3 * Media-handling base classes and generic functionality.
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.
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.
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
23 use MediaWiki\MediaWikiServices
;
26 * Base media handler class
30 abstract class MediaHandler
{
31 const TRANSFORM_LATER
= 1;
32 const METADATA_GOOD
= true;
33 const METADATA_BAD
= false;
34 const METADATA_COMPATIBLE
= 2; // for old but backwards compatible.
36 * Max length of error logged by logErrorForExternalProcess()
38 const MAX_ERR_LOG_SIZE
= 65535;
41 * Get a MediaHandler for a given MIME type from the instance cache
44 * @return MediaHandler|bool
46 static function getHandler( $type ) {
47 return MediaWikiServices
::getInstance()
48 ->getMediaHandlerFactory()->getHandler( $type );
52 * Get an associative array mapping magic word IDs to parameter names.
53 * Will be used by the parser to identify parameters.
55 abstract public function getParamMap();
58 * Validate a thumbnail parameter at parse time.
59 * Return true to accept the parameter, and false to reject it.
60 * If you return false, the parser will do something quiet and forgiving.
65 abstract public function validateParam( $name, $value );
68 * Merge a parameter array into a string appropriate for inclusion in filenames
70 * @param array $params Array of parameters that have been through normaliseParams.
73 abstract public function makeParamString( $params );
76 * Parse a param string made with makeParamString back into an array
78 * @param string $str The parameter string without file name (e.g. 122px)
79 * @return array|bool Array of parameters or false on failure.
81 abstract public function parseParamString( $str );
84 * Changes the parameter array as necessary, ready for transformation.
85 * Should be idempotent.
86 * Returns false if the parameters are unacceptable and the transform should fail
88 * @param array &$params
90 abstract function normaliseParams( $image, &$params );
93 * Get an image size array like that returned by getimagesize(), or false if it
94 * can't be determined.
96 * This function is used for determining the width, height and bitdepth directly
97 * from an image. The results are stored in the database in the img_width,
98 * img_height, img_bits fields.
100 * @note If this is a multipage file, return the width and height of the
103 * @param File|FSFile $image The image object, or false if there isn't one.
104 * Warning, FSFile::getPropsFromPath might pass an FSFile instead of File (!)
105 * @param string $path The filename
106 * @return array|bool Follow the format of PHP getimagesize() internal function.
107 * See https://secure.php.net/getimagesize. MediaWiki will only ever use the
108 * first two array keys (the width and height), and the 'bits' associative
109 * key. All other array keys are ignored. Returning a 'bits' key is optional
110 * as not all formats have a notion of "bitdepth". Returns false on failure.
112 abstract function getImageSize( $image, $path );
115 * Get handler-specific metadata which will be saved in the img_metadata field.
117 * @param File|FSFile $image The image object, or false if there isn't one.
118 * Warning, FSFile::getPropsFromPath might pass an FSFile instead of File (!)
119 * @param string $path The filename
120 * @return string A string of metadata in php serialized form (Run through serialize())
122 function getMetadata( $image, $path ) {
127 * Get metadata version.
129 * This is not used for validating metadata, this is used for the api when returning
130 * metadata, since api content formats should stay the same over time, and so things
131 * using ForeignApiRepo can keep backwards compatibility
133 * All core media handlers share a common version number, and extensions can
134 * use the GetMetadataVersion hook to append to the array (they should append a unique
135 * string so not to get confusing). If there was a media handler named 'foo' with metadata
136 * version 3 it might add to the end of the array the element 'foo=3'. if the core metadata
137 * version is 2, the end version string would look like '2;foo=3'.
139 * @return string Version string
141 static function getMetadataVersion() {
142 $version = [ '2' ]; // core metadata version
143 Hooks
::run( 'GetMetadataVersion', [ &$version ] );
145 return implode( ';', $version );
149 * Convert metadata version.
151 * By default just returns $metadata, but can be used to allow
152 * media handlers to convert between metadata versions.
154 * @param string|array $metadata Metadata array (serialized if string)
155 * @param int $version Target version
156 * @return array Serialized metadata in specified version, or $metadata on fail.
158 function convertMetadataVersion( $metadata, $version = 1 ) {
159 if ( !is_array( $metadata ) ) {
160 // unserialize to keep return parameter consistent.
161 Wikimedia\
suppressWarnings();
162 $ret = unserialize( $metadata );
163 Wikimedia\restoreWarnings
();
172 * Get a string describing the type of metadata, for display purposes.
174 * @note This method is currently unused.
