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         static $CRC_size;
  36         static $text_chunks;
  37
  38         const VERSION = 1;
  39         const MAX_CHUNK_SIZE = 3145728; // 3 megabytes
  40
  41         static function getMetadata( $filename ) {
  42                 self::$png_sig = pack( "C8", 137, 80, 78, 71, 13, 10, 26, 10 );
  43                 self::$CRC_size = 4;
  44                 /* based on list at http://owl.phy.queensu.ca/~phil/exiftool/TagNames/PNG.html#TextualData
  45                  * and http://www.w3.org/TR/PNG/#11keywords
  46                  */
  47                 self::$text_chunks = array(
  48                         'xml:com.adobe.xmp' => 'xmp',
  49                         # Artist is unofficial. Author is the recommended
  50                         # keyword in the PNG spec. However some people output
  51                         # Artist so support both.
  52                         'artist'      => 'Artist',
  53                         'model'       => 'Model',
  54                         'make'        => 'Make',
  55                         'author'      => 'Artist',
  56                         'comment'     => 'PNGFileComment',
  57                         'description' => 'ImageDescription',
  58                         'title'       => 'ObjectName',
  59                         'copyright'   => 'Copyright',
  60                         # Source as in original device used to make image
  61                         # not as in who gave you the image
  62                         'source'      => 'Model',
  63                         'software'    => 'Software',
  64                         'disclaimer'  => 'Disclaimer',
  65                         'warning'     => 'ContentWarning',
  66                         'url'         => 'Identifier', # Not sure if this is best mapping. Maybe WebStatement.
  67                         'label'       => 'Label',
  68                         'creation time' => 'DateTimeDigitized',
  69                         /* Other potentially useful things - Document */
  70                 );
  71
  72                 $frameCount = 0;
  73                 $loopCount = 1;
  74                 $text = array();
  75                 $duration = 0.0;
  76                 $bitDepth = 0;
  77                 $colorType = 'unknown';
  78
  79                 if ( !$filename ) {
  80                         throw new Exception( __METHOD__ . ": No file name specified" );
  81                 } elseif ( !file_exists( $filename ) || is_dir( $filename ) ) {
  82                         throw new Exception( __METHOD__ . ": File $filename does not exist" );
  83                 }
  84
  85                 $fh = fopen( $filename, 'rb' );
  86
  87                 if ( !$fh ) {
  88                         throw new Exception( __METHOD__ . ": Unable to open file $filename" );
  89                 }
  90
  91                 // Check for the PNG header
  92                 $buf = fread( $fh, 8 );
  93                 if ( $buf != self::$png_sig ) {
  94                         throw new Exception( __METHOD__ . ": Not a valid PNG file; header: $buf" );
  95                 }
  96
  97                 // Read chunks
  98                 while ( !feof( $fh ) ) {
  99                         $buf = fread( $fh, 4 );
 100                         if ( !$buf || strlen( $buf ) < 4 ) {
 101                                 throw new Exception( __METHOD__ . ": Read error" );
 102                         }
 103                         $chunk_size = unpack( "N", $buf );
 104                         $chunk_size = $chunk_size[1];
 105
 106                         if ( $chunk_size < 0 ) {
 107                                 throw new Exception( __METHOD__ . ": Chunk size too big for unpack" );
 108                         }
 109
 110                         $chunk_type = fread( $fh, 4 );
 111                         if ( !$chunk_type || strlen( $chunk_type ) < 4 ) {
 112                                 throw new Exception( __METHOD__ . ": Read error" );
 113                         }
 114
 115                         if ( $chunk_type == "IHDR" ) {
 116                                 $buf = self::read( $fh, $chunk_size );
 117                                 if ( !$buf || strlen( $buf ) < $chunk_size ) {
 118                                         throw new Exception( __METHOD__ . ": Read error" );
 119                                 }
 120                                 $bitDepth = ord( substr( $buf, 8, 1 ) );
 121                                 // Detect the color type in British English as per the spec
 122                                 // http://www.w3.org/TR/PNG/#11IHDR
 123                                 switch ( ord( substr( $buf, 9, 1 ) ) ) {
 124                                         case 0:
 125                                                 $colorType = 'greyscale';
 126                                                 break;
 127                                         case 2:
 128                                                 $colorType = 'truecolour';
 129                                                 break;
 130                                         case 3:
 131                                                 $colorType = 'index-coloured';
 132                                                 break;
 133                                         case 4:
 134                                                 $colorType = 'greyscale-alpha';
 135                                                 break;
 136                                         case 6:
 137                                                 $colorType = 'truecolour-alpha';
 138                                                 break;
 139                                         default:
 140                                                 $colorType = 'unknown';
 141                                                 break;
 142                                 }
 143                         } elseif ( $chunk_type == "acTL" ) {
