includes/media/MediaHandler.php

   1 <?php
   2 /**
   3  * Media-handling base classes and generic functionality.
   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  * Base media handler class
  26  *
  27  * @ingroup Media
  28  */
  29 abstract class MediaHandler {
  30         const TRANSFORM_LATER = 1;
  31         const METADATA_GOOD = true;
  32         const METADATA_BAD = false;
  33         const METADATA_COMPATIBLE = 2; // for old but backwards compatible.
  34         /**
  35          * Instance cache
  36          */
  37         static $handlers = array();
  38
  39         /**
  40          * Get a MediaHandler for a given MIME type from the instance cache
  41          *
  42          * @param $type string
  43          *
  44          * @return MediaHandler
  45          */
  46         static function getHandler( $type ) {
  47                 global $wgMediaHandlers;
  48                 if ( !isset( $wgMediaHandlers[$type] ) ) {
  49                         wfDebug( __METHOD__ . ": no handler found for $type.\n" );
  50                         return false;
  51                 }
  52                 $class = $wgMediaHandlers[$type];
  53                 if ( !isset( self::$handlers[$class] ) ) {
  54                         self::$handlers[$class] = new $class;
  55                         if ( !self::$handlers[$class]->isEnabled() ) {
  56                                 self::$handlers[$class] = false;
  57                         }
  58                 }
  59                 return self::$handlers[$class];
  60         }
  61
  62         /**
  63          * Get an associative array mapping magic word IDs to parameter names.
  64          * Will be used by the parser to identify parameters.
  65          */
  66         abstract function getParamMap();
  67
  68         /**
  69          * Validate a thumbnail parameter at parse time.
  70          * Return true to accept the parameter, and false to reject it.
  71          * If you return false, the parser will do something quiet and forgiving.
  72          *
  73          * @param $name
  74          * @param $value
  75          */
  76         abstract function validateParam( $name, $value );
  77
  78         /**
  79          * Merge a parameter array into a string appropriate for inclusion in filenames
  80          *
  81          * @param $params array Array of parameters that have been through normaliseParams.
  82          */
  83         abstract function makeParamString( $params );
  84
  85         /**
  86          * Parse a param string made with makeParamString back into an array
  87          *
  88          * @param $str string
  89          */
  90         abstract function parseParamString( $str );
  91
  92         /**
  93          * Changes the parameter array as necessary, ready for transformation.
  94          * Should be idempotent.
  95          * Returns false if the parameters are unacceptable and the transform should fail
  96          * @param $image
  97          * @param $params
  98          */
  99         abstract function normaliseParams( $image, &$params );
 100
 101         /**
 102          * Get an image size array like that returned by getimagesize(), or false if it
 103          * can't be determined.
 104          *
 105          * @param $image File: the image object, or false if there isn't one
 106          * @param string $path the filename
 107          * @return Array Follow the format of PHP getimagesize() internal function. See http://www.php.net/getimagesize
 108          */
 109         abstract function getImageSize( $image, $path );
 110
 111         /**
 112          * Get handler-specific metadata which will be saved in the img_metadata field.
 113          *
 114          * @param $image File: the image object, or false if there isn't one.
 115          *   Warning, FSFile::getPropsFromPath might pass an (object)array() instead (!)
 116          * @param string $path the filename
 117          * @return String
 118          */
 119         function getMetadata( $image, $path ) {
 120                 return '';
 121         }
 122
 123         /**
 124          * Get metadata version.
 125          *
 126          * This is not used for validating metadata, this is used for the api when returning
 127          * metadata, since api content formats should stay the same over time, and so things
 128          * using ForiegnApiRepo can keep backwards compatibility
 129          *
 130          * All core media handlers share a common version number, and extensions can
 131          * use the GetMetadataVersion hook to append to the array (they should append a unique
 132          * string so not to get confusing). If there was a media handler named 'foo' with metadata
 133          * version 3 it might add to the end of the array the element 'foo=3'. if the core metadata
 134          * version is 2, the end version string would look like '2;foo=3'.
 135          *
 136          * @return string version string
 137          */
 138         static function getMetadataVersion() {
 139                 $version = Array( '2' ); // core metadata version
 140                 wfRunHooks( 'GetMetadataVersion', Array( &$version ) );
 141                 return implode( ';', $version );
 142         }
 143
 144         /**
 145          * Convert metadata version.
