includes/media/Exif.php

   1 <?php
   2 /**
   3  * This program is free software; you can redistribute it and/or modify
   4  * it under the terms of the GNU General Public License as published by
   5  * the Free Software Foundation; either version 2 of the License, or
   6  * (at your option) any later version.
   7  *
   8  * This program is distributed in the hope that it will be useful,
   9  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  11  * GNU General Public License for more details.
  12  *
  13  * You should have received a copy of the GNU General Public License along
  14  * with this program; if not, write to the Free Software Foundation, Inc.,
  15  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  16  * http://www.gnu.org/copyleft/gpl.html
  17  *
  18  * @ingroup Media
  19  * @author Ævar Arnfjörð Bjarmason <avarab@gmail.com>
  20  * @copyright Copyright © 2005, Ævar Arnfjörð Bjarmason, 2009 Brent Garber
  21  * @license http://www.gnu.org/copyleft/gpl.html GNU General Public License
  22  * @see http://exif.org/Exif2-2.PDF The Exif 2.2 specification
  23  * @file
  24  */
  25
  26 /**
  27  * Class to extract and validate Exif data from jpeg (and possibly tiff) files.
  28  * @ingroup Media
  29  */
  30 class Exif {
  31
  32         const BYTE      = 1;    //!< An 8-bit (1-byte) unsigned integer.
  33         const ASCII     = 2;    //!< An 8-bit byte containing one 7-bit ASCII code. The final byte is terminated with NULL.
  34         const SHORT     = 3;    //!< A 16-bit (2-byte) unsigned integer.
  35         const LONG      = 4;    //!< A 32-bit (4-byte) unsigned integer.
  36         const RATIONAL  = 5;    //!< Two LONGs. The first LONG is the numerator and the second LONG expresses the denominator
  37         const UNDEFINED = 7;    //!< An 8-bit byte that can take any value depending on the field definition
  38         const SLONG     = 9;    //!< A 32-bit (4-byte) signed integer (2's complement notation),
  39         const SRATIONAL = 10;   //!< Two SLONGs. The first SLONG is the numerator and the second SLONG is the denominator.
  40         const IGNORE    = -1;   // A fake value for things we don't want or don't support.
  41
  42         //@{
  43         /* @var array
  44          * @private
  45          */
  46
  47         /**
  48          * Exif tags grouped by category, the tagname itself is the key and the type
  49          * is the value, in the case of more than one possible value type they are
  50          * separated by commas.
  51          */
  52         var $mExifTags;
  53
  54         /**
  55          * The raw Exif data returned by exif_read_data()
  56          */
  57         var $mRawExifData;
  58
  59         /**
  60          * A Filtered version of $mRawExifData that has been pruned of invalid
  61          * tags and tags that contain content they shouldn't contain according
  62          * to the Exif specification
  63          */
  64         var $mFilteredExifData;
  65
  66         /**
  67          * Filtered and formatted Exif data, see FormatMetadata::getFormattedData()
  68          */
  69         var $mFormattedExifData;
  70
  71         //@}
  72
  73         //@{
  74         /* @var string
  75          * @private
  76          */
  77
  78         /**
  79          * The file being processed
  80          */
  81         var $file;
  82
  83         /**
  84          * The basename of the file being processed
  85          */
  86         var $basename;
  87
  88         /**
  89          * The private log to log to, e.g. 'exif'
  90          */
  91         var $log = false;
  92
  93         /**
  94          * The byte order of the file. Needed because php's
  95          * extension doesn't fully process some obscure props.
  96          */
  97         private $byteOrder;
  98         //@}
  99
 100         /**
 101          * Constructor
 102          *
 103          * @param $file String: filename.
 104          * @param $byteOrder String Type of byte ordering either 'BE' (Big Endian) or 'LE' (Little Endian). Default ''.
 105          * @todo FIXME: The following are broke:
 106          * SubjectArea. Need to test the more obscure tags.
 107          *
 108          * DigitalZoomRatio = 0/0 is rejected. need to determine if that's valid.
 109          * possibly should treat 0/0 = 0. need to read exif spec on that.
 110          */
 111         function __construct( $file, $byteOrder = '' ) {
 112                 /**
 113                  * Page numbers here refer to pages in the EXIF 2.2 standard
 114                  *
 115                  * Note, Exif::UNDEFINED is treated as a string, not as an array of bytes
 116                  * so don't put a count parameter for any UNDEFINED values.
