includes/media/XMPValidate.php

   1 <?php
   2 /**
   3  * Methods for validating XMP properties.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup Media
  22  */
  23
  24 /**
  25  * This contains some static methods for
  26  * validating XMP properties. See XMPInfo and XMPReader classes.
  27  *
  28  * Each of these functions take the same parameters
  29  * * an info array which is a subset of the XMPInfo::items array
  30  * * A value (passed as reference) to validate. This can be either a
  31  *      simple value or an array
  32  * * A boolean to determine if this is validating a simple or complex values
  33  *
  34  * It should be noted that when an array is being validated, typically the validation
  35  * function is called once for each value, and then once at the end for the entire array.
  36  *
  37  * These validation functions can also be used to modify the data. See the gps and flash one's
  38  * for example.
  39  *
  40  * @see http://www.adobe.com/devnet/xmp/pdfs/XMPSpecificationPart1.pdf starting at pg 28
  41  * @see http://www.adobe.com/devnet/xmp/pdfs/XMPSpecificationPart2.pdf starting at pg 11
  42  */
  43 class XMPValidate {
  44         /**
  45          * function to validate boolean properties ( True or False )
  46          *
  47          * @param array $info information about current property
  48          * @param &$val Mixed current value to validate
  49          * @param $standalone Boolean if this is a simple property or array
  50          */
  51         public static function validateBoolean( $info, &$val, $standalone ) {
  52                 if ( !$standalone ) {
  53                         // this only validates standalone properties, not arrays, etc
  54                         return;
  55                 }
  56                 if ( $val !== 'True' && $val !== 'False' ) {
  57                         wfDebugLog( 'XMP', __METHOD__ . " Expected True or False but got $val" );
  58                         $val = null;
  59                 }
  60
  61         }
  62
  63         /**
  64          * function to validate rational properties ( 12/10 )
  65          *
  66          * @param array $info information about current property
  67          * @param &$val Mixed current value to validate
  68          * @param $standalone Boolean if this is a simple property or array
  69          */
  70         public static function validateRational( $info, &$val, $standalone ) {
  71                 if ( !$standalone ) {
  72                         // this only validates standalone properties, not arrays, etc
  73                         return;
  74                 }
  75                 if ( !preg_match( '/^(?:-?\d+)\/(?:\d+[1-9]|[1-9]\d*)$/D', $val ) ) {
  76                         wfDebugLog( 'XMP', __METHOD__ . " Expected rational but got $val" );
  77                         $val = null;
  78                 }
  79
  80         }
  81
  82         /**
  83          * function to validate rating properties -1, 0-5
  84          *
  85          * if its outside of range put it into range.
  86          *
  87          * @see MWG spec
  88          * @param array $info information about current property
  89          * @param &$val Mixed current value to validate
  90          * @param $standalone Boolean if this is a simple property or array
  91          */
  92         public static function validateRating( $info, &$val, $standalone ) {
  93                 if ( !$standalone ) {
  94                         // this only validates standalone properties, not arrays, etc
  95                         return;
  96                 }
  97                 if ( !preg_match( '/^[-+]?\d*(?:\.?\d*)$/D', $val )
  98                         || !is_numeric( $val )
  99                 ) {
 100                         wfDebugLog( 'XMP', __METHOD__ . " Expected rating but got $val" );
 101                         $val = null;
 102                         return;
 103                 } else {
 104                         $nVal = (float) $val;
 105                         if ( $nVal < 0 ) {
 106                                 // We do < 0 here instead of < -1 here, since
 107                                 // the values between 0 and -1 are also illegal
 108                                 // as -1 is meant as a special reject rating.