 146          *
 147          * By default just returns $metadata, but can be used to allow
 148          * media handlers to convert between metadata versions.
 149          *
 150          * @param $metadata Mixed String or Array metadata array (serialized if string)
 151          * @param $version Integer target version
 152          * @return Array serialized metadata in specified version, or $metadata on fail.
 153          */
 154         function convertMetadataVersion( $metadata, $version = 1 ) {
 155                 if ( !is_array( $metadata ) ) {
 156
 157                         //unserialize to keep return parameter consistent.
 158                         wfSuppressWarnings();
 159                         $ret = unserialize( $metadata );
 160                         wfRestoreWarnings();
 161                         return $ret;
 162                 }
 163                 return $metadata;
 164         }
 165
 166         /**
 167          * Get a string describing the type of metadata, for display purposes.
 168          *
 169          * @return string
 170          */
 171         function getMetadataType( $image ) {
 172                 return false;
 173         }
 174
 175         /**
 176          * Check if the metadata string is valid for this handler.
 177          * If it returns MediaHandler::METADATA_BAD (or false), Image
 178          * will reload the metadata from the file and update the database.
 179          * MediaHandler::METADATA_GOOD for if the metadata is a-ok,
 180          * MediaHanlder::METADATA_COMPATIBLE if metadata is old but backwards
 181          * compatible (which may or may not trigger a metadata reload).
 182          * @return bool
 183          */
 184         function isMetadataValid( $image, $metadata ) {
 185                 return self::METADATA_GOOD;
 186         }
 187
 188         /**
 189          * Get a MediaTransformOutput object representing an alternate of the transformed
 190          * output which will call an intermediary thumbnail assist script.
 191          *
 192          * Used when the repository has a thumbnailScriptUrl option configured.
 193          *
 194          * Return false to fall back to the regular getTransform().
 195          * @return bool
 196          */
 197         function getScriptedTransform( $image, $script, $params ) {
 198                 return false;
 199         }
 200
 201         /**
 202          * Get a MediaTransformOutput object representing the transformed output. Does not
 203          * actually do the transform.
 204          *
 205          * @param $image File: the image object
 206          * @param string $dstPath filesystem destination path
 207          * @param string $dstUrl Destination URL to use in output HTML
 208          * @param array $params Arbitrary set of parameters validated by $this->validateParam()
 209          * @return MediaTransformOutput
 210          */
 211         final function getTransform( $image, $dstPath, $dstUrl, $params ) {
 212                 return $this->doTransform( $image, $dstPath, $dstUrl, $params, self::TRANSFORM_LATER );
 213         }
 214
 215         /**
 216          * Get a MediaTransformOutput object representing the transformed output. Does the
 217          * transform unless $flags contains self::TRANSFORM_LATER.
 218          *
 219          * @param $image File: the image object
 220          * @param string $dstPath filesystem destination path
 221          * @param string $dstUrl destination URL to use in output HTML
 222          * @param array $params arbitrary set of parameters validated by $this->validateParam()
 223          *   Note: These parameters have *not* gone through $this->normaliseParams()
 224          * @param $flags Integer: a bitfield, may contain self::TRANSFORM_LATER
 225          *
 226          * @return MediaTransformOutput
 227          */
 228         abstract function doTransform( $image, $dstPath, $dstUrl, $params, $flags = 0 );
 229
 230         /**
 231          * Get the thumbnail extension and MIME type for a given source MIME type
 232          *
 233          * @param String $ext Extension of original file
 234          * @param String $mime Mime type of original file
 235          * @param Array $params Handler specific rendering parameters
 236          * @return array thumbnail extension and MIME type
 237          */
 238         function getThumbType( $ext, $mime, $params = null ) {
 239                 $magic = MimeMagic::singleton();
 240                 if ( !$ext || $magic->isMatchingExtension( $ext, $mime ) === false ) {
 241                         // The extension is not valid for this mime type and we do
 242                         // recognize the mime type
 243                         $extensions = $magic->getExtensionsForType( $mime );
 244                         if ( $extensions ) {
 245                                 return array( strtok( $extensions, ' ' ), $mime );
 246                         }
 247                 }
 248
 249                 // The extension is correct (true) or the mime type is unknown to
 250                 // MediaWiki (null)
 251                 return array( $ext, $mime );
 252         }
 253
 254         /**
 255          * Get useful response headers for GET/HEAD requests for a file with the given metadata
 256          * @param $metadata mixed Result of the getMetadata() function of this handler for a file
 257          * @return Array
 258          */
 259         public function getStreamHeaders( $metadata ) {
 260                 return array();
 261         }
 262
 263         /**
 264          * True if the handled types can be transformed
 265          * @return bool
 266          */
 267         function canRender( $file ) {
 268                 return true;
 269         }
 270
 271         /**
 272          * True if handled types cannot be displayed directly in a browser
 273          * but can be rendered
 274          * @return bool
 275          */
 276         function mustRender( $file ) {
 277                 return false;
 278         }
 279
 280         /**
 281          * True if the type has multi-page capabilities
 282          * @return bool
 283          */
 284         function isMultiPage( $file ) {
 285                 return false;
 286         }
 287
 288         /**
 289          * Page count for a multi-page document, false if unsupported or unknown
 290          * @return bool
 291          */
 292         function pageCount( $file ) {
 293                 return false;
 294         }
 295
 296         /**
 297          * The material is vectorized and thus scaling is lossless
 298          * @return bool
 299          */
 300         function isVectorized( $file ) {
 301                 return false;
 302         }
 303
 304         /**
 305          * The material is an image, and is animated.