 117                  *
 118                  * @link http://exif.org/Exif2-2.PDF The Exif 2.2 specification
 119                  */
 120                 $this->mExifTags = array(
 121                         # TIFF Rev. 6.0 Attribute Information (p22)
 122                         'IFD0' => array(
 123                                 # Tags relating to image structure
 124                                 'ImageWidth' => Exif::SHORT.','.Exif::LONG,             # Image width
 125                                 'ImageLength' => Exif::SHORT.','.Exif::LONG,            # Image height
 126                                 'BitsPerSample' => array( Exif::SHORT, 3 ),             # Number of bits per component
 127                                 # "When a primary image is JPEG compressed, this designation is not"
 128                                 # "necessary and is omitted." (p23)
 129                                 'Compression' => Exif::SHORT,                           # Compression scheme #p23
 130                                 'PhotometricInterpretation' => Exif::SHORT,             # Pixel composition #p23
 131                                 'Orientation' => Exif::SHORT,                           # Orientation of image #p24
 132                                 'SamplesPerPixel' => Exif::SHORT,                       # Number of components
 133                                 'PlanarConfiguration' => Exif::SHORT,                   # Image data arrangement #p24
 134                                 'YCbCrSubSampling' => array( Exif::SHORT, 2),           # Subsampling ratio of Y to C #p24
 135                                 'YCbCrPositioning' => Exif::SHORT,                      # Y and C positioning #p24-25
 136                                 'XResolution' => Exif::RATIONAL,                        # Image resolution in width direction
 137                                 'YResolution' => Exif::RATIONAL,                        # Image resolution in height direction
 138                                 'ResolutionUnit' => Exif::SHORT,                        # Unit of X and Y resolution #(p26)
 139
 140                                 # Tags relating to recording offset
 141                                 'StripOffsets' => Exif::SHORT.','.Exif::LONG,                   # Image data location
 142                                 'RowsPerStrip' => Exif::SHORT.','.Exif::LONG,                   # Number of rows per strip
 143                                 'StripByteCounts' => Exif::SHORT.','.Exif::LONG,                # Bytes per compressed strip
 144                                 'JPEGInterchangeFormat' => Exif::SHORT.','.Exif::LONG,          # Offset to JPEG SOI
 145                                 'JPEGInterchangeFormatLength' => Exif::SHORT.','.Exif::LONG,    # Bytes of JPEG data
 146
 147                                 # Tags relating to image data characteristics
 148                                 'TransferFunction' => Exif::IGNORE,                     # Transfer function
 149                                 'WhitePoint' => array( Exif::RATIONAL, 2),              # White point chromaticity
 150                                 'PrimaryChromaticities' => array( Exif::RATIONAL, 6),   # Chromaticities of primarities
 151                                 'YCbCrCoefficients' => array( Exif::RATIONAL, 3),       # Color space transformation matrix coefficients #p27
 152                                 'ReferenceBlackWhite' => array( Exif::RATIONAL, 6),     # Pair of black and white reference values
 153
 154                                 # Other tags
 155                                 'DateTime' => Exif::ASCII,                              # File change date and time
 156                                 'ImageDescription' => Exif::ASCII,                      # Image title
 157                                 'Make' => Exif::ASCII,                                  # Image input equipment manufacturer
 158                                 'Model' => Exif::ASCII,                                 # Image input equipment model
 159                                 'Software' => Exif::ASCII,                              # Software used
 160                                 'Artist' => Exif::ASCII,                                # Person who created the image
 161                                 'Copyright' => Exif::ASCII,                             # Copyright holder
 162                         ),
 163
 164                         # Exif IFD Attribute Information (p30-31)
 165                         'EXIF' => array(
 166                                 # TODO: NOTE: Nonexistence of this field is taken to mean nonconformance
 167                                 # to the EXIF 2.1 AND 2.2 standards
 168                                 'ExifVersion' =>  Exif::UNDEFINED,                      # Exif version
 169                                 'FlashPixVersion' => Exif::UNDEFINED,                   # Supported Flashpix version #p32
 170
 171                                 # Tags relating to Image Data Characteristics
 172                                 'ColorSpace' => Exif::SHORT,                            # Color space information #p32
 173
 174                                 # Tags relating to image configuration
 175                                 'ComponentsConfiguration' => Exif::UNDEFINED,                   # Meaning of each component #p33
 176                                 'CompressedBitsPerPixel' => Exif::RATIONAL,                     # Image compression mode
 177                                 'PixelYDimension' => Exif::SHORT.','.Exif::LONG,                # Valid image width
 178                                 'PixelXDimension' => Exif::SHORT.','.Exif::LONG,                # Valid image height
 179
 180                                 # Tags relating to related user information
 181                                 'MakerNote' => Exif::IGNORE,                            # Manufacturer notes
 182                                 'UserComment' => Exif::UNDEFINED,                       # User comments #p34
 183
 184                                 # Tags relating to related file information
 185                                 'RelatedSoundFile' => Exif::ASCII,                      # Related audio file
 186
 187                                 # Tags relating to date and time
 188                                 'DateTimeOriginal' => Exif::ASCII,                      # Date and time of original data generation #p36
 189                                 'DateTimeDigitized' => Exif::ASCII,                     # Date and time of original data generation
 190                                 'SubSecTime' => Exif::ASCII,                            # DateTime subseconds
 191                                 'SubSecTimeOriginal' => Exif::ASCII,                    # DateTimeOriginal subseconds
 192                                 'SubSecTimeDigitized' => Exif::ASCII,                   # DateTimeDigitized subseconds
 193
 194                                 # Tags relating to picture-taking conditions (p31)
 195                                 'ExposureTime' => Exif::RATIONAL,                       # Exposure time
 196                                 'FNumber' => Exif::RATIONAL,                            # F Number
 197                                 'ExposureProgram' => Exif::SHORT,                       # Exposure Program #p38
 198                                 'SpectralSensitivity' => Exif::ASCII,                   # Spectral sensitivity
 199                                 'ISOSpeedRatings' => Exif::SHORT,                       # ISO speed rating
 200                                 'OECF' => Exif::IGNORE,
 201                                 # Optoelectronic conversion factor. Note: We don't have support for this atm.
