includes/media/BitmapMetadataHandler.php

   1 <?php
   2 /**
   3  * Extraction of metadata from different bitmap image types.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup Media
  22  */
  23
  24 use MediaWiki\Logger\LoggerFactory;
  25
  26 /**
  27  * Class to deal with reconciling and extracting metadata from bitmap images.
  28  * This is meant to comply with http://www.metadataworkinggroup.org/pdf/mwg_guidance.pdf
  29  *
  30  * This sort of acts as an intermediary between MediaHandler::getMetadata
  31  * and the various metadata extractors.
  32  *
  33  * @todo Other image formats.
  34  * @ingroup Media
  35  */
  36 class BitmapMetadataHandler {
  37         /** @var array */
  38         private $metadata = [];
  39
  40         /** @var array Metadata priority */
  41         private $metaPriority = [
  42                 20 => [ 'other' ],
  43                 40 => [ 'native' ],
  44                 60 => [ 'iptc-good-hash', 'iptc-no-hash' ],
  45                 70 => [ 'xmp-deprecated' ],
  46                 80 => [ 'xmp-general' ],
  47                 90 => [ 'xmp-exif' ],
  48                 100 => [ 'iptc-bad-hash' ],
  49                 120 => [ 'exif' ],
  50         ];
  51
  52         /** @var string */
  53         private $iptcType = 'iptc-no-hash';
  54
  55         /**
  56          * This does the photoshop image resource app13 block
  57          * of interest, IPTC-IIM metadata is stored here.
  58          *
  59          * Mostly just calls doPSIR and doIPTC
  60          *
  61          * @param string $app13 String containing app13 block from jpeg file
  62          */
  63         private function doApp13( $app13 ) {
  64                 try {
  65                         $this->iptcType = JpegMetadataExtractor::doPSIR( $app13 );
  66                 } catch ( Exception $e ) {
  67                         // Error reading the iptc hash information.
  68                         // This probably means the App13 segment is something other than what we expect.
  69                         // However, still try to read it, and treat it as if the hash didn't exist.
  70                         wfDebug( "Error parsing iptc data of file: " . $e->getMessage() . "\n" );
  71                         $this->iptcType = 'iptc-no-hash';
  72                 }
  73
  74                 $iptc = IPTC::parse( $app13 );
  75                 $this->addMetadata( $iptc, $this->iptcType );
  76         }
  77
  78         /**
  79          * Get exif info using exif class.
  80          * Basically what used to be in BitmapHandler::getMetadata().
  81          * Just calls stuff in the Exif class.
  82          *
  83          * Parameters are passed to the Exif class.
  84          *
  85          * @param string $filename
  86          * @param string $byteOrder
  87          */
  88         function getExif( $filename, $byteOrder ) {
  89                 global $wgShowEXIF;
  90                 if ( file_exists( $filename ) && $wgShowEXIF ) {
  91                         $exif = new Exif( $filename, $byteOrder );
  92                         $data = $exif->getFilteredData();
  93                         if ( $data ) {
  94                                 $this->addMetadata( $data, 'exif' );
  95                         }
  96                 }
  97         }
  98
  99         /** Add misc metadata. Warning: atm if the metadata category
 100          * doesn't have a priority, it will be silently discarded.
 101          *
 102          * @param array $metaArray Array of metadata values
 103          * @param string $type Type. defaults to other. if two things have the same type they're merged
 104          */
 105         function addMetadata( $metaArray, $type = 'other' ) {
 106                 if ( isset( $this->metadata[$type] ) ) {
 107                         /* merge with old data */
 108                         $metaArray = $metaArray + $this->metadata[$type];
 109                 }
 110
 111                 $this->metadata[$type] = $metaArray;
 112         }
 113
 114         /**
 115          * Merge together the various types of metadata
 116          * the different types have different priorites,
 117          * and are merged in order.
 118          *
 119          * This function is generally called by the media handlers' getMetadata()
 120          *
 121          * @return array Metadata array
 122          */
 123         function getMetadataArray() {
 124                 // this seems a bit ugly... This is all so its merged in right order
 125                 // based on the MWG recomendation.
