includes/media/PNGMetadataExtractor.php

   1 <?php
   2 /**
   3  * PNG frame counter and metadata extractor.
   4  *
   5  * Slightly derived from GIFMetadataExtractor.php
   6  * Deliberately not using MWExceptions to avoid external dependencies, encouraging
   7  * redistribution.
   8  *
   9  * This program is free software; you can redistribute it and/or modify
  10  * it under the terms of the GNU General Public License as published by
  11  * the Free Software Foundation; either version 2 of the License, or
  12  * (at your option) any later version.
  13  *
  14  * This program is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  17  * GNU General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU General Public License along
  20  * with this program; if not, write to the Free Software Foundation, Inc.,
  21  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  22  * http://www.gnu.org/copyleft/gpl.html
  23  *
  24  * @file
  25  * @ingroup Media
  26  */
  27
  28 /**
  29  * PNG frame counter.
  30  *
  31  * @ingroup Media
  32  */
  33 class PNGMetadataExtractor {
  34         /** @var string */
  35         private static $pngSig;
  36
  37         /** @var int */
  38         private static $crcSize;
  39
  40         /** @var array */
  41         private static $textChunks;
  42
  43         const VERSION = 1;
  44         const MAX_CHUNK_SIZE = 3145728; // 3 megabytes
  45
  46         static function getMetadata( $filename ) {
  47                 self::$pngSig = pack( "C8", 137, 80, 78, 71, 13, 10, 26, 10 );
  48                 self::$crcSize = 4;
  49                 /* based on list at http://owl.phy.queensu.ca/~phil/exiftool/TagNames/PNG.html#TextualData
  50                  * and https://www.w3.org/TR/PNG/#11keywords
  51                  */
  52                 self::$textChunks = [
  53                         'xml:com.adobe.xmp' => 'xmp',
  54                         # Artist is unofficial. Author is the recommended
  55                         # keyword in the PNG spec. However some people output
  56                         # Artist so support both.
  57                         'artist' => 'Artist',
  58                         'model' => 'Model',
  59                         'make' => 'Make',
  60                         'author' => 'Artist',
  61                         'comment' => 'PNGFileComment',
  62                         'description' => 'ImageDescription',
  63                         'title' => 'ObjectName',
  64                         'copyright' => 'Copyright',
  65                         # Source as in original device used to make image
  66                         # not as in who gave you the image
  67                         'source' => 'Model',
  68                         'software' => 'Software',
  69                         'disclaimer' => 'Disclaimer',
  70                         'warning' => 'ContentWarning',
  71                         'url' => 'Identifier', # Not sure if this is best mapping. Maybe WebStatement.
  72                         'label' => 'Label',
  73                         'creation time' => 'DateTimeDigitized',
  74                         /* Other potentially useful things - Document */
  75                 ];
  76
  77                 $frameCount = 0;
  78                 $loopCount = 1;
  79                 $text = [];
  80                 $duration = 0.0;
  81                 $bitDepth = 0;
  82                 $colorType = 'unknown';
  83
  84                 if ( !$filename ) {
  85                         throw new Exception( __METHOD__ . ": No file name specified" );
  86                 } elseif ( !file_exists( $filename ) || is_dir( $filename ) ) {
  87                         throw new Exception( __METHOD__ . ": File $filename does not exist" );
  88                 }
  89
  90                 $fh = fopen( $filename, 'rb' );
  91
  92                 if ( !$fh ) {
  93                         throw new Exception( __METHOD__ . ": Unable to open file $filename" );
  94                 }
  95
  96                 // Check for the PNG header
  97                 $buf = fread( $fh, 8 );
  98                 if ( $buf != self::$pngSig ) {
  99                         throw new Exception( __METHOD__ . ": Not a valid PNG file; header: $buf" );
 100                 }
 101
 102                 // Read chunks
 103                 while ( !feof( $fh ) ) {
 104                         $buf = fread( $fh, 4 );
 105                         if ( !$buf || strlen( $buf ) < 4 ) {
 106                                 throw new Exception( __METHOD__ . ": Read error" );
 107                         }
 108                         $chunk_size = unpack( "N", $buf )[1];
 109
 110                         if ( $chunk_size < 0 ) {
 111                                 throw new Exception( __METHOD__ . ": Chunk size too big for unpack" );
 112                         }
 113
 114                         $chunk_type = fread( $fh, 4 );
 115                         if ( !$chunk_type || strlen( $chunk_type ) < 4 ) {
 116                                 throw new Exception( __METHOD__ . ": Read error" );
 117                         }
 118
 119                         if ( $chunk_type == "IHDR" ) {
 120                                 $buf = self::read( $fh, $chunk_size );
 121                                 if ( !$buf || strlen( $buf ) < $chunk_size ) {
 122                                         throw new Exception( __METHOD__ . ": Read error" );
 123                                 }
 124                                 $bitDepth = ord( substr( $buf, 8, 1 ) );
 125                                 // Detect the color type in British English as per the spec
 126                                 // https://www.w3.org/TR/PNG/#11IHDR
 127                                 switch ( ord( substr( $buf, 9, 1 ) ) ) {
 128                                         case 0:
 129                                                 $colorType = 'greyscale';
 130                                                 break;
 131                                         case 2:
 132                                                 $colorType = 'truecolour';
 133                                                 break;
 134                                         case 3:
 135                                                 $colorType = 'index-coloured';
 136                                                 break;
 137                                         case 4:
 138                                                 $colorType = 'greyscale-alpha';