 202                                 'ShutterSpeedValue' => Exif::SRATIONAL,                 # Shutter speed
 203                                 'ApertureValue' => Exif::RATIONAL,                      # Aperture
 204                                 'BrightnessValue' => Exif::SRATIONAL,                   # Brightness
 205                                 'ExposureBiasValue' => Exif::SRATIONAL,                 # Exposure bias
 206                                 'MaxApertureValue' => Exif::RATIONAL,                   # Maximum land aperture
 207                                 'SubjectDistance' => Exif::RATIONAL,                    # Subject distance
 208                                 'MeteringMode' => Exif::SHORT,                          # Metering mode #p40
 209                                 'LightSource' => Exif::SHORT,                           # Light source #p40-41
 210                                 'Flash' => Exif::SHORT,                                 # Flash #p41-42
 211                                 'FocalLength' => Exif::RATIONAL,                        # Lens focal length
 212                                 'SubjectArea' => array( Exif::SHORT, 4 ),               # Subject area
 213                                 'FlashEnergy' => Exif::RATIONAL,                        # Flash energy
 214                                 'SpatialFrequencyResponse' => Exif::IGNORE,             # Spatial frequency response. Not supported atm.
 215                                 'FocalPlaneXResolution' => Exif::RATIONAL,              # Focal plane X resolution
 216                                 'FocalPlaneYResolution' => Exif::RATIONAL,              # Focal plane Y resolution
 217                                 'FocalPlaneResolutionUnit' => Exif::SHORT,              # Focal plane resolution unit #p46
 218                                 'SubjectLocation' => array( Exif::SHORT, 2),            # Subject location
 219                                 'ExposureIndex' => Exif::RATIONAL,                      # Exposure index
 220                                 'SensingMethod' => Exif::SHORT,                         # Sensing method #p46
 221                                 'FileSource' => Exif::UNDEFINED,                        # File source #p47
 222                                 'SceneType' => Exif::UNDEFINED,                         # Scene type #p47
 223                                 'CFAPattern' => Exif::IGNORE,                           # CFA pattern. not supported atm.
 224                                 'CustomRendered' => Exif::SHORT,                        # Custom image processing #p48
 225                                 'ExposureMode' => Exif::SHORT,                          # Exposure mode #p48
 226                                 'WhiteBalance' => Exif::SHORT,                          # White Balance #p49
 227                                 'DigitalZoomRatio' => Exif::RATIONAL,                   # Digital zoom ration
 228                                 'FocalLengthIn35mmFilm' => Exif::SHORT,                 # Focal length in 35 mm film
 229                                 'SceneCaptureType' => Exif::SHORT,                      # Scene capture type #p49
 230                                 'GainControl' => Exif::SHORT,                           # Scene control #p49-50
 231                                 'Contrast' => Exif::SHORT,                              # Contrast #p50
 232                                 'Saturation' => Exif::SHORT,                            # Saturation #p50
 233                                 'Sharpness' => Exif::SHORT,                             # Sharpness #p50
 234                                 'DeviceSettingDescription' => Exif::IGNORE,
 235                                 # Device settings description. This could maybe be supported. Need to find an
 236                                 # example file that uses this to see if it has stuff of interest in it.
 237                                 'SubjectDistanceRange' => Exif::SHORT,                  # Subject distance range #p51
 238
 239                                 'ImageUniqueID' => Exif::ASCII,                         # Unique image ID
 240                         ),
 241
 242                         # GPS Attribute Information (p52)
 243                         'GPS' => array(
 244                                 'GPSVersion' => Exif::UNDEFINED,
 245                                 # Should be an array of 4 Exif::BYTE's. However php treats it as an undefined
 246                                 # Note exif standard calls this GPSVersionID, but php doesn't like the id suffix
 247                                 'GPSLatitudeRef' => Exif::ASCII,                        # North or South Latitude #p52-53
 248                                 'GPSLatitude' => array( Exif::RATIONAL, 3 ),            # Latitude
 249                                 'GPSLongitudeRef' => Exif::ASCII,                       # East or West Longitude #p53
 250                                 'GPSLongitude' => array( Exif::RATIONAL, 3),            # Longitude
 251                                 'GPSAltitudeRef' => Exif::UNDEFINED,
 252                                 # Altitude reference. Note, the exif standard says this should be an EXIF::Byte,
 253                                 # but php seems to disagree.