178 function getMetadataType( $image ) {
183 * Check if the metadata string is valid for this handler.
184 * If it returns MediaHandler::METADATA_BAD (or false), Image
185 * will reload the metadata from the file and update the database.
186 * MediaHandler::METADATA_GOOD for if the metadata is a-ok,
187 * MediaHandler::METADATA_COMPATIBLE if metadata is old but backwards
188 * compatible (which may or may not trigger a metadata reload).
190 * @note Returning self::METADATA_BAD will trigger a metadata reload from
191 * file on page view. Always returning this from a broken file, or suddenly
192 * triggering as bad metadata for a large number of files can cause
193 * performance problems.
195 * @param string $metadata The metadata in serialized form
198 function isMetadataValid( $image, $metadata ) {
199 return self
::METADATA_GOOD
;
203 * Get an array of standard (FormatMetadata type) metadata values.
205 * The returned data is largely the same as that from getMetadata(),
206 * but formatted in a standard, stable, handler-independent way.
207 * The idea being that some values like ImageDescription or Artist
208 * are universal and should be retrievable in a handler generic way.
210 * The specific properties are the type of properties that can be
211 * handled by the FormatMetadata class. These values are exposed to the
212 * user via the filemetadata parser function.
214 * Details of the response format of this function can be found at
215 * https://www.mediawiki.org/wiki/Manual:File_metadata_handling
216 * tl/dr: the response is an associative array of
217 * properties keyed by name, but the value can be complex. You probably
218 * want to call one of the FormatMetadata::flatten* functions on the
219 * property values before using them, or call
220 * FormatMetadata::getFormattedData() on the full response array, which
221 * transforms all values into prettified, human-readable text.
223 * Subclasses overriding this function must return a value which is a
224 * valid API response fragment (all associative array keys are valid
227 * Note, if the file simply has no metadata, but the handler supports
228 * this interface, it should return an empty array, not false.
231 * @return array|bool False if interface not supported
234 public function getCommonMetaArray( File
$file ) {
239 * Get a MediaTransformOutput object representing an alternate of the transformed
240 * output which will call an intermediary thumbnail assist script.
242 * Used when the repository has a thumbnailScriptUrl option configured.
244 * Return false to fall back to the regular getTransform().
246 * @param string $script
247 * @param array $params
248 * @return bool|ThumbnailImage
250 function getScriptedTransform( $image, $script, $params ) {
255 * Get a MediaTransformOutput object representing the transformed output. Does not
256 * actually do the transform.
259 * @param string $dstPath Filesystem destination path
260 * @param string $dstUrl Destination URL to use in output HTML
261 * @param array $params Arbitrary set of parameters validated by $this->validateParam()
262 * @return MediaTransformOutput
264 final function getTransform( $image, $dstPath, $dstUrl, $params ) {
265 return $this->doTransform( $image, $dstPath, $dstUrl, $params, self
::TRANSFORM_LATER
);
269 * Get a MediaTransformOutput object representing the transformed output. Does the
270 * transform unless $flags contains self::TRANSFORM_LATER.
273 * @param string $dstPath Filesystem destination path
274 * @param string $dstUrl Destination URL to use in output HTML
275 * @param array $params Arbitrary set of parameters validated by $this->validateParam()
276 * Note: These parameters have *not* gone through $this->normaliseParams()
277 * @param int $flags A bitfield, may contain self::TRANSFORM_LATER
278 * @return MediaTransformOutput
280 abstract function doTransform( $image, $dstPath, $dstUrl, $params, $flags = 0 );
283 * Get the thumbnail extension and MIME type for a given source MIME type
285 * @param string $ext Extension of original file
286 * @param string $mime MIME type of original file
287 * @param array|null $params Handler specific rendering parameters
288 * @return array Thumbnail extension and MIME type
290 function getThumbType( $ext, $mime, $params = null ) {
291 $magic = MediaWiki\MediaWikiServices
::getInstance()->getMimeAnalyzer();
292 if ( !$ext ||
$magic->isMatchingExtension( $ext, $mime ) === false ) {
293 // The extension is not valid for this MIME type and we do
294 // recognize the MIME type
295 $extensions = $magic->getExtensionsForType( $mime );
297 return [ strtok( $extensions, ' ' ), $mime ];
301 // The extension is correct (true) or the MIME type is unknown to
303 return [ $ext, $mime ];
307 * True if the handled types can be transformed
312 public function canRender( $file ) {
317 * True if handled types cannot be displayed directly in a browser
318 * but can be rendered
323 public function mustRender( $file ) {
328 * True if the type has multi-page capabilities
333 public function isMultiPage( $file ) {
338 * Page count for a multi-page document, false if unsupported or unknown
343 function pageCount( File
$file ) {
348 * The material is vectorized and thus scaling is lossless
353 function isVectorized( $file ) {
358 * The material is an image, and is animated.
