Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
78.57% |
110 / 140 |
|
38.89% |
7 / 18 |
CRAP | |
0.00% |
0 / 1 |
| WebPHandler | |
78.57% |
110 / 140 |
|
38.89% |
7 / 18 |
76.59 | |
0.00% |
0 / 1 |
| getSizeAndMetadata | |
100.00% |
10 / 10 |
|
100.00% |
1 / 1 |
2 | |||
| getMetadataType | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| isFileMetadataValid | |
0.00% |
0 / 11 |
|
0.00% |
0 / 1 |
42 | |||
| extractMetadata | |
100.00% |
14 / 14 |
|
100.00% |
1 / 1 |
4 | |||
| extractMetadataFromChunks | |
100.00% |
32 / 32 |
|
100.00% |
1 / 1 |
9 | |||
| decodeMediaMetadata | |
100.00% |
20 / 20 |
|
100.00% |
1 / 1 |
8 | |||
| extractChunk | |
66.67% |
2 / 3 |
|
0.00% |
0 / 1 |
3.33 | |||
| decodeLossyChunkHeader | |
100.00% |
11 / 11 |
|
100.00% |
1 / 1 |
2 | |||
| decodeLosslessChunkHeader | |
100.00% |
11 / 11 |
|
100.00% |
1 / 1 |
2 | |||
| decodeExtendedChunkHeader | |
100.00% |
10 / 10 |
|
100.00% |
1 / 1 |
1 | |||
| mustRender | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| canRender | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
6 | |||
| isAnimatedImage | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
12 | |||
| canAnimateThumbnail | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getThumbType | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| hasGDSupport | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
6 | |||
| getCommonMetaArray | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| formatMetadata | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
6 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * Handler for Google's WebP format <https://developers.google.com/speed/webp/> |
| 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 MediaWiki\FileRepo\File\File; |
| 25 | use MediaWiki\Logger\LoggerFactory; |
| 26 | use MediaWiki\MediaWikiServices; |
| 27 | use Wikimedia\XMPReader\Reader as XMPReader; |
| 28 | |
| 29 | /** |
| 30 | * Handler for Google's WebP format <https://developers.google.com/speed/webp/> |
| 31 | * |
| 32 | * @ingroup Media |
| 33 | */ |
| 34 | class WebPHandler extends BitmapHandler { |
| 35 | /** |
| 36 | * Value to store in img_metadata if there was an error extracting metadata |
| 37 | */ |
| 38 | private const BROKEN_FILE = '0'; |
| 39 | /** |
| 40 | * Minimum chunk header size to be able to read all header types |
| 41 | */ |
| 42 | private const MINIMUM_CHUNK_HEADER_LENGTH = 18; |
| 43 | /** |
| 44 | * Max size of metadata chunk to extract |
| 45 | */ |
| 46 | private const MAX_METADATA_CHUNK_SIZE = 1024 * 1024 * 2; |
| 47 | /** |
| 48 | * Version of the metadata stored in db records |
| 49 | */ |
| 50 | private const _MW_WEBP_VERSION = 2; |
| 51 | |
| 52 | private const VP8X_ICC = 32; |
| 53 | private const VP8X_ALPHA = 16; |
| 54 | private const VP8X_EXIF = 8; |
| 55 | private const VP8X_XMP = 4; |
| 56 | private const VP8X_ANIM = 2; |
| 57 | |
| 58 | public function getSizeAndMetadata( $state, $filename ) { |
| 59 | $parsedWebPData = self::extractMetadata( $filename ); |
| 60 | if ( !$parsedWebPData ) { |
