includes/filerepo/ForeignAPIRepo.php

   1 <?php
   2 /**
   3  * Foreign repository accessible through api.php requests.
   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 FileRepo
  22  */
  23
  24 /**
  25  * A foreign repository with a remote MediaWiki with an API thingy
  26  *
  27  * Example config:
  28  *
  29  * $wgForeignFileRepos[] = array(
  30  *   'class'                  => 'ForeignAPIRepo',
  31  *   'name'                   => 'shared',
  32  *   'apibase'                => 'http://en.wikipedia.org/w/api.php',
  33  *   'fetchDescription'       => true, // Optional
  34  *   'descriptionCacheExpiry' => 3600,
  35  * );
  36  *
  37  * @ingroup FileRepo
  38  */
  39 class ForeignAPIRepo extends FileRepo {
  40         /* This version string is used in the user agent for requests and will help
  41          * server maintainers in identify ForeignAPI usage.
  42          * Update the version every time you make breaking or significant changes. */
  43         const VERSION = "2.1";
  44
  45         var $fileFactory = array( 'ForeignAPIFile', 'newFromTitle' );
  46         /* Check back with Commons after a day */
  47         var $apiThumbCacheExpiry = 86400; /* 24*60*60 */
  48         /* Redownload thumbnail files after a month */
  49         var $fileCacheExpiry = 2592000; /* 86400*30 */
  50
  51         protected $mQueryCache = array();
  52         protected $mFileExists = array();
  53
  54         /**
  55          * @param $info array|null
  56          */
  57         function __construct( $info ) {
  58                 global $wgLocalFileRepo;
  59                 parent::__construct( $info );
  60
  61                 // http://commons.wikimedia.org/w/api.php
  62                 $this->mApiBase = isset( $info['apibase'] ) ? $info['apibase'] : null;
  63
  64                 if( isset( $info['apiThumbCacheExpiry'] ) ) {
  65                         $this->apiThumbCacheExpiry = $info['apiThumbCacheExpiry'];
  66                 }
  67                 if( isset( $info['fileCacheExpiry'] ) ) {
  68                         $this->fileCacheExpiry = $info['fileCacheExpiry'];
  69                 }
  70                 if( !$this->scriptDirUrl ) {
  71                         // hack for description fetches
  72                         $this->scriptDirUrl = dirname( $this->mApiBase );
  73                 }
  74                 // If we can cache thumbs we can guess sane defaults for these
  75                 if( $this->canCacheThumbs() && !$this->url ) {
  76                         $this->url = $wgLocalFileRepo['url'];
  77                 }
  78                 if( $this->canCacheThumbs() && !$this->thumbUrl ) {
  79                         $this->thumbUrl = $this->url . '/thumb';
  80                 }
  81         }
  82
  83         /**
  84          * Per docs in FileRepo, this needs to return false if we don't support versioned
  85          * files. Well, we don't.
  86          *
  87          * @param $title Title
  88          * @param $time string|bool
  89          * @return File
  90          */
  91         function newFile( $title, $time = false ) {
  92                 if ( $time ) {
  93                         return false;
  94                 }
  95                 return parent::newFile( $title, $time );
  96         }
  97
  98         /**
  99          * @param $files array
 100          * @return array
 101          */
 102         function fileExistsBatch( array $files ) {
 103                 $results = array();
 104                 foreach ( $files as $k => $f ) {
 105                         if ( isset( $this->mFileExists[$k] ) ) {
 106                                 $results[$k] = true;
 107                                 unset( $files[$k] );
 108                         } elseif( self::isVirtualUrl( $f ) ) {
 109                                 # @todo FIXME: We need to be able to handle virtual
 110                                 # URLs better, at least when we know they refer to the
 111                                 # same repo.
