includes/api/ApiParse.php

   1 <?php
   2 /**
   3  *
   4  *
   5  * Created on Dec 01, 2007
   6  *
   7  * Copyright © 2007 Yuri Astrakhan <Firstname><Lastname>@gmail.com
   8  *
   9  * This program is free software; you can redistribute it and/or modify
  10  * it under the terms of the GNU General Public License as published by
  11  * the Free Software Foundation; either version 2 of the License, or
  12  * (at your option) any later version.
  13  *
  14  * This program is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  17  * GNU General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU General Public License along
  20  * with this program; if not, write to the Free Software Foundation, Inc.,
  21  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  22  * http://www.gnu.org/copyleft/gpl.html
  23  *
  24  * @file
  25  */
  26
  27 if ( !defined( 'MEDIAWIKI' ) ) {
  28         // Eclipse helper - will be ignored in production
  29         require_once( "ApiBase.php" );
  30 }
  31
  32 /**
  33  * @ingroup API
  34  */
  35 class ApiParse extends ApiBase {
  36
  37         private $section, $text, $pstText = null;
  38
  39         public function __construct( $main, $action ) {
  40                 parent::__construct( $main, $action );
  41         }
  42
  43         public function execute() {
  44                 // The data is hot but user-dependent, like page views, so we set vary cookies
  45                 $this->getMain()->setCacheMode( 'anon-public-user-private' );
  46
  47                 // Get parameters
  48                 $params = $this->extractRequestParams();
  49                 $text = $params['text'];
  50                 $title = $params['title'];
  51                 $page = $params['page'];
  52                 $pageid = $params['pageid'];
  53                 $oldid = $params['oldid'];
  54
  55                 if ( !is_null( $page ) && ( !is_null( $text ) || $title != 'API' ) ) {
  56                         $this->dieUsage( 'The page parameter cannot be used together with the text and title parameters', 'params' );
  57                 }
  58                 $prop = array_flip( $params['prop'] );
  59
  60                 if ( isset( $params['section'] ) ) {
  61                         $this->section = $params['section'];
  62                 } else {
  63                         $this->section = false;
  64                 }
  65
  66                 // The parser needs $wgTitle to be set, apparently the
  67                 // $title parameter in Parser::parse isn't enough *sigh*
  68                 global $wgParser, $wgUser, $wgTitle, $wgLang;
  69
  70                 // Currently unnecessary, code to act as a safeguard against any change in current behaviour of uselang breaks
  71                 $oldLang = null;
  72                 if ( isset( $params['uselang'] ) && $params['uselang'] != $wgLang->getCode() ) {
  73                         $oldLang = $wgLang; // Backup wgLang
  74                         $wgLang = Language::factory( $params['uselang'] );
  75                 }
  76
  77                 $popts = new ParserOptions();
  78                 $popts->setTidy( true );
  79                 $popts->enableLimitReport( !$params['disablepp'] );
  80
  81                 $redirValues = null;
  82
  83                 if ( !is_null( $oldid ) || !is_null( $pageid ) || !is_null( $page ) ) {
  84
  85                         if ( !is_null( $oldid ) ) {
  86                                 // Don't use the parser cache
  87                                 $rev = Revision::newFromID( $oldid );
  88                                 if ( !$rev ) {
  89                                         $this->dieUsage( "There is no revision ID $oldid", 'missingrev' );
  90                                 }
  91                                 if ( !$rev->userCan( Revision::DELETED_TEXT ) ) {
  92                                         $this->dieUsage( "You don't have permission to view deleted revisions", 'permissiondenied' );
  93                                 }
  94
  95                                 $titleObj = $rev->getTitle();
  96
  97                                 $wgTitle = $titleObj;
  98
  99                                 // If for some reason the "oldid" is actually the current revision, it may be cached
 100                                 if ( $titleObj->getLatestRevID() === intval( $oldid ) )  {
 101                                         $articleObj = new Article( $titleObj, 0 );
 102
 103                                         $p_result = $this->getParsedSectionOrText( $articleObj, $titleObj, $popts, $pageid,
 104                                                  isset( $prop['wikitext'] ) ) ;
 105
 106                                 } else { // This is an old revision, so get the text differently
 107                                         $this->text = $rev->getText( Revision::FOR_THIS_USER );
 108
 109                                         $wgTitle = $titleObj;
 110
 111                                         if ( $this->section !== false ) {
 112                                                 $this->text = $this->getSectionText( $this->text, 'r' . $rev->getId() );