 254                                 'GPSAltitude' => Exif::RATIONAL,                        # Altitude
 255                                 'GPSTimeStamp' => array( Exif::RATIONAL, 3),            # GPS time (atomic clock)
 256                                 'GPSSatellites' => Exif::ASCII,                         # Satellites used for measurement
 257                                 'GPSStatus' => Exif::ASCII,                             # Receiver status #p54
 258                                 'GPSMeasureMode' => Exif::ASCII,                        # Measurement mode #p54-55
 259                                 'GPSDOP' => Exif::RATIONAL,                             # Measurement precision
 260                                 'GPSSpeedRef' => Exif::ASCII,                           # Speed unit #p55
 261                                 'GPSSpeed' => Exif::RATIONAL,                           # Speed of GPS receiver
 262                                 'GPSTrackRef' => Exif::ASCII,                           # Reference for direction of movement #p55
 263                                 'GPSTrack' => Exif::RATIONAL,                           # Direction of movement
 264                                 'GPSImgDirectionRef' => Exif::ASCII,                    # Reference for direction of image #p56
 265                                 'GPSImgDirection' => Exif::RATIONAL,                    # Direction of image
 266                                 'GPSMapDatum' => Exif::ASCII,                           # Geodetic survey data used
 267                                 'GPSDestLatitudeRef' => Exif::ASCII,                    # Reference for latitude of destination #p56
 268                                 'GPSDestLatitude' => array( Exif::RATIONAL, 3 ),        # Latitude destination
 269                                 'GPSDestLongitudeRef' => Exif::ASCII,                   # Reference for longitude of destination #p57
 270                                 'GPSDestLongitude' => array( Exif::RATIONAL, 3 ),       # Longitude of destination
 271                                 'GPSDestBearingRef' => Exif::ASCII,                     # Reference for bearing of destination #p57
 272                                 'GPSDestBearing' => Exif::RATIONAL,                     # Bearing of destination
 273                                 'GPSDestDistanceRef' => Exif::ASCII,                    # Reference for distance to destination #p57-58
 274                                 'GPSDestDistance' => Exif::RATIONAL,                    # Distance to destination
 275                                 'GPSProcessingMethod' => Exif::UNDEFINED,               # Name of GPS processing method
 276                                 'GPSAreaInformation' => Exif::UNDEFINED,                # Name of GPS area
 277                                 'GPSDateStamp' => Exif::ASCII,                          # GPS date
 278                                 'GPSDifferential' => Exif::SHORT,                       # GPS differential correction
 279                         ),
 280                 );
 281
 282                 $this->file = $file;
 283                 $this->basename = wfBaseName( $this->file );
 284                 if ( $byteOrder === 'BE' || $byteOrder === 'LE' ) {
 285                         $this->byteOrder = $byteOrder;
 286                 } else {
 287                         // Only give a warning for b/c, since originally we didn't
 288                         // require this. The number of things affected by this is
 289                         // rather small.
 290                         wfWarn( 'Exif class did not have byte order specified. '
 291                          . 'Some properties may be decoded incorrectly.' );
 292                         $this->byteOrder = 'BE'; // BE seems about twice as popular as LE in jpg's.
 293                 }
 294
 295                 $this->debugFile( $this->basename, __FUNCTION__, true );
 296                 if( function_exists( 'exif_read_data' ) ) {
 297                         wfSuppressWarnings();
 298                         $data = exif_read_data( $this->file, 0, true );
 299                         wfRestoreWarnings();
 300                 } else {
 301                         throw new MWException( "Internal error: exif_read_data not present. \$wgShowEXIF may be incorrectly set or not checked by an extension." );
 302                 }
 303                 /**
 304                  * exif_read_data() will return false on invalid input, such as
 305                  * when somebody uploads a file called something.jpeg
 306                  * containing random gibberish.
 307                  */
 308                 $this->mRawExifData = $data ? $data : array();
 309                 $this->makeFilteredData();
 310                 $this->collapseData();
 311                 $this->debugFile( __FUNCTION__, false );
 312         }
 313
 314         /**
 315          * Make $this->mFilteredExifData
 316          */
 317         function makeFilteredData() {
 318                 $this->mFilteredExifData = Array();
 319
 320                 foreach ( array_keys( $this->mRawExifData ) as $section ) {
 321                         if ( !in_array( $section, array_keys( $this->mExifTags ) ) ) {
 322                                 $this->debug( $section , __FUNCTION__, "'$section' is not a valid Exif section" );
 323                                 continue;
 324                         }
 325
 326                         foreach ( array_keys( $this->mRawExifData[$section] ) as $tag ) {
 327                                 if ( !in_array( $tag, array_keys( $this->mExifTags[$section] ) ) ) {
 328                                         $this->debug( $tag, __FUNCTION__, "'$tag' is not a valid tag in '$section'" );
 329                                         continue;
 330                                 }
 331
 332                                 $this->mFilteredExifData[$tag] = $this->mRawExifData[$section][$tag];
 333                                 // This is ok, as the tags in the different sections do not conflict.
 334                                 // except in computed and thumbnail section, which we don't use.