 126                 $temp = [];
 127                 krsort( $this->metaPriority );
 128                 foreach ( $this->metaPriority as $pri ) {
 129                         foreach ( $pri as $type ) {
 130                                 if ( isset( $this->metadata[$type] ) ) {
 131                                         // Do some special casing for multilingual values.
 132                                         // Don't discard translations if also as a simple value.
 133                                         foreach ( $this->metadata[$type] as $itemName => $item ) {
 134                                                 if ( is_array( $item ) && isset( $item['_type'] ) && $item['_type'] === 'lang' ) {
 135                                                         if ( isset( $temp[$itemName] ) && !is_array( $temp[$itemName] ) ) {
 136                                                                 $default = $temp[$itemName];
 137                                                                 $temp[$itemName] = $item;
 138                                                                 $temp[$itemName]['x-default'] = $default;
 139                                                                 unset( $this->metadata[$type][$itemName] );
 140                                                         }
 141                                                 }
 142                                         }
 143
 144                                         $temp = $temp + $this->metadata[$type];
 145                                 }
 146                         }
 147                 }
 148
 149                 return $temp;
 150         }
 151
 152         /** Main entry point for jpeg's.
 153          *
 154          * @param string $filename Filename (with full path)
 155          * @return array Metadata result array.
 156          * @throws MWException On invalid file.
 157          */
 158         static function Jpeg( $filename ) {
 159                 $showXMP = XMPReader::isSupported();
 160                 $meta = new self();
 161
 162                 $seg = JpegMetadataExtractor::segmentSplitter( $filename );
 163
 164                 if ( isset( $seg['COM'] ) && isset( $seg['COM'][0] ) ) {
 165                         $meta->addMetadata( [ 'JPEGFileComment' => $seg['COM'] ], 'native' );
 166                 }
 167                 if ( isset( $seg['PSIR'] ) && count( $seg['PSIR'] ) > 0 ) {
 168                         foreach ( $seg['PSIR'] as $curPSIRValue ) {
 169                                 $meta->doApp13( $curPSIRValue );
 170                         }
 171                 }
 172                 if ( isset( $seg['XMP'] ) && $showXMP ) {
 173                         $xmp = new XMPReader( LoggerFactory::getInstance( 'XMP' ), $filename );
 174                         $xmp->parse( $seg['XMP'] );
 175                         foreach ( $seg['XMP_ext'] as $xmpExt ) {
 176                                 /* Support for extended xmp in jpeg files
 177                                  * is not well tested and a bit fragile.
 178                                  */
 179                                 $xmp->parseExtended( $xmpExt );
 180                         }
 181                         $res = $xmp->getResults();
 182                         foreach ( $res as $type => $array ) {
 183                                 $meta->addMetadata( $array, $type );
 184                         }
 185                 }
 186
 187                 $meta->getExif( $filename, isset( $seg['byteOrder'] ) ? $seg['byteOrder'] : 'BE' );
 188
 189                 return $meta->getMetadataArray();
 190         }
 191
 192         /** Entry point for png
 193          * At some point in the future this might
 194          * merge the png various tEXt chunks to that
 195          * are interesting, but for now it only does XMP
 196          *
 197          * @param string $filename Full path to file
 198          * @return array Array for storage in img_metadata.
 199          */
 200         public static function PNG( $filename ) {
 201                 $showXMP = XMPReader::isSupported();
 202
 203                 $meta = new self();
 204                 $array = PNGMetadataExtractor::getMetadata( $filename );
 205                 if ( isset( $array['text']['xmp']['x-default'] )
 206                         && $array['text']['xmp']['x-default'] !== '' && $showXMP
 207                 ) {
 208                         $xmp = new XMPReader( LoggerFactory::getInstance( 'XMP' ), $filename );
 209                         $xmp->parse( $array['text']['xmp']['x-default'] );
 210                         $xmpRes = $xmp->getResults();
 211                         foreach ( $xmpRes as $type => $xmpSection ) {
 212                                 $meta->addMetadata( $xmpSection, $type );
 213                         }
 214                 }
 215                 unset( $array['text']['xmp'] );
 216                 $meta->addMetadata( $array['text'], 'native' );
 217                 unset( $array['text'] );
 218                 $array['metadata'] = $meta->getMetadataArray();
 219                 $array['metadata']['_MW_PNG_VERSION'] = PNGMetadataExtractor::VERSION;
 220
 221                 return $array;
 222         }
 223
 224         /** function for gif images.