 109                                 wfDebugLog( 'XMP', __METHOD__ . " Rating too low, setting to -1 (Rejected)" );
 110                                 $val = '-1';
 111                                 return;
 112                         }
 113                         if ( $nVal > 5 ) {
 114                                 wfDebugLog( 'XMP', __METHOD__ . " Rating too high, setting to 5" );
 115                                 $val = '5';
 116                                 return;
 117                         }
 118                 }
 119         }
 120
 121         /**
 122          * function to validate integers
 123          *
 124          * @param array $info information about current property
 125          * @param &$val Mixed current value to validate
 126          * @param $standalone Boolean if this is a simple property or array
 127          */
 128         public static function validateInteger( $info, &$val, $standalone ) {
 129                 if ( !$standalone ) {
 130                         // this only validates standalone properties, not arrays, etc
 131                         return;
 132                 }
 133                 if ( !preg_match( '/^[-+]?\d+$/D', $val ) ) {
 134                         wfDebugLog( 'XMP', __METHOD__ . " Expected integer but got $val" );
 135                         $val = null;
 136                 }
 137
 138         }
 139
 140         /**
 141          * function to validate properties with a fixed number of allowed
 142          * choices. (closed choice)
 143          *
 144          * @param array $info information about current property
 145          * @param &$val Mixed current value to validate
 146          * @param $standalone Boolean if this is a simple property or array
 147          */
 148         public static function validateClosed( $info, &$val, $standalone ) {
 149                 if ( !$standalone ) {
 150                         // this only validates standalone properties, not arrays, etc
 151                         return;
 152                 }
 153
 154                 //check if its in a numeric range
 155                 $inRange = false;
 156                 if ( isset( $info['rangeLow'] )
 157                         && isset( $info['rangeHigh'] )
 158                         && is_numeric( $val )
 159                         && ( intval( $val ) <= $info['rangeHigh'] )
 160                         && ( intval( $val ) >= $info['rangeLow'] )
 161                 ) {
 162                         $inRange = true;
 163                 }
 164
 165                 if ( !isset( $info['choices'][$val] ) && !$inRange ) {
 166                         wfDebugLog( 'XMP', __METHOD__ . " Expected closed choice, but got $val" );
 167                         $val = null;
 168                 }
 169         }
 170
 171         /**
 172          * function to validate and modify flash structure
 173          *
 174          * @param array $info information about current property
 175          * @param &$val Mixed current value to validate
 176          * @param $standalone Boolean if this is a simple property or array
 177          */
 178         public static function validateFlash( $info, &$val, $standalone ) {
 179                 if ( $standalone ) {
 180                         // this only validates flash structs, not individual properties
 181                         return;
 182                 }
 183                 if ( !( isset( $val['Fired'] )
 184                         && isset( $val['Function'] )
 185                         && isset( $val['Mode'] )
 186                         && isset( $val['RedEyeMode'] )
 187                         && isset( $val['Return'] )
 188                 ) ) {
 189                         wfDebugLog( 'XMP', __METHOD__ . " Flash structure did not have all the required components" );
 190                         $val = null;
 191                 } else {
 192                         $val = ( "\0" | ( $val['Fired'] === 'True' )
 193                                 | ( intval( $val['Return'] ) << 1 )
 194                                 | ( intval( $val['Mode'] ) << 3 )
 195                                 | ( ( $val['Function'] === 'True' ) << 5 )
 196                                 | ( ( $val['RedEyeMode'] === 'True' ) << 6 ) );
 197                 }
 198         }
 199
 200         /**
 201          * function to validate LangCode properties ( en-GB, etc )
 202          *
 203          * This is just a naive check to make sure it somewhat looks like a lang code.
 204          *
 205          * @see rfc 3066
 206          * @see http://www.adobe.com/devnet/xmp/pdfs/XMPSpecificationPart1.pdf page 30 (section 8.2.2.5)
 207          *
 208          * @param array $info information about current property
 209          * @param &$val Mixed current value to validate
 210          * @param $standalone Boolean if this is a simple property or array
 211          */
 212         public static function validateLangCode( $info, &$val, $standalone ) {
 213                 if ( !$standalone ) {
 214                         // this only validates standalone properties, not arrays, etc
 215                         return;
 216                 }
 217                 if ( !preg_match( '/^[-A-Za-z0-9]{2,}$/D', $val ) ) {
 218                         //this is a rather naive check.
 219                         wfDebugLog( 'XMP', __METHOD__ . " Expected Lang code but got $val" );
 220                         $val = null;
 221                 }
 222
 223         }
 224
 225         /**
 226          * function to validate date properties, and convert to (partial) Exif format.
 227          *
 228          * Dates can be one of the following formats:
 229          * YYYY
 230          * YYYY-MM
 231          * YYYY-MM-DD
 232          * YYYY-MM-DDThh:mmTZD
 233          * YYYY-MM-DDThh:mm:ssTZD
 234          * YYYY-MM-DDThh:mm:ss.sTZD
 235          *
 236          * @param array $info information about current property
 237          * @param &$val Mixed current value to validate. Converts to TS_EXIF as a side-effect.
 238          *      in cases where there's only a partial date, it will give things like
 239          *      2011:04.
 240          * @param $standalone Boolean if this is a simple property or array
 241          */
 242         public static function validateDate( $info, &$val, $standalone ) {
 243                 if ( !$standalone ) {
 244                         // this only validates standalone properties, not arrays, etc
 245                         return;
 246                 }
 247                 $res = array();
 248                 if ( !preg_match(
 249                         /* ahh! scary regex... */
 250                         '/^([0-3]\d{3})(?:-([01]\d)(?:-([0-3]\d)(?:T([0-2]\d):([0-6]\d)(?::([0-6]\d)(?:\.\d+)?)?([-+]\d{2}:\d{2}|Z)?)?)?)?$/D',
 251                         $val, $res )
 252                 ) {
 253                         wfDebugLog( 'XMP', __METHOD__ . " Expected date but got $val" );
 254                         $val = null;
 255                 } else {
 256                         /*
 257                          * $res is formatted as follows:
 258                          * 0 -> full date.