 335
 336                                 $value = $this->mRawExifData[$section][$tag];
 337                                 if ( !$this->validate( $section, $tag, $value ) ) {
 338                                         $this->debug( $value, __FUNCTION__, "'$tag' contained invalid data" );
 339                                         unset( $this->mFilteredExifData[$tag] );
 340                                 }
 341                         }
 342                 }
 343         }
 344
 345         /**
 346         * Collapse some fields together.
 347         * This converts some fields from exif form, to a more friendly form.
 348         * For example GPS latitude to a single number.
 349         *
 350         * The rationale behind this is that we're storing data, not presenting to the user
 351         * For example a longitude is a single number describing how far away you are from
 352         * the prime meridian. Well it might be nice to split it up into minutes and seconds
 353         * for the user, it doesn't really make sense to split a single number into 4 parts
 354         * for storage. (degrees, minutes, second, direction vs single floating point number).
 355         *
 356         * Other things this might do (not really sure if they make sense or not):
 357         * Dates -> mediawiki date format.
 358         * convert values that can be in different units to be in one standardized unit.
 359         *
 360         * As an alternative approach, some of this could be done in the validate phase
 361         * if we make up our own types like Exif::DATE.
 362         */
 363         function collapseData( ) {
 364
 365                 $this->exifGPStoNumber( 'GPSLatitude' );
 366                 $this->exifGPStoNumber( 'GPSDestLatitude' );
 367                 $this->exifGPStoNumber( 'GPSLongitude' );
 368                 $this->exifGPStoNumber( 'GPSDestLongitude' );
 369
 370                 if ( isset( $this->mFilteredExifData['GPSAltitude'] ) && isset( $this->mFilteredExifData['GPSAltitudeRef'] ) ) {
 371                         if ( $this->mFilteredExifData['GPSAltitudeRef'] === "\1" ) {
 372                                 $this->mFilteredExifData['GPSAltitude'] *= - 1;
 373                         }
 374                         unset( $this->mFilteredExifData['GPSAltitudeRef'] );
 375                 }
 376
 377                 $this->exifPropToOrd( 'FileSource' );
 378                 $this->exifPropToOrd( 'SceneType' );
 379
 380                 $this->charCodeString( 'UserComment' );
 381                 $this->charCodeString( 'GPSProcessingMethod');
 382                 $this->charCodeString( 'GPSAreaInformation' );
 383
 384                 //ComponentsConfiguration should really be an array instead of a string...
 385                 //This turns a string of binary numbers into an array of numbers.
 386
 387                 if ( isset ( $this->mFilteredExifData['ComponentsConfiguration'] ) ) {
 388                         $val = $this->mFilteredExifData['ComponentsConfiguration'];
 389                         $ccVals = array();
 390                         for ($i = 0; $i < strlen($val); $i++) {
 391                                 $ccVals[$i] = ord( substr($val, $i, 1) );
 392                         }
 393                         $ccVals['_type'] = 'ol'; //this is for formatting later.
 394                         $this->mFilteredExifData['ComponentsConfiguration'] = $ccVals;
 395                 }
 396
 397                 //GPSVersion(ID) is treated as the wrong type by php exif support.
 398                 //Go through each byte turning it into a version string.
 399                 //For example: "\x02\x02\x00\x00" -> "2.2.0.0"
 400
 401                 //Also change exif tag name from GPSVersion (what php exif thinks it is)
 402                 //to GPSVersionID (what the exif standard thinks it is).
 403
 404                 if ( isset ( $this->mFilteredExifData['GPSVersion'] ) ) {
 405                         $val = $this->mFilteredExifData['GPSVersion'];
 406                         $newVal = '';
 407                         for ($i = 0; $i < strlen($val); $i++) {
 408                                 if ( $i !== 0 ) {
 409                                         $newVal .= '.';
 410                                 }
 411                                 $newVal .= ord( substr($val, $i, 1) );
 412                         }
 413                         if ( $this->byteOrder === 'LE' ) {
 414                                 // Need to reverse the string
 415                                 $newVal2 = '';
 416                                 for ( $i = strlen( $newVal ) - 1; $i >= 0; $i-- ) {
 417                                         $newVal2 .= substr( $newVal, $i, 1 );
 418                                 }
 419                                 $this->mFilteredExifData['GPSVersionID'] = $newVal2;
 420                         } else {
 421                                 $this->mFilteredExifData['GPSVersionID'] = $newVal;
 422                         }
 423                         unset( $this->mFilteredExifData['GPSVersion'] );
 424                 }
 425
 426         }
 427         /**
 428         * Do userComment tags and similar. See pg. 34 of exif standard.
 429         * basically first 8 bytes is charset, rest is value.
 430         * This has not been tested on any shift-JIS strings.
 431         * @param $prop String prop name.
 432         */
 433         private function charCodeString ( $prop ) {
 434                 if ( isset( $this->mFilteredExifData[$prop] ) ) {
 435
 436                         if ( strlen($this->mFilteredExifData[$prop]) <= 8 ) {
 437                                 //invalid. Must be at least 9 bytes long.
