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          */
 157         public function getDescriptionUnescaped() {
 158                 return $this->description;
 159         }
 160
 161         /**
 162          * Get the language of this item
 163          *
 164          * @return string
 165          */
 166         public function getLanguage() {
 167                 global $wgLanguageCode;
 168                 return LanguageCode::bcp47( $wgLanguageCode );
 169         }
 170
 171         /**
 172          * Get the date of this item
 173          *
 174          * @return string
 175          */
 176         public function getDate() {
 177                 return $this->date;
 178         }
 179
 180         /**
 181          * Get the author of this item; already xml-encoded
 182          *
 183          * @return string
 184          */
 185         public function getAuthor() {
 186                 return $this->xmlEncode( $this->author );
 187         }
 188
 189         /**
 190          * Get the author of this item without any escaping
 191          *
 192          * @return string
 193          */
 194         public function getAuthorUnescaped() {
 195                 return $this->author;
 196         }
 197
 198         /**
 199          * Get the comment of this item; already xml-encoded
 200          *
 201          * @return string
 202          */
 203         public function getComments() {
 204                 return $this->xmlEncode( $this->comments );
 205         }
 206
 207         /**
 208          * Get the comment of this item without any escaping
 209          *
 210          * @return string
 211          */
 212         public function getCommentsUnescaped() {
 213                 return $this->comments;
 214         }
 215
 216         /**
 217          * Quickie hack... strip out wikilinks to more legible form from the comment.
 218          *
 219          * @param string $text Wikitext
 220          * @return string
 221          */
 222         public static function stripComment( $text ) {
 223                 return preg_replace( '/\[\[([^]]*\|)?([^]]+)\]\]/', '\2', $text );
 224         }
 225         /**#@-*/
 226 }
 227
 228 /**
 229  * Class to support the outputting of syndication feeds in Atom and RSS format.
 230  *
 231  * @ingroup Feed
 232  */
 233 abstract class ChannelFeed extends FeedItem {
 234
 235         /** @var TemplateParser */
 236         protected $templateParser;
 237
 238         /**
 239          * @param string|Title $title Feed's title
 240          * @param string $description
 241          * @param string $url URL uniquely designating the feed.
 242          * @param string $date Feed's date
 243          * @param string $author Author's user name
 244          * @param string $comments
 245          */
 246         function __construct( $title, $description, $url, $date = '', $author = '', $comments = '' ) {
 247                 parent::__construct( $title, $description, $url, $date, $author, $comments );
 248                 $this->templateParser = new TemplateParser();
 249         }
 250
 251         /**
 252          * Generate Header of the feed
 253          * @par Example:
 254          * @code
 255          * print "<feed>";
 256          * @endcode
 257          */
 258         abstract public function outHeader();
 259
 260         /**
 261          * Generate an item
 262          * @par Example:
 263          * @code
 264          * print "<item>...</item>";
 265          * @endcode
 266          * @param FeedItem $item
 267          */
 268         abstract public function outItem( $item );
 269
 270         /**
 271          * Generate Footer of the feed
 272          * @par Example:
 273          * @code
 274          * print "</feed>";
 275          * @endcode
 276          */
 277         abstract public function outFooter();
 278
 279         /**
 280          * Setup and send HTTP headers. Don't send any content;
 281          * content might end up being cached and re-sent with
 282          * these same headers later.
 283          *
 284          * This should be called from the outHeader() method,
 285          * but can also be called separately.
