includes/api/ApiParse.php

   1 <?php
   2 /**
   3  * Created on Dec 01, 2007
   4  *
   5  * Copyright © 2007 Yuri Astrakhan <Firstname><Lastname>@gmail.com
   6  *
   7  * This program is free software; you can redistribute it and/or modify
   8  * it under the terms of the GNU General Public License as published by
   9  * the Free Software Foundation; either version 2 of the License, or
  10  * (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  15  * GNU General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU General Public License along
  18  * with this program; if not, write to the Free Software Foundation, Inc.,
  19  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  20  * http://www.gnu.org/copyleft/gpl.html
  21  *
  22  * @file
  23  */
  24
  25 if ( !defined( 'MEDIAWIKI' ) ) {
  26         // Eclipse helper - will be ignored in production
  27         require_once( "ApiBase.php" );
  28 }
  29
  30 /**
  31  * @ingroup API
  32  */
  33 class ApiParse extends ApiBase {
  34         private $section, $text, $pstText = null;
  35
  36         public function __construct( $main, $action ) {
  37                 parent::__construct( $main, $action );
  38         }
  39
  40         public function execute() {
  41                 // The data is hot but user-dependent, like page views, so we set vary cookies
  42                 $this->getMain()->setCacheMode( 'anon-public-user-private' );
  43
  44                 // Get parameters
  45                 $params = $this->extractRequestParams();
  46                 $text = $params['text'];
  47                 $title = $params['title'];
  48                 $page = $params['page'];
  49                 $pageid = $params['pageid'];
  50                 $oldid = $params['oldid'];
  51
  52                 if ( !is_null( $page ) && ( !is_null( $text ) || $title != 'API' ) ) {
  53                         $this->dieUsage( 'The page parameter cannot be used together with the text and title parameters', 'params' );
  54                 }
  55                 $prop = array_flip( $params['prop'] );
  56
  57                 if ( isset( $params['section'] ) ) {
  58                         $this->section = $params['section'];
  59                 } else {
  60                         $this->section = false;
  61                 }
  62
  63                 // The parser needs $wgTitle to be set, apparently the
  64                 // $title parameter in Parser::parse isn't enough *sigh*
  65                 global $wgParser, $wgUser, $wgTitle, $wgLang;
  66
  67                 // Currently unnecessary, code to act as a safeguard against any change in current behaviour of uselang breaks
  68                 $oldLang = null;
  69                 if ( isset( $params['uselang'] ) && $params['uselang'] != $wgLang->getCode() ) {
  70                         $oldLang = $wgLang; // Backup wgLang
  71                         $wgLang = Language::factory( $params['uselang'] );
  72                 }
  73
  74                 $popts = new ParserOptions();
  75                 $popts->setTidy( true );
  76                 $popts->enableLimitReport( !$params['disablepp'] );
  77
  78                 $redirValues = null;
  79
  80                 if ( !is_null( $oldid ) || !is_null( $pageid ) || !is_null( $page ) ) {
  81                         if ( !is_null( $oldid ) ) {
  82                                 // Don't use the parser cache
  83                                 $rev = Revision::newFromID( $oldid );
  84                                 if ( !$rev ) {
  85                                         $this->dieUsage( "There is no revision ID $oldid", 'missingrev' );
  86                                 }
  87                                 if ( !$rev->userCan( Revision::DELETED_TEXT ) ) {
  88                                         $this->dieUsage( "You don't have permission to view deleted revisions", 'permissiondenied' );
  89                                 }
  90
  91                                 $titleObj = $rev->getTitle();
  92
  93                                 $wgTitle = $titleObj;
  94
  95                                 // If for some reason the "oldid" is actually the current revision, it may be cached
  96                                 if ( $titleObj->getLatestRevID() === intval( $oldid ) )  {
  97                                         $articleObj = new Article( $titleObj, 0 );
  98
  99                                         $p_result = $this->getParsedSectionOrText( $articleObj, $titleObj, $popts, $pageid,
 100                                                  isset( $prop['wikitext'] ) ) ;
 101                                 } else { // This is an old revision, so get the text differently
 102                                         $this->text = $rev->getText( Revision::FOR_THIS_USER );
 103
 104                                         $wgTitle = $titleObj;
 105
 106                                         if ( $this->section !== false ) {
 107                                                 $this->text = $this->getSectionText( $this->text, 'r' . $rev->getId() );