 438
 439                                 $this->debug( $this->mFilteredExifData[$prop] , __FUNCTION__, false );
 440                                 unset($this->mFilteredExifData[$prop]);
 441                                 return;
 442                         }
 443                         $charCode = substr( $this->mFilteredExifData[$prop], 0, 8);
 444                         $val = substr( $this->mFilteredExifData[$prop], 8);
 445
 446
 447                         switch ($charCode) {
 448                                 case "\x4A\x49\x53\x00\x00\x00\x00\x00":
 449                                         //JIS
 450                                         $charset = "Shift-JIS";
 451                                         break;
 452                                 case "UNICODE\x00":
 453                                         $charset = "UTF-16" . $this->byteOrder;
 454                                         break;
 455                                 default: //ascii or undefined.
 456                                         $charset = "";
 457                                         break;
 458                         }
 459                         // This could possibly check to see if iconv is really installed
 460                         // or if we're using the compatibility wrapper in globalFunctions.php
 461                         if ($charset) {
 462                                 wfSuppressWarnings();
 463                                 $val = iconv($charset, 'UTF-8//IGNORE', $val);
 464                                 wfRestoreWarnings();
 465                         } else {
 466                                 // if valid utf-8, assume that, otherwise assume windows-1252
 467                                 $valCopy = $val;
 468                                 UtfNormal::quickIsNFCVerify( $valCopy ); //validates $valCopy.
 469                                 if ( $valCopy !== $val ) {
 470                                         wfSuppressWarnings();
 471                                         $val = iconv('Windows-1252', 'UTF-8//IGNORE', $val);
 472                                         wfRestoreWarnings();
 473                                 }
 474                         }
 475
 476                         //trim and check to make sure not only whitespace.
 477                         $val = trim($val);
 478                         if ( strlen( $val ) === 0 ) {
 479                                 //only whitespace.
 480                                 $this->debug( $this->mFilteredExifData[$prop] , __FUNCTION__, "$prop: Is only whitespace" );
 481                                 unset($this->mFilteredExifData[$prop]);
 482                                 return;
 483                         }
 484
 485                         //all's good.
 486                         $this->mFilteredExifData[$prop] = $val;
 487                 }
 488         }
 489         /**
 490         * Convert an Exif::UNDEFINED from a raw binary string
 491         * to its value. This is sometimes needed depending on
 492         * the type of UNDEFINED field
 493         * @param $prop String name of property
 494         */
 495         private function exifPropToOrd ( $prop ) {
 496                 if ( isset( $this->mFilteredExifData[$prop] ) ) {
 497                         $this->mFilteredExifData[$prop] = ord( $this->mFilteredExifData[$prop] );
 498                 }
 499         }
 500         /**
 501         * Convert gps in exif form to a single floating point number
 502         * for example 10 degress 20`40`` S -> -10.34444
 503         * @param String $prop a gps coordinate exif tag name (like GPSLongitude)
 504         */
 505         private function exifGPStoNumber ( $prop ) {
 506                 $loc =& $this->mFilteredExifData[$prop];
 507                 $dir =& $this->mFilteredExifData[$prop . 'Ref'];
 508                 $res = false;
 509
 510                 if ( isset( $loc ) && isset( $dir ) && ( $dir === 'N' || $dir === 'S' || $dir === 'E' || $dir === 'W' ) ) {
 511                         list( $num, $denom ) = explode( '/', $loc[0] );
 512                         $res = $num / $denom;
 513                         list( $num, $denom ) = explode( '/', $loc[1] );
 514                         $res += ( $num / $denom ) * ( 1 / 60 );
 515                         list( $num, $denom ) = explode( '/', $loc[2] );
 516                         $res += ( $num / $denom ) * ( 1 / 3600 );
 517
 518                         if ( $dir === 'S' || $dir === 'W' ) {
 519                                 $res *= - 1; // make negative
 520                         }
 521                 }
 522
 523                 // update the exif records.
 524
 525                 if ( $res !== false ) { // using !== as $res could potentially be 0
 526                         $this->mFilteredExifData[$prop] = $res;
 527                         unset( $this->mFilteredExifData[$prop . 'Ref'] );
 528                 } else { // if invalid
 529                         unset( $this->mFilteredExifData[$prop] );
 530                         unset( $this->mFilteredExifData[$prop . 'Ref'] );
 531                 }
 532         }
 533
 534         /**
 535          * Use FormatMetadata to create formatted values for display to user
 536          * (is this ever used?)
 537          *
 538          * @deprecated since 1.18
 539          */
 540         function makeFormattedData( ) {
 541                 wfDeprecated( __METHOD__, '1.18' );
 542                 $this->mFormattedExifData = FormatMetadata::getFormattedData(
 543                         $this->mFilteredExifData );
 544         }
 545         /**#@-*/
 546
 547         /**#@+
 548          * @return array
 549          */
 550         /**
 551          * Get $this->mRawExifData
 552          * @return array
 553          */
 554         function getData() {
 555                 return $this->mRawExifData;
 556         }
 557
 558         /**
 559          * Get $this->mFilteredExifData
 560          */
 561         function getFilteredData() {
 562                 return $this->mFilteredExifData;
 563         }
 564
 565         /**
 566          * Get $this->mFormattedExifData
 567          *
 568          * This returns the data for display to user.
