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