 113                                         }
 114
 115                                         $p_result = $wgParser->parse( $this->text, $titleObj, $popts );
 116                                 }
 117
 118                         } else { // Not $oldid
 119
 120                                 if ( !is_null ( $pageid ) ) {
 121                                         $titleObj = Title::newFromID( $pageid );
 122
 123                                         if ( !$titleObj ) {
 124                                                 $this->dieUsageMsg( array( 'nosuchpageid', $pageid ) );
 125                                         }
 126                                 } else { // $page
 127
 128                                         if ( $params['redirects'] ) {
 129                                                 $req = new FauxRequest( array(
 130                                                         'action' => 'query',
 131                                                         'redirects' => '',
 132                                                         'titles' => $page
 133                                                 ) );
 134                                                 $main = new ApiMain( $req );
 135                                                 $main->execute();
 136                                                 $data = $main->getResultData();
 137                                                 $redirValues = @$data['query']['redirects'];
 138                                                 $to = $page;
 139                                                 foreach ( (array)$redirValues as $r ) {
 140                                                         $to = $r['to'];
 141                                                 }
 142                                         } else {
 143                                                 $to = $page;
 144                                         }
 145                                         $titleObj = Title::newFromText( $to );
 146                                         if ( !$titleObj || !$titleObj->exists() ) {
 147                                                 $this->dieUsage( "The page you specified doesn't exist", 'missingtitle' );
 148                                         }
 149                                 }
 150                                 $wgTitle = $titleObj;
 151
 152                                 $articleObj = new Article( $titleObj, 0 );
 153                                 if ( isset( $prop['revid'] ) ) {
 154                                         $oldid = $articleObj->getRevIdFetched();
 155                                 }
 156
 157                                 $p_result = $this->getParsedSectionOrText( $articleObj, $titleObj, $popts, $pageid,
 158                                          isset( $prop['wikitext'] ) ) ;
 159                         }
 160
 161                 } else { // Not $oldid, $pageid, $page. Hence based on $text
 162
 163                         $this->text = $text;
 164                         $titleObj = Title::newFromText( $title );
 165                         if ( !$titleObj ) {
 166                                 $titleObj = Title::newFromText( 'API' );
 167                         }
 168                         $wgTitle = $titleObj;
 169
 170                         if ( $this->section !== false ) {
 171                                 $this->text = $this->getSectionText( $this->text, $titleObj->getText() );
 172                         }
 173
 174                         if ( $params['pst'] || $params['onlypst'] ) {
 175                                 $this->pstText = $wgParser->preSaveTransform( $this->text, $titleObj, $wgUser, $popts );
 176                         }
 177                         if ( $params['onlypst'] ) {
 178                                 // Build a result and bail out
 179                                 $result_array['text'] = array();
 180                                 $this->getResult()->setContent( $result_array['text'], $this->pstText );
 181                                 if ( isset( $prop['wikitext'] ) ) {
 182                                         $result_array['wikitext'] = array();
 183                                         $this->getResult()->setContent( $result_array['wikitext'], $this->text );
 184                                 }
 185                                 $this->getResult()->addValue( null, $this->getModuleName(), $result_array );
 186                                 return;
 187                         }
 188                         $p_result = $wgParser->parse( $params['pst'] ? $this->pstText : $this->text, $titleObj, $popts );
 189                 }
 190
 191                 // Return result
 192                 $result = $this->getResult();
 193                 $result_array = array();
 194
 195                 $result_array['title'] = $titleObj->getPrefixedText();
 196
 197                 if ( !is_null( $oldid ) ) {
 198                         $result_array['revid'] = intval( $oldid );
 199                 }
 200
 201                 if ( $params['redirects'] && !is_null( $redirValues ) ) {
