includes/Feed.php

   1 <?php
   2 /**
   3  * Basic support for outputting syndication feeds in RSS, other formats.
   4  *
   5  * Contain a feed class as well as classes to build rss / atom ... feeds
   6  * Available feeds are defined in Defines.php
   7  *
   8  * Copyright © 2004 Brion Vibber <brion@pobox.com>
   9  * https://www.mediawiki.org/
  10  *
  11  * This program is free software; you can redistribute it and/or modify
  12  * it under the terms of the GNU General Public License as published by
  13  * the Free Software Foundation; either version 2 of the License, or
  14  * (at your option) any later version.
  15  *
  16  * This program is distributed in the hope that it will be useful,
  17  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  18  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  19  * GNU General Public License for more details.
  20  *
  21  * You should have received a copy of the GNU General Public License along
  22  * with this program; if not, write to the Free Software Foundation, Inc.,
  23  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  24  * http://www.gnu.org/copyleft/gpl.html
  25  *
  26  * @file
  27  */
  28
  29 /**
  30  * @defgroup Feed Feed
  31  */
  32
  33 /**
  34  * A base class for basic support for outputting syndication feeds in RSS and other formats.
  35  *
  36  * @ingroup Feed
  37  */
  38 class FeedItem {
  39         /** @var Title */
  40         public $title;
  41
  42         public $description;
  43
  44         public $url;
  45
  46         public $date;
  47
  48         public $author;
  49
  50         public $uniqueId;
  51
  52         public $comments;
  53
  54         public $rssIsPermalink = false;
  55
  56         /**
  57          * @param string|Title $title Item's title
  58          * @param string $description
  59          * @param string $url URL uniquely designating the item.
  60          * @param string $date Item's date
  61          * @param string $author Author's user name
  62          * @param string $comments
  63          */
  64         function __construct( $title, $description, $url, $date = '', $author = '', $comments = '' ) {
  65                 $this->title = $title;
  66                 $this->description = $description;
  67                 $this->url = $url;
  68                 $this->uniqueId = $url;
  69                 $this->date = $date;
  70                 $this->author = $author;
  71                 $this->comments = $comments;
  72         }
  73
  74         /**
  75          * Encode $string so that it can be safely embedded in a XML document
  76          *
  77          * @param string $string String to encode
  78          * @return string
  79          */
  80         public function xmlEncode( $string ) {
  81                 $string = str_replace( "\r\n", "\n", $string );
  82                 $string = preg_replace( '/[\x00-\x08\x0b\x0c\x0e-\x1f]/', '', $string );
  83                 return htmlspecialchars( $string );
  84         }
  85
  86         /**
  87          * Get the unique id of this item; already xml-encoded
  88          * @return string
  89          */
  90         public function getUniqueID() {
  91                 $id = $this->getUniqueIDUnescaped();
  92                 if ( $id ) {
  93                         return $this->xmlEncode( $id );
  94                 }
  95         }
  96
  97         /**
  98          * Get the unique id of this item, without any escaping
  99          * @return string
 100          */
 101         public function getUniqueIdUnescaped() {
 102                 if ( $this->uniqueId ) {
 103                         return wfExpandUrl( $this->uniqueId, PROTO_CURRENT );
 104                 }
 105         }
 106
 107         /**
 108          * Set the unique id of an item
 109          *
 110          * @param string $uniqueId Unique id for the item
 111          * @param bool $rssIsPermalink Set to true if the guid (unique id) is a permalink (RSS feeds only)
 112          */
 113         public function setUniqueId( $uniqueId, $rssIsPermalink = false ) {
 114                 $this->uniqueId = $uniqueId;
 115                 $this->rssIsPermalink = $rssIsPermalink;
 116         }
 117
 118         /**
 119          * Get the title of this item; already xml-encoded
 120          *
 121          * @return string
 122          */
 123         public function getTitle() {
 124                 return $this->xmlEncode( $this->title );
 125         }
 126
 127         /**
 128          * Get the URL of this item; already xml-encoded
 129          *
 130          * @return string
 131          */
 132         public function getUrl() {
 133                 return $this->xmlEncode( $this->url );
 134         }
 135
