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 use Psr\Log\LoggerInterface;
  25 use Psr\Log\LoggerAwareInterface;
  26
  27 /**
  28  * This contains some static methods for
  29  * validating XMP properties. See XMPInfo and XMPReader classes.
  30  *
  31  * Each of these functions take the same parameters
  32  * * an info array which is a subset of the XMPInfo::items array
  33  * * A value (passed as reference) to validate. This can be either a
  34  *    simple value or an array
  35  * * A boolean to determine if this is validating a simple or complex values
  36  *
  37  * It should be noted that when an array is being validated, typically the validation
  38  * function is called once for each value, and then once at the end for the entire array.
  39  *
  40  * These validation functions can also be used to modify the data. See the gps and flash one's
  41  * for example.
  42  *
  43  * @see http://www.adobe.com/devnet/xmp/pdfs/XMPSpecificationPart1.pdf starting at pg 28
  44  * @see http://www.adobe.com/devnet/xmp/pdfs/XMPSpecificationPart2.pdf starting at pg 11
  45  */
  46 class XMPValidate implements LoggerAwareInterface {
  47
  48         /**
  49          * @var LoggerInterface
  50          */
  51         private $logger;
  52
  53         public function __construct( LoggerInterface $logger ) {
  54                 $this->setLogger( $logger );
  55         }
  56
  57         public function setLogger( LoggerInterface $logger ) {
  58                 $this->logger = $logger;
  59         }
  60         /**
  61          * Function to validate boolean properties ( True or False )
  62          *
  63          * @param array $info Information about current property
  64          * @param mixed &$val Current value to validate
  65          * @param bool $standalone If this is a simple property or array
  66          */
  67         public function validateBoolean( $info, &$val, $standalone ) {
  68                 if ( !$standalone ) {
  69                         // this only validates standalone properties, not arrays, etc
  70                         return;
  71                 }
  72                 if ( $val !== 'True' && $val !== 'False' ) {
  73                         $this->logger->info( __METHOD__ . " Expected True or False but got $val" );
  74                         $val = null;
  75                 }
  76         }
  77
  78         /**
  79          * function to validate rational properties ( 12/10 )
  80          *
  81          * @param array $info Information about current property
  82          * @param mixed &$val Current value to validate
  83          * @param bool $standalone If this is a simple property or array
  84          */
  85         public function validateRational( $info, &$val, $standalone ) {
  86                 if ( !$standalone ) {
  87                         // this only validates standalone properties, not arrays, etc
  88                         return;
  89                 }
  90                 if ( !preg_match( '/^(?:-?\d+)\/(?:\d+[1-9]|[1-9]\d*)$/D', $val ) ) {
  91                         $this->logger->info( __METHOD__ . " Expected rational but got $val" );
  92                         $val = null;
  93                 }
  94         }
  95
  96         /**
  97          * function to validate rating properties -1, 0-5
  98          *
  99          * if its outside of range put it into range.
 100          *
 101          * @see MWG spec
 102          * @param array $info Information about current property
 103          * @param mixed &$val Current value to validate
 104          * @param bool $standalone If this is a simple property or array
 105          */
 106         public function validateRating( $info, &$val, $standalone ) {
 107                 if ( !$standalone ) {
 108                         // this only validates standalone properties, not arrays, etc
 109                         return;
 110                 }
 111                 if ( !preg_match( '/^[-+]?\d*(?:\.?\d*)$/D', $val )
 112                         || !is_numeric( $val )
 113                 ) {
 114                         $this->logger->info( __METHOD__ . " Expected rating but got $val" );
 115                         $val = null;
 116
 117                         return;
 118                 } else {
 119                         $nVal = (float)$val;
 120                         if ( $nVal < 0 ) {
 121                                 // We do < 0 here instead of < -1 here, since
 122                                 // the values between 0 and -1 are also illegal
 123                                 // as -1 is meant as a special reject rating.