 108                                         }
 109
 110                                         $p_result = $wgParser->parse( $this->text, $titleObj, $popts );
 111                                 }
 112                         } else { // Not $oldid
 113                                 if ( $params['redirects'] ) {
 114                                         $reqParams = array(
 115                                                 'action' => 'query',
 116                                                 'redirects' => '',
 117                                         );
 118                                         if ( !is_null ( $pageid ) ) {
 119                                                 $reqParams['pageids'] = $pageid;
 120                                         } else { // $page
 121                                                 $reqParams['titles'] = $page;
 122                                         }
 123                                         $req = new FauxRequest( $reqParams );
 124                                         $main = new ApiMain( $req );
 125                                         $main->execute();
 126                                         $data = $main->getResultData();
 127                                         $redirValues = isset( $data['query']['redirects'] )
 128                                                 ? $data['query']['redirects']
 129                                                 : array();
 130                                         $to = $page;
 131                                         foreach ( (array)$redirValues as $r ) {
 132                                                 $to = $r['to'];
 133                                         }
 134                                         $titleObj = Title::newFromText( $to );
 135                                 } else {
 136                                         if ( !is_null ( $pageid ) ) {
 137                                                 $reqParams['pageids'] = $pageid;
 138                                                 $titleObj = Title::newFromID( $pageid );
 139                                         } else { // $page
 140                                                 $to = $page;
 141                                                 $titleObj = Title::newFromText( $to );
 142                                         }
 143                                 }
 144                                 if ( !is_null ( $pageid ) ) {
 145                                         if ( !$titleObj ) {
 146                                                 // Still throw nosuchpageid error if pageid was provided
 147                                                 $this->dieUsageMsg( array( 'nosuchpageid', $pageid ) );
 148                                         }
 149                                 } elseif ( !$titleObj || !$titleObj->exists() ) {
 150                                         $this->dieUsage( "The page you specified doesn't exist", 'missingtitle' );
 151                                 }
 152                                 $wgTitle = $titleObj;
 153
 154                                 $articleObj = new Article( $titleObj, 0 );
 155                                 if ( isset( $prop['revid'] ) ) {
 156                                         $oldid = $articleObj->getRevIdFetched();
 157                                 }
 158
 159                                 $p_result = $this->getParsedSectionOrText( $articleObj, $titleObj, $popts, $pageid,
 160                                          isset( $prop['wikitext'] ) ) ;
 161                         }
 162                 } else { // Not $oldid, $pageid, $page. Hence based on $text
 163
 164                         $this->text = $text;
 165                         $titleObj = Title::newFromText( $title );
 166                         if ( !$titleObj ) {
 167                                 $this->dieUsageMsg( array( 'invalidtitle', $title ) );
 168                         }
 169                         $wgTitle = $titleObj;
 170
 171                         if ( $this->section !== false ) {
 172                                 $this->text = $this->getSectionText( $this->text, $titleObj->getText() );
 173                         }
 174
 175                         if ( $params['pst'] || $params['onlypst'] ) {
 176                                 $this->pstText = $wgParser->preSaveTransform( $this->text, $titleObj, $wgUser, $popts );
 177                         }
 178                         if ( $params['onlypst'] ) {
 179                                 // Build a result and bail out
 180                                 $result_array['text'] = array();
 181                                 $this->getResult()->setContent( $result_array['text'], $this->pstText );
 182                                 if ( isset( $prop['wikitext'] ) ) {
 183                                         $result_array['wikitext'] = array();
 184                                         $this->getResult()->setContent( $result_array['wikitext'], $this->text );