| 61 | return [ 'metadata' => [ '_error' => self::BROKEN_FILE ] ]; |
| 62 | } |
| 63 | |
| 64 | $parsedWebPData['metadata']['_MW_WEBP_VERSION'] = self::_MW_WEBP_VERSION; |
| 65 | $info = [ |
| 66 | 'width' => $parsedWebPData['width'], |
| 67 | 'height' => $parsedWebPData['height'], |
| 68 | 'metadata' => $parsedWebPData |
| 69 | ]; |
| 70 | return $info; |
| 71 | } |
| 72 | |
| 73 | public function getMetadataType( $image ) { |
| 74 | return 'parsed-webp'; |
| 75 | } |
| 76 | |
| 77 | public function isFileMetadataValid( $image ) { |
| 78 | $data = $image->getMetadataArray(); |
| 79 | if ( $data === [ '_error' => self::BROKEN_FILE ] ) { |
| 80 | // Do not repetitivly regenerate metadata on broken file. |
| 81 | return self::METADATA_GOOD; |
| 82 | } |
| 83 | |
| 84 | if ( !$data || !isset( $data['_error'] ) ) { |
| 85 | wfDebug( __METHOD__ . " invalid WebP metadata" ); |
| 86 | |
| 87 | return self::METADATA_BAD; |
| 88 | } |
| 89 | |
| 90 | if ( !isset( $data['metadata']['_MW_WEBP_VERSION'] ) |
| 91 | || $data['metadata']['_MW_WEBP_VERSION'] != self::_MW_WEBP_VERSION |
| 92 | ) { |
| 93 | wfDebug( __METHOD__ . " old but compatible WebP metadata" ); |
| 94 | |
| 95 | return self::METADATA_COMPATIBLE; |
| 96 | } |
| 97 | return self::METADATA_GOOD; |
| 98 | } |
| 99 | |
| 100 | /** |
| 101 | * Extracts the image size and WebP type from a file |
| 102 | * |
| 103 | * @param string $filename |
| 104 | * @return array|false Header data array with entries 'compression', 'width' and 'height', |
| 105 | * where 'compression' can be 'lossy', 'lossless', 'animated' or 'unknown'. False if |
| 106 | * file is not a valid WebP file. |
| 107 | */ |
| 108 | public static function extractMetadata( $filename ) { |
| 109 | wfDebugLog( 'WebP', __METHOD__ . ": Extracting metadata from $filename" ); |
| 110 | |
| 111 | $info = RiffExtractor::findChunksFromFile( $filename, 100 ); |
| 112 | if ( $info === false ) { |
| 113 | wfDebugLog( 'WebP', __METHOD__ . ": Not a valid RIFF file" ); |
| 114 | return false; |
| 115 | } |
| 116 | |
| 117 | if ( $info['fourCC'] !== 'WEBP' ) { |
| 118 | wfDebugLog( 'WebP', __METHOD__ . ': FourCC was not WEBP: ' . |
| 119 | bin2hex( $info['fourCC'] ) ); |
| 120 | return false; |
| 121 | } |
| 122 | $metadata = self::extractMetadataFromChunks( $info['chunks'], $filename ); |
| 123 | if ( !$metadata ) { |
| 124 | wfDebugLog( 'WebP', __METHOD__ . ": No VP8 chunks found" ); |
| 125 | return false; |
| 126 | } |
| 127 | |
| 128 | return $metadata; |
| 129 | } |
| 130 | |
| 131 | /** |
| 132 | * Extracts the image size and WebP type from a file based on the chunk list |
| 133 | * @param array[] $chunks Chunks as extracted by RiffExtractor |
| 134 | * @param string $filename |
| 135 | * @return array Header data array with entries 'compression', 'width' and 'height', where |
| 136 | * 'compression' can be 'lossy', 'lossless', 'animated' or 'unknown' |
| 137 | */ |
| 138 | public static function extractMetadataFromChunks( $chunks, $filename ) { |
| 139 | $vp8Info = []; |
| 140 | $exifData = null; |
| 141 | $xmpData = null; |
| 142 | |