 136         /** Get the URL of this item without any escaping
 137          *
 138          * @return string
 139          */
 140         public function getUrlUnescaped() {
 141                 return $this->url;
 142         }
 143
 144         /**
 145          * Get the description of this item; already xml-encoded
 146          *
 147          * @return string
 148          */
 149         public function getDescription() {
 150                 return $this->xmlEncode( $this->description );
 151         }
 152
 153         /**
 154          * Get the description of this item without any escaping
 155          *
 156          * @return string
 157          */
 158         public function getDescriptionUnescaped() {
 159                 return $this->description;
 160         }
 161
 162         /**
 163          * Get the language of this item
 164          *
 165          * @return string
 166          */
 167         public function getLanguage() {
 168                 global $wgLanguageCode;
 169                 return LanguageCode::bcp47( $wgLanguageCode );
 170         }
 171
 172         /**
 173          * Get the date of this item
 174          *
 175          * @return string
 176          */
 177         public function getDate() {
 178                 return $this->date;
 179         }
 180
 181         /**
 182          * Get the author of this item; already xml-encoded
 183          *
 184          * @return string
 185          */
 186         public function getAuthor() {
 187                 return $this->xmlEncode( $this->author );
 188         }
 189
 190         /**
 191          * Get the author of this item without any escaping
 192          *
 193          * @return string
 194          */
 195         public function getAuthorUnescaped() {
 196                 return $this->author;
 197         }
 198
 199         /**
 200          * Get the comment of this item; already xml-encoded
 201          *
 202          * @return string
 203          */
 204         public function getComments() {
 205                 return $this->xmlEncode( $this->comments );
 206         }
 207
 208         /**
 209          * Get the comment of this item without any escaping
 210          *
 211          * @return string
 212          */
 213         public function getCommentsUnescaped() {
 214                 return $this->comments;
 215         }
 216
 217         /**
 218          * Quickie hack... strip out wikilinks to more legible form from the comment.
 219          *
 220          * @param string $text Wikitext
 221          * @return string
 222          */
 223         public static function stripComment( $text ) {
 224                 return preg_replace( '/\[\[([^]]*\|)?([^]]+)\]\]/', '\2', $text );
 225         }
 226         /**#@-*/
 227 }
 228
 229 /**
 230  * Class to support the outputting of syndication feeds in Atom and RSS format.
 231  *
 232  * @ingroup Feed
 233  */
 234 abstract class ChannelFeed extends FeedItem {
 235
 236         /** @var TemplateParser */
 237         protected $templateParser;
 238
 239         /**
 240          * @param string|Title $title Feed's title
 241          * @param string $description
 242          * @param string $url URL uniquely designating the feed.
 243          * @param string $date Feed's date
 244          * @param string $author Author's user name
 245          * @param string $comments
 246          */
 247         function __construct( $title, $description, $url, $date = '', $author = '', $comments = '' ) {
 248                 parent::__construct( $title, $description, $url, $date, $author, $comments );
 249                 $this->templateParser = new TemplateParser();
 250         }
 251
 252         /**
 253          * Generate Header of the feed
 254          * @par Example:
 255          * @code
 256          * print "<feed>";
 257          * @endcode
 258          */
 259         abstract public function outHeader();
 260
 261         /**
 262          * Generate an item
 263          * @par Example:
 264          * @code
 265          * print "<item>...</item>";
 266          * @endcode
 267          * @param FeedItem $item
 268          */
 269         abstract public function outItem( $item );
 270
 271         /**
 272          * Generate Footer of the feed
 273          * @par Example:
 274          * @code
 275          * print "</feed>";
 276          * @endcode
 277          */
 278         abstract public function outFooter();
 279
 280         /**
 281          * Setup and send HTTP headers. Don't send any content;
 282          * content might end up being cached and re-sent with
 283          * these same headers later.
 284          *
 285          * This should be called from the outHeader() method,
 286          * but can also be called separately.
