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