Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
100.00% |
131 / 131 |
|
100.00% |
11 / 11 |
CRAP | |
100.00% |
1 / 1 |
| Validate | |
100.00% |
131 / 131 |
|
100.00% |
11 / 11 |
67 | |
100.00% |
1 / 1 |
| __construct | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| validateBoolean | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
4 | |||
| validateRational | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
3 | |||
| validateRating | |
100.00% |
14 / 14 |
|
100.00% |
1 / 1 |
6 | |||
| validateInteger | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
3 | |||
| validateClosed | |
100.00% |
9 / 9 |
|
100.00% |
1 / 1 |
9 | |||
| validateReal | |
100.00% |
13 / 13 |
|
100.00% |
1 / 1 |
8 | |||
| validateFlash | |
100.00% |
15 / 15 |
|
100.00% |
1 / 1 |
7 | |||
| validateLangCode | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
3 | |||
| validateDate | |
100.00% |
39 / 39 |
|
100.00% |
1 / 1 |
15 | |||
| validateGPS | |
100.00% |
24 / 24 |
|
100.00% |
1 / 1 |
8 | |||
| 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 | namespace Wikimedia\XMPReader; |
| 25 | |
| 26 | use Psr\Log\LoggerAwareInterface; |
| 27 | use Psr\Log\LoggerAwareTrait; |
| 28 | use Psr\Log\LoggerInterface; |
| 29 | use Wikimedia\Timestamp\ConvertibleTimestamp; |
| 30 | |
| 31 | /** |
| 32 | * This contains some static methods for |
| 33 | * validating XMP properties. See XMPInfo and Reader classes. |
| 34 | * |
| 35 | * Each of these functions take the same parameters |
| 36 | * * an info array which is a subset of the XMPInfo::items array |
| 37 | * * A value (passed as reference) to validate. This can be either a |
| 38 | * simple value or an array |
| 39 | * * A boolean to determine if this is validating a simple or complex values |
| 40 | * |
| 41 | * It should be noted that when an array is being validated, typically the validation |
| 42 | * function is called once for each value, and then once at the end for the entire array. |
| 43 | * |
| 44 | * These validation functions can also be used to modify the data. See the gps and flash one's |
| 45 | * for example. |
| 46 | * |
| 47 | * @see https://www.adobe.com/devnet/xmp.html |
| 48 | * @see https://wwwimages2.adobe.com/content/dam/acom/en/devnet/xmp/pdfs/ |
| 49 | * XMP%20SDK%20Release%20cc-2016-08/XMPSpecificationPart1.pdf starting at pg 28 |
| 50 | * @see https://wwwimages2.adobe.com/content/dam/acom/en/devnet/xmp/pdfs/ |
| 51 | * XMP%20SDK%20Release%20cc-2016-08/XMPSpecificationPart2.pdf starting at pg 11 |
| 52 | */ |
| 53 | class Validate implements LoggerAwareInterface { |
| 54 | use LoggerAwareTrait; |
| 55 | |
| 56 | /** |
| 57 | * Create new instance, with a logger |
| 58 | * |
| 59 | * @param LoggerInterface $logger |
| 60 | */ |
| 61 | public function __construct( LoggerInterface $logger ) { |
| 62 | $this->setLogger( $logger ); |
| 63 | } |
| 64 | |
| 65 | /** |
| 66 | * Function to validate boolean properties ( True or False ) |
| 67 | * |
| 68 | * @param array $info Information about current property |
| 69 | * @param mixed &$val Current value to validate |
| 70 | * @param bool $standalone If this is a simple property or array |
| 71 | */ |
| 72 | public function validateBoolean( $info, &$val, $standalone ): void { |
| 73 | if ( !$standalone ) { |
| 74 | // this only validates standalone properties, not arrays, etc |
| 75 | // @codeCoverageIgnoreStart |
| 76 | return; |
| 77 | // @codeCoverageIgnoreEnd |
| 78 | } |
| 79 | if ( $val !== 'True' && $val !== 'False' ) { |