| 143 | foreach ( $chunks as $chunk ) { |
| 144 | // Note, spec says it should be 'XMP ' but some real life files use "XMP\0" |
| 145 | if ( !in_array( $chunk['fourCC'], [ 'VP8 ', 'VP8L', 'VP8X', 'EXIF', 'XMP ', "XMP\0" ] ) ) { |
| 146 | // Not a chunk containing interesting metadata |
| 147 | continue; |
| 148 | } |
| 149 | |
| 150 | $chunkHeader = file_get_contents( $filename, false, null, |
| 151 | $chunk['start'], self::MINIMUM_CHUNK_HEADER_LENGTH ); |
| 152 | wfDebugLog( 'WebP', __METHOD__ . ": {$chunk['fourCC']}" ); |
| 153 | |
| 154 | switch ( $chunk['fourCC'] ) { |
| 155 | case 'VP8 ': |
| 156 | $vp8Info = array_merge( $vp8Info, |
| 157 | self::decodeLossyChunkHeader( $chunkHeader ) ); |
| 158 | break; |
| 159 | case 'VP8L': |
| 160 | $vp8Info = array_merge( $vp8Info, |
| 161 | self::decodeLosslessChunkHeader( $chunkHeader ) ); |
| 162 | break; |
| 163 | case 'VP8X': |
| 164 | $vp8Info = array_merge( $vp8Info, |
| 165 | self::decodeExtendedChunkHeader( $chunkHeader ) ); |
| 166 | // Continue looking for other chunks to improve the metadata |
| 167 | break; |
| 168 | case 'EXIF': |
| 169 | // Spec says ignore all but first one |
| 170 | $exifData ??= self::extractChunk( $chunk, $filename ); |
| 171 | break; |
| 172 | case 'XMP ': |
| 173 | case "XMP\0": |
| 174 | $xmpData ??= self::extractChunk( $chunk, $filename ); |
| 175 | break; |
| 176 | } |
| 177 | } |
| 178 | $vp8Info = array_merge( $vp8Info, |
| 179 | self::decodeMediaMetadata( $exifData, $xmpData, $filename ) ); |
| 180 | return $vp8Info; |
| 181 | } |
| 182 | |
| 183 | /** |
| 184 | * Decode metadata about the file (XMP & Exif). |
| 185 | * |
| 186 | * @param string|null $exifData Binary exif data from file |
| 187 | * @param string|null $xmpData XMP data from file |
| 188 | * @param string|null $filename (Used for logging only) |
| 189 | * @return array |
| 190 | */ |
| 191 | private static function decodeMediaMetadata( $exifData, $xmpData, $filename ) { |
| 192 | if ( $exifData === null && $xmpData === null ) { |
| 193 | // Nothing to do |
| 194 | return []; |
| 195 | } |
| 196 | $bitmapMetadataHandler = new BitmapMetadataHandler; |
| 197 | |
| 198 | if ( $xmpData && XMPReader::isSupported() ) { |
| 199 | $xmpReader = new XMPReader( LoggerFactory::getInstance( 'XMP' ), $filename ); |
| 200 | $xmpReader->parse( $xmpData ); |
| 201 | $res = $xmpReader->getResults(); |
| 202 | foreach ( $res as $type => $array ) { |
| 203 | $bitmapMetadataHandler->addMetadata( $array, $type ); |
| 204 | } |
| 205 | } |
| 206 | |
| 207 | if ( $exifData ) { |
| 208 | // The Exif section of a webp file is basically a tiff file without an image. |
| 209 | // Some files start with an Exif\0\0. This is wrong according to standard and |
| 210 | // will prevent us from reading file, so remove for compatibility. |
| 211 | if ( substr( $exifData, 0, 6 ) === "Exif\x00\x00" ) { |
| 212 | $exifData = substr( $exifData, 6 ); |
| 213 | } |
| 214 | $tmpFile = MediaWikiServices::getInstance()-> |
| 215 | getTempFSFileFactory()-> |
| 216 | newTempFSFile( 'webp-exif_', 'tiff' ); |
| 217 | |