 185                                 }
 186                                 $this->getResult()->addValue( null, $this->getModuleName(), $result_array );
 187                                 return;
 188                         }
 189                         $p_result = $wgParser->parse( $params['pst'] ? $this->pstText : $this->text, $titleObj, $popts );
 190                 }
 191
 192                 // Return result
 193                 $result = $this->getResult();
 194                 $result_array = array();
 195
 196                 $result_array['title'] = $titleObj->getPrefixedText();
 197
 198                 if ( !is_null( $oldid ) ) {
 199                         $result_array['revid'] = intval( $oldid );
 200                 }
 201
 202                 if ( $params['redirects'] && !is_null( $redirValues ) ) {
 203                         $result_array['redirects'] = $redirValues;
 204                 }
 205
 206                 if ( isset( $prop['text'] ) ) {
 207                         $result_array['text'] = array();
 208                         $result->setContent( $result_array['text'], $p_result->getText() );
 209                 }
 210
 211                 if ( !is_null( $params['summary'] ) ) {
 212                         $result_array['parsedsummary'] = array();
 213                         $result->setContent( $result_array['parsedsummary'], $wgUser->getSkin()->formatComment( $params['summary'], $titleObj ) );
 214                 }
 215
 216                 if ( isset( $prop['langlinks'] ) ) {
 217                         $result_array['langlinks'] = $this->formatLangLinks( $p_result->getLanguageLinks() );
 218                 }
 219                 if ( isset( $prop['languageshtml'] ) ) {
 220                         $languagesHtml = $this->languagesHtml( $p_result->getLanguageLinks() );
 221                         $result_array['languageshtml'] = array();
 222                         $result->setContent( $result_array['languageshtml'], $languagesHtml );
 223                 }
 224                 if ( isset( $prop['categories'] ) ) {
 225                         $result_array['categories'] = $this->formatCategoryLinks( $p_result->getCategories() );
 226                 }
 227                 if ( isset( $prop['categorieshtml'] ) ) {
 228                         $categoriesHtml = $this->categoriesHtml( $p_result->getCategories() );
 229                         $result_array['categorieshtml'] = array();
 230                         $result->setContent( $result_array['categorieshtml'], $categoriesHtml );
 231                 }
 232                 if ( isset( $prop['links'] ) ) {
 233                         $result_array['links'] = $this->formatLinks( $p_result->getLinks() );
 234                 }
 235                 if ( isset( $prop['templates'] ) ) {
 236                         $result_array['templates'] = $this->formatLinks( $p_result->getTemplates() );
 237                 }
 238                 if ( isset( $prop['images'] ) ) {
 239                         $result_array['images'] = array_keys( $p_result->getImages() );
 240                 }
 241                 if ( isset( $prop['externallinks'] ) ) {
 242                         $result_array['externallinks'] = array_keys( $p_result->getExternalLinks() );
 243                 }
 244                 if ( isset( $prop['sections'] ) ) {
 245                         $result_array['sections'] = $p_result->getSections();
 246                 }
 247
 248                 if ( isset( $prop['displaytitle'] ) ) {
 249                         $result_array['displaytitle'] = $p_result->getDisplayTitle() ?
 250                                                         $p_result->getDisplayTitle() :
 251                                                         $titleObj->getPrefixedText();
 252                 }
 253
 254                 if ( isset( $prop['headitems'] ) || isset( $prop['headhtml'] ) ) {
 255                         $context = new RequestContext;
 256                         $context->getOutput()->addParserOutputNoText( $p_result );
 257
 258                         if ( isset( $prop['headitems'] ) ) {
 259                                 $headItems = $this->formatHeadItems( $p_result->getHeadItems() );
 260
 261                                 $context->getSkin()->setupUserCss( $context->getOutput() );
 262                                 $css = $this->formatCss( $context->getOutput()->buildCssLinksArray() );
 263
 264                                 $scripts = array( $context->getOutput()->getHeadScripts( $context->getSkin() ) );
 265
 266                                 $result_array['headitems'] = array_merge( $headItems, $css, $scripts );
 267                         }
 268
 269                         if ( isset( $prop['headhtml'] ) ) {
 270                                 $result_array['headhtml'] = array();
 271                                 $result->setContent( $result_array['headhtml'], $context->getOutput()->headElement( $context->getSkin() ) );
 272                         }
 273                 }
 274
