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