 306          * In particular, video material need not return true.
 307          * @note Before 1.20, this was a method of ImageHandler only
 308          * @return bool
 309          */
 310         function isAnimatedImage( $file ) {
 311                 return false;
 312         }
 313
 314         /**
 315          * If the material is animated, we can animate the thumbnail
 316          * @since 1.20
 317          * @return bool If material is not animated, handler may return any value.
 318          */
 319         function canAnimateThumbnail( $file ) {
 320                 return true;
 321         }
 322
 323         /**
 324          * False if the handler is disabled for all files
 325          * @return bool
 326          */
 327         function isEnabled() {
 328                 return true;
 329         }
 330
 331         /**
 332          * Get an associative array of page dimensions
 333          * Currently "width" and "height" are understood, but this might be
 334          * expanded in the future.
 335          * Returns false if unknown.
 336          *
 337          * It is expected that handlers for paged media (e.g. DjVuHandler)
 338          * will override this method so that it gives the correct results
 339          * for each specific page of the file, using the $page argument.
 340          *
 341          * @note For non-paged media, use getImageSize.
 342          *
 343          * @param $image File
 344          * @param $page What page to get dimensions of
 345          * @return array|bool
 346          */
 347         function getPageDimensions( $image, $page ) {
 348                 $gis = $this->getImageSize( $image, $image->getLocalRefPath() );
 349                 if ( $gis ) {
 350                         return array(
 351                                 'width' => $gis[0],
 352                                 'height' => $gis[1]
 353                         );
 354                 } else {
 355                         return false;
 356                 }
 357         }
 358
 359         /**
 360          * Generic getter for text layer.
 361          * Currently overloaded by PDF and DjVu handlers
 362          * @return bool
 363          */
 364         function getPageText( $image, $page ) {
 365                 return false;
 366         }
 367
 368         /**
 369          * Get an array structure that looks like this:
 370          *
 371          * array(
 372          *    'visible' => array(
 373          *       'Human-readable name' => 'Human readable value',
 374          *       ...
 375          *    ),
 376          *    'collapsed' => array(
 377          *       'Human-readable name' => 'Human readable value',
 378          *       ...
 379          *    )
 380          * )
 381          * The UI will format this into a table where the visible fields are always
 382          * visible, and the collapsed fields are optionally visible.
 383          *
 384          * The function should return false if there is no metadata to display.
 385          */
 386
 387         /**
 388          * @todo FIXME: I don't really like this interface, it's not very flexible
 389          * I think the media handler should generate HTML instead. It can do
 390          * all the formatting according to some standard. That makes it possible
 391          * to do things like visual indication of grouped and chained streams
 392          * in ogg container files.
 393          * @return bool
 394          */
 395         function formatMetadata( $image ) {
 396                 return false;
 397         }
 398
 399         /** sorts the visible/invisible field.
 400          * Split off from ImageHandler::formatMetadata, as used by more than
 401          * one type of handler.
 402          *
 403          * This is used by the media handlers that use the FormatMetadata class
 404          *
 405          * @param array $metadataArray metadata array
 406          * @return array for use displaying metadata.