 287          */
 288         public function httpHeaders() {
 289                 global $wgOut, $wgVaryOnXFP;
 290
 291                 # We take over from $wgOut, excepting its cache header info
 292                 $wgOut->disable();
 293                 $mimetype = $this->contentType();
 294                 header( "Content-type: $mimetype; charset=UTF-8" );
 295
 296                 // Set a sane filename
 297                 $exts = MediaWiki\MediaWikiServices::getInstance()->getMimeAnalyzer()
 298                         ->getExtensionsForType( $mimetype );
 299                 $ext = $exts ? strtok( $exts, ' ' ) : 'xml';
 300                 header( "Content-Disposition: inline; filename=\"feed.{$ext}\"" );
 301
 302                 if ( $wgVaryOnXFP ) {
 303                         $wgOut->addVaryHeader( 'X-Forwarded-Proto' );
 304                 }
 305                 $wgOut->sendCacheControl();
 306         }
 307
 308         /**
 309          * Return an internet media type to be sent in the headers.
 310          *
 311          * @return string
 312          */
 313         private function contentType() {
 314                 global $wgRequest;
 315
 316                 $ctype = $wgRequest->getVal( 'ctype', 'application/xml' );
 317                 $allowedctypes = [
 318                         'application/xml',
 319                         'text/xml',
 320                         'application/rss+xml',
 321                         'application/atom+xml'
 322                 ];
 323
 324                 return ( in_array( $ctype, $allowedctypes ) ? $ctype : 'application/xml' );
 325         }
 326
 327         /**
 328          * Output the initial XML headers.
 329          */
 330         protected function outXmlHeader() {
 331                 $this->httpHeaders();
 332                 echo '<?xml version="1.0"?>' . "\n";
 333         }
 334 }
 335
 336 /**
 337  * Generate a RSS feed
 338  *
 339  * @ingroup Feed
 340  */
 341 class RSSFeed extends ChannelFeed {
 342
 343         /**
 344          * Format a date given a timestamp. If a timestamp is not given, nothing is returned
 345          *
 346          * @param int|null $ts Timestamp
 347          * @return string|null Date string
 348          */
 349         function formatTime( $ts ) {
 350                 if ( $ts ) {
 351                         return gmdate( 'D, d M Y H:i:s \G\M\T', wfTimestamp( TS_UNIX, $ts ) );
 352                 }
 353         }
 354
 355         /**
 356          * Output an RSS 2.0 header
 357          */
 358         function outHeader() {
 359                 global $wgVersion;
 360
 361                 $this->outXmlHeader();
 362                 // Manually escaping rather than letting Mustache do it because Mustache
 363                 // uses htmlentities, which does not work with XML
 364                 $templateParams = [
 365                         'title' => $this->getTitle(),
 366                         'url' => $this->xmlEncode( wfExpandUrl( $this->getUrlUnescaped(), PROTO_CURRENT ) ),
 367                         'description' => $this->getDescription(),
 368                         'language' => $this->xmlEncode( $this->getLanguage() ),
 369                         'version' => $this->xmlEncode( $wgVersion ),
 370                         'timestamp' => $this->xmlEncode( $this->formatTime( wfTimestampNow() ) )
 371                 ];
 372                 print $this->templateParser->processTemplate( 'RSSHeader', $templateParams );
 373         }
 374
 375         /**
 376          * Output an RSS 2.0 item
 377          * @param FeedItem $item Item to be output
 378          */
 379         function outItem( $item ) {
 380                 // Manually escaping rather than letting Mustache do it because Mustache
 381                 // uses htmlentities, which does not work with XML
 382                 $templateParams = [
 383                         "title" => $item->getTitle(),
 384                         "url" => $this->xmlEncode( wfExpandUrl( $item->getUrlUnescaped(), PROTO_CURRENT ) ),
 385                         "permalink" => $item->rssIsPermalink,
 386                         "uniqueID" => $item->getUniqueId(),
 387                         "description" => $item->getDescription(),
 388                         "date" => $this->xmlEncode( $this->formatTime( $item->getDate() ) ),
 389                         "author" => $item->getAuthor()
 390                 ];
 391                 $comments = $item->getCommentsUnescaped();
 392                 if ( $comments ) {
 393                         $commentsEscaped = $this->xmlEncode( wfExpandUrl( $comments, PROTO_CURRENT ) );
 394                         $templateParams["comments"] = $commentsEscaped;
 395                 }
 396                 print $this->templateParser->processTemplate( 'RSSItem', $templateParams );
 397         }
 398
 399         /**
 400          * Output an RSS 2.0 footer
 401          */
 402         function outFooter() {
 403                 print "</channel></rss>";
 404         }
 405 }
 406
 407 /**
 408  * Generate an Atom feed
 409  *
 410  * @ingroup Feed
 411  */
 412 class AtomFeed extends ChannelFeed {
 413         /**
 414          * Format a date given timestamp, if one is given.