 112                                 $results[$k] = false;
 113                                 unset( $files[$k] );
 114                         } elseif ( FileBackend::isStoragePath( $f ) ) {
 115                                 $results[$k] = false;
 116                                 unset( $files[$k] );
 117                                 wfWarn( "Got mwstore:// path '$f'." );
 118                         }
 119                 }
 120
 121                 $data = $this->fetchImageQuery( array( 'titles' => implode( $files, '|' ),
 122                                                                                         'prop' => 'imageinfo' ) );
 123                 if( isset( $data['query']['pages'] ) ) {
 124                         $i = 0;
 125                         foreach( $files as $key => $file ) {
 126                                 $results[$key] = $this->mFileExists[$key] = !isset( $data['query']['pages'][$i]['missing'] );
 127                                 $i++;
 128                         }
 129                 }
 130                 return $results;
 131         }
 132
 133         /**
 134          * @param $virtualUrl string
 135          * @return bool
 136          */
 137         function getFileProps( $virtualUrl ) {
 138                 return false;
 139         }
 140
 141         /**
 142          * @param $query array
 143          * @return string
 144          */
 145         function fetchImageQuery( $query ) {
 146                 global $wgMemc;
 147
 148                 $query = array_merge( $query,
 149                         array(
 150                                 'format'    => 'json',
 151                                 'action'    => 'query',
 152                                 'redirects' => 'true'
 153                         ) );
 154                 if ( $this->mApiBase ) {
 155                         $url = wfAppendQuery( $this->mApiBase, $query );
 156                 } else {
 157                         $url = $this->makeUrl( $query, 'api' );
 158                 }
 159
 160                 if( !isset( $this->mQueryCache[$url] ) ) {
 161                         $key = $this->getLocalCacheKey( 'ForeignAPIRepo', 'Metadata', md5( $url ) );
 162                         $data = $wgMemc->get( $key );
 163                         if( !$data ) {
 164                                 $data = self::httpGet( $url );
 165                                 if ( !$data ) {
 166                                         return null;
 167                                 }
 168                                 $wgMemc->set( $key, $data, 3600 );
 169                         }
 170
 171                         if( count( $this->mQueryCache ) > 100 ) {
 172                                 // Keep the cache from growing infinitely
 173                                 $this->mQueryCache = array();
 174                         }
 175                         $this->mQueryCache[$url] = $data;
 176                 }
 177                 return FormatJson::decode( $this->mQueryCache[$url], true );
 178         }
 179
 180         /**
 181          * @param $data array
 182          * @return bool|array
 183          */
 184         function getImageInfo( $data ) {
 185                 if( $data && isset( $data['query']['pages'] ) ) {
 186                         foreach( $data['query']['pages'] as $info ) {
 187                                 if( isset( $info['imageinfo'][0] ) ) {
 188                                         return $info['imageinfo'][0];
 189                                 }
 190                         }
 191                 }
 192                 return false;
 193         }
 194
 195         /**
 196          * @param $hash string
 197          * @return array
 198          */
 199         function findBySha1( $hash ) {
 200                 $results = $this->fetchImageQuery( array(
 201                                                                                 'aisha1base36' => $hash,
 202                                                                                 'aiprop'       => ForeignAPIFile::getProps(),
 203                                                                                 'list'         => 'allimages', ) );
 204                 $ret = array();
 205                 if ( isset( $results['query']['allimages'] ) ) {
 206                         foreach ( $results['query']['allimages'] as $img ) {
 207                                 // 1.14 was broken, doesn't return name attribute
 208                                 if( !isset( $img['name'] ) ) {
 209                                         continue;
 210                                 }
 211                                 $ret[] = new ForeignAPIFile( Title::makeTitle( NS_FILE, $img['name'] ), $this, $img );
 212                         }
 213                 }
 214                 return $ret;
 215         }
 216
 217         /**
 218          * @param $name string
 219          * @param $width int
 220          * @param $height int
 221          * @param $result null
 222          * @param $otherParams string
 223          * @return bool
 224          */
 225         function getThumbUrl( $name, $width = -1, $height = -1, &$result = null, $otherParams = '' ) {
 226                 $data = $this->fetchImageQuery( array(
 227                         'titles' => 'File:' . $name,
 228                         'iiprop' => 'url|timestamp',
 229                         'iiurlwidth' => $width,
 230                         'iiurlheight' => $height,
 231                         'iiurlparam'  => $otherParams,
 232                         'prop' => 'imageinfo' ) );
 233                 $info = $this->getImageInfo( $data );
 234
 235                 if( $data && $info && isset( $info['thumburl'] ) ) {
 236                         wfDebug( __METHOD__ . " got remote thumb " . $info['thumburl'] . "\n" );
 237                         $result = $info;
 238                         return $info['thumburl'];
 239                 } else {
 240                         return false;
 241                 }
 242         }
 243
 244         /**
 245          * Return the imageurl from cache if possible
 246          *
 247          * If the url has been requested today, get it from cache
 248          * Otherwise retrieve remote thumb url, check for local file.
