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