 124                                 $this->logger->info( __METHOD__ . " Rating too low, setting to -1 (Rejected)" );
 125                                 $val = '-1';
 126
 127                                 return;
 128                         }
 129                         if ( $nVal > 5 ) {
 130                                 $this->logger->info( __METHOD__ . " Rating too high, setting to 5" );
 131                                 $val = '5';
 132
 133                                 return;
 134                         }
 135                 }
 136         }
 137
 138         /**
 139          * function to validate integers
 140          *
 141          * @param array $info Information about current property
 142          * @param mixed &$val Current value to validate
 143          * @param bool $standalone If this is a simple property or array
 144          */
 145         public function validateInteger( $info, &$val, $standalone ) {
 146                 if ( !$standalone ) {
 147                         // this only validates standalone properties, not arrays, etc
 148                         return;
 149                 }
 150                 if ( !preg_match( '/^[-+]?\d+$/D', $val ) ) {
 151                         $this->logger->info( __METHOD__ . " Expected integer but got $val" );
 152                         $val = null;
 153                 }
 154         }
 155
 156         /**
 157          * function to validate properties with a fixed number of allowed
 158          * choices. (closed choice)
 159          *
 160          * @param array $info Information about current property
 161          * @param mixed &$val Current value to validate
 162          * @param bool $standalone If this is a simple property or array
 163          */
 164         public function validateClosed( $info, &$val, $standalone ) {
 165                 if ( !$standalone ) {
 166                         // this only validates standalone properties, not arrays, etc
 167                         return;
 168                 }
 169
 170                 // check if its in a numeric range
 171                 $inRange = false;
 172                 if ( isset( $info['rangeLow'] )
 173                         && isset( $info['rangeHigh'] )
 174                         && is_numeric( $val )
 175                         && ( intval( $val ) <= $info['rangeHigh'] )
 176                         && ( intval( $val ) >= $info['rangeLow'] )
 177                 ) {
 178                         $inRange = true;
 179                 }
 180
 181                 if ( !isset( $info['choices'][$val] ) && !$inRange ) {
 182                         $this->logger->info( __METHOD__ . " Expected closed choice, but got $val" );
 183                         $val = null;
 184                 }
 185         }
 186
 187         /**
 188          * function to validate and modify flash structure
 189          *
 190          * @param array $info Information about current property
 191          * @param mixed &$val Current value to validate
 192          * @param bool $standalone If this is a simple property or array
 193          */
 194         public function validateFlash( $info, &$val, $standalone ) {
 195                 if ( $standalone ) {
 196                         // this only validates flash structs, not individual properties
 197                         return;
 198                 }
 199                 if ( !( isset( $val['Fired'] )
 200                         && isset( $val['Function'] )
 201                         && isset( $val['Mode'] )
 202                         && isset( $val['RedEyeMode'] )
 203                         && isset( $val['Return'] )
 204                 ) ) {
 205                         $this->logger->info( __METHOD__ . " Flash structure did not have all the required components" );
 206                         $val = null;
 207                 } else {
 208                         $val = ( "\0" | ( $val['Fired'] === 'True' )
 209                                 | ( intval( $val['Return'] ) << 1 )
 210                                 | ( intval( $val['Mode'] ) << 3 )
 211                                 | ( ( $val['Function'] === 'True' ) << 5 )
 212                                 | ( ( $val['RedEyeMode'] === 'True' ) << 6 ) );
 213                 }
 214         }
 215
 216         /**
 217          * function to validate LangCode properties ( en-GB, etc )
 218          *
 219          * This is just a naive check to make sure it somewhat looks like a lang code.
 220          *
 221          * @see BCP 47
 222          * @see https://wwwimages2.adobe.com/content/dam/Adobe/en/devnet/xmp/pdfs/
 223          *      XMP%20SDK%20Release%20cc-2014-12/XMPSpecificationPart1.pdf page 22 (section 8.2.2.4)
 224          *
 225          * @param array $info Information about current property
 226          * @param mixed &$val Current value to validate
 227          * @param bool $standalone If this is a simple property or array
 228          */
 229         public function validateLangCode( $info, &$val, $standalone ) {
 230                 if ( !$standalone ) {
 231                         // this only validates standalone properties, not arrays, etc
 232                         return;
 233                 }
 234                 if ( !preg_match( '/^[-A-Za-z0-9]{2,}$/D', $val ) ) {
 235                         // this is a rather naive check.
 236                         $this->logger->info( __METHOD__ . " Expected Lang code but got $val" );
 237                         $val = null;
 238                 }
 239         }
 240
 241         /**
 242          * function to validate date properties, and convert to (partial) Exif format.
 243          *
 244          * Dates can be one of the following formats:
 245          * YYYY
 246          * YYYY-MM
 247          * YYYY-MM-DD
 248          * YYYY-MM-DDThh:mmTZD
 249          * YYYY-MM-DDThh:mm:ssTZD
 250          * YYYY-MM-DDThh:mm:ss.sTZD
 251          *
 252          * @param array $info Information about current property
 253          * @param mixed &$val Current value to validate. Converts to TS_EXIF as a side-effect.
 254          *    in cases where there's only a partial date, it will give things like
 255          *    2011:04.
 256          * @param bool $standalone If this is a simple property or array
 257          */
 258         public function validateDate( $info, &$val, $standalone ) {
 259                 if ( !$standalone ) {
 260                         // this only validates standalone properties, not arrays, etc
 261                         return;
 262                 }
 263                 $res = [];
 264                 // @codingStandardsIgnoreStart Long line that cannot be broken
 265                 if ( !preg_match(
 266                         /* ahh! scary regex... */
 267                         '/^([0-3]\d{3})(?:-([01]\d)(?:-([0-3]\d)(?:T([0-2]\d):([0-6]\d)(?::([0-6]\d)(?:\.\d+)?)?([-+]\d{2}:\d{2}|Z)?)?)?)?$/D',
 268                         $val, $res )
 269                 ) {
 270                         // @codingStandardsIgnoreEnd
 271
 272                         $this->logger->info( __METHOD__ . " Expected date but got $val" );
 273                         $val = null;
 274                 } else {
 275                         /*
 276                          * $res is formatted as follows:
 277                          * 0 -> full date.