 275                 if ( isset( $prop['iwlinks'] ) ) {
 276                         $result_array['iwlinks'] = $this->formatIWLinks( $p_result->getInterwikiLinks() );
 277                 }
 278
 279                 if ( isset( $prop['wikitext'] ) ) {
 280                         $result_array['wikitext'] = array();
 281                         $result->setContent( $result_array['wikitext'], $this->text );
 282                         if ( !is_null( $this->pstText ) ) {
 283                                 $result_array['psttext'] = array();
 284                                 $result->setContent( $result_array['psttext'], $this->pstText );
 285                         }
 286                 }
 287
 288                 $result_mapping = array(
 289                         'redirects' => 'r',
 290                         'langlinks' => 'll',
 291                         'categories' => 'cl',
 292                         'links' => 'pl',
 293                         'templates' => 'tl',
 294                         'images' => 'img',
 295                         'externallinks' => 'el',
 296                         'iwlinks' => 'iw',
 297                         'sections' => 's',
 298                         'headitems' => 'hi',
 299                 );
 300                 $this->setIndexedTagNames( $result_array, $result_mapping );
 301                 $result->addValue( null, $this->getModuleName(), $result_array );
 302
 303                 if ( !is_null( $oldLang ) ) {
 304                         $wgLang = $oldLang; // Reset $wgLang to $oldLang
 305                 }
 306         }
 307
 308         /**
 309          * @param $articleObj Article
 310          * @param $titleObj Title
 311          * @param $popts ParserOptions
 312          * @param $pageId Int
 313          * @param $getWikitext Bool
 314          * @return ParserOutput
 315          */
 316         private function getParsedSectionOrText( $articleObj, $titleObj, $popts, $pageId = null, $getWikitext = false ) {
 317                 if ( $this->section !== false ) {
 318                         global $wgParser;
 319
 320                         $this->text = $this->getSectionText( $articleObj->getRawText(), !is_null ( $pageId )
 321                                         ? 'page id ' . $pageId : $titleObj->getText() );
 322
 323                         return $wgParser->parse( $this->text, $titleObj, $popts );
 324                 } else {
 325                         // Try the parser cache first
 326                         $pout = $articleObj->getParserOutput();
 327                         if ( $getWikitext ) {
 328                                 $rev = Revision::newFromTitle( $titleObj );
 329                                 if ( $rev ) {
 330                                         $this->text = $rev->getText();
 331                                 }
 332                         }
 333                         return $pout;
 334                 }
 335         }
 336
 337         private function getSectionText( $text, $what ) {
 338                 global $wgParser;
 339                 $text = $wgParser->getSection( $text, $this->section, false );
 340                 if ( $text === false ) {
 341                         $this->dieUsage( "There is no section {$this->section} in " . $what, 'nosuchsection' );
 342                 }
 343                 return $text;
 344         }
 345
 346         private function formatLangLinks( $links ) {
 347                 $result = array();
 348                 foreach ( $links as $link ) {
 349                         $entry = array();
 350                         $bits = explode( ':', $link, 2 );
 351                         $title = Title::newFromText( $link );
 352
 353                         $entry['lang'] = $bits[0];
 354                         if ( $title ) {
 355                                 $entry['url'] = $title->getFullURL();
 356                         }
 357                         $this->getResult()->setContent( $entry, $bits[1] );
 358                         $result[] = $entry;
 359                 }
 360                 return $result;
 361         }
 362
 363         private function formatCategoryLinks( $links ) {
 364                 $result = array();
 365                 foreach ( $links as $link => $sortkey ) {
 366                         $entry = array();
 367                         $entry['sortkey'] = $sortkey;
 368                         $this->getResult()->setContent( $entry, $link );
 369                         $result[] = $entry;
 370                 }
 371                 return $result;
 372         }
 373
 374         private function categoriesHtml( $categories ) {
 375                 global $wgOut, $wgUser;
 376                 $wgOut->addCategoryLinks( $categories );
 377                 return $wgUser->getSkin()->getCategories();
 378         }
 379
 380         /**
 381          * @deprecated since 1.18 No modern skin generates language links this way, please use language links
 382          *                        data to generate your own HTML.
