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