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 use Wikimedia\XMPReader\Reader as XMPReader;
  26
  27 /**
  28  * Class to deal with reconciling and extracting metadata from bitmap images.
  29  * This is meant to comply with http://www.metadataworkinggroup.org/pdf/mwg_guidance.pdf
  30  *
  31  * This sort of acts as an intermediary between MediaHandler::getMetadata
  32  * and the various metadata extractors.
  33  *
  34  * @todo Other image formats.
  35  * @ingroup Media
  36  */
  37 class BitmapMetadataHandler {
  38         /** @var array */
  39         private $metadata = [];
  40
  41         /** @var array Metadata priority */
  42         private $metaPriority = [
  43                 20 => [ 'other' ],
  44                 40 => [ 'native' ],
  45                 60 => [ 'iptc-good-hash', 'iptc-no-hash' ],
  46                 70 => [ 'xmp-deprecated' ],
  47                 80 => [ 'xmp-general' ],
  48                 90 => [ 'xmp-exif' ],
  49                 100 => [ 'iptc-bad-hash' ],
  50                 120 => [ 'exif' ],
  51         ];
  52
  53         /** @var string */
  54         private $iptcType = 'iptc-no-hash';
  55
  56         /**
  57          * This does the photoshop image resource app13 block
  58          * of interest, IPTC-IIM metadata is stored here.
  59          *
  60          * Mostly just calls doPSIR and doIPTC
  61          *
  62          * @param string $app13 String containing app13 block from jpeg file
  63          */
  64         private function doApp13( $app13 ) {
  65                 try {
  66                         $this->iptcType = JpegMetadataExtractor::doPSIR( $app13 );
  67                 } catch ( Exception $e ) {
  68                         // Error reading the iptc hash information.
  69                         // This probably means the App13 segment is something other than what we expect.
  70                         // However, still try to read it, and treat it as if the hash didn't exist.
  71                         wfDebug( "Error parsing iptc data of file: " . $e->getMessage() . "\n" );
  72                         $this->iptcType = 'iptc-no-hash';
  73                 }
  74
  75                 $iptc = IPTC::parse( $app13 );
  76                 $this->addMetadata( $iptc, $this->iptcType );
  77         }
  78
  79         /**
  80          * Get exif info using exif class.
  81          * Basically what used to be in BitmapHandler::getMetadata().
  82          * Just calls stuff in the Exif class.
  83          *
  84          * Parameters are passed to the Exif class.
  85          *
  86          * @param string $filename
  87          * @param string $byteOrder
  88          */
  89         function getExif( $filename, $byteOrder ) {
  90                 global $wgShowEXIF;
  91                 if ( file_exists( $filename ) && $wgShowEXIF ) {
  92                         $exif = new Exif( $filename, $byteOrder );
  93                         $data = $exif->getFilteredData();
  94                         if ( $data ) {
  95                                 $this->addMetadata( $data, 'exif' );
  96                         }
  97                 }
  98         }
  99
 100         /** Add misc metadata. Warning: atm if the metadata category
 101          * doesn't have a priority, it will be silently discarded.
 102          *
 103          * @param array $metaArray Array of metadata values
 104          * @param string $type Type. defaults to other. if two things have the same type they're merged
 105          */
 106         function addMetadata( $metaArray, $type = 'other' ) {
 107                 if ( isset( $this->metadata[$type] ) ) {
 108                         /* merge with old data */
 109                         $metaArray = $metaArray + $this->metadata[$type];
 110                 }
 111
 112                 $this->metadata[$type] = $metaArray;
 113         }
 114
 115         /**
 116          * Merge together the various types of metadata
 117          * the different types have different priorites,
 118          * and are merged in order.
 119          *
 120          * This function is generally called by the media handlers' getMetadata()
 121          *
 122          * @return array Metadata array
 123          */
 124         function getMetadataArray() {
 125                 // this seems a bit ugly... This is all so its merged in right order
 126                 // based on the MWG recomendation.