 383          */
 384         private function languagesHtml( $languages ) {
 385                 global $wgContLang, $wgHideInterlanguageLinks;
 386
 387                 if ( $wgHideInterlanguageLinks || count( $languages ) == 0 ) {
 388                         return '';
 389                 }
 390
 391                 $s = htmlspecialchars( wfMsg( 'otherlanguages' ) . wfMsg( 'colon-separator' ) );
 392
 393                 $langs = array();
 394                 foreach ( $languages as $l ) {
 395                         $nt = Title::newFromText( $l );
 396                         $text = $wgContLang->getLanguageName( $nt->getInterwiki() );
 397
 398                         $langs[] = Html::element( 'a',
 399                                 array( 'href' => $nt->getFullURL(), 'title' => $nt->getText(), 'class' => "external" ),
 400                                 $text == '' ? $l : $text );
 401                 }
 402
 403                 $s .= implode( htmlspecialchars( wfMsgExt( 'pipe-separator', 'escapenoentities' ) ), $langs );
 404
 405                 if ( $wgContLang->isRTL() ) {
 406                         $s = Html::rawElement( 'span', array( 'dir' => "LTR" ), $s );
 407                 }
 408
 409                 return $s;
 410         }
 411
 412         private function formatLinks( $links ) {
 413                 $result = array();
 414                 foreach ( $links as $ns => $nslinks ) {
 415                         foreach ( $nslinks as $title => $id ) {
 416                                 $entry = array();
 417                                 $entry['ns'] = $ns;
 418                                 $this->getResult()->setContent( $entry, Title::makeTitle( $ns, $title )->getFullText() );
 419                                 if ( $id != 0 ) {
 420                                         $entry['exists'] = '';
 421                                 }
 422                                 $result[] = $entry;
 423                         }
 424                 }
 425                 return $result;
 426         }
 427
 428         private function formatIWLinks( $iw ) {
 429                 $result = array();
 430                 foreach ( $iw as $prefix => $titles ) {
 431                         foreach ( array_keys( $titles ) as $title ) {
 432                                 $entry = array();
 433                                 $entry['prefix'] = $prefix;
 434
 435                                 $title = Title::newFromText( "{$prefix}:{$title}" );
 436                                 if ( $title ) {
 437                                         $entry['url'] = $title->getFullURL();
 438                                 }
 439
 440                                 $this->getResult()->setContent( $entry, $title->getFullText() );
 441                                 $result[] = $entry;
 442                         }
 443                 }
 444                 return $result;
 445         }
 446
 447         private function formatHeadItems( $headItems ) {
 448                 $result = array();
 449                 foreach ( $headItems as $tag => $content ) {
 450                         $entry = array();
 451                         $entry['tag'] = $tag;
 452                         $this->getResult()->setContent( $entry, $content );
 453                         $result[] = $entry;
 454                 }
 455                 return $result;
 456         }
 457
 458         private function formatCss( $css ) {
 459                 $result = array();
 460                 foreach ( $css as $file => $link ) {
 461                         $entry = array();
 462                         $entry['file'] = $file;
 463                         $this->getResult()->setContent( $entry, $link );
 464                         $result[] = $entry;
 465                 }
 466                 return $result;
 467         }
 468
 469         private function setIndexedTagNames( &$array, $mapping ) {
 470                 foreach ( $mapping as $key => $name ) {
 471                         if ( isset( $array[$key] ) ) {
 472                                 $this->getResult()->setIndexedTagName( $array[$key], $name );
 473                         }
 474                 }
 475         }
 476
 477         public function getAllowedParams() {
 478                 return array(
 479                         'title' => array(
 480                                 ApiBase::PARAM_DFLT => 'API',
 481                         ),
 482                         'text' => null,
 483                         'summary' => null,
 484                         'page' => null,
 485                         'pageid' => array(
 486                                 ApiBase::PARAM_TYPE => 'integer',
 487                         ),
 488                         'redirects' => false,
 489                         'oldid' => array(
 490                                 ApiBase::PARAM_TYPE => 'integer',
 491                         ),
 492                         'prop' => array(
 493                                 ApiBase::PARAM_DFLT => 'text|langlinks|categories|links|templates|images|externallinks|sections|revid|displaytitle',
 494                                 ApiBase::PARAM_ISMULTI => true,
 495                                 ApiBase::PARAM_TYPE => array(
 496                                         'text',
 497                                         'langlinks',
 498                                         'languageshtml',
 499                                         'categories',
 500                                         'categorieshtml',
 501                                         'links',
 502                                         'templates',