 144                                 $buf = fread( $fh, $chunk_size );
 145                                 if ( !$buf || strlen( $buf ) < $chunk_size || $chunk_size < 4 ) {
 146                                         throw new Exception( __METHOD__ . ": Read error" );
 147                                 }
 148
 149                                 $actl = unpack( "Nframes/Nplays", $buf );
 150                                 $frameCount = $actl['frames'];
 151                                 $loopCount = $actl['plays'];
 152                         } elseif ( $chunk_type == "fcTL" ) {
 153                                 $buf = self::read( $fh, $chunk_size );
 154                                 if ( !$buf || strlen( $buf ) < $chunk_size ) {
 155                                         throw new Exception( __METHOD__ . ": Read error" );
 156                                 }
 157                                 $buf = substr( $buf, 20 );
 158                                 if ( strlen( $buf ) < 4 ) {
 159                                         throw new Exception( __METHOD__ . ": Read error" );
 160                                 }
 161
 162                                 $fctldur = unpack( "ndelay_num/ndelay_den", $buf );
 163                                 if ( $fctldur['delay_den'] == 0 ) {
 164                                         $fctldur['delay_den'] = 100;
 165                                 }
 166                                 if ( $fctldur['delay_num'] ) {
 167                                         $duration += $fctldur['delay_num'] / $fctldur['delay_den'];
 168                                 }
 169                         } elseif ( $chunk_type == "iTXt" ) {
 170                                 // Extracts iTXt chunks, uncompressing if necessary.
 171                                 $buf = self::read( $fh, $chunk_size );
 172                                 $items = array();
 173                                 if ( preg_match(
 174                                         '/^([^\x00]{1,79})\x00(\x00|\x01)\x00([^\x00]*)(.)[^\x00]*\x00(.*)$/Ds',
 175                                         $buf, $items )
 176                                 ) {
 177                                         /* $items[1] = text chunk name, $items[2] = compressed flag,
 178                                          * $items[3] = lang code (or ""), $items[4]= compression type.
 179                                          * $items[5] = content
 180                                          */
 181
 182                                         // Theoretically should be case-sensitive, but in practise...
 183                                         $items[1] = strtolower( $items[1] );
 184                                         if ( !isset( self::$text_chunks[$items[1]] ) ) {
 185                                                 // Only extract textual chunks on our list.
 186                                                 fseek( $fh, self::$CRC_size, SEEK_CUR );
 187                                                 continue;
 188                                         }
 189
 190                                         $items[3] = strtolower( $items[3] );
 191                                         if ( $items[3] == '' ) {
 192                                                 // if no lang specified use x-default like in xmp.
 193                                                 $items[3] = 'x-default';
 194                                         }
 195
 196                                         // if compressed
 197                                         if ( $items[2] == "\x01" ) {
 198                                                 if ( function_exists( 'gzuncompress' ) && $items[4] === "\x00" ) {
 199                                                         wfSuppressWarnings();
 200                                                         $items[5] = gzuncompress( $items[5] );
 201                                                         wfRestoreWarnings();
 202
 203                                                         if ( $items[5] === false ) {
 204                                                                 // decompression failed
 205                                                                 wfDebug( __METHOD__ . ' Error decompressing iTxt chunk - ' . $items[1] );
 206                                                                 fseek( $fh, self::$CRC_size, SEEK_CUR );
 207                                                                 continue;
 208                                                         }
 209
 210                                                 } else {
 211                                                         wfDebug( __METHOD__ . ' Skipping compressed png iTXt chunk due to lack of zlib,'
 212                                                                 . ' or potentially invalid compression method' );
 213                                                         fseek( $fh, self::$CRC_size, SEEK_CUR );
 214                                                         continue;
 215                                                 }
 216                                         }
 217                                         $finalKeyword = self::$text_chunks[$items[1]];
 218                                         $text[$finalKeyword][$items[3]] = $items[5];
 219                                         $text[$finalKeyword]['_type'] = 'lang';
 220
 221                                 } else {
 222                                         // Error reading iTXt chunk
 223                                         throw new Exception( __METHOD__ . ": Read error on iTXt chunk" );
 224                                 }
 225
 226                         } elseif ( $chunk_type == 'tEXt' ) {
 227                                 $buf = self::read( $fh, $chunk_size );
 228
 229                                 // In case there is no \x00 which will make explode fail.