| 80 | $this->logger->info( __METHOD__ . " Expected True or False but got $val" ); |
| 81 | $val = null; |
| 82 | } |
| 83 | } |
| 84 | |
| 85 | /** |
| 86 | * function to validate rational properties ( 12/10 ) |
| 87 | * |
| 88 | * @param array $info Information about current property |
| 89 | * @param mixed &$val Current value to validate |
| 90 | * @param bool $standalone If this is a simple property or array |
| 91 | */ |
| 92 | public function validateRational( $info, &$val, $standalone ): void { |
| 93 | if ( !$standalone ) { |
| 94 | // this only validates standalone properties, not arrays, etc |
| 95 | // @codeCoverageIgnoreStart |
| 96 | return; |
| 97 | // @codeCoverageIgnoreEnd |
| 98 | } |
| 99 | if ( !preg_match( '/^(?:-?\d+)\/(?:\d+[1-9]|[1-9]\d*)$/D', $val ) ) { |
| 100 | $this->logger->info( __METHOD__ . " Expected rational but got $val" ); |
| 101 | $val = null; |
| 102 | } |
| 103 | } |
| 104 | |
| 105 | /** |
| 106 | * function to validate rating properties -1, 0-5 |
| 107 | * |
| 108 | * if its outside of range put it into range. |
| 109 | * |
| 110 | * @see MWG spec |
| 111 | * @param array $info Information about current property |
| 112 | * @param mixed &$val Current value to validate |
| 113 | * @param bool $standalone If this is a simple property or array |
| 114 | */ |
| 115 | public function validateRating( $info, &$val, $standalone ): void { |
| 116 | if ( !$standalone ) { |
| 117 | // this only validates standalone properties, not arrays, etc |
| 118 | // @codeCoverageIgnoreStart |
| 119 | return; |
| 120 | // @codeCoverageIgnoreEnd |
| 121 | } |
| 122 | if ( !preg_match( '/^[-+]?\d*(?:\.?\d*)$/D', $val ) |
| 123 | || !is_numeric( $val ) |
| 124 | ) { |
| 125 | $this->logger->info( __METHOD__ . " Expected rating but got $val" ); |
| 126 | $val = null; |
| 127 | |
| 128 | return; |
| 129 | } |
| 130 | |
| 131 | $nVal = (float)$val; |
| 132 | if ( $nVal < 0 ) { |
| 133 | // We do < 0 here instead of < -1 here, since |
| 134 | // the values between 0 and -1 are also illegal |
| 135 | // as -1 is meant as a special reject rating. |
| 136 | $this->logger->info( __METHOD__ . " Rating too low, setting to -1 (Rejected)" ); |
| 137 | $val = '-1'; |
| 138 | |
| 139 | return; |
| 140 | } |
| 141 | if ( $nVal > 5 ) { |
| 142 | $this->logger->info( __METHOD__ . " Rating too high, setting to 5" ); |
| 143 | $val = '5'; |
| 144 | } |
| 145 | } |
| 146 | |
| 147 | /** |
| 148 | * function to validate integers |
| 149 | * |
| 150 | * @param array $info Information about current property |
| 151 | * @param mixed &$val Current value to validate |
| 152 | * @param bool $standalone If this is a simple property or array |
| 153 | */ |
| 154 | public function validateInteger( $info, &$val, $standalone ): void { |
| 155 | if ( !$standalone ) { |
| 156 | // this only validates standalone properties, not arrays, etc |
| 157 | // @codeCoverageIgnoreStart |
| 158 | return; |
| 159 | // @codeCoverageIgnoreEnd |
| 160 | } |
| 161 | if ( !preg_match( '/^[-+]?\d+$/D', $val ) ) { |
| 162 | $this->logger->info( __METHOD__ . " Expected integer but got $val" ); |
| 163 | $val = null; |
| 164 | } |
| 165 | } |
| 166 | |
| 167 | /** |
| 168 | * function to validate properties with a fixed number of allowed |
| 169 | * choices. (closed choice) |
| 170 | * |
| 171 | * @param array $info Information about current property |
| 172 | * @param mixed &$val Current value to validate |
| 173 | * @param bool $standalone If this is a simple property or array |
| 174 | */ |