 202                         $result_array['redirects'] = $redirValues;
 203                 }
 204
 205                 if ( isset( $prop['text'] ) ) {
 206                         $result_array['text'] = array();
 207                         $result->setContent( $result_array['text'], $p_result->getText() );
 208                 }
 209
 210                 if ( !is_null( $params['summary'] ) ) {
 211                         $result_array['parsedsummary'] = array();
 212                         $result->setContent( $result_array['parsedsummary'], $wgUser->getSkin()->formatComment( $params['summary'], $titleObj ) );
 213                 }
 214
 215                 if ( isset( $prop['langlinks'] ) ) {
 216                         $result_array['langlinks'] = $this->formatLangLinks( $p_result->getLanguageLinks() );
 217                 }
 218                 if ( isset( $prop['languageshtml'] ) ) {
 219                         $languagesHtml = $this->languagesHtml( $p_result->getLanguageLinks() );
 220                         $result_array['languageshtml'] = array();
 221                         $result->setContent( $result_array['languageshtml'], $languagesHtml );
 222                 }
 223                 if ( isset( $prop['categories'] ) ) {
 224                         $result_array['categories'] = $this->formatCategoryLinks( $p_result->getCategories() );
 225                 }
 226                 if ( isset( $prop['categorieshtml'] ) ) {
 227                         $categoriesHtml = $this->categoriesHtml( $p_result->getCategories() );
 228                         $result_array['categorieshtml'] = array();
 229                         $result->setContent( $result_array['categorieshtml'], $categoriesHtml );
 230                 }
 231                 if ( isset( $prop['links'] ) ) {
 232                         $result_array['links'] = $this->formatLinks( $p_result->getLinks() );
 233                 }
 234                 if ( isset( $prop['templates'] ) ) {
 235                         $result_array['templates'] = $this->formatLinks( $p_result->getTemplates() );
 236                 }
 237                 if ( isset( $prop['images'] ) ) {
 238                         $result_array['images'] = array_keys( $p_result->getImages() );
 239                 }
 240                 if ( isset( $prop['externallinks'] ) ) {
 241                         $result_array['externallinks'] = array_keys( $p_result->getExternalLinks() );
 242                 }
 243                 if ( isset( $prop['sections'] ) ) {
 244                         $result_array['sections'] = $p_result->getSections();
 245                 }
 246
 247                 if ( isset( $prop['displaytitle'] ) ) {
 248                         $result_array['displaytitle'] = $p_result->getDisplayTitle() ?
 249                                                         $p_result->getDisplayTitle() :
 250                                                         $titleObj->getPrefixedText();
 251                 }
 252
 253                 if ( isset( $prop['headitems'] ) || isset( $prop['headhtml'] ) ) {
 254                         $out = new OutputPage;
 255                         $out->setUser( $wgUser );
 256                         $out->addParserOutputNoText( $p_result );
 257                         $userSkin = $wgUser->getSkin();
 258
 259                         if ( isset( $prop['headitems'] ) ) {
 260                                 $headItems = $this->formatHeadItems( $p_result->getHeadItems() );
 261
 262                                 $userSkin->setupUserCss( $out );
 263                                 $css = $this->formatCss( $out->buildCssLinksArray() );
 264
 265                                 $scripts = array( $out->getHeadScripts( $userSkin ) );
 266
 267                                 $result_array['headitems'] = array_merge( $headItems, $css, $scripts );
 268                         }
 269
 270                         if ( isset( $prop['headhtml'] ) ) {
 271                                 $result_array['headhtml'] = array();
 272                                 $result->setContent( $result_array['headhtml'], $out->headElement( $userSkin ) );
 273                         }
 274                 }
 275
 276                 if ( isset( $prop['iwlinks'] ) ) {
 277                         $result_array['iwlinks'] = $this->formatIWLinks( $p_result->getInterwikiLinks() );
 278                 }
 279
 280                 if ( isset( $prop['wikitext'] ) ) {
 281                         $result_array['wikitext'] = array();
 282                         $result->setContent( $result_array['wikitext'], $this->text );
 283                         if ( !is_null( $this->pstText ) ) {
 284                                 $result_array['psttext'] = array();
 285                                 $result->setContent( $result_array['psttext'], $this->pstText );
 286                         }
 287                 }
 288
 289                 $result_mapping = array(
 290                         'redirects' => 'r',
 291                         'langlinks' => 'll',
 292                         'categories' => 'cl',
 293                         'links' => 'pl',