 569          * Its unclear if this is ever used.
 570          *
 571          * @deprecated since 1.18
 572          */
 573         function getFormattedData() {
 574                 wfDeprecated( __METHOD__, '1.18' );
 575                 if (!$this->mFormattedExifData) {
 576                         $this->makeFormattedData();
 577                 }
 578                 return $this->mFormattedExifData;
 579         }
 580         /**#@-*/
 581
 582         /**
 583          * The version of the output format
 584          *
 585          * Before the actual metadata information is saved in the database we
 586          * strip some of it since we don't want to save things like thumbnails
 587          * which usually accompany Exif data. This value gets saved in the
 588          * database along with the actual Exif data, and if the version in the
 589          * database doesn't equal the value returned by this function the Exif
 590          * data is regenerated.
 591          *
 592          * @return int
 593          */
 594         public static function version() {
 595                 return 2; // We don't need no bloddy constants!
 596         }
 597
 598         /**#@+
 599          * Validates if a tag value is of the type it should be according to the Exif spec
 600          *
 601          * @private
 602          *
 603          * @param $in Mixed: the input value to check
 604          * @return bool
 605          */
 606         private function isByte( $in ) {
 607                 if ( !is_array( $in ) && sprintf('%d', $in) == $in && $in >= 0 && $in <= 255 ) {
 608                         $this->debug( $in, __FUNCTION__, true );
 609                         return true;
 610                 } else {
 611                         $this->debug( $in, __FUNCTION__, false );
 612                         return false;
 613                 }
 614         }
 615
 616         /**
 617          * @param $in
 618          * @return bool
 619          */
 620         private function isASCII( $in ) {
 621                 if ( is_array( $in ) ) {
 622                         return false;
 623                 }
 624
 625                 if ( preg_match( "/[^\x0a\x20-\x7e]/", $in ) ) {
 626                         $this->debug( $in, __FUNCTION__, 'found a character not in our whitelist' );
 627                         return false;
 628                 }
 629
 630                 if ( preg_match( '/^\s*$/', $in ) ) {
 631                         $this->debug( $in, __FUNCTION__, 'input consisted solely of whitespace' );
 632                         return false;
 633                 }
 634
 635                 return true;
 636         }
 637
 638         /**
 639          * @param $in
 640          * @return bool
 641          */
 642         private function isShort( $in ) {
 643                 if ( !is_array( $in ) && sprintf('%d', $in) == $in && $in >= 0 && $in <= 65536 ) {
 644                         $this->debug( $in, __FUNCTION__, true );
 645                         return true;
 646                 } else {
 647                         $this->debug( $in, __FUNCTION__, false );
 648                         return false;
 649                 }
 650         }
 651
 652         /**
 653          * @param $in
 654          * @return bool
 655          */
 656         private function isLong( $in ) {
 657                 if ( !is_array( $in ) && sprintf('%d', $in) == $in && $in >= 0 && $in <= 4294967296 ) {
 658                         $this->debug( $in, __FUNCTION__, true );
 659                         return true;
 660                 } else {
 661                         $this->debug( $in, __FUNCTION__, false );
 662                         return false;
 663                 }
 664         }
 665
 666         /**
 667          * @param $in
 668          * @return bool
 669          */
 670         private function isRational( $in ) {
 671                 $m = array();
 672                 if ( !is_array( $in ) && @preg_match( '/^(\d+)\/(\d+[1-9]|[1-9]\d*)$/', $in, $m ) ) { # Avoid division by zero
 673                         return $this->isLong( $m[1] ) && $this->isLong( $m[2] );
 674                 } else {
 675                         $this->debug( $in, __FUNCTION__, 'fed a non-fraction value' );
 676                         return false;
 677                 }
 678         }
 679
 680         /**
 681          * @param $in
 682          * @return bool
 683          */
 684         private function isUndefined( $in ) {
 685                 $this->debug( $in, __FUNCTION__, true );
 686                 return true;
 687         }
 688
 689         /**
 690          * @param $in
 691          * @return bool
 692          */
 693         private function isSlong( $in ) {
 694                 if ( $this->isLong( abs( $in ) ) ) {
 695                         $this->debug( $in, __FUNCTION__, true );
 696                         return true;
 697                 } else {
 698                         $this->debug( $in, __FUNCTION__, false );
 699                         return false;
 700                 }
 701         }
 702
 703         /**
 704          * @param $in
 705          * @return bool
 706          */
 707         private function isSrational( $in ) {
 708                 $m = array();
 709                 if ( !is_array( $in ) && preg_match( '/^(-?\d+)\/(\d+[1-9]|[1-9]\d*)$/', $in, $m ) ) { # Avoid division by zero
 710                         return $this->isSlong( $m[0] ) && $this->isSlong( $m[1] );
 711                 } else {
 712                         $this->debug( $in, __FUNCTION__, 'fed a non-fraction value' );
 713                         return false;
 714                 }
 715         }
 716         /**#@-*/
 717
 718         /**
 719          * Validates if a tag has a legal value according to the Exif spec
 720          *
 721          * @private
 722          * @param $section String: section where tag is located.