 259                          * 1 -> year, 2-> month, 3-> day, 4-> hour, 5-> minute, 6->second
 260                          * 7-> Timezone specifier (Z or something like +12:30 )
 261                          * many parts are optional, some aren't. For example if you specify
 262                          * minute, you must specify hour, day, month, and year but not second or TZ.
 263                          */
 264
 265                         /*
 266                          * First of all, if year = 0000, Something is wrongish,
 267                          * so don't extract. This seems to happen when
 268                          * some programs convert between metadata formats.
 269                          */
 270                         if ( $res[1] === '0000' ) {
 271                                 wfDebugLog( 'XMP', __METHOD__ . " Invalid date (year 0): $val" );
 272                                 $val = null;
 273                                 return;
 274                         }
 275
 276                         if ( !isset( $res[4] ) ) { //hour
 277                                 //just have the year month day (if that)
 278                                 $val = $res[1];
 279                                 if ( isset( $res[2] ) ) {
 280                                         $val .= ':' . $res[2];
 281                                 }
 282                                 if ( isset( $res[3] ) ) {
 283                                         $val .= ':' . $res[3];
 284                                 }
 285                                 return;
 286                         }
 287
 288                         if ( !isset( $res[7] ) || $res[7] === 'Z' ) {
 289                                 //if hour is set, then minute must also be or regex above will fail.
 290                                 $val = $res[1] . ':' . $res[2] . ':' . $res[3]
 291                                         . ' ' . $res[4] . ':' . $res[5];
 292                                 if ( isset( $res[6] ) && $res[6] !== '' ) {
 293                                         $val .= ':' . $res[6];
 294                                 }
 295                                 return;
 296                         }
 297
 298                         // Extra check for empty string necessary due to TZ but no second case.
 299                         $stripSeconds = false;
 300                         if ( !isset( $res[6] ) || $res[6] === '' ) {
 301                                 $res[6] = '00';
 302                                 $stripSeconds = true;
 303                         }
 304
 305                         // Do timezone processing. We've already done the case that tz = Z.
 306
 307                         // We know that if we got to this step, year, month day hour and min must be set
 308                         // by virtue of regex not failing.
 309
 310                         $unix = wfTimestamp( TS_UNIX, $res[1] . $res[2] . $res[3] . $res[4] . $res[5] . $res[6] );
 311                         $offset = intval( substr( $res[7], 1, 2 ) ) * 60 * 60;
 312                         $offset += intval( substr( $res[7], 4, 2 ) ) * 60;
 313                         if ( substr( $res[7], 0, 1 ) === '-' ) {
 314                                 $offset = -$offset;
 315                         }
 316                         $val = wfTimestamp( TS_EXIF, $unix + $offset );
 317
 318                         if ( $stripSeconds ) {
 319                                 // If seconds weren't specified, remove the trailing ':00'.
 320                                 $val = substr( $val, 0, -3 );
 321                         }
 322                 }
 323
 324         }
 325
 326         /** function to validate, and more importantly
 327          * translate the XMP DMS form of gps coords to
 328          * the decimal form we use.
 329          *
 330          * @see http://www.adobe.com/devnet/xmp/pdfs/XMPSpecificationPart2.pdf
 331          *        section 1.2.7.4 on page 23
 332          *
 333          * @param array $info unused (info about prop)
 334          * @param &$val String GPS string in either DDD,MM,SSk or
 335          *           or DDD,MM.mmk form
 336          * @param $standalone Boolean if its a simple prop (should always be true)
 337          */
 338         public static function validateGPS( $info, &$val, $standalone ) {
 339                 if ( !$standalone ) {
 340                         return;
 341                 }
 342
 343                 $m = array();
 344                 if ( preg_match(
 345                         '/(\d{1,3}),(\d{1,2}),(\d{1,2})([NWSE])/D',
 346                         $val, $m )
 347                 ) {
 348                         $coord = intval( $m[1] );
 349                         $coord += intval( $m[2] ) * ( 1 / 60 );
 350                         $coord += intval( $m[3] ) * ( 1 / 3600 );
 351                         if ( $m[4] === 'S' || $m[4] === 'W' ) {
 352                                 $coord = -$coord;
 353                         }
 354                         $val = $coord;
 355                         return;
 356                 } elseif ( preg_match(
 357                         '/(\d{1,3}),(\d{1,2}(?:.\d*)?)([NWSE])/D',
 358                         $val, $m )
 359                 ) {
 360                         $coord = intval( $m[1] );
 361                         $coord += floatval( $m[2] ) * ( 1 / 60 );
 362                         if ( $m[3] === 'S' || $m[3] === 'W' ) {
 363                                 $coord = -$coord;
 364                         }
 365                         $val = $coord;
 366                         return;
 367
 368                 } else {
 369                         wfDebugLog( 'XMP', __METHOD__
 370                                 . " Expected GPSCoordinate, but got $val." );
 371                         $val = null;
 372                         return;
 373                 }
 374         }
 375 }