 139                                                 break;
 140                                         case 6:
 141                                                 $colorType = 'truecolour-alpha';
 142                                                 break;
 143                                         default:
 144                                                 $colorType = 'unknown';
 145                                                 break;
 146                                 }
 147                         } elseif ( $chunk_type == "acTL" ) {
 148                                 $buf = fread( $fh, $chunk_size );
 149                                 if ( !$buf || strlen( $buf ) < $chunk_size || $chunk_size < 4 ) {
 150                                         throw new Exception( __METHOD__ . ": Read error" );
 151                                 }
 152
 153                                 $actl = unpack( "Nframes/Nplays", $buf );
 154                                 $frameCount = $actl['frames'];
 155                                 $loopCount = $actl['plays'];
 156                         } elseif ( $chunk_type == "fcTL" ) {
 157                                 $buf = self::read( $fh, $chunk_size );
 158                                 if ( !$buf || strlen( $buf ) < $chunk_size ) {
 159                                         throw new Exception( __METHOD__ . ": Read error" );
 160                                 }
 161                                 $buf = substr( $buf, 20 );
 162                                 if ( strlen( $buf ) < 4 ) {
 163                                         throw new Exception( __METHOD__ . ": Read error" );
 164                                 }
 165
 166                                 $fctldur = unpack( "ndelay_num/ndelay_den", $buf );
 167                                 if ( $fctldur['delay_den'] == 0 ) {
 168                                         $fctldur['delay_den'] = 100;
 169                                 }
 170                                 if ( $fctldur['delay_num'] ) {
 171                                         $duration += $fctldur['delay_num'] / $fctldur['delay_den'];
 172                                 }
 173                         } elseif ( $chunk_type == "iTXt" ) {
 174                                 // Extracts iTXt chunks, uncompressing if necessary.
 175                                 $buf = self::read( $fh, $chunk_size );
 176                                 $items = [];
 177                                 if ( preg_match(
 178                                         '/^([^\x00]{1,79})\x00(\x00|\x01)\x00([^\x00]*)(.)[^\x00]*\x00(.*)$/Ds',
 179                                         $buf, $items )
 180                                 ) {
 181                                         /* $items[1] = text chunk name, $items[2] = compressed flag,
 182                                          * $items[3] = lang code (or ""), $items[4]= compression type.
 183                                          * $items[5] = content
 184                                          */
 185
 186                                         // Theoretically should be case-sensitive, but in practise...
 187                                         $items[1] = strtolower( $items[1] );
 188                                         if ( !isset( self::$textChunks[$items[1]] ) ) {
 189                                                 // Only extract textual chunks on our list.
 190                                                 fseek( $fh, self::$crcSize, SEEK_CUR );
 191                                                 continue;
 192                                         }
 193
 194                                         $items[3] = strtolower( $items[3] );
 195                                         if ( $items[3] == '' ) {
 196                                                 // if no lang specified use x-default like in xmp.
 197                                                 $items[3] = 'x-default';
 198                                         }
 199
 200                                         // if compressed
 201                                         if ( $items[2] == "\x01" ) {
 202                                                 if ( function_exists( 'gzuncompress' ) && $items[4] === "\x00" ) {
 203                                                         MediaWiki\suppressWarnings();
 204                                                         $items[5] = gzuncompress( $items[5] );
 205                                                         MediaWiki\restoreWarnings();
 206
 207                                                         if ( $items[5] === false ) {
 208                                                                 // decompression failed
 209                                                                 wfDebug( __METHOD__ . ' Error decompressing iTxt chunk - ' . $items[1] . "\n" );
 210                                                                 fseek( $fh, self::$crcSize, SEEK_CUR );
 211                                                                 continue;
 212                                                         }
 213                                                 } else {
 214                                                         wfDebug( __METHOD__ . ' Skipping compressed png iTXt chunk due to lack of zlib,'
 215                                                                 . " or potentially invalid compression method\n" );
 216                                                         fseek( $fh, self::$crcSize, SEEK_CUR );
 217                                                         continue;
 218                                                 }
 219                                         }
 220                                         $finalKeyword = self::$textChunks[$items[1]];
 221                                         $text[$finalKeyword][$items[3]] = $items[5];
 222                                         $text[$finalKeyword]['_type'] = 'lang';
 223                                 } else {
 224                                         // Error reading iTXt chunk
 225                                         throw new Exception( __METHOD__ . ": Read error on iTXt chunk" );
 226                                 }
 227                         } elseif ( $chunk_type == 'tEXt' ) {
 228                                 $buf = self::read( $fh, $chunk_size );
 229
 230                                 // In case there is no \x00 which will make explode fail.