 127                 $temp = [];
 128                 krsort( $this->metaPriority );
 129                 foreach ( $this->metaPriority as $pri ) {
 130                         foreach ( $pri as $type ) {
 131                                 if ( isset( $this->metadata[$type] ) ) {
 132                                         // Do some special casing for multilingual values.
 133                                         // Don't discard translations if also as a simple value.
 134                                         foreach ( $this->metadata[$type] as $itemName => $item ) {
 135                                                 if ( is_array( $item ) && isset( $item['_type'] ) && $item['_type'] === 'lang' ) {
 136                                                         if ( isset( $temp[$itemName] ) && !is_array( $temp[$itemName] ) ) {
 137                                                                 $default = $temp[$itemName];
 138                                                                 $temp[$itemName] = $item;
 139                                                                 $temp[$itemName]['x-default'] = $default;
 140                                                                 unset( $this->metadata[$type][$itemName] );
 141                                                         }
 142                                                 }
 143                                         }
 144
 145                                         $temp = $temp + $this->metadata[$type];
 146                                 }
 147                         }
 148                 }
 149
 150                 return $temp;
 151         }
 152
 153         /** Main entry point for jpeg's.
 154          *
 155          * @param string $filename Filename (with full path)
 156          * @return array Metadata result array.
 157          * @throws MWException On invalid file.
 158          */
 159         static function Jpeg( $filename ) {
 160                 $showXMP = XMPReader::isSupported();
 161                 $meta = new self();
 162
 163                 $seg = JpegMetadataExtractor::segmentSplitter( $filename );
 164
 165                 if ( isset( $seg['COM'] ) && isset( $seg['COM'][0] ) ) {
 166                         $meta->addMetadata( [ 'JPEGFileComment' => $seg['COM'] ], 'native' );
 167                 }
 168                 if ( isset( $seg['PSIR'] ) && count( $seg['PSIR'] ) > 0 ) {
 169                         foreach ( $seg['PSIR'] as $curPSIRValue ) {
 170                                 $meta->doApp13( $curPSIRValue );
 171                         }
 172                 }
 173                 if ( isset( $seg['XMP'] ) && $showXMP ) {
 174                         $xmp = new XMPReader( LoggerFactory::getInstance( 'XMP' ), $filename );
 175                         $xmp->parse( $seg['XMP'] );
 176                         foreach ( $seg['XMP_ext'] as $xmpExt ) {
 177                                 /* Support for extended xmp in jpeg files
 178                                  * is not well tested and a bit fragile.
 179                                  */
 180                                 $xmp->parseExtended( $xmpExt );
 181                         }
 182                         $res = $xmp->getResults();
 183                         foreach ( $res as $type => $array ) {
 184                                 $meta->addMetadata( $array, $type );
 185                         }
 186                 }
 187
 188                 $meta->getExif( $filename, $seg['byteOrder'] ?? 'BE' );
 189
 190                 return $meta->getMetadataArray();
 191         }
 192
 193         /** Entry point for png
 194          * At some point in the future this might
 195          * merge the png various tEXt chunks to that
 196          * are interesting, but for now it only does XMP
 197          *
 198          * @param string $filename Full path to file
 199          * @return array Array for storage in img_metadata.
 200          */
 201         public static function PNG( $filename ) {
 202                 $showXMP = XMPReader::isSupported();
 203
 204                 $meta = new self();
 205                 $array = PNGMetadataExtractor::getMetadata( $filename );
 206                 if ( isset( $array['text']['xmp']['x-default'] )
 207                         && $array['text']['xmp']['x-default'] !== '' && $showXMP
 208                 ) {
 209                         $xmp = new XMPReader( LoggerFactory::getInstance( 'XMP' ), $filename );
 210                         $xmp->parse( $array['text']['xmp']['x-default'] );
 211                         $xmpRes = $xmp->getResults();
 212                         foreach ( $xmpRes as $type => $xmpSection ) {
 213                                 $meta->addMetadata( $xmpSection, $type );
 214                         }
 215                 }
 216                 unset( $array['text']['xmp'] );
 217                 $meta->addMetadata( $array['text'], 'native' );
 218                 unset( $array['text'] );
 219                 $array['metadata'] = $meta->getMetadataArray();
 220                 $array['metadata']['_MW_PNG_VERSION'] = PNGMetadataExtractor::VERSION;
 221
 222                 return $array;
 223         }
 224
 225         /** function for gif images.