 230                                 if ( strpos( $buf, "\x00" ) === false ) {
 231                                         throw new Exception( __METHOD__ . ": Read error on tEXt chunk" );
 232                                 }
 233
 234                                 list( $keyword, $content ) = explode( "\x00", $buf, 2 );
 235                                 if ( $keyword === '' || $content === '' ) {
 236                                         throw new Exception( __METHOD__ . ": Read error on tEXt chunk" );
 237                                 }
 238
 239                                 // Theoretically should be case-sensitive, but in practise...
 240                                 $keyword = strtolower( $keyword );
 241                                 if ( !isset( self::$text_chunks[ $keyword ] ) ) {
 242                                         // Don't recognize chunk, so skip.
 243                                         fseek( $fh, self::$CRC_size, SEEK_CUR );
 244                                         continue;
 245                                 }
 246                                 wfSuppressWarnings();
 247                                 $content = iconv( 'ISO-8859-1', 'UTF-8', $content );
 248                                 wfRestoreWarnings();
 249
 250                                 if ( $content === false ) {
 251                                         throw new Exception( __METHOD__ . ": Read error (error with iconv)" );
 252                                 }
 253
 254                                 $finalKeyword = self::$text_chunks[$keyword];
 255                                 $text[$finalKeyword]['x-default'] = $content;
 256                                 $text[$finalKeyword]['_type'] = 'lang';
 257
 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( $keyword );
 273
 274                                         if ( !isset( self::$text_chunks[ $keyword ] ) ) {
 275                                                 // Don't recognize chunk, so skip.
 276                                                 fseek( $fh, self::$CRC_size, 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." );
 283                                                 fseek( $fh, self::$CRC_size, SEEK_CUR );
 284                                                 continue;
 285                                         }
 286
 287                                         wfSuppressWarnings();
 288                                         $content = gzuncompress( $content );
 289                                         wfRestoreWarnings();
 290
 291                                         if ( $content === false ) {
 292                                                 // decompression failed
 293                                                 wfDebug( __METHOD__ . ' Error decompressing zTXt chunk - ' . $keyword );
 294                                                 fseek( $fh, self::$CRC_size, SEEK_CUR );
 295                                                 continue;
 296                                         }
 297
 298                                         wfSuppressWarnings();
 299                                         $content = iconv( 'ISO-8859-1', 'UTF-8', $content );
 300                                         wfRestoreWarnings();
 301
 302                                         if ( $content === false ) {
 303                                                 throw new Exception( __METHOD__ . ": Read error (error with iconv)" );
 304                                         }
 305
 306                                         $finalKeyword = self::$text_chunks[$keyword];
 307                                         $text[$finalKeyword]['x-default'] = $content;
 308                                         $text[$finalKeyword]['_type'] = 'lang';
 309
 310                                 } else {
 311                                         wfDebug( __METHOD__ . " Cannot decompress zTXt chunk due to lack of zlib. Skipping." );
 312                                         fseek( $fh, $chunk_size, SEEK_CUR );
 313                                 }
 314                         } elseif ( $chunk_type == 'tIME' ) {
 315                                 // last mod timestamp.
 316                                 if ( $chunk_size !== 7 ) {
 317                                         throw new Exception( __METHOD__ . ": tIME wrong size" );
 318                                 }
 319                                 $buf = self::read( $fh, $chunk_size );
 320                                 if ( !$buf || strlen( $buf ) < $chunk_size ) {
 321                                         throw new Exception( __METHOD__ . ": Read error" );
 322                                 }
 323
 324                                 // Note: spec says this should be UTC.