359 * In particular, video material need not return true.
360 * @note Before 1.20, this was a method of ImageHandler only
365 function isAnimatedImage( $file ) {
370 * If the material is animated, we can animate the thumbnail
374 * @return bool If material is not animated, handler may return any value.
376 function canAnimateThumbnail( $file ) {
381 * False if the handler is disabled for all files
384 function isEnabled() {
389 * Get an associative array of page dimensions
390 * Currently "width" and "height" are understood, but this might be
391 * expanded in the future.
392 * Returns false if unknown.
394 * It is expected that handlers for paged media (e.g. DjVuHandler)
395 * will override this method so that it gives the correct results
396 * for each specific page of the file, using the $page argument.
398 * @note For non-paged media, use getImageSize.
401 * @param int $page What page to get dimensions of
404 function getPageDimensions( File
$image, $page ) {
405 $gis = $this->getImageSize( $image, $image->getLocalRefPath() );
417 * Generic getter for text layer.
418 * Currently overloaded by PDF and DjVu handlers
420 * @param int $page Page number to get information for
421 * @return bool|string Page text or false when no text found or if
424 function getPageText( File
$image, $page ) {
429 * Get the text of the entire document.
431 * @return bool|string The text of the document or false if unsupported.
433 public function getEntireText( File
$file ) {
434 $numPages = $file->pageCount();
436 // Not a multipage document
437 return $this->getPageText( $file, 1 );
440 for ( $i = 1; $i <= $numPages; $i++
) {
441 $curPage = $this->getPageText( $file, $i );
442 if ( is_string( $curPage ) ) {
443 $document .= $curPage . "\n";
446 if ( $document !== '' ) {
453 * Get an array structure that looks like this:
457 * 'Human-readable name' => 'Human readable value',
461 * 'Human-readable name' => 'Human readable value',
465 * The UI will format this into a table where the visible fields are always
466 * visible, and the collapsed fields are optionally visible.
468 * The function should return false if there is no metadata to display.
472 * @todo FIXME: This interface is not very flexible. The media handler
473 * should generate HTML instead. It can do all the formatting according
474 * to some standard. That makes it possible to do things like visual
475 * indication of grouped and chained streams in ogg container files.
477 * @param bool|IContextSource $context Context to use (optional)
480 function formatMetadata( $image, $context = false ) {
484 /** sorts the visible/invisible field.
485 * Split off from ImageHandler::formatMetadata, as used by more than
486 * one type of handler.
488 * This is used by the media handlers that use the FormatMetadata class
490 * @param array $metadataArray
491 * @param bool|IContextSource $context Context to use (optional)
492 * @return array Array for use displaying metadata.
494 function formatMetadataHelper( $metadataArray, $context = false ) {
500 $formatted = FormatMetadata
::getFormattedData( $metadataArray, $context );
501 // Sort fields into visible and collapsed
502 $visibleFields = $this->visibleMetadataFields();
503 foreach ( $formatted as $name => $value ) {
504 $tag = strtolower( $name );
505 self
::addMeta( $result,
506 in_array( $tag, $visibleFields ) ?
'visible' : 'collapsed',
517 * Get a list of metadata items which should be displayed when
518 * the metadata table is collapsed.
520 * @return array Array of strings
522 protected function visibleMetadataFields() {
523 return FormatMetadata
::getVisibleFields();
527 * This is used to generate an array element for each metadata value
528 * That array is then used to generate the table of metadata values
531 * @param array &$array An array containing elements for each type of visibility
532 * and each of those elements being an array of metadata items. This function adds
533 * a value to that array.
534 * @param string $visibility ('visible' or 'collapsed') if this value is hidden
536 * @param string $type Type of metadata tag (currently always 'exif')
537 * @param string $id The name of the metadata tag (like 'artist' for example).
538 * its name in the table displayed is the message "$type-$id" (Ex exif-artist ).
539 * @param string $value Thingy goes into a wikitext table; it used to be escaped but
540 * that was incompatible with previous practise of customized display
541 * with wikitext formatting via messages such as 'exif-model-value'.
542 * So the escaping is taken back out, but generally this seems a confusing
544 * @param bool|string $param Value to pass to the message for the name of the field
545 * as $1. Currently this parameter doesn't seem to ever be used.