| 175 | public function validateClosed( $info, &$val, $standalone ): void { |
| 176 | if ( !$standalone ) { |
| 177 | // this only validates standalone properties, not arrays, etc |
| 178 | // @codeCoverageIgnoreStart |
| 179 | return; |
| 180 | // @codeCoverageIgnoreEnd |
| 181 | } |
| 182 | |
| 183 | // check if it's in a numeric range |
| 184 | $inRange = false; |
| 185 | if ( is_numeric( $val ) |
| 186 | && isset( $info['rangeLow'] ) && isset( $info['rangeHigh'] ) |
| 187 | && ( (int)$val <= $info['rangeHigh'] ) && ( (int)$val >= $info['rangeLow'] ) |
| 188 | ) { |
| 189 | $inRange = true; |
| 190 | } |
| 191 | |
| 192 | if ( !isset( $info['choices'][$val] ) && !$inRange ) { |
| 193 | $this->logger->info( __METHOD__ . " Expected closed choice, but got $val" ); |
| 194 | $val = null; |
| 195 | } |
| 196 | } |
| 197 | |
| 198 | /** |
| 199 | * function to validate and modify real numbers, with ranges |
| 200 | * |
| 201 | * @param array $info Information about current property |
| 202 | * @param mixed &$val Current value to validate |
| 203 | * @param bool $standalone If this is a simple property or array |
| 204 | */ |
| 205 | public function validateReal( $info, &$val, $standalone ): void { |
| 206 | if ( !$standalone ) { |
| 207 | // this only validates standalone properties, not arrays, etc |
| 208 | // @codeCoverageIgnoreStart |
| 209 | return; |
| 210 | // @codeCoverageIgnoreEnd |
| 211 | } |
| 212 | |
| 213 | $isReal = is_numeric( $val ) && (float)$val; |
| 214 | if ( !$isReal ) { |
| 215 | $this->logger->info( __METHOD__ . " Expected real, but got $val" ); |
| 216 | $val = null; |
| 217 | return; |
| 218 | } |
| 219 | |
| 220 | // check if it's in a numeric range |
| 221 | if ( isset( $info['rangeLow'] ) && isset( $info['rangeHigh'] ) |
| 222 | && ( (float)$val > $info['rangeHigh'] || (float)$val < $info['rangeLow'] ) |
| 223 | ) { |
| 224 | $this->logger->info( |
| 225 | __METHOD__ |
| 226 | . " Expected value within range of {$info['rangeLow']}-{$info['rangeHigh']}, but got $val" |
| 227 | ); |
| 228 | $val = null; |
| 229 | } |
| 230 | } |
| 231 | |
| 232 | /** |
| 233 | * function to validate and modify flash structure |
| 234 | * |
| 235 | * @param array $info Information about current property |
| 236 | * @param mixed &$val Current value to validate |
| 237 | * @param bool $standalone If this is a simple property or array |
| 238 | */ |
| 239 | public function validateFlash( $info, &$val, $standalone ): void { |
| 240 | if ( $standalone ) { |
| 241 | // this only validates flash structs, not individual properties |
| 242 | // @codeCoverageIgnoreStart |
| 243 | return; |
| 244 | // @codeCoverageIgnoreEnd |
| 245 | } |
| 246 | if ( !isset( $val['Fired'] ) || |
| 247 | !isset( $val['Function'] ) || |
| 248 | !isset( $val['Mode'] ) || |
| 249 | !isset( $val['RedEyeMode'] ) || |
| 250 | !isset( $val['Return'] ) |
| 251 | ) { |
| 252 | $this->logger->info( __METHOD__ . ' Flash structure did not have all the required components' ); |
| 253 | $val = null; |
| 254 | } else { |
| 255 | // @phan-suppress-next-line PhanTypeInvalidRightOperandOfBitwiseOp |
| 256 | $val = ( 0 | ( $val['Fired'] === 'True' ) |
| 257 | | ( (int)$val['Return'] << 1 ) |
| 258 | | ( (int)$val['Mode'] << 3 ) |
| 259 | // @phan-suppress-next-line PhanTypeInvalidLeftOperandOfIntegerOp |
| 260 | | ( ( $val['Function'] === 'True' ) << 5 ) |
| 261 | // @phan-suppress-next-line PhanTypeInvalidLeftOperandOfIntegerOp |
| 262 | | ( ( $val['RedEyeMode'] === 'True' ) << 6 ) ); |