 225          *
 226          * They don't really have native metadata, so just merges together
 227          * XMP and image comment.
 228          *
 229          * @param string $filename Full path to file
 230          * @return array Metadata array
 231          */
 232         public static function GIF( $filename ) {
 233                 $meta = new self();
 234                 $baseArray = GIFMetadataExtractor::getMetadata( $filename );
 235
 236                 if ( count( $baseArray['comment'] ) > 0 ) {
 237                         $meta->addMetadata( [ 'GIFFileComment' => $baseArray['comment'] ], 'native' );
 238                 }
 239
 240                 if ( $baseArray['xmp'] !== '' && XMPReader::isSupported() ) {
 241                         $xmp = new XMPReader( LoggerFactory::getInstance( 'XMP' ), $filename );
 242                         $xmp->parse( $baseArray['xmp'] );
 243                         $xmpRes = $xmp->getResults();
 244                         foreach ( $xmpRes as $type => $xmpSection ) {
 245                                 $meta->addMetadata( $xmpSection, $type );
 246                         }
 247                 }
 248
 249                 unset( $baseArray['comment'] );
 250                 unset( $baseArray['xmp'] );
 251
 252                 $baseArray['metadata'] = $meta->getMetadataArray();
 253                 $baseArray['metadata']['_MW_GIF_VERSION'] = GIFMetadataExtractor::VERSION;
 254
 255                 return $baseArray;
 256         }
 257
 258         /**
 259          * This doesn't do much yet, but eventually I plan to add
 260          * XMP support for Tiff. (PHP's exif support already extracts
 261          * but needs some further processing because PHP's exif support
 262          * is stupid...)
 263          *
 264          * @todo Add XMP support, so this function actually makes sense to put here.
 265          *
 266          * The various exceptions this throws are caught later.
 267          * @param string $filename
 268          * @throws MWException
 269          * @return array The metadata.
 270          */
 271         public static function Tiff( $filename ) {
 272                 if ( file_exists( $filename ) ) {
 273                         $byteOrder = self::getTiffByteOrder( $filename );
 274                         if ( !$byteOrder ) {
 275                                 throw new MWException( "Error determining byte order of $filename" );
 276                         }
 277                         $exif = new Exif( $filename, $byteOrder );
 278                         $data = $exif->getFilteredData();
 279                         if ( $data ) {
 280                                 $data['MEDIAWIKI_EXIF_VERSION'] = Exif::version();
 281
 282                                 return $data;
 283                         } else {
 284                                 throw new MWException( "Could not extract data from tiff file $filename" );
 285                         }
 286                 } else {
 287                         throw new MWException( "File doesn't exist - $filename" );
 288                 }
 289         }
 290
 291         /**
 292          * Read the first 2 bytes of a tiff file to figure out
 293          * Little Endian or Big Endian. Needed for exif stuff.
 294          *
 295          * @param string $filename The filename
 296          * @return string 'BE' or 'LE' or false
 297          */
 298         static function getTiffByteOrder( $filename ) {
 299                 $fh = fopen( $filename, 'rb' );
 300                 if ( !$fh ) {
 301                         return false;
 302                 }
 303                 $head = fread( $fh, 2 );
 304                 fclose( $fh );
 305
 306                 switch ( $head ) {
 307                         case 'II':
 308                                 return 'LE'; // II for intel.
 309                         case 'MM':
 310                                 return 'BE'; // MM for motorla.
 311                         default:
 312                                 return false; // Something went wrong.
 313
 314                 }
 315         }
 316 }