 226          *
 227          * They don't really have native metadata, so just merges together
 228          * XMP and image comment.
 229          *
 230          * @param string $filename Full path to file
 231          * @return array Metadata array
 232          */
 233         public static function GIF( $filename ) {
 234                 $meta = new self();
 235                 $baseArray = GIFMetadataExtractor::getMetadata( $filename );
 236
 237                 if ( count( $baseArray['comment'] ) > 0 ) {
 238                         $meta->addMetadata( [ 'GIFFileComment' => $baseArray['comment'] ], 'native' );
 239                 }
 240
 241                 if ( $baseArray['xmp'] !== '' && XMPReader::isSupported() ) {
 242                         $xmp = new XMPReader( LoggerFactory::getInstance( 'XMP' ), $filename );
 243                         $xmp->parse( $baseArray['xmp'] );
 244                         $xmpRes = $xmp->getResults();
 245                         foreach ( $xmpRes as $type => $xmpSection ) {
 246                                 $meta->addMetadata( $xmpSection, $type );
 247                         }
 248                 }
 249
 250                 unset( $baseArray['comment'] );
 251                 unset( $baseArray['xmp'] );
 252
 253                 $baseArray['metadata'] = $meta->getMetadataArray();
 254                 $baseArray['metadata']['_MW_GIF_VERSION'] = GIFMetadataExtractor::VERSION;
 255
 256                 return $baseArray;
 257         }
 258
 259         /**
 260          * This doesn't do much yet, but eventually I plan to add
 261          * XMP support for Tiff. (PHP's exif support already extracts
 262          * but needs some further processing because PHP's exif support
 263          * is stupid...)
 264          *
 265          * @todo Add XMP support, so this function actually makes sense to put here.
 266          *
 267          * The various exceptions this throws are caught later.
 268          * @param string $filename
 269          * @throws MWException
 270          * @return array The metadata.
 271          */
 272         public static function Tiff( $filename ) {
 273                 if ( file_exists( $filename ) ) {
 274                         $byteOrder = self::getTiffByteOrder( $filename );
 275                         if ( !$byteOrder ) {
 276                                 throw new MWException( "Error determining byte order of $filename" );
 277                         }
 278                         $exif = new Exif( $filename, $byteOrder );
 279                         $data = $exif->getFilteredData();
 280                         if ( $data ) {
 281                                 $data['MEDIAWIKI_EXIF_VERSION'] = Exif::version();
 282
 283                                 return $data;
 284                         } else {
 285                                 throw new MWException( "Could not extract data from tiff file $filename" );
 286                         }
 287                 } else {
 288                         throw new MWException( "File doesn't exist - $filename" );
 289                 }
 290         }
 291
 292         /**
 293          * Read the first 2 bytes of a tiff file to figure out
 294          * Little Endian or Big Endian. Needed for exif stuff.
 295          *
 296          * @param string $filename
 297          * @return string 'BE' or 'LE' or false
 298          */
 299         static function getTiffByteOrder( $filename ) {
 300                 $fh = fopen( $filename, 'rb' );
 301                 if ( !$fh ) {
 302                         return false;
 303                 }
 304                 $head = fread( $fh, 2 );
 305                 fclose( $fh );
 306
 307                 switch ( $head ) {
 308                         case 'II':
 309                                 return 'LE'; // II for intel.
 310                         case 'MM':
 311                                 return 'BE'; // MM for motorla.
 312                         default:
 313                                 return false; // Something went wrong.
 314
 315                 }
 316         }
 317 }