| 263 | } |
| 264 | } |
| 265 | |
| 266 | /** |
| 267 | * function to validate LangCode properties ( en-GB, etc. ) |
| 268 | * |
| 269 | * This is just a naive check to make sure it somewhat looks like a lang code. |
| 270 | * |
| 271 | * @see BCP 47 |
| 272 | * @see https://wwwimages2.adobe.com/content/dam/Adobe/en/devnet/xmp/pdfs/ |
| 273 | * XMP%20SDK%20Release%20cc-2014-12/XMPSpecificationPart1.pdf page 22 (section 8.2.2.4) |
| 274 | * |
| 275 | * @param array $info Information about current property |
| 276 | * @param mixed &$val Current value to validate |
| 277 | * @param bool $standalone If this is a simple property or array |
| 278 | */ |
| 279 | public function validateLangCode( $info, &$val, $standalone ): void { |
| 280 | if ( !$standalone ) { |
| 281 | // this only validates standalone properties, not arrays, etc |
| 282 | // @codeCoverageIgnoreStart |
| 283 | return; |
| 284 | // @codeCoverageIgnoreEnd |
| 285 | } |
| 286 | if ( !preg_match( '/^[-A-Za-z0-9]{2,}$/D', $val ) ) { |
| 287 | // this is a rather naive check. |
| 288 | $this->logger->info( __METHOD__ . " Expected Lang code but got $val" ); |
| 289 | $val = null; |
| 290 | } |
| 291 | } |
| 292 | |
| 293 | /** |
| 294 | * function to validate date properties, and convert to (partial) Exif format. |
| 295 | * |
| 296 | * Dates can be one of the following formats: |
| 297 | * YYYY |
| 298 | * YYYY-MM |
| 299 | * YYYY-MM-DD |
| 300 | * YYYY-MM-DDThh:mmTZD |
| 301 | * YYYY-MM-DDThh:mm:ssTZD |
| 302 | * YYYY-MM-DDThh:mm:ss.sTZD |
| 303 | * |
| 304 | * @param array $info Information about current property |
| 305 | * @param mixed &$val Current value to validate. Converts to TS_EXIF as a side effect. |
| 306 | * in cases where there's only a partial date, it will give things like |
| 307 | * 2011:04. |
| 308 | * @param bool $standalone If this is a simple property or array |
| 309 | */ |
| 310 | public function validateDate( $info, &$val, $standalone ): void { |
| 311 | if ( !$standalone ) { |
| 312 | // this only validates standalone properties, not arrays, etc |
| 313 | // @codeCoverageIgnoreStart |
| 314 | return; |
| 315 | // @codeCoverageIgnoreEnd |
| 316 | } |
| 317 | $res = []; |
| 318 | if ( !preg_match( |
| 319 | /* ahh! scary regex... */ |
| 320 | // phpcs:ignore Generic.Files.LineLength |
| 321 | '/^([0-3]\d{3})(?:-([01]\d)(?:-([0-3]\d)(?:T([0-2]\d):([0-6]\d)(?::([0-6]\d)(?:\.\d+)?)?([-+]\d{2}:\d{2}|Z)?)?)?)?$/D', |
| 322 | $val, $res ) |
| 323 | ) { |
| 324 | $this->logger->info( __METHOD__ . " Expected date but got $val" ); |
| 325 | $val = null; |
| 326 | return; |
| 327 | } |
| 328 | |
| 329 | /* |
| 330 | * $res is formatted as follows: |
| 331 | * 0 -> full date. |
| 332 | * 1 -> year, 2-> month, 3-> day, 4-> hour, 5-> minute, 6->second |
| 333 | * 7-> Timezone specifier (Z or something like +12:30 ) |
| 334 | * many parts are optional, some aren't. For example if you specify |
| 335 | * minute, you must specify hour, day, month, and year but not second or TZ. |
| 336 | */ |
| 337 | |
| 338 | /* |
| 339 | * First of all, if year = 0000, Something is wrong-ish, |
| 340 | * so don't extract. This seems to happen when |
| 341 | * some programs convert between metadata formats. |
| 342 | */ |
| 343 | if ( $res[1] === '0000' ) { |
| 344 | $this->logger->info( __METHOD__ . " Invalid date (year 0): $val" ); |
| 345 | $val = null; |
| 346 | |
| 347 | return; |
| 348 | } |
| 349 | |
| 350 | // hour |
| 351 | if ( !isset( $res[4] ) ) { |
| 352 | // just have the year month day (if that) |