 407          */
 408         function formatMetadataHelper( $metadataArray ) {
 409                 $result = array(
 410                         'visible' => array(),
 411                         'collapsed' => array()
 412                 );
 413
 414                 $formatted = FormatMetadata::getFormattedData( $metadataArray );
 415                 // Sort fields into visible and collapsed
 416                 $visibleFields = $this->visibleMetadataFields();
 417                 foreach ( $formatted as $name => $value ) {
 418                         $tag = strtolower( $name );
 419                         self::addMeta( $result,
 420                                 in_array( $tag, $visibleFields ) ? 'visible' : 'collapsed',
 421                                 'exif',
 422                                 $tag,
 423                                 $value
 424                         );
 425                 }
 426                 return $result;
 427         }
 428
 429         /**
 430          * Get a list of metadata items which should be displayed when
 431          * the metadata table is collapsed.
 432          *
 433          * @return array of strings
 434          * @access protected
 435          */
 436         function visibleMetadataFields() {
 437                 $fields = array();
 438                 $lines = explode( "\n", wfMessage( 'metadata-fields' )->inContentLanguage()->text() );
 439                 foreach ( $lines as $line ) {
 440                         $matches = array();
 441                         if ( preg_match( '/^\\*\s*(.*?)\s*$/', $line, $matches ) ) {
 442                                 $fields[] = $matches[1];
 443                         }
 444                 }
 445                 $fields = array_map( 'strtolower', $fields );
 446                 return $fields;
 447         }
 448
 449         /**
 450          * This is used to generate an array element for each metadata value
 451          * That array is then used to generate the table of metadata values
 452          * on the image page
 453          *
 454          * @param &$array Array An array containing elements for each type of visibility
 455          * and each of those elements being an array of metadata items. This function adds
 456          * a value to that array.
 457          * @param string $visibility ('visible' or 'collapsed') if this value is hidden
 458          * by default.
 459          * @param string $type type of metadata tag (currently always 'exif')
 460          * @param string $id the name of the metadata tag (like 'artist' for example).
 461          * its name in the table displayed is the message "$type-$id" (Ex exif-artist ).
 462          * @param string $value thingy goes into a wikitext table; it used to be escaped but
 463          * that was incompatible with previous practise of customized display
 464          * with wikitext formatting via messages such as 'exif-model-value'.
 465          * So the escaping is taken back out, but generally this seems a confusing
 466          * interface.
 467          * @param string $param value to pass to the message for the name of the field
 468          * as $1. Currently this parameter doesn't seem to ever be used.
 469          *
 470          * Note, everything here is passed through the parser later on (!)
 471          */
 472         protected static function addMeta( &$array, $visibility, $type, $id, $value, $param = false ) {
 473                 $msg = wfMessage( "$type-$id", $param );
 474                 if ( $msg->exists() ) {
 475                         $name = $msg->text();
 476                 } else {
 477                         // This is for future compatibility when using instant commons.
 478                         // So as to not display as ugly a name if a new metadata
 479                         // property is defined that we don't know about
 480                         // (not a major issue since such a property would be collapsed
 481                         // by default).
 482                         wfDebug( __METHOD__ . ' Unknown metadata name: ' . $id . "\n" );
 483                         $name = wfEscapeWikiText( $id );
 484                 }
 485                 $array[$visibility][] = array(
 486                         'id' => "$type-$id",
 487                         'name' => $name,
 488                         'value' => $value
 489                 );
 490         }
 491
 492         /**
 493          * Used instead of getLongDesc if there is no handler registered for file.
 494          *
 495          * @param $file File
 496          * @return string
 497          */
 498         function getShortDesc( $file ) {
 499                 global $wgLang;
 500                 return htmlspecialchars( $wgLang->formatSize( $file->getSize() ) );
 501         }
 502
 503         /**
 504          * Short description. Shown on Special:Search results.
 505          *
 506          * @param $file File
 507          * @return string
 508          */
 509         function getLongDesc( $file ) {
 510                 global $wgLang;
 511                 return wfMessage( 'file-info', htmlspecialchars( $wgLang->formatSize( $file->getSize() ) ),
 512                         $file->getMimeType() )->parse();
 513         }
 514
 515         /**
 516          * Long description. Shown under image on image description page surounded by ().
 517          *
 518          * @param $file File
 519          * @return string
 520          */
 521         static function getGeneralShortDesc( $file ) {
 522                 global $wgLang;
 523                 return $wgLang->formatSize( $file->getSize() );
 524         }
 525
 526         /**
 527          * Used instead of getShortDesc if there is no handler registered for file.