 415          *
 416          * @param string|int|null $timestamp
 417          * @return string|null
 418          */
 419         function formatTime( $timestamp ) {
 420                 if ( $timestamp ) {
 421                         // need to use RFC 822 time format at least for rss2.0
 422                         return gmdate( 'Y-m-d\TH:i:s', wfTimestamp( TS_UNIX, $timestamp ) );
 423                 }
 424         }
 425
 426         /**
 427          * Outputs a basic header for Atom 1.0 feeds.
 428          */
 429         function outHeader() {
 430                 global $wgVersion;
 431                 $this->outXmlHeader();
 432                 // Manually escaping rather than letting Mustache do it because Mustache
 433                 // uses htmlentities, which does not work with XML
 434                 $templateParams = [
 435                         'language' => $this->xmlEncode( $this->getLanguage() ),
 436                         'feedID' => $this->getFeedID(),
 437                         'title' => $this->getTitle(),
 438                         'url' => $this->xmlEncode( wfExpandUrl( $this->getUrlUnescaped(), PROTO_CURRENT ) ),
 439                         'selfUrl' => $this->getSelfUrl(),
 440                         'timestamp' => $this->xmlEncode( $this->formatTime( wfTimestampNow() ) ),
 441                         'description' => $this->getDescription(),
 442                         'version' => $this->xmlEncode( $wgVersion ),
 443                 ];
 444                 print $this->templateParser->processTemplate( 'AtomHeader', $templateParams );
 445         }
 446
 447         /**
 448          * Atom 1.0 requires a unique, opaque IRI as a unique identifier
 449          * for every feed we create. For now just use the URL, but who
 450          * can tell if that's right? If we put options on the feed, do we
 451          * have to change the id? Maybe? Maybe not.
 452          *
 453          * @return string
 454          */
 455         private function getFeedId() {
 456                 return $this->getSelfUrl();
 457         }
 458
 459         /**
 460          * Atom 1.0 requests a self-reference to the feed.
 461          * @return string
 462          */
 463         private function getSelfUrl() {
 464                 global $wgRequest;
 465                 return htmlspecialchars( $wgRequest->getFullRequestURL() );
 466         }
 467
 468         /**
 469          * Output a given item.
 470          * @param FeedItem $item
 471          */
 472         function outItem( $item ) {
 473                 global $wgMimeType;
 474                 // Manually escaping rather than letting Mustache do it because Mustache
 475                 // uses htmlentities, which does not work with XML
 476                 $templateParams = [
 477                         "uniqueID" => $item->getUniqueId(),
 478                         "title" => $item->getTitle(),
 479                         "mimeType" => $this->xmlEncode( $wgMimeType ),
 480                         "url" => $this->xmlEncode( wfExpandUrl( $item->getUrlUnescaped(), PROTO_CURRENT ) ),
 481                         "date" => $this->xmlEncode( $this->formatTime( $item->getDate() ) ),
 482                         "description" => $item->getDescription(),
 483                         "author" => $item->getAuthor()
 484                 ];
 485                 print $this->templateParser->processTemplate( 'AtomItem', $templateParams );
 486         }
 487
 488         /**
 489          * Outputs the footer for Atom 1.0 feed (basically '\</feed\>').
 490          */
 491         function outFooter() {
 492                 print "</feed>";
 493         }
 494 }