547 * Note, everything here is passed through the parser later on (!)
549 protected static function addMeta( &$array, $visibility, $type, $id, $value, $param = false ) {
550 $msg = wfMessage( "$type-$id", $param );
551 if ( $msg->exists() ) {
552 $name = $msg->text();
554 // This is for future compatibility when using instant commons.
555 // So as to not display as ugly a name if a new metadata
556 // property is defined that we don't know about
557 // (not a major issue since such a property would be collapsed
559 wfDebug( __METHOD__
. ' Unknown metadata name: ' . $id . "\n" );
560 $name = wfEscapeWikiText( $id );
562 $array[$visibility][] = [
570 * Short description. Shown on Special:Search results.
575 function getShortDesc( $file ) {
576 return self
::getGeneralShortDesc( $file );
580 * Long description. Shown under image on image description page surounded by ().
585 function getLongDesc( $file ) {
586 return self
::getGeneralLongDesc( $file );
590 * Used instead of getShortDesc if there is no handler registered for file.
595 static function getGeneralShortDesc( $file ) {
598 return htmlspecialchars( $wgLang->formatSize( $file->getSize() ) );
602 * Used instead of getLongDesc if there is no handler registered for file.
607 static function getGeneralLongDesc( $file ) {
608 return wfMessage( 'file-info' )->sizeParams( $file->getSize() )
609 ->params( '<span class="mime-type">' . $file->getMimeType() . '</span>' )->parse();
613 * Calculate the largest thumbnail width for a given original file size
614 * such that the thumbnail's height is at most $maxHeight.
615 * @param int $boxWidth Width of the thumbnail box.
616 * @param int $boxHeight Height of the thumbnail box.
617 * @param int $maxHeight Maximum height expected for the thumbnail.
620 public static function fitBoxWidth( $boxWidth, $boxHeight, $maxHeight ) {
621 $idealWidth = $boxWidth * $maxHeight / $boxHeight;
622 $roundedUp = ceil( $idealWidth );
623 if ( round( $roundedUp * $boxHeight / $boxWidth ) > $maxHeight ) {
624 return floor( $idealWidth );
631 * Shown in file history box on image description page.
634 * @return string Dimensions
636 function getDimensionsString( $file ) {
641 * Modify the parser object post-transform.
643 * This is often used to do $parser->addOutputHook(),
644 * in order to add some javascript to render a viewer.
645 * See TimedMediaHandler or OggHandler for an example.
647 * @param Parser $parser
650 function parserTransformHook( $parser, $file ) {
654 * File validation hook called on upload.
656 * If the file at the given local path is not valid, or its MIME type does not
657 * match the handler class, a Status object should be returned containing
660 * @param string $fileName The local path to the file.
663 function verifyUpload( $fileName ) {
664 return Status
::newGood();
668 * Check for zero-sized thumbnails. These can be generated when
669 * no disk space is available or some other error occurs
671 * @param string $dstPath The location of the suspect file
672 * @param int $retval Return value of some shell process, file will be deleted if this is non-zero
673 * @return bool True if removed, false otherwise
675 function removeBadFile( $dstPath, $retval = 0 ) {
676 if ( file_exists( $dstPath ) ) {
677 $thumbstat = stat( $dstPath );
678 if ( $thumbstat['size'] == 0 ||
$retval != 0 ) {
679 $result = unlink( $dstPath );
682 wfDebugLog( 'thumbnail',
683 sprintf( 'Removing bad %d-byte thumbnail "%s". unlink() succeeded',
684 $thumbstat['size'], $dstPath ) );
686 wfDebugLog( 'thumbnail',
687 sprintf( 'Removing bad %d-byte thumbnail "%s". unlink() failed',
688 $thumbstat['size'], $dstPath ) );
699 * Remove files from the purge list.
701 * This is used by some video handlers to prevent ?action=purge
702 * from removing a transcoded video, which is expensive to
705 * @see LocalFile::purgeThumbnails
707 * @param array &$files
708 * @param array $options Purge options. Currently will always be
709 * an array with a single key 'forThumbRefresh' set to true.
711 public function filterThumbnailPurgeList( &$files, $options ) {
716 * True if the handler can rotate the media
717 * @since 1.24 non-static. From 1.21-1.23 was static
720 public function canRotate() {
725 * On supporting image formats, try to read out the low-level orientation
726 * of the file and return the angle that the file needs to be rotated to
729 * This information is only useful when manipulating the original file;
730 * the width and height we normally work with is logical, and will match
731 * any produced output views.