 294                         'templates' => 'tl',
 295                         'images' => 'img',
 296                         'externallinks' => 'el',
 297                         'iwlinks' => 'iw',
 298                         'sections' => 's',
 299                         'headitems' => 'hi',
 300                 );
 301                 $this->setIndexedTagNames( $result_array, $result_mapping );
 302                 $result->addValue( null, $this->getModuleName(), $result_array );
 303
 304                 if ( !is_null( $oldLang ) ) {
 305                         $wgLang = $oldLang; // Reset $wgLang to $oldLang
 306                 }
 307         }
 308
 309         /**
 310          * @param  $articleObj Article
 311          * @param  $titleObj Title
 312          * @param  $popts ParserOptions
 313          * @param  $pageId Int
 314          * @param  $getWikitext Bool
 315          * @return ParserOutput
 316          */
 317         private function getParsedSectionOrText( $articleObj, $titleObj, $popts, $pageId = null, $getWikitext = false ) {
 318                 if ( $this->section !== false ) {
 319                         global $wgParser;
 320
 321                         $this->text = $this->getSectionText( $articleObj->getRawText(), !is_null ( $pageId )
 322                                         ? 'page id ' . $pageId : $titleObj->getText() );
 323
 324                         return $wgParser->parse( $this->text, $titleObj, $popts );
 325                 } else {
 326                         // Try the parser cache first
 327                         $pout = $articleObj->getParserOutput();
 328                         if ( $getWikitext ) {
 329                                 $rev = Revision::newFromTitle( $titleObj );
 330                                 if ( $rev ) {
 331                                         $this->text = $rev->getText();
 332                                 }
 333                         }
 334                         return $pout;
 335                 }
 336         }
 337
 338         private function getSectionText( $text, $what ) {
 339                 global $wgParser;
 340                 $text = $wgParser->getSection( $text, $this->section, false );
 341                 if ( $text === false ) {
 342                         $this->dieUsage( "There is no section {$this->section} in " . $what, 'nosuchsection' );
 343                 }
 344                 return $text;
 345         }
 346
 347         private function formatLangLinks( $links ) {
 348                 $result = array();
 349                 foreach ( $links as $link ) {
 350                         $entry = array();
 351                         $bits = explode( ':', $link, 2 );
 352                         $title = Title::newFromText( $link );
 353
 354                         $entry['lang'] = $bits[0];
 355                         if ( $title ) {
 356                                 $entry['url'] = $title->getFullURL();
 357                         }
 358                         $this->getResult()->setContent( $entry, $bits[1] );
 359                         $result[] = $entry;
 360                 }
 361                 return $result;
 362         }
 363
 364         private function formatCategoryLinks( $links ) {
 365                 $result = array();
 366                 foreach ( $links as $link => $sortkey ) {
 367                         $entry = array();
 368                         $entry['sortkey'] = $sortkey;
 369                         $this->getResult()->setContent( $entry, $link );
 370                         $result[] = $entry;
 371                 }
 372                 return $result;
 373         }
 374
 375         private function categoriesHtml( $categories ) {
 376                 global $wgOut, $wgUser;
 377                 $wgOut->addCategoryLinks( $categories );
 378                 $sk = $wgUser->getSkin();
 379                 return $sk->getCategories();
 380         }
 381
 382         /**
 383          * @deprecated No modern skin generates langlinks this way, please use langlinks data to generate your own html
 384          */
 385         private function languagesHtml( $languages ) {
 386                 global $wgContLang, $wgHideInterlanguageLinks;
 387
 388                 if ( $wgHideInterlanguageLinks || count( $languages ) == 0 ) {
 389                         return '';
 390                 }
 391
 392                 $s = htmlspecialchars( wfMsg( 'otherlanguages' ) . wfMsg( 'colon-separator' ) );
 393
 394                 $langs = array();
 395                 foreach ( $languages as $l ) {
 396                         $nt = Title::newFromText( $l );
 397                         $text = $wgContLang->getLanguageName( $nt->getInterwiki() );
 398
 399                         $langs[] = Html::element( 'a',
 400                                 array( 'href' => $nt->getFullURL(), 'title' => $nt->getText(), 'class' => "external" ),
 401                                 $text == '' ? $l : $text );
 402                 }
 403
 404                 $s .= implode( htmlspecialchars( wfMsgExt( 'pipe-separator', 'escapenoentities' ) ), $langs );
 405