 325                                 $t = unpack( "ny/Cm/Cd/Ch/Cmin/Cs", $buf );
 326                                 $strTime = sprintf( "%04d%02d%02d%02d%02d%02d",
 327                                         $t['y'], $t['m'], $t['d'], $t['h'],
 328                                         $t['min'], $t['s'] );
 329
 330                                 $exifTime = wfTimestamp( TS_EXIF, $strTime );
 331
 332                                 if ( $exifTime ) {
 333                                         $text['DateTime'] = $exifTime;
 334                                 }
 335
 336                         } elseif ( $chunk_type == 'pHYs' ) {
 337                                 // how big pixels are (dots per meter).
 338                                 if ( $chunk_size !== 9 ) {
 339                                         throw new Exception( __METHOD__ . ": pHYs wrong size" );
 340                                 }
 341
 342                                 $buf = self::read( $fh, $chunk_size );
 343                                 if ( !$buf || strlen( $buf ) < $chunk_size ) {
 344                                         throw new Exception( __METHOD__ . ": Read error" );
 345                                 }
 346
 347                                 $dim = unpack( "Nwidth/Nheight/Cunit", $buf );
 348                                 if ( $dim['unit'] == 1 ) {
 349                                         // Need to check for negative because php
 350                                         // doesn't deal with super-large unsigned 32-bit ints well
 351                                         if ( $dim['width'] > 0 && $dim['height'] > 0 ) {
 352                                                 // unit is meters
 353                                                 // (as opposed to 0 = undefined )
 354                                                 $text['XResolution'] = $dim['width']
 355                                                         . '/100';
 356                                                 $text['YResolution'] = $dim['height']
 357                                                         . '/100';
 358                                                 $text['ResolutionUnit'] = 3;
 359                                                 // 3 = dots per cm (from Exif).
 360                                         }
 361                                 }
 362
 363                         } elseif ( $chunk_type == "IEND" ) {
 364                                 break;
 365                         } else {
 366                                 fseek( $fh, $chunk_size, SEEK_CUR );
 367                         }
 368                         fseek( $fh, self::$CRC_size, SEEK_CUR );
 369                 }
 370                 fclose( $fh );
 371
 372                 if ( $loopCount > 1 ) {
 373                         $duration *= $loopCount;
 374                 }
 375
 376                 if ( isset( $text['DateTimeDigitized'] ) ) {
 377                         // Convert date format from rfc2822 to exif.
 378                         foreach ( $text['DateTimeDigitized'] as $name => &$value ) {
 379                                 if ( $name === '_type' ) {
 380                                         continue;
 381                                 }
 382
 383                                 // @todo FIXME: Currently timezones are ignored.
 384                                 // possibly should be wfTimestamp's
 385                                 // responsibility. (at least for numeric TZ)
 386                                 $formatted = wfTimestamp( TS_EXIF, $value );
 387                                 if ( $formatted ) {
 388                                         // Only change if we could convert the
 389                                         // date.
 390                                         // The png standard says it should be
 391                                         // in rfc2822 format, but not required.
 392                                         // In general for the exif stuff we
 393                                         // prettify the date if we can, but we
 394                                         // display as-is if we cannot or if
 395                                         // it is invalid.
 396                                         // So do the same here.
 397
 398                                         $value = $formatted;
 399                                 }
 400                         }
 401                 }
 402                 return array(
 403                         'frameCount' => $frameCount,
 404                         'loopCount' => $loopCount,
 405                         'duration' => $duration,
 406                         'text' => $text,
 407                         'bitDepth' => $bitDepth,
 408                         'colorType' => $colorType,
 409                 );
 410
 411         }
 412         /**
 413          * Read a chunk, checking to make sure its not too big.
 414          *
 415          * @param $fh resource The file handle
 416          * @param $size Integer size in bytes.
 417          * @throws Exception if too big.
 418          * @return String The chunk.
 419          */
 420         private static function read( $fh, $size ) {
 421                 if ( $size > self::MAX_CHUNK_SIZE ) {
 422                         throw new Exception( __METHOD__ . ': Chunk size of ' . $size .
 423                                 ' too big. Max size is: ' . self::MAX_CHUNK_SIZE );
 424                 }
 425                 return fread( $fh, $size );
 426         }
 427 }