 231                                 if ( strpos( $buf, "\x00" ) === false ) {
 232                                         throw new Exception( __METHOD__ . ": Read error on tEXt chunk" );
 233                                 }
 234
 235                                 list( $keyword, $content ) = explode( "\x00", $buf, 2 );
 236                                 if ( $keyword === '' || $content === '' ) {
 237                                         throw new Exception( __METHOD__ . ": Read error on tEXt chunk" );
 238                                 }
 239
 240                                 // Theoretically should be case-sensitive, but in practise...
 241                                 $keyword = strtolower( $keyword );
 242                                 if ( !isset( self::$textChunks[$keyword] ) ) {
 243                                         // Don't recognize chunk, so skip.
 244                                         fseek( $fh, self::$crcSize, SEEK_CUR );
 245                                         continue;
 246                                 }
 247                                 MediaWiki\suppressWarnings();
 248                                 $content = iconv( 'ISO-8859-1', 'UTF-8', $content );
 249                                 MediaWiki\restoreWarnings();
 250
 251                                 if ( $content === false ) {
 252                                         throw new Exception( __METHOD__ . ": Read error (error with iconv)" );
 253                                 }
 254
 255                                 $finalKeyword = self::$textChunks[$keyword];
 256                                 $text[$finalKeyword]['x-default'] = $content;
 257                                 $text[$finalKeyword]['_type'] = 'lang';
 258                         } elseif ( $chunk_type == 'zTXt' ) {
 259                                 if ( function_exists( 'gzuncompress' ) ) {
 260                                         $buf = self::read( $fh, $chunk_size );
 261
 262                                         // In case there is no \x00 which will make explode fail.
 263                                         if ( strpos( $buf, "\x00" ) === false ) {
 264                                                 throw new Exception( __METHOD__ . ": Read error on zTXt chunk" );
 265                                         }
 266
 267                                         list( $keyword, $postKeyword ) = explode( "\x00", $buf, 2 );
 268                                         if ( $keyword === '' || $postKeyword === '' ) {
 269                                                 throw new Exception( __METHOD__ . ": Read error on zTXt chunk" );
 270                                         }
 271                                         // Theoretically should be case-sensitive, but in practise...
 272                                         $keyword = strtolower( $keyword );
 273
 274                                         if ( !isset( self::$textChunks[$keyword] ) ) {
 275                                                 // Don't recognize chunk, so skip.
 276                                                 fseek( $fh, self::$crcSize, SEEK_CUR );
 277                                                 continue;
 278                                         }
 279                                         $compression = substr( $postKeyword, 0, 1 );
 280                                         $content = substr( $postKeyword, 1 );
 281                                         if ( $compression !== "\x00" ) {
 282                                                 wfDebug( __METHOD__ . " Unrecognized compression method in zTXt ($keyword). Skipping.\n" );
 283                                                 fseek( $fh, self::$crcSize, SEEK_CUR );
 284                                                 continue;
 285                                         }
 286
 287                                         MediaWiki\suppressWarnings();
 288                                         $content = gzuncompress( $content );
 289                                         MediaWiki\restoreWarnings();
 290
 291                                         if ( $content === false ) {
 292                                                 // decompression failed
 293                                                 wfDebug( __METHOD__ . ' Error decompressing zTXt chunk - ' . $keyword . "\n" );
 294                                                 fseek( $fh, self::$crcSize, SEEK_CUR );
 295                                                 continue;
 296                                         }
 297
 298                                         MediaWiki\suppressWarnings();
 299                                         $content = iconv( 'ISO-8859-1', 'UTF-8', $content );
 300                                         MediaWiki\restoreWarnings();
 301
 302                                         if ( $content === false ) {
 303                                                 throw new Exception( __METHOD__ . ": Read error (error with iconv)" );
 304                                         }
 305
 306                                         $finalKeyword = self::$textChunks[$keyword];
 307                                         $text[$finalKeyword]['x-default'] = $content;
 308                                         $text[$finalKeyword]['_type'] = 'lang';
 309                                 } else {
 310                                         wfDebug( __METHOD__ . " Cannot decompress zTXt chunk due to lack of zlib. Skipping.\n" );
 311                                         fseek( $fh, $chunk_size, SEEK_CUR );
 312                                 }
 313                         } elseif ( $chunk_type == 'tIME' ) {
 314                                 // last mod timestamp.