 286          */
 287         public function httpHeaders() {
 288                 global $wgOut, $wgVaryOnXFP;
 289
 290                 # We take over from $wgOut, excepting its cache header info
 291                 $wgOut->disable();
 292                 $mimetype = $this->contentType();
 293                 header( "Content-type: $mimetype; charset=UTF-8" );
 294
 295                 // Set a sane filename
 296                 $exts = MediaWiki\MediaWikiServices::getInstance()->getMimeAnalyzer()
 297                         ->getExtensionsForType( $mimetype );
 298                 $ext = $exts ? strtok( $exts, ' ' ) : 'xml';
 299                 header( "Content-Disposition: inline; filename=\"feed.{$ext}\"" );
 300
 301                 if ( $wgVaryOnXFP ) {
 302                         $wgOut->addVaryHeader( 'X-Forwarded-Proto' );
 303                 }
 304                 $wgOut->sendCacheControl();
 305         }
 306
 307         /**
 308          * Return an internet media type to be sent in the headers.
 309          *
 310          * @return string
 311          */
 312         private function contentType() {
 313                 global $wgRequest;
 314
 315                 $ctype = $wgRequest->getVal( 'ctype', 'application/xml' );
 316                 $allowedctypes = [
 317                         'application/xml',
 318                         'text/xml',
 319                         'application/rss+xml',
 320                         'application/atom+xml'
 321                 ];
 322
 323                 return ( in_array( $ctype, $allowedctypes ) ? $ctype : 'application/xml' );
 324         }
 325
 326         /**
 327          * Output the initial XML headers.
 328          */
 329         protected function outXmlHeader() {
 330                 $this->httpHeaders();
 331                 echo '<?xml version="1.0"?>' . "\n";
 332         }
 333 }
 334
 335 /**
 336  * Generate a RSS feed
 337  *
 338  * @ingroup Feed
 339  */
 340 class RSSFeed extends ChannelFeed {
 341
 342         /**
 343          * Format a date given a timestamp. If a timestamp is not given, nothing is returned
 344          *
 345          * @param int|null $ts Timestamp
 346          * @return string|null Date string
 347          */
 348         function formatTime( $ts ) {
 349                 if ( $ts ) {
 350                         return gmdate( 'D, d M Y H:i:s \G\M\T', wfTimestamp( TS_UNIX, $ts ) );
 351                 }
 352         }
 353
 354         /**
 355          * Output an RSS 2.0 header
 356          */
 357         function outHeader() {
 358                 global $wgVersion;
 359
 360                 $this->outXmlHeader();
 361                 // Manually escaping rather than letting Mustache do it because Mustache
 362                 // uses htmlentities, which does not work with XML
 363                 $templateParams = [
 364                         'title' => $this->getTitle(),
 365                         'url' => $this->xmlEncode( wfExpandUrl( $this->getUrlUnescaped(), PROTO_CURRENT ) ),
 366                         'description' => $this->getDescription(),
 367                         'language' => $this->xmlEncode( $this->getLanguage() ),
 368                         'version' => $this->xmlEncode( $wgVersion ),
 369                         'timestamp' => $this->xmlEncode( $this->formatTime( wfTimestampNow() ) )
 370                 ];
 371                 print $this->templateParser->processTemplate( 'RSSHeader', $templateParams );
 372         }
 373
 374         /**
 375          * Output an RSS 2.0 item
 376          * @param FeedItem $item Item to be output
 377          */
 378         function outItem( $item ) {
 379                 // Manually escaping rather than letting Mustache do it because Mustache
 380                 // uses htmlentities, which does not work with XML
 381                 $templateParams = [
 382                         "title" => $item->getTitle(),
 383                         "url" => $this->xmlEncode( wfExpandUrl( $item->getUrlUnescaped(), PROTO_CURRENT ) ),
 384                         "permalink" => $item->rssIsPermalink,
 385                         "uniqueID" => $item->getUniqueId(),
 386                         "description" => $item->getDescription(),
 387                         "date" => $this->xmlEncode( $this->formatTime( $item->getDate() ) ),
 388                         "author" => $item->getAuthor()
 389                 ];
 390                 $comments = $item->getCommentsUnescaped();
 391                 if ( $comments ) {
 392                         $commentsEscaped = $this->xmlEncode( wfExpandUrl( $comments, PROTO_CURRENT ) );
 393                         $templateParams["comments"] = $commentsEscaped;
 394                 }
 395                 print $this->templateParser->processTemplate( 'RSSItem', $templateParams );
 396         }
 397
 398         /**
 399          * Output an RSS 2.0 footer
 400          */
 401         function outFooter() {
 402                 print "</channel></rss>";
 403         }
 404 }
 405
 406 /**
 407  * Generate an Atom feed
 408  *
 409  * @ingroup Feed
 410  */
 411 class AtomFeed extends ChannelFeed {
 412         /**
 413          * Format a date given timestamp, if one is given.
