $bpp > 0 ) {
 267                         for ( $i = 1; $i <= pow( 2, $bpp ); ++$i ) {
 268                                 fread( $fh, 3 );
 269                         }
 270                 }
 271         }
 272
 273         /**
 274          * @param $data
 275          * @throws Exception
 276          * @return int
 277          */
 278         static function decodeBPP( $data ) {
 279                 if ( strlen( $data ) < 1 ) {
 280                         throw new Exception( "Ran out of input" );
 281                 }
 282                 $buf = unpack( 'C', $data );
 283                 $buf = $buf[1];
 284                 $bpp = ( $buf & 7 ) + 1;
 285                 $buf >>= 7;
 286
 287                 $have_map = $buf & 1;
 288
 289                 return $have_map ? $bpp : 0;
 290         }
 291
 292         /**
 293          * @param $fh
 294          * @throws Exception
 295          */
 296         static function skipBlock( $fh ) {
 297                 while ( !feof( $fh ) ) {
 298                         $buf = fread( $fh, 1 );
 299                         if ( strlen( $buf ) < 1 ) {
 300                                 throw new Exception( "Ran out of input" );
 301                         }
 302                         $block_len = unpack( 'C', $buf );
 303                         $block_len = $block_len[1];
 304                         if ( $block_len == 0 ) {
 305                                 return;
 306                         }
 307                         fread( $fh, $block_len );
 308                 }
 309         }
 310
 311         /**
 312          * Read a block. In the GIF format, a block is made up of
 313          * several sub-blocks. Each sub block starts with one byte
 314          * saying how long the sub-block is, followed by the sub-block.
 315          * The entire block is terminated by a sub-block of length
 316          * 0.
 317          * @param $fh FileHandle
 318          * @param $includeLengths Boolean Include the length bytes of the
 319          *  sub-blocks in the returned value. Normally this is false,
 320          *  except XMP is weird and does a hack where you need to keep
 321          *  these length bytes.
 322          * @throws Exception
 323          * @return string The data.
 324          */
 325         static function readBlock( $fh, $includeLengths = false ) {
 326                 $data = '';
 327                 $subLength = fread( $fh, 1 );
 328                 $blocks = 0;
 329
 330                 while ( $subLength !== "\0" ) {
 331                         $blocks++;
 332                         if ( $blocks > self::MAX_SUBBLOCKS ) {
 333                                 throw new Exception( "MAX_SUBBLOCKS exceeded (over $blocks sub-blocks)" );
 334                         }
 335                         if ( feof( $fh ) ) {
 336                                 throw new Exception( "Read error: Unexpected EOF." );
 337                         }
 338                         if ( $includeLengths ) {
 339                                 $data .= $subLength;
 340                         }
 341
 342                         $data .= fread( $fh, ord( $subLength ) );
 343                         $subLength = fread( $fh, 1 );
 344                 }
 345
 346                 return $data;
 347         }
 348 }