 406                 if ( $wgContLang->isRTL() ) {
 407                         $s = Html::rawElement( 'span', array( 'dir' => "LTR" ), $s );
 408                 }
 409
 410                 return $s;
 411         }
 412
 413         private function formatLinks( $links ) {
 414                 $result = array();
 415                 foreach ( $links as $ns => $nslinks ) {
 416                         foreach ( $nslinks as $title => $id ) {
 417                                 $entry = array();
 418                                 $entry['ns'] = $ns;
 419                                 $this->getResult()->setContent( $entry, Title::makeTitle( $ns, $title )->getFullText() );
 420                                 if ( $id != 0 ) {
 421                                         $entry['exists'] = '';
 422                                 }
 423                                 $result[] = $entry;
 424                         }
 425                 }
 426                 return $result;
 427         }
 428
 429         private function formatIWLinks( $iw ) {
 430                 $result = array();
 431                 foreach ( $iw as $prefix => $titles ) {
 432                         foreach ( array_keys( $titles ) as $title ) {
 433                                 $entry = array();
 434                                 $entry['prefix'] = $prefix;
 435
 436                                 $title = Title::newFromText( "{$prefix}:{$title}" );
 437                                 if ( $title ) {
 438                                         $entry['url'] = $title->getFullURL();
 439                                 }
 440
 441                                 $this->getResult()->setContent( $entry, $title->getFullText() );
 442                                 $result[] = $entry;
 443                         }
 444                 }
 445                 return $result;
 446         }
 447
 448         private function formatHeadItems( $headItems ) {
 449                 $result = array();
 450                 foreach ( $headItems as $tag => $content ) {
 451                         $entry = array();
 452                         $entry['tag'] = $tag;
 453                         $this->getResult()->setContent( $entry, $content );
 454                         $result[] = $entry;
 455                 }
 456                 return $result;
 457         }
 458
 459         private function formatCss( $css ) {
 460                 $result = array();
 461                 foreach ( $css as $file => $link ) {
 462                         $entry = array();
 463                         $entry['file'] = $file;
 464                         $this->getResult()->setContent( $entry, $link );
 465                         $result[] = $entry;
 466                 }
 467                 return $result;
 468         }
 469
 470         private function setIndexedTagNames( &$array, $mapping ) {
 471                 foreach ( $mapping as $key => $name ) {
 472                         if ( isset( $array[$key] ) ) {
 473                                 $this->getResult()->setIndexedTagName( $array[$key], $name );
 474                         }
 475                 }
 476         }
 477
 478         public function getAllowedParams() {
 479                 return array(
 480                         'title' => array(
 481                                 ApiBase::PARAM_DFLT => 'API',
 482                         ),
 483                         'text' => null,
 484                         'summary' => null,
 485                         'page' => null,
 486                         'pageid' => array(
 487                                 ApiBase::PARAM_TYPE => 'integer',
 488                         ),
 489                         'redirects' => false,
 490                         'oldid' => array(
 491                                 ApiBase::PARAM_TYPE => 'integer',
 492                         ),
 493                         'prop' => array(
 494                                 ApiBase::PARAM_DFLT => 'text|langlinks|categories|links|templates|images|externallinks|sections|revid|displaytitle',
 495                                 ApiBase::PARAM_ISMULTI => true,
 496                                 ApiBase::PARAM_TYPE => array(
 497                                         'text',
 498                                         'langlinks',
 499                                         'languageshtml',
 500                                         'categories',
 501                                         'categorieshtml',
 502                                         'links',
 503                                         'templates',
 504                                         'images',
 505                                         'externallinks',
 506                                         'sections',
 507                                         'revid',
 508                                         'displaytitle',
 509                                         'headitems',
 510                                         'headhtml',
 511                                         'iwlinks',
 512                                         'wikitext',
 513                                 )