 528          *
 529          * @param $file File
 530          * @return string
 531          */
 532         static function getGeneralLongDesc( $file ) {
 533                 global $wgLang;
 534                 return wfMessage( 'file-info', $wgLang->formatSize( $file->getSize() ),
 535                         $file->getMimeType() )->parse();
 536         }
 537
 538         /**
 539          * Calculate the largest thumbnail width for a given original file size
 540          * such that the thumbnail's height is at most $maxHeight.
 541          * @param $boxWidth Integer Width of the thumbnail box.
 542          * @param $boxHeight Integer Height of the thumbnail box.
 543          * @param $maxHeight Integer Maximum height expected for the thumbnail.
 544          * @return Integer.
 545          */
 546         public static function fitBoxWidth( $boxWidth, $boxHeight, $maxHeight ) {
 547                 $idealWidth = $boxWidth * $maxHeight / $boxHeight;
 548                 $roundedUp = ceil( $idealWidth );
 549                 if ( round( $roundedUp * $boxHeight / $boxWidth ) > $maxHeight ) {
 550                         return floor( $idealWidth );
 551                 } else {
 552                         return $roundedUp;
 553                 }
 554         }
 555
 556         /**
 557          * Shown in file history box on image description page.
 558          *
 559          * @param File $file
 560          * @return String Dimensions
 561          */
 562         function getDimensionsString( $file ) {
 563                 return '';
 564         }
 565
 566         /**
 567          * Modify the parser object post-transform.
 568          *
 569          * This is often used to do $parser->addOutputHook(),
 570          * in order to add some javascript to render a viewer.
 571          * See TimedMediaHandler or OggHandler for an example.
 572          *
 573          * @param Parser $parser
 574          * @param File $file
 575          */
 576         function parserTransformHook( $parser, $file ) {}
 577
 578         /**
 579          * File validation hook called on upload.
 580          *
 581          * If the file at the given local path is not valid, or its MIME type does not
 582          * match the handler class, a Status object should be returned containing
 583          * relevant errors.
 584          *
 585          * @param string $fileName The local path to the file.
 586          * @return Status object
 587          */
 588         function verifyUpload( $fileName ) {
 589                 return Status::newGood();
 590         }
 591
 592         /**
 593          * Check for zero-sized thumbnails. These can be generated when
 594          * no disk space is available or some other error occurs
 595          *
 596          * @param string $dstPath The location of the suspect file
 597          * @param int $retval Return value of some shell process, file will be deleted if this is non-zero
 598          * @return bool True if removed, false otherwise
 599          */
 600         function removeBadFile( $dstPath, $retval = 0 ) {
 601                 if ( file_exists( $dstPath ) ) {
 602                         $thumbstat = stat( $dstPath );
 603                         if ( $thumbstat['size'] == 0 || $retval != 0 ) {
 604                                 $result = unlink( $dstPath );
 605
 606                                 if ( $result ) {
 607                                         wfDebugLog( 'thumbnail',
 608                                                 sprintf( 'Removing bad %d-byte thumbnail "%s". unlink() succeeded',
 609                                                         $thumbstat['size'], $dstPath ) );
 610                                 } else {
 611                                         wfDebugLog( 'thumbnail',
 612                                                 sprintf( 'Removing bad %d-byte thumbnail "%s". unlink() failed',
 613                                                         $thumbstat['size'], $dstPath ) );
 614                                 }
 615                                 return true;
 616                         }
 617                 }
 618                 return false;
 619         }
 620
 621         /**
 622          * Remove files from the purge list.
 623          *
 624          * This is used by some video handlers to prevent ?action=purge
 625          * from removing a transcoded video, which is expensive to
 626          * regenerate.
 627          *
 628          * @see LocalFile::purgeThumbnails
 629          *
 630          * @param array $files
 631          * @param array $options Purge options. Currently will always be
 632          *  an array with a single key 'forThumbRefresh' set to true.
 633          */
 634         public function filterThumbnailPurgeList( &$files, $options ) {
 635                 // Do nothing
 636         }
 637
 638         /*
 639          * True if the handler can rotate the media
 640          * @since 1.21
 641          * @return bool
 642          */
 643         public static function canRotate() {
 644                 return false;
 645         }
 646 }