 503                                         'images',
 504                                         'externallinks',
 505                                         'sections',
 506                                         'revid',
 507                                         'displaytitle',
 508                                         'headitems',
 509                                         'headhtml',
 510                                         'iwlinks',
 511                                         'wikitext',
 512                                 )
 513                         ),
 514                         'pst' => false,
 515                         'onlypst' => false,
 516                         'uselang' => null,
 517                         'section' => null,
 518                         'disablepp' => false,
 519                 );
 520         }
 521
 522         public function getParamDescription() {
 523                 $p = $this->getModulePrefix();
 524                 return array(
 525                         'text' => 'Wikitext to parse',
 526                         'summary' => 'Summary to parse',
 527                         'redirects' => "If the {$p}page or the {$p}pageid parameter is set to a redirect, resolve it",
 528                         'title' => 'Title of page the text belongs to',
 529                         'page' => "Parse the content of this page. Cannot be used together with {$p}text and {$p}title",
 530                         'pageid' => "Parse the content of this page. Overrides {$p}page",
 531                         'oldid' => "Parse the content of this revision. Overrides {$p}page and {$p}pageid",
 532                         'prop' => array(
 533                                 'Which pieces of information to get',
 534                                 ' text           - Gives the parsed text of the wikitext',
 535                                 ' langlinks      - Gives the language links in the parsed wikitext',
 536                                 ' categories     - Gives the categories in the parsed wikitext',
 537                                 ' categorieshtml - Gives the HTML version of the categories',
 538                                 ' languageshtml  - Gives the HTML version of the language links',
 539                                 ' links          - Gives the internal links in the parsed wikitext',
 540                                 ' templates      - Gives the templates in the parsed wikitext',
 541                                 ' images         - Gives the images in the parsed wikitext',
 542                                 ' externallinks  - Gives the external links in the parsed wikitext',
 543                                 ' sections       - Gives the sections in the parsed wikitext',
 544                                 ' revid          - Adds the revision ID of the parsed page',
 545                                 ' displaytitle   - Adds the title of the parsed wikitext',
 546                                 ' headitems      - Gives items to put in the <head> of the page',
 547                                 ' headhtml       - Gives parsed <head> of the page',
 548                                 ' iwlinks        - Gives interwiki links in the parsed wikitext',
 549                                 ' wikitext       - Gives the original wikitext that was parsed',
 550                         ),
 551                         'pst' => array(
 552                                 'Do a pre-save transform on the input before parsing it',
 553                                 'Ignored if page, pageid or oldid is used'
 554                         ),
 555                         'onlypst' => array(
 556                                 'Do a pre-save transform (PST) on the input, but don\'t parse it',
 557                                 'Returns the same wikitext, after a PST has been applied. Ignored if page, pageid or oldid is used'
 558                         ),
 559                         'uselang' => 'Which language to parse the request in',
 560                         'section' => 'Only retrieve the content of this section number',
 561                         'disablepp' => 'Disable the PP Report from the parser output',
 562                 );
 563         }
 564
 565         public function getDescription() {
 566                 return 'Parses wikitext and returns parser output';
 567         }
 568
 569         public function getPossibleErrors() {
 570                 return array_merge( parent::getPossibleErrors(), array(
 571                         array( 'code' => 'params', 'info' => 'The page parameter cannot be used together with the text and title parameters' ),
 572                         array( 'code' => 'missingrev', 'info' => 'There is no revision ID oldid' ),
 573                         array( 'code' => 'permissiondenied', 'info' => 'You don\'t have permission to view deleted revisions' ),
 574                         array( 'code' => 'missingtitle', 'info' => 'The page you specified doesn\'t exist' ),
 575                         array( 'code' => 'nosuchsection', 'info' => 'There is no section sectionnumber in page' ),
 576                         array( 'nosuchpageid' ),
 577                         array( 'invalidtitle', 'title' ),
 578                 ) );
 579         }
 580
 581         protected function getExamples() {
 582                 return array(
 583                         'api.php?action=parse&text={{Project:Sandbox}}'
 584                 );
 585         }
 586
 587         public function getVersion() {
 588                 return __CLASS__ . ': $Id$';
 589         }
 590 }