 278                          * 1 -> year, 2-> month, 3-> day, 4-> hour, 5-> minute, 6->second
 279                          * 7-> Timezone specifier (Z or something like +12:30 )
 280                          * many parts are optional, some aren't. For example if you specify
 281                          * minute, you must specify hour, day, month, and year but not second or TZ.
 282                          */
 283
 284                         /*
 285                          * First of all, if year = 0000, Something is wrongish,
 286                          * so don't extract. This seems to happen when
 287                          * some programs convert between metadata formats.
 288                          */
 289                         if ( $res[1] === '0000' ) {
 290                                 $this->logger->info( __METHOD__ . " Invalid date (year 0): $val" );
 291                                 $val = null;
 292
 293                                 return;
 294                         }
 295
 296                         if ( !isset( $res[4] ) ) { // hour
 297                                 // just have the year month day (if that)
 298                                 $val = $res[1];
 299                                 if ( isset( $res[2] ) ) {
 300                                         $val .= ':' . $res[2];
 301                                 }
 302                                 if ( isset( $res[3] ) ) {
 303                                         $val .= ':' . $res[3];
 304                                 }
 305
 306                                 return;
 307                         }
 308
 309                         if ( !isset( $res[7] ) || $res[7] === 'Z' ) {
 310                                 // if hour is set, then minute must also be or regex above will fail.
 311                                 $val = $res[1] . ':' . $res[2] . ':' . $res[3]
 312                                         . ' ' . $res[4] . ':' . $res[5];
 313                                 if ( isset( $res[6] ) && $res[6] !== '' ) {
 314                                         $val .= ':' . $res[6];
 315                                 }
 316
 317                                 return;
 318                         }
 319
 320                         // Extra check for empty string necessary due to TZ but no second case.
 321                         $stripSeconds = false;
 322                         if ( !isset( $res[6] ) || $res[6] === '' ) {
 323                                 $res[6] = '00';
 324                                 $stripSeconds = true;
 325                         }
 326
 327                         // Do timezone processing. We've already done the case that tz = Z.
 328
 329                         // We know that if we got to this step, year, month day hour and min must be set
 330                         // by virtue of regex not failing.
 331
 332                         $unix = wfTimestamp( TS_UNIX, $res[1] . $res[2] . $res[3] . $res[4] . $res[5] . $res[6] );
 333                         $offset = intval( substr( $res[7], 1, 2 ) ) * 60 * 60;
 334                         $offset += intval( substr( $res[7], 4, 2 ) ) * 60;
 335                         if ( substr( $res[7], 0, 1 ) === '-' ) {
 336                                 $offset = -$offset;
 337                         }
 338                         $val = wfTimestamp( TS_EXIF, $unix + $offset );
 339
 340                         if ( $stripSeconds ) {
 341                                 // If seconds weren't specified, remove the trailing ':00'.
 342                                 $val = substr( $val, 0, -3 );
 343                         }
 344                 }
 345         }
 346
 347         /** function to validate, and more importantly
 348          * translate the XMP DMS form of gps coords to
 349          * the decimal form we use.
 350          *
 351          * @see http://www.adobe.com/devnet/xmp/pdfs/XMPSpecificationPart2.pdf
 352          *        section 1.2.7.4 on page 23
 353          *
 354          * @param array $info Unused (info about prop)
 355          * @param string &$val GPS string in either DDD,MM,SSk or
 356          *   or DDD,MM.mmk form
 357          * @param bool $standalone If its a simple prop (should always be true)
 358          */
 359         public function validateGPS( $info, &$val, $standalone ) {
 360                 if ( !$standalone ) {
 361                         return;
 362                 }
 363
 364                 $m = [];
 365                 if ( preg_match(
 366                         '/(\d{1,3}),(\d{1,2}),(\d{1,2})([NWSE])/D',
 367                         $val, $m )
 368                 ) {
 369                         $coord = intval( $m[1] );
 370                         $coord += intval( $m[2] ) * ( 1 / 60 );
 371                         $coord += intval( $m[3] ) * ( 1 / 3600 );
 372                         if ( $m[4] === 'S' || $m[4] === 'W' ) {
 373                                 $coord = -$coord;
 374                         }
 375                         $val = $coord;
 376
 377                         return;
 378                 } elseif ( preg_match(
 379                         '/(\d{1,3}),(\d{1,2}(?:.\d*)?)([NWSE])/D',
 380                         $val, $m )
 381                 ) {
 382                         $coord = intval( $m[1] );
 383                         $coord += floatval( $m[2] ) * ( 1 / 60 );
 384                         if ( $m[3] === 'S' || $m[3] === 'W' ) {
 385                                 $coord = -$coord;
 386                         }
 387                         $val = $coord;
 388
 389                         return;
 390                 } else {
 391                         $this->logger->info( __METHOD__
 392                                 . " Expected GPSCoordinate, but got $val." );
 393                         $val = null;
 394
 395                         return;
 396                 }
 397         }
 398 }