 315                                 if ( $chunk_size !== 7 ) {
 316                                         throw new Exception( __METHOD__ . ": tIME wrong size" );
 317                                 }
 318                                 $buf = self::read( $fh, $chunk_size );
 319                                 if ( !$buf || strlen( $buf ) < $chunk_size ) {
 320                                         throw new Exception( __METHOD__ . ": Read error" );
 321                                 }
 322
 323                                 // Note: spec says this should be UTC.
 324                                 $t = unpack( "ny/Cm/Cd/Ch/Cmin/Cs", $buf );
 325                                 $strTime = sprintf( "%04d%02d%02d%02d%02d%02d",
 326                                         $t['y'], $t['m'], $t['d'], $t['h'],
 327                                         $t['min'], $t['s'] );
 328
 329                                 $exifTime = wfTimestamp( TS_EXIF, $strTime );
 330
 331                                 if ( $exifTime ) {
 332                                         $text['DateTime'] = $exifTime;
 333                                 }
 334                         } elseif ( $chunk_type == 'pHYs' ) {
 335                                 // how big pixels are (dots per meter).
 336                                 if ( $chunk_size !== 9 ) {
 337                                         throw new Exception( __METHOD__ . ": pHYs wrong size" );
 338                                 }
 339
 340                                 $buf = self::read( $fh, $chunk_size );
 341                                 if ( !$buf || strlen( $buf ) < $chunk_size ) {
 342                                         throw new Exception( __METHOD__ . ": Read error" );
 343                                 }
 344
 345                                 $dim = unpack( "Nwidth/Nheight/Cunit", $buf );
 346                                 if ( $dim['unit'] == 1 ) {
 347                                         // Need to check for negative because php
 348                                         // doesn't deal with super-large unsigned 32-bit ints well
 349                                         if ( $dim['width'] > 0 && $dim['height'] > 0 ) {
 350                                                 // unit is meters
 351                                                 // (as opposed to 0 = undefined )
 352                                                 $text['XResolution'] = $dim['width']
 353                                                         . '/100';
 354                                                 $text['YResolution'] = $dim['height']
 355                                                         . '/100';
 356                                                 $text['ResolutionUnit'] = 3;
 357                                                 // 3 = dots per cm (from Exif).
 358                                         }
 359                                 }
 360                         } elseif ( $chunk_type == "IEND" ) {
 361                                 break;
 362                         } else {
 363                                 fseek( $fh, $chunk_size, SEEK_CUR );
 364                         }
 365                         fseek( $fh, self::$crcSize, SEEK_CUR );
 366                 }
 367                 fclose( $fh );
 368
 369                 if ( $loopCount > 1 ) {
 370                         $duration *= $loopCount;
 371                 }
 372
 373                 if ( isset( $text['DateTimeDigitized'] ) ) {
 374                         // Convert date format from rfc2822 to exif.
 375                         foreach ( $text['DateTimeDigitized'] as $name => &$value ) {
 376                                 if ( $name === '_type' ) {
 377                                         continue;
 378                                 }
 379
 380                                 // @todo FIXME: Currently timezones are ignored.
 381                                 // possibly should be wfTimestamp's
 382                                 // responsibility. (at least for numeric TZ)
 383                                 $formatted = wfTimestamp( TS_EXIF, $value );
 384                                 if ( $formatted ) {
 385                                         // Only change if we could convert the
 386                                         // date.
 387                                         // The png standard says it should be
 388                                         // in rfc2822 format, but not required.
 389                                         // In general for the exif stuff we
 390                                         // prettify the date if we can, but we
 391                                         // display as-is if we cannot or if
 392                                         // it is invalid.
 393                                         // So do the same here.
 394
 395                                         $value = $formatted;
 396                                 }
 397                         }
 398                 }
 399
 400                 return [
 401                         'frameCount' => $frameCount,
 402                         'loopCount' => $loopCount,
 403                         'duration' => $duration,
 404                         'text' => $text,
 405                         'bitDepth' => $bitDepth,
 406                         'colorType' => $colorType,
 407                 ];
 408         }
 409
 410         /**
 411          * Read a chunk, checking to make sure its not too big.
 412          *
 413          * @param resource $fh The file handle
 414          * @param int $size Size in bytes.
 415          * @throws Exception If too big
 416          * @return string The chunk.
 417          */
 418         private static function read( $fh, $size ) {
 419                 if ( $size > self::MAX_CHUNK_SIZE ) {
 420                         throw new Exception( __METHOD__ . ': Chunk size of ' . $size .
 421                                 ' too big. Max size is: ' . self::MAX_CHUNK_SIZE );
 422                 }
 423
 424                 return fread( $fh, $size );
 425         }
 426 }