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