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         /** @var string */
  36         private static $gifFrameSep;
  37
  38         /** @var string */
  39         private static $gifExtensionSep;
  40
  41         /** @var string */
  42         private static $gifTerm;
  43
  44         const VERSION = 1;
  45
  46         // Each sub-block is less than or equal to 255 bytes.
  47         // Most of the time its 255 bytes, except for in XMP
  48         // blocks, where it's usually between 32-127 bytes each.
  49         const MAX_SUBBLOCKS = 262144; // 5mb divided by 20.
  50
  51         /**
  52          * @throws Exception
  53          * @param string $filename
  54          * @return array
  55          */
  56         static function getMetadata( $filename ) {
  57                 self::$gifFrameSep = pack( "C", ord( "," ) ); // 2C
  58                 self::$gifExtensionSep = pack( "C", ord( "!" ) ); // 21
  59                 self::$gifTerm = pack( "C", ord( ";" ) ); // 3B
  60
  61                 $frameCount = 0;
  62                 $duration = 0.0;
  63                 $isLooped = false;
  64                 $xmp = "";
  65                 $comment = [];
  66
  67                 if ( !$filename ) {
  68                         throw new Exception( "No file name specified" );
  69                 } elseif ( !file_exists( $filename ) || is_dir( $filename ) ) {
  70                         throw new Exception( "File $filename does not exist" );
  71                 }
  72
  73                 $fh = fopen( $filename, 'rb' );
  74
  75                 if ( !$fh ) {
  76                         throw new Exception( "Unable to open file $filename" );
  77                 }
  78
  79                 // Check for the GIF header
  80                 $buf = fread( $fh, 6 );
  81                 if ( !( $buf == 'GIF87a' || $buf == 'GIF89a' ) ) {
  82                         throw new Exception( "Not a valid GIF file; header: $buf" );
  83                 }
  84
  85                 // Read width and height.
  86                 $buf = fread( $fh, 2 );
  87                 $width = unpack( 'v', $buf )[1];
  88                 $buf = fread( $fh, 2 );
  89                 $height = unpack( 'v', $buf )[1];
  90
  91                 // Read BPP
  92                 $buf = fread( $fh, 1 );
  93                 $bpp = self::decodeBPP( $buf );
  94
  95                 // Skip over background and aspect ratio
  96                 fread( $fh, 2 );
  97
  98                 // Skip over the GCT
  99                 self::readGCT( $fh, $bpp );
 100
 101                 while ( !feof( $fh ) ) {
 102                         $buf = fread( $fh, 1 );
 103
 104                         if ( $buf == self::$gifFrameSep ) {
 105                                 // Found a frame
 106                                 $frameCount++;
 107
 108                                 # # Skip bounding box
 109                                 fread( $fh, 8 );
 110
 111                                 # # Read BPP
 112                                 $buf = fread( $fh, 1 );
 113                                 $bpp = self::decodeBPP( $buf );
 114
 115                                 # # Read GCT
 116                                 self::readGCT( $fh, $bpp );
 117                                 fread( $fh, 1 );
 118                                 self::skipBlock( $fh );
 119                         } elseif ( $buf == self::$gifExtensionSep ) {
 120                                 $buf = fread( $fh, 1 );
 121                                 if ( strlen( $buf ) < 1 ) {
 122                                         throw new Exception( "Ran out of input" );
 123                                 }
 124                                 $extension_code = unpack( 'C', $buf )[1];
 125
 126                                 if ( $extension_code == 0xF9 ) {
 127                                         // Graphics Control Extension.
 128                                         fread( $fh, 1 ); // Block size
 129
 130                                         fread( $fh, 1 ); // Transparency, disposal method, user input
 131
 132                                         $buf = fread( $fh, 2 ); // Delay, in hundredths of seconds.
