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