| 218 | $exifDataFile = $tmpFile->getPath(); |
| 219 | file_put_contents( $exifDataFile, $exifData ); |
| 220 | $byteOrder = BitmapMetadataHandler::getTiffByteOrder( $exifDataFile ); |
| 221 | $bitmapMetadataHandler->getExif( $exifDataFile, $byteOrder ); |
| 222 | } |
| 223 | return [ 'media-metadata' => $bitmapMetadataHandler->getMetadataArray() ]; |
| 224 | } |
| 225 | |
| 226 | /** |
| 227 | * @param array $chunk Information about chunk |
| 228 | * @param string $filename |
| 229 | * @return null|string Contents of chunk (excluding fourcc, size and padding) |
| 230 | */ |
| 231 | private static function extractChunk( $chunk, $filename ) { |
| 232 | if ( $chunk['size'] > self::MAX_METADATA_CHUNK_SIZE || $chunk['size'] < 1 ) { |
| 233 | return null; |
| 234 | } |
| 235 | |
| 236 | // Skip first 8 bytes as that is the fourCC header followed by size of chunk. |
| 237 | return file_get_contents( $filename, false, null, $chunk['start'] + 8, $chunk['size'] ); |
| 238 | } |
| 239 | |
| 240 | /** |
| 241 | * Decodes a lossy chunk header |
| 242 | * @param string $header First few bytes of the header, expected to be at least 18 bytes long |
| 243 | * @return bool|array See WebPHandler::decodeHeader |
| 244 | */ |
| 245 | protected static function decodeLossyChunkHeader( $header ) { |
| 246 | // Bytes 0-3 are 'VP8 ' |
| 247 | // Bytes 4-7 are the VP8 stream size |
| 248 | // Bytes 8-10 are the frame tag |
| 249 | // Bytes 11-13 are 0x9D 0x01 0x2A called the sync code |
| 250 | $syncCode = substr( $header, 11, 3 ); |
| 251 | if ( $syncCode !== "\x9D\x01\x2A" ) { |
| 252 | wfDebugLog( 'WebP', __METHOD__ . ': Invalid sync code: ' . |
| 253 | bin2hex( $syncCode ) ); |
| 254 | return []; |
| 255 | } |
| 256 | // Bytes 14-17 are image size |
| 257 | $imageSize = unpack( 'v2', substr( $header, 14, 4 ) ); |
| 258 | // Image sizes are 14 bit, 2 MSB are scaling parameters which are ignored here |
| 259 | return [ |
| 260 | 'compression' => 'lossy', |
| 261 | 'width' => $imageSize[1] & 0x3FFF, |
| 262 | 'height' => $imageSize[2] & 0x3FFF |
| 263 | ]; |
| 264 | } |
| 265 | |
| 266 | /** |
| 267 | * Decodes a lossless chunk header |
| 268 | * @param string $header First few bytes of the header, expected to be at least 13 bytes long |
| 269 | * @return bool|array See WebPHandler::decodeHeader |
| 270 | */ |
| 271 | public static function decodeLosslessChunkHeader( $header ) { |
| 272 | // Bytes 0-3 are 'VP8L' |
| 273 | // Bytes 4-7 are chunk stream size |
| 274 | // Byte 8 is 0x2F called the signature |
| 275 | if ( $header[8] !== "\x2F" ) { |
| 276 | wfDebugLog( 'WebP', __METHOD__ . ': Invalid signature: ' . |
| 277 | bin2hex( $header[8] ) ); |
| 278 | return []; |
| 279 | } |
| 280 | // Bytes 9-12 contain the image size |
| 281 | // Bits 0-13 are width-1; bits 15-27 are height-1 |
| 282 | $imageSize = unpack( 'C4', substr( $header, 9, 4 ) ); |
| 283 | return [ |
| 284 | 'compression' => 'lossless', |
| 285 | 'width' => ( $imageSize[1] | ( ( $imageSize[2] & 0x3F ) << 8 ) ) + 1, |
| 286 | 'height' => ( ( ( $imageSize[2] & 0xC0 ) >> 6 ) | |