 133                                         if ( strlen( $buf ) < 2 ) {
 134                                                 throw new Exception( "Ran out of input" );
 135                                         }
 136                                         $delay = unpack( 'v', $buf )[1];
 137                                         $duration += $delay * 0.01;
 138
 139                                         fread( $fh, 1 ); // Transparent colour index
 140
 141                                         $term = fread( $fh, 1 ); // Should be a terminator
 142                                         if ( strlen( $term ) < 1 ) {
 143                                                 throw new Exception( "Ran out of input" );
 144                                         }
 145                                         $term = unpack( 'C', $term )[1];
 146                                         if ( $term != 0 ) {
 147                                                 throw new Exception( "Malformed Graphics Control Extension block" );
 148                                         }
 149                                 } elseif ( $extension_code == 0xFE ) {
 150                                         // Comment block(s).
 151                                         $data = self::readBlock( $fh );
 152                                         if ( $data === "" ) {
 153                                                 throw new Exception( 'Read error, zero-length comment block' );
 154                                         }
 155
 156                                         // The standard says this should be ASCII, however its unclear if
 157                                         // thats true in practise. Check to see if its valid utf-8, if so
 158                                         // assume its that, otherwise assume its windows-1252 (iso-8859-1)
 159                                         $dataCopy = $data;
 160                                         // quickIsNFCVerify has the side effect of replacing any invalid characters
 161                                         UtfNormal\Validator::quickIsNFCVerify( $dataCopy );
 162
 163                                         if ( $dataCopy !== $data ) {
 164                                                 Wikimedia\suppressWarnings();
 165                                                 $data = iconv( 'windows-1252', 'UTF-8', $data );
 166                                                 Wikimedia\restoreWarnings();
 167                                         }
 168
 169                                         $commentCount = count( $comment );
 170                                         if ( $commentCount === 0
 171                                                 || $comment[$commentCount - 1] !== $data
 172                                         ) {
 173                                                 // Some applications repeat the same comment on each
 174                                                 // frame of an animated GIF image, so if this comment
 175                                                 // is identical to the last, only extract once.
 176                                                 $comment[] = $data;
 177                                         }
 178                                 } elseif ( $extension_code == 0xFF ) {
 179                                         // Application extension (Netscape info about the animated gif)
 180                                         // or XMP (or theoretically any other type of extension block)
 181                                         $blockLength = fread( $fh, 1 );
 182                                         if ( strlen( $blockLength ) < 1 ) {
 183                                                 throw new Exception( "Ran out of input" );
 184                                         }
 185                                         $blockLength = unpack( 'C', $blockLength )[1];
 186                                         $data = fread( $fh, $blockLength );
 187
 188                                         if ( $blockLength != 11 ) {
 189                                                 wfDebug( __METHOD__ . " GIF application block with wrong length\n" );
 190                                                 fseek( $fh, -( $blockLength + 1 ), SEEK_CUR );
 191                                                 self::skipBlock( $fh );
 192                                                 continue;
 193                                         }
 194
 195                                         // NETSCAPE2.0 (application name for animated gif)
 196                                         if ( $data == 'NETSCAPE2.0' ) {
 197                                                 $data = fread( $fh, 2 ); // Block length and introduction, should be 03 01
 198
 199                                                 if ( $data != "\x03\x01" ) {
 200                                                         throw new Exception( "Expected \x03\x01, got $data" );
 201                                                 }
 202
 203                                                 // Unsigned little-endian integer, loop count or zero for "forever"
 204                                                 $loopData = fread( $fh, 2 );
 205                                                 if ( strlen( $loopData ) < 2 ) {
 206                                                         throw new Exception( "Ran out of input" );
 207                                                 }
 208                                                 $loopCount = unpack( 'v', $loopData )[1];
 209
 210                                                 if ( $loopCount != 1 ) {
 211                                                         $isLooped = true;
 212                                                 }
 213
 214                                                 // Read out terminator byte
 215                                                 fread( $fh, 1 );
 216                                         } elseif ( $data == 'XMP DataXMP' ) {
 217                                                 // application name for XMP data.
 218                                                 // see pg 18 of XMP spec part 3.