 514                         ),
 515                         'pst' => false,
 516                         'onlypst' => false,
 517                         'uselang' => null,
 518                         'section' => null,
 519                         'disablepp' => false,
 520                 );
 521         }
 522
 523         public function getParamDescription() {
 524                 $p = $this->getModulePrefix();
 525                 return array(
 526                         'text' => 'Wikitext to parse',
 527                         'summary' => 'Summary to parse',
 528                         'redirects' => "If the {$p}page parameter is set to a redirect, resolve it",
 529                         'title' => 'Title of page the text belongs to',
 530                         'page' => "Parse the content of this page. Cannot be used together with {$p}text and {$p}title",
 531                         'pageid' => "Parse the content of this page. Overrides {$p}page",
 532                         'oldid' => "Parse the content of this revision. Overrides {$p}page and {$p}pageid",
 533                         'prop' => array(
 534                                 'Which pieces of information to get',
 535                                 ' text           - Gives the parsed text of the wikitext',
 536                                 ' langlinks      - Gives the language links in the parsed wikitext',
 537                                 ' categories     - Gives the categories in the parsed wikitext',
 538                                 ' categorieshtml - Gives the HTML version of the categories',
 539                                 ' languageshtml  - Gives the HTML version of the language links',
 540                                 ' links          - Gives the internal links in the parsed wikitext',
 541                                 ' templates      - Gives the templates in the parsed wikitext',
 542                                 ' images         - Gives the images in the parsed wikitext',
 543                                 ' externallinks  - Gives the external links in the parsed wikitext',
 544                                 ' sections       - Gives the sections in the parsed wikitext',
 545                                 ' revid          - Adds the revision ID of the parsed page',
 546                                 ' displaytitle   - Adds the title of the parsed wikitext',
 547                                 ' headitems      - Gives items to put in the <head> of the page',
 548                                 ' headhtml       - Gives parsed <head> of the page',
 549                                 ' iwlinks        - Gives interwiki links in the parsed wikitext',
 550                                 ' wikitext       - Gives the original wikitext that was parsed',
 551                                 'NOTE: Section tree is only generated if there are more than 4 sections, or if the __TOC__ keyword is present'
 552                         ),
 553                         'pst' => array(
 554                                 'Do a pre-save transform on the input before parsing it',
 555                                 'Ignored if page, pageid or oldid is used'
 556                         ),
 557                         'onlypst' => array(
 558                                 'Do a pre-save transform (PST) on the input, but don\'t parse it',
 559                                 'Returns the same wikitext, after a PST has been applied. Ignored if page, pageid or oldid is used'
 560                         ),
 561                         'uselang' => 'Which language to parse the request in',
 562                         'section' => 'Only retrieve the content of this section number',
 563                         'disablepp' => 'Disable the PP Report from the parser output',
 564                 );
 565         }
 566
 567         public function getDescription() {
 568                 return 'This module parses wikitext and returns parser output';
 569         }
 570
 571         public function getPossibleErrors() {
 572                 return array_merge( parent::getPossibleErrors(), array(
 573                         array( 'code' => 'params', 'info' => 'The page parameter cannot be used together with the text and title parameters' ),
 574                         array( 'code' => 'missingrev', 'info' => 'There is no revision ID oldid' ),
 575                         array( 'code' => 'permissiondenied', 'info' => 'You don\'t have permission to view deleted revisions' ),
 576                         array( 'code' => 'missingtitle', 'info' => 'The page you specified doesn\'t exist' ),
 577                         array( 'code' => 'nosuchsection', 'info' => 'There is no section sectionnumber in page' ),
 578                         array( 'nosuchpageid' ),
 579                 ) );
 580         }
 581
 582         protected function getExamples() {
 583                 return array(
 584                         'api.php?action=parse&text={{Project:Sandbox}}'
 585                 );
 586         }
 587
 588         public function getVersion() {
 589                 return __CLASS__ . ': $Id$';
 590         }
 591 }