 249          *
 250          * @param $name String is a dbkey form of a title
 251          * @param $width
 252          * @param $height
 253          * @param String $params Other rendering parameters (page number, etc) from handler's makeParamString.
 254          * @return bool|string
 255          */
 256         function getThumbUrlFromCache( $name, $width, $height, $params = "" ) {
 257                 global $wgMemc;
 258                 // We can't check the local cache using FileRepo functions because
 259                 // we override fileExistsBatch(). We have to use the FileBackend directly.
 260                 $backend = $this->getBackend(); // convenience
 261
 262                 if ( !$this->canCacheThumbs() ) {
 263                         $result = null; // can't pass "null" by reference, but it's ok as default value
 264                         return $this->getThumbUrl( $name, $width, $height, $result, $params );
 265                 }
 266                 $key = $this->getLocalCacheKey( 'ForeignAPIRepo', 'ThumbUrl', $name );
 267                 $sizekey = "$width:$height:$params";
 268
 269                 /* Get the array of urls that we already know */
 270                 $knownThumbUrls = $wgMemc->get($key);
 271                 if( !$knownThumbUrls ) {
 272                         /* No knownThumbUrls for this file */
 273                         $knownThumbUrls = array();
 274                 } else {
 275                         if( isset( $knownThumbUrls[$sizekey] ) ) {
 276                                 wfDebug( __METHOD__ . ': Got thumburl from local cache: ' .
 277                                         "{$knownThumbUrls[$sizekey]} \n");
 278                                 return $knownThumbUrls[$sizekey];
 279                         }
 280                         /* This size is not yet known */
 281                 }
 282
 283                 $metadata = null;
 284                 $foreignUrl = $this->getThumbUrl( $name, $width, $height, $metadata, $params );
 285
 286                 if( !$foreignUrl ) {
 287                         wfDebug( __METHOD__ . " Could not find thumburl\n" );
 288                         return false;
 289                 }
 290
 291                 // We need the same filename as the remote one :)
 292                 $fileName = rawurldecode( pathinfo( $foreignUrl, PATHINFO_BASENAME ) );
 293                 if( !$this->validateFilename( $fileName ) ) {
 294                         wfDebug( __METHOD__ . " The deduced filename $fileName is not safe\n" );
 295                         return false;
 296                 }
 297                 $localPath =  $this->getZonePath( 'thumb' ) . "/" . $this->getHashPath( $name ) . $name;
 298                 $localFilename = $localPath . "/" . $fileName;
 299                 $localUrl =  $this->getZoneUrl( 'thumb' ) . "/" . $this->getHashPath( $name ) . rawurlencode( $name ) . "/" . rawurlencode( $fileName );
 300
 301                 if( $backend->fileExists( array( 'src' => $localFilename ) )
 302                         && isset( $metadata['timestamp'] ) )
 303                 {
 304                         wfDebug( __METHOD__ . " Thumbnail was already downloaded before\n" );
 305                         $modified = $backend->getFileTimestamp( array( 'src' => $localFilename ) );
 306                         $remoteModified = strtotime( $metadata['timestamp'] );
 307                         $current = time();
 308                         $diff = abs( $modified - $current );
 309                         if( $remoteModified < $modified && $diff < $this->fileCacheExpiry ) {
 310                                 /* Use our current and already downloaded thumbnail */
 311                                 $knownThumbUrls[$sizekey] = $localUrl;
 312                                 $wgMemc->set( $key, $knownThumbUrls, $this->apiThumbCacheExpiry );
 313                                 return $localUrl;
 314                         }
 315                         /* There is a new Commons file, or existing thumbnail older than a month */
 316                 }
 317                 $thumb = self::httpGet( $foreignUrl );
 318                 if( !$thumb ) {
 319                         wfDebug( __METHOD__ . " Could not download thumb\n" );
 320                         return false;
 321                 }
 322
 323                 # @todo FIXME: Delete old thumbs that aren't being used. Maintenance script?