 219
 220                                                 $xmp = self::readBlock( $fh, true );
 221
 222                                                 if ( substr( $xmp, -257, 3 ) !== "\x01\xFF\xFE"
 223                                                         || substr( $xmp, -4 ) !== "\x03\x02\x01\x00"
 224                                                 ) {
 225                                                         // this is just a sanity check.
 226                                                         throw new Exception( "XMP does not have magic trailer!" );
 227                                                 }
 228
 229                                                 // strip out trailer.
 230                                                 $xmp = substr( $xmp, 0, -257 );
 231                                         } else {
 232                                                 // unrecognized extension block
 233                                                 fseek( $fh, -( $blockLength + 1 ), SEEK_CUR );
 234                                                 self::skipBlock( $fh );
 235                                                 continue;
 236                                         }
 237                                 } else {
 238                                         self::skipBlock( $fh );
 239                                 }
 240                         } elseif ( $buf == self::$gifTerm ) {
 241                                 break;
 242                         } else {
 243                                 if ( strlen( $buf ) < 1 ) {
 244                                         throw new Exception( "Ran out of input" );
 245                                 }
 246                                 $byte = unpack( 'C', $buf )[1];
 247                                 throw new Exception( "At position: " . ftell( $fh ) . ", Unknown byte " . $byte );
 248                         }
 249                 }
 250
 251                 return [
 252                         'frameCount' => $frameCount,
 253                         'looped' => $isLooped,
 254                         'duration' => $duration,
 255                         'xmp' => $xmp,
 256                         'comment' => $comment,
 257                 ];
 258         }
 259
 260         /**
 261          * @param resource $fh
 262          * @param int $bpp
 263          * @return void
 264          */
 265         static function readGCT( $fh, $bpp ) {
 266                 if ( $bpp > 0 ) {
 267                         $max = pow( 2, $bpp );
 268                         for ( $i = 1; $i <= $max; ++$i ) {
 269                                 fread( $fh, 3 );
 270                         }
 271                 }
 272         }
 273
 274         /**
 275          * @param string $data
 276          * @throws Exception
 277          * @return int
 278          */
 279         static function decodeBPP( $data ) {
 280                 if ( strlen( $data ) < 1 ) {
 281                         throw new Exception( "Ran out of input" );
 282                 }
 283                 $buf = unpack( 'C', $data )[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 resource $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 )[1];
 303                         if ( $block_len == 0 ) {
 304                                 return;
 305                         }
 306                         fread( $fh, $block_len );
 307                 }
 308         }
 309
 310         /**
 311          * Read a block. In the GIF format, a block is made up of
 312          * several sub-blocks. Each sub block starts with one byte
 313          * saying how long the sub-block is, followed by the sub-block.
 314          * The entire block is terminated by a sub-block of length
 315          * 0.
 316          * @param resource $fh File handle
 317          * @param bool $includeLengths Include the length bytes of the
 318          *  sub-blocks in the returned value. Normally this is false,
 319          *  except XMP is weird and does a hack where you need to keep
 320          *  these length bytes.
 321          * @throws Exception
 322          * @return string The data.
 323          */
 324         static function readBlock( $fh, $includeLengths = false ) {
 325                 $data = '';
 326                 $subLength = fread( $fh, 1 );
 327                 $blocks = 0;
 328
 329                 while ( $subLength !== "\0" ) {
 330                         $blocks++;
 331                         if ( $blocks > self::MAX_SUBBLOCKS ) {
 332                                 throw new Exception( "MAX_SUBBLOCKS exceeded (over $blocks sub-blocks)" );
 333                         }
 334                         if ( feof( $fh ) ) {
 335                                 throw new Exception( "Read error: Unexpected EOF." );
 336                         }
 337                         if ( $includeLengths ) {
 338                                 $data .= $subLength;
 339                         }
 340
 341                         $data .= fread( $fh, ord( $subLength ) );
 342                         $subLength = fread( $fh, 1 );
 343                 }
 344
 345                 return $data;
 346         }
 347 }