 414          *
 415          * @param string|int|null $timestamp
 416          * @return string|null
 417          */
 418         function formatTime( $timestamp ) {
 419                 if ( $timestamp ) {
 420                         // need to use RFC 822 time format at least for rss2.0
 421                         return gmdate( 'Y-m-d\TH:i:s', wfTimestamp( TS_UNIX, $timestamp ) );
 422                 }
 423         }
 424
 425         /**
 426          * Outputs a basic header for Atom 1.0 feeds.
 427          */
 428         function outHeader() {
 429                 global $wgVersion;
 430                 $this->outXmlHeader();
 431                 // Manually escaping rather than letting Mustache do it because Mustache
 432                 // uses htmlentities, which does not work with XML
 433                 $templateParams = [
 434                         'language' => $this->xmlEncode( $this->getLanguage() ),
 435                         'feedID' => $this->getFeedID(),
 436                         'title' => $this->getTitle(),
 437                         'url' => $this->xmlEncode( wfExpandUrl( $this->getUrlUnescaped(), PROTO_CURRENT ) ),
 438                         'selfUrl' => $this->getSelfUrl(),
 439                         'timestamp' => $this->xmlEncode( $this->formatTime( wfTimestampNow() ) ),
 440                         'description' => $this->getDescription(),
 441                         'version' => $this->xmlEncode( $wgVersion ),
 442                 ];
 443                 print $this->templateParser->processTemplate( 'AtomHeader', $templateParams );
 444         }
 445
 446         /**
 447          * Atom 1.0 requires a unique, opaque IRI as a unique identifier
 448          * for every feed we create. For now just use the URL, but who
 449          * can tell if that's right? If we put options on the feed, do we
 450          * have to change the id? Maybe? Maybe not.
 451          *
 452          * @return string
 453          */
 454         private function getFeedId() {
 455                 return $this->getSelfUrl();
 456         }
 457
 458         /**
 459          * Atom 1.0 requests a self-reference to the feed.
 460          * @return string
 461          */
 462         private function getSelfUrl() {
 463                 global $wgRequest;
 464                 return htmlspecialchars( $wgRequest->getFullRequestURL() );
 465         }
 466
 467         /**
 468          * Output a given item.
 469          * @param FeedItem $item
 470          */
 471         function outItem( $item ) {
 472                 global $wgMimeType;
 473                 // Manually escaping rather than letting Mustache do it because Mustache
 474                 // uses htmlentities, which does not work with XML
 475                 $templateParams = [
 476                         "uniqueID" => $item->getUniqueId(),
 477                         "title" => $item->getTitle(),
 478                         "mimeType" => $this->xmlEncode( $wgMimeType ),
 479                         "url" => $this->xmlEncode( wfExpandUrl( $item->getUrlUnescaped(), PROTO_CURRENT ) ),
 480                         "date" => $this->xmlEncode( $this->formatTime( $item->getDate() ) ),
 481                         "description" => $item->getDescription(),
 482                         "author" => $item->getAuthor()
 483                 ];
 484                 print $this->templateParser->processTemplate( 'AtomItem', $templateParams );
 485         }
 486
 487         /**
 488          * Outputs the footer for Atom 1.0 feed (basically '\</feed\>').
 489          */
 490         function outFooter() {
 491                 print "</feed>";
 492         }
 493 }