 324                 $backend->prepare( array( 'dir' => dirname( $localFilename ) ) );
 325                 $op = array( 'op' => 'create', 'dst' => $localFilename, 'content' => $thumb );
 326                 if( !$backend->doOperation( $op )->isOK() ) {
 327                         wfRestoreWarnings();
 328                         wfDebug( __METHOD__ . " could not write to thumb path '$localFilename'\n" );
 329                         return $foreignUrl;
 330                 }
 331                 $knownThumbUrls[$sizekey] = $localUrl;
 332                 $wgMemc->set( $key, $knownThumbUrls, $this->apiThumbCacheExpiry );
 333                 wfDebug( __METHOD__ . " got local thumb $localUrl, saving to cache \n" );
 334                 return $localUrl;
 335         }
 336
 337         /**
 338          * @see FileRepo::getZoneUrl()
 339          * @param $zone String
 340          * @return String
 341          */
 342         function getZoneUrl( $zone ) {
 343                 switch ( $zone ) {
 344                         case 'public':
 345                                 return $this->url;
 346                         case 'thumb':
 347                                 return $this->thumbUrl;
 348                         default:
 349                                 return parent::getZoneUrl( $zone );
 350                 }
 351         }
 352
 353         /**
 354          * Get the local directory corresponding to one of the basic zones
 355          * @param $zone string
 356          * @return bool|null|string
 357          */
 358         function getZonePath( $zone ) {
 359                 $supported = array( 'public', 'thumb' );
 360                 if ( in_array( $zone, $supported ) ) {
 361                         return parent::getZonePath( $zone );
 362                 }
 363                 return false;
 364         }
 365
 366         /**
 367          * Are we locally caching the thumbnails?
 368          * @return bool
 369          */
 370         public function canCacheThumbs() {
 371                 return ( $this->apiThumbCacheExpiry > 0 );
 372         }
 373
 374         /**
 375          * The user agent the ForeignAPIRepo will use.
 376          * @return string
 377          */
 378         public static function getUserAgent() {
 379                 return Http::userAgent() . " ForeignAPIRepo/" . self::VERSION;
 380         }
 381
 382         /**
 383          * Like a Http:get request, but with custom User-Agent.
 384          * @see Http:get
 385          * @param $url string
 386          * @param $timeout string
 387          * @param $options array
 388          * @return bool|String
 389          */
 390         public static function httpGet( $url, $timeout = 'default', $options = array() ) {
 391                 $options['timeout'] = $timeout;
 392                 /* Http::get */
 393                 $url = wfExpandUrl( $url, PROTO_HTTP );
 394                 wfDebug( "ForeignAPIRepo: HTTP GET: $url\n" );
 395                 $options['method'] = "GET";
 396
 397                 if ( !isset( $options['timeout'] ) ) {
 398                         $options['timeout'] = 'default';
 399                 }
 400
 401                 $req = MWHttpRequest::factory( $url, $options );
 402                 $req->setUserAgent( ForeignAPIRepo::getUserAgent() );
 403                 $status = $req->execute();
 404
 405                 if ( $status->isOK() ) {
 406                         return $req->getContent();
 407                 } else {
 408                         return false;
 409                 }
 410         }
 411
 412         /**
 413          * @param $callback Array|string
 414          * @throws MWException
 415          */
 416         function enumFiles( $callback ) {
 417                 throw new MWException( 'enumFiles is not supported by ' . get_class( $this ) );
 418         }
 419
 420         /**
 421          * @throws MWException
 422          */
 423         protected function assertWritableRepo() {
 424                 throw new MWException( get_class( $this ) . ': write operations are not supported.' );
 425         }
 426 }