733 * For files we don't know, we return 0.
736 * @return int 0, 90, 180 or 270
738 public function getRotation( $file ) {
743 * Log an error that occurred in an external process
745 * Moved from BitmapHandler to MediaHandler with MediaWiki 1.23
749 * @param string $err Error reported by command. Anything longer than
750 * MediaHandler::MAX_ERR_LOG_SIZE is stripped off.
753 protected function logErrorForExternalProcess( $retval, $err, $cmd ) {
754 # Keep error output limited (T59985)
755 $errMessage = trim( substr( $err, 0, self
::MAX_ERR_LOG_SIZE
) );
757 wfDebugLog( 'thumbnail',
758 sprintf( 'thumbnail failed on %s: error %d "%s" from "%s"',
759 wfHostname(), $retval, $errMessage, $cmd ) );
763 * Get list of languages file can be viewed in.
766 * @return string[] Array of language codes, or empty array if unsupported.
769 public function getAvailableLanguages( File
$file ) {
774 * When overridden in a descendant class, returns a language code most suiting
778 * @param string $userPreferredLanguage Language code requesed
779 * @param string[] $availableLanguages Languages present in the file
780 * @return string|null Language code picked or null if not supported/available
782 public function getMatchedLanguage( $userPreferredLanguage, array $availableLanguages ) {
787 * On file types that support renderings in multiple languages,
788 * which language is used by default if unspecified.
790 * If getAvailableLanguages returns a non-empty array, this must return
791 * a valid language code. Otherwise can return null if files of this
792 * type do not support alternative language renderings.
795 * @return string|null Language code or null if multi-language not supported for filetype.
798 public function getDefaultRenderLanguage( File
$file ) {
803 * If its an audio file, return the length of the file. Otherwise 0.
805 * File::getLength() existed for a long time, but was calling a method
806 * that only existed in some subclasses of this class (The TMH ones).
809 * @return float Length in seconds
812 public function getLength( $file ) {
817 * True if creating thumbnails from the file is large or otherwise resource-intensive.
821 public function isExpensiveToThumbnail( $file ) {
826 * Returns whether or not this handler supports the chained generation of thumbnails according
831 public function supportsBucketing() {
836 * Returns a normalised params array for which parameters have been cleaned up for bucketing
838 * @param array $params
841 public function sanitizeParamsForBucketing( $params ) {
846 * Gets configuration for the file warning message. Return value of
847 * the following structure:
849 * // Required, module with messages loaded for the client
850 * 'module' => 'example.filewarning.messages',
851 * // Required, array of names of messages
853 * // Required, main warning message
854 * 'main' => 'example-filewarning-main',
855 * // Optional, header for warning dialog
856 * 'header' => 'example-filewarning-header',
857 * // Optional, footer for warning dialog
858 * 'footer' => 'example-filewarning-footer',
859 * // Optional, text for more-information link (see below)
860 * 'info' => 'example-filewarning-info',
862 * // Optional, link for more information
863 * 'link' => 'http://example.com',
866 * Returns null if no warning is necessary.
870 public function getWarningConfig( $file ) {
875 * Converts a dimensions array about a potentially multipage document from an
876 * exhaustive list of ordered page numbers to a list of page ranges
877 * @param array $pagesByDimensions
881 public static function getPageRangesByDimensions( $pagesByDimensions ) {
882 $pageRangesByDimensions = [];
884 foreach ( $pagesByDimensions as $dimensions => $pageList ) {
886 $firstPage = $pageList[0];
887 $lastPage = $firstPage - 1;
889 foreach ( $pageList as $page ) {
890 if ( $page > $lastPage +
1 ) {
891 if ( $firstPage != $lastPage ) {
892 $ranges[] = "$firstPage-$lastPage";
894 $ranges[] = "$firstPage";
903 if ( $firstPage != $lastPage ) {
904 $ranges[] = "$firstPage-$lastPage";
906 $ranges[] = "$firstPage";
909 $pageRangesByDimensions[ $dimensions ] = $ranges;
912 $dimensionsString = [];
913 foreach ( $pageRangesByDimensions as $dimensions => $pageRanges ) {
914 $dimensionsString[] = "$dimensions:" . implode( ',', $pageRanges );
917 return implode( '/', $dimensionsString );
921 * Get useful response headers for GET/HEAD requests for a file with the given metadata
922 * @param array $metadata Contains this handler's unserialized getMetadata() for a file
926 public function getContentHeaders( $metadata ) {
927 return [ 'X-Content-Dimensions' => '' ]; // T175689