| 353 | $val = $res[1]; |
| 354 | if ( isset( $res[2] ) ) { |
| 355 | $val .= ':' . $res[2]; |
| 356 | } |
| 357 | if ( isset( $res[3] ) ) { |
| 358 | $val .= ':' . $res[3]; |
| 359 | } |
| 360 | |
| 361 | return; |
| 362 | } |
| 363 | |
| 364 | if ( !isset( $res[7] ) || $res[7] === 'Z' ) { |
| 365 | // if hour is set, then minute must also be or regex above will fail. |
| 366 | $val = $res[1] . ':' . $res[2] . ':' . $res[3] |
| 367 | . ' ' . $res[4] . ':' . $res[5]; |
| 368 | if ( isset( $res[6] ) && $res[6] !== '' ) { |
| 369 | $val .= ':' . $res[6]; |
| 370 | } |
| 371 | |
| 372 | return; |
| 373 | } |
| 374 | |
| 375 | // Extra check for empty string necessary due to TZ but no second case. |
| 376 | $stripSeconds = false; |
| 377 | if ( !isset( $res[6] ) || $res[6] === '' ) { |
| 378 | $res[6] = '00'; |
| 379 | $stripSeconds = true; |
| 380 | } |
| 381 | |
| 382 | // Do timezone processing. We've already done the case that tz = Z. |
| 383 | |
| 384 | // We know that if we got to this step, year, month day hour and min must be set |
| 385 | // by virtue of regex not failing. |
| 386 | |
| 387 | $unix = ConvertibleTimestamp::convert( TS_UNIX, |
| 388 | $res[1] . $res[2] . $res[3] . $res[4] . $res[5] . $res[6] |
| 389 | ); |
| 390 | $offset = (int)substr( $res[7], 1, 2 ) * 60 * 60; |
| 391 | $offset += (int)substr( $res[7], 4, 2 ) * 60; |
| 392 | if ( substr( $res[7], 0, 1 ) === '-' ) { |
| 393 | $offset = -$offset; |
| 394 | } |
| 395 | $val = ConvertibleTimestamp::convert( TS_EXIF, (int)$unix + $offset ); |
| 396 | |
| 397 | if ( $stripSeconds ) { |
| 398 | // If seconds weren't specified, remove the trailing ':00'. |
| 399 | $val = substr( $val, 0, -3 ); |
| 400 | } |
| 401 | } |
| 402 | |
| 403 | /** function to validate, and more importantly |
| 404 | * translate the XMP DMS form of gps coords to |
| 405 | * the decimal form we use. |
| 406 | * |
| 407 | * @see http://www.adobe.com/devnet/xmp/pdfs/XMPSpecificationPart2.pdf |
| 408 | * section 1.2.7.4 on page 23 |
| 409 | * |
| 410 | * @param array $info Unused (info about prop) |
| 411 | * @param string &$val GPS string in either DDD,MM,SSk or |
| 412 | * or DDD,MM.mmk form |
| 413 | * @param bool $standalone If it's a simple prop (should always be true) |
| 414 | */ |
| 415 | public function validateGPS( $info, &$val, $standalone ): void { |
| 416 | if ( !$standalone ) { |
| 417 | // this only validates standalone properties, not arrays, etc |
| 418 | // @codeCoverageIgnoreStart |
| 419 | return; |
| 420 | // @codeCoverageIgnoreEnd |
| 421 | } |
| 422 | |
| 423 | $m = []; |
| 424 | if ( preg_match( |
| 425 | '/(\d{1,3}),(\d{1,2}),(\d{1,2})([NWSE])/D', |
| 426 | $val, $m ) |
| 427 | ) { |
| 428 | $coord = (int)$m[1]; |
| 429 | $coord += (int)$m[2] * ( 1 / 60 ); |
| 430 | $coord += (int)$m[3] * ( 1 / 3600 ); |
| 431 | if ( $m[4] === 'S' || $m[4] === 'W' ) { |
| 432 | $coord = -$coord; |
| 433 | } |
| 434 | $val = $coord; |
| 435 | |
| 436 | return; |
| 437 | } |
| 438 | |
| 439 | if ( preg_match( |
| 440 | '/(\d{1,3}),(\d{1,2}(?:.\d*)?)([NWSE])/D', |
| 441 | $val, $m ) |
| 442 | ) { |
| 443 | $coord = (int)$m[1]; |
| 444 | $coord += (float)$m[2] * ( 1 / 60 ); |
| 445 | if ( $m[3] === 'S' || $m[3] === 'W' ) { |
| 446 | $coord = -$coord; |
| 447 | } |
| 448 | $val = $coord; |
| 449 | |
| 450 | return; |
| 451 | } |
| 452 | |
| 453 | $this->logger->info( __METHOD__ |
| 454 | . " Expected GPSCoordinate, but got $val." ); |
| 455 | $val = null; |
| 456 | } |
| 457 | } |