| 287 | ( $imageSize[3] << 2 ) | ( ( $imageSize[4] & 0x03 ) << 10 ) ) + 1 |
| 288 | ]; |
| 289 | } |
| 290 | |
| 291 | /** |
| 292 | * Decodes an extended chunk header |
| 293 | * @param string $header First few bytes of the header, expected to be at least 18 bytes long |
| 294 | * @return bool|array See WebPHandler::decodeHeader |
| 295 | */ |
| 296 | public static function decodeExtendedChunkHeader( $header ) { |
| 297 | // Bytes 0-3 are 'VP8X' |
| 298 | // Byte 4-7 are chunk length |
| 299 | // Byte 8-11 are a flag bytes |
| 300 | $flags = unpack( 'c', substr( $header, 8, 1 ) ); |
| 301 | |
| 302 | // Byte 12-17 are image size (24 bits) |
| 303 | $width = unpack( 'V', substr( $header, 12, 3 ) . "\x00" ); |
| 304 | $height = unpack( 'V', substr( $header, 15, 3 ) . "\x00" ); |
| 305 | |
| 306 | return [ |
| 307 | 'compression' => 'unknown', |
| 308 | 'animated' => ( $flags[1] & self::VP8X_ANIM ) === self::VP8X_ANIM, |
| 309 | 'transparency' => ( $flags[1] & self::VP8X_ALPHA ) === self::VP8X_ALPHA, |
| 310 | 'width' => ( $width[1] & 0xFFFFFF ) + 1, |
| 311 | 'height' => ( $height[1] & 0xFFFFFF ) + 1 |
| 312 | ]; |
| 313 | } |
| 314 | |
| 315 | /** |
| 316 | * @param File $file |
| 317 | * @return bool True, not all browsers support WebP |
| 318 | */ |
| 319 | public function mustRender( $file ) { |
| 320 | return true; |
| 321 | } |
| 322 | |
| 323 | /** |
| 324 | * @param File $file |
| 325 | * @return bool False if we are unable to render this image |
| 326 | */ |
| 327 | public function canRender( $file ) { |
| 328 | if ( $this->isAnimatedImage( $file ) ) { |
| 329 | return false; |
| 330 | } |
| 331 | return true; |
| 332 | } |
| 333 | |
| 334 | /** |
| 335 | * @param File $image |
| 336 | * @return bool |
| 337 | */ |
| 338 | public function isAnimatedImage( $image ) { |
| 339 | $metadata = $image->getMetadataArray(); |
| 340 | if ( isset( $metadata['animated'] ) && $metadata['animated'] === true ) { |
| 341 | return true; |
| 342 | } |
| 343 | |
| 344 | return false; |
| 345 | } |
| 346 | |
| 347 | public function canAnimateThumbnail( $file ) { |
| 348 | return false; |
| 349 | } |
| 350 | |
| 351 | /** |
| 352 | * Render files as PNG |
| 353 | * |
| 354 | * @param string $ext |
| 355 | * @param string $mime |
| 356 | * @param array|null $params |
| 357 | * @return array |
| 358 | */ |
| 359 | public function getThumbType( $ext, $mime, $params = null ) { |
| 360 | return [ 'png', 'image/png' ]; |
| 361 | } |
| 362 | |
| 363 | protected function hasGDSupport() { |
| 364 | return function_exists( 'gd_info' ) && ( gd_info()['WebP Support'] ?? false ); |
| 365 | } |
| 366 | |
| 367 | public function getCommonMetaArray( File $image ) { |
| 368 | $meta = $image->getMetadataArray(); |
| 369 | return $meta['media-metadata'] ?? []; |
| 370 | } |
| 371 | |
| 372 | public function formatMetadata( $image, $context = false ) { |
| 373 | $meta = $this->getCommonMetaArray( $image ); |
| 374 | if ( !$meta ) { |
| 375 | return false; |
| 376 | } |
| 377 | |
| 378 | return $this->formatMetadataHelper( $meta, $context ); |
| 379 | } |
| 380 | } |