 723          * @param $tag String: the tag to check.
 724          * @param $val Mixed: the value of the tag.
 725          * @param $recursive Boolean: true if called recursively for array types.
 726          * @return bool
 727          */
 728         private function validate( $section, $tag, $val, $recursive = false ) {
 729                 $debug = "tag is '$tag'";
 730                 $etype = $this->mExifTags[$section][$tag];
 731                 $ecount = 1;
 732                 if( is_array( $etype ) ) {
 733                         list( $etype, $ecount ) = $etype;
 734                         if ( $recursive )
 735                                 $ecount = 1; // checking individual elements
 736                 }
 737                 $count = count( $val );
 738                 if( $ecount != $count ) {
 739                         $this->debug( $val, __FUNCTION__, "Expected $ecount elements for $tag but got $count" );
 740                         return false;
 741                 }
 742                 if( $count > 1 ) {
 743                         foreach( $val as $v ) {
 744                                 if( !$this->validate( $section, $tag, $v, true ) ) {
 745                                         return false;
 746                                 }
 747                         }
 748                         return true;
 749                 }
 750                 // Does not work if not typecast
 751                 switch( (string)$etype ) {
 752                         case (string)Exif::BYTE:
 753                                 $this->debug( $val, __FUNCTION__, $debug );
 754                                 return $this->isByte( $val );
 755                         case (string)Exif::ASCII:
 756                                 $this->debug( $val, __FUNCTION__, $debug );
 757                                 return $this->isASCII( $val );
 758                         case (string)Exif::SHORT:
 759                                 $this->debug( $val, __FUNCTION__, $debug );
 760                                 return $this->isShort( $val );
 761                         case (string)Exif::LONG:
 762                                 $this->debug( $val, __FUNCTION__, $debug );
 763                                 return $this->isLong( $val );
 764                         case (string)Exif::RATIONAL:
 765                                 $this->debug( $val, __FUNCTION__, $debug );
 766                                 return $this->isRational( $val );
 767                         case (string)Exif::UNDEFINED:
 768                                 $this->debug( $val, __FUNCTION__, $debug );
 769                                 return $this->isUndefined( $val );
 770                         case (string)Exif::SLONG:
 771                                 $this->debug( $val, __FUNCTION__, $debug );
 772                                 return $this->isSlong( $val );
 773                         case (string)Exif::SRATIONAL:
 774                                 $this->debug( $val, __FUNCTION__, $debug );
 775                                 return $this->isSrational( $val );
 776                         case (string)Exif::SHORT.','.Exif::LONG:
 777                                 $this->debug( $val, __FUNCTION__, $debug );
 778                                 return $this->isShort( $val ) || $this->isLong( $val );
 779                         case (string)Exif::IGNORE:
 780                                 $this->debug( $val, __FUNCTION__, $debug );
 781                                 return false;
 782                         default:
 783                                 $this->debug( $val, __FUNCTION__, "The tag '$tag' is unknown" );
 784                                 return false;
 785                 }
 786         }
 787
 788         /**
 789          * Convenience function for debugging output
 790          *
 791          * @private
 792          *
 793          * @param $in Mixed:
 794          * @param $fname String:
 795          * @param $action Mixed: , default NULL.
 796          */
 797         private function debug( $in, $fname, $action = null ) {
 798                 if ( !$this->log ) {
 799                         return;
 800                 }
 801                 $type = gettype( $in );
 802                 $class = ucfirst( __CLASS__ );
 803                 if ( $type === 'array' ) {
 804                         $in = print_r( $in, true );
 805                 }
 806
 807                 if ( $action === true ) {
 808                         wfDebugLog( $this->log, "$class::$fname: accepted: '$in' (type: $type)\n");
 809                 } elseif ( $action === false ) {
 810                         wfDebugLog( $this->log, "$class::$fname: rejected: '$in' (type: $type)\n");
 811                 } elseif ( $action === null ) {
 812                         wfDebugLog( $this->log, "$class::$fname: input was: '$in' (type: $type)\n");
 813                 } else {
 814                         wfDebugLog( $this->log, "$class::$fname: $action (type: $type; content: '$in')\n");
 815                 }
 816         }
 817
 818         /**
 819          * Convenience function for debugging output
 820          *
 821          * @private
 822          *
 823          * @param $fname String: the name of the function calling this function
 824          * @param $io Boolean: Specify whether we're beginning or ending
 825          */
 826         private function debugFile( $fname, $io ) {
 827                 if ( !$this->log ) {
 828                         return;
 829                 }
 830                 $class = ucfirst( __CLASS__ );
 831                 if ( $io ) {
 832                         wfDebugLog( $this->log, "$class::$fname: begin processing: '{$this->basename}'\n" );
 833                 } else {
 834                         wfDebugLog( $this->log, "$class::$fname: end processing: '{$this->basename}'\n" );
 835                 }
 836         }
 837 }
 838