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                         $context = new RequestContext;
 255                         $context->output->addParserOutputNoText( $p_result );
 256
 257                         if ( isset( $prop['headitems'] ) ) {
 258                                 $headItems = $this->formatHeadItems( $p_result->getHeadItems() );
 259
 260                                 $context->skin->setupUserCss( $context->output );
 261                                 $css = $this->formatCss( $context->output->buildCssLinksArray() );
 262
 263                                 $scripts = array( $context->output->getHeadScripts( $context->skin ) );
 264
 265                                 $result_array['headitems'] = array_merge( $headItems, $css, $scripts );
 266                         }
 267
 268                         if ( isset( $prop['headhtml'] ) ) {
 269                                 $result_array['headhtml'] = array();
 270                                 $result->setContent( $result_array['headhtml'], $context->output->headElement( $context->skin ) );
 271                         }
 272                 }
 273
 274                 if ( isset( $prop['iwlinks'] ) ) {
 275                         $result_array['iwlinks'] = $this->formatIWLinks( $p_result->getInterwikiLinks() );
 276                 }
 277
 278                 if ( isset( $prop['wikitext'] ) ) {
 279                         $result_array['wikitext'] = array();
 280                         $result->setContent( $result_array['wikitext'], $this->text );
 281                         if ( !is_null( $this->pstText ) ) {
 282                                 $result_array['psttext'] = array();
 283                                 $result->setContent( $result_array['psttext'], $this->pstText );
 284                         }
 285                 }
 286
 287                 $result_mapping = array(
 288                         'redirects' => 'r',
 289                         'langlinks' => 'll',
 290                         'categories' => 'cl',
 291                         'links' => 'pl',
 292                         'templates' => 'tl',
 293                         'images' => 'img',
 294                         'externallinks' => 'el',
 295                         'iwlinks' => 'iw',
 296                         'sections' => 's',
 297                         'headitems' => 'hi',
 298                 );
 299                 $this->setIndexedTagNames( $result_array, $result_mapping );
 300                 $result->addValue( null, $this->getModuleName(), $result_array );
 301
 302                 if ( !is_null( $oldLang ) ) {
 303                         $wgLang = $oldLang; // Reset $wgLang to $oldLang
 304                 }
 305         }
 306
 307         /**
 308          * @param  $articleObj Article
 309          * @param  $titleObj Title
 310          * @param  $popts ParserOptions
 311          * @param  $pageId Int
 312          * @param  $getWikitext Bool
 313          * @return ParserOutput
 314          */
 315         private function getParsedSectionOrText( $articleObj, $titleObj, $popts, $pageId = null, $getWikitext = false ) {
 316                 if ( $this->section !== false ) {
 317                         global $wgParser;
 318
 319                         $this->text = $this->getSectionText( $articleObj->getRawText(), !is_null ( $pageId )
 320                                         ? 'page id ' . $pageId : $titleObj->getText() );
 321
 322                         return $wgParser->parse( $this->text, $titleObj, $popts );
 323                 } else {
 324                         // Try the parser cache first
 325                         $pout = $articleObj->getParserOutput();
 326                         if ( $getWikitext ) {
 327                                 $rev = Revision::newFromTitle( $titleObj );
 328                                 if ( $rev ) {
 329                                         $this->text = $rev->getText();
 330                                 }
 331                         }
 332                         return $pout;
 333                 }
 334         }
 335
 336         private function getSectionText( $text, $what ) {
 337                 global $wgParser;
 338                 $text = $wgParser->getSection( $text, $this->section, false );
 339                 if ( $text === false ) {
 340                         $this->dieUsage( "There is no section {$this->section} in " . $what, 'nosuchsection' );
 341                 }
 342                 return $text;
 343         }
 344
 345         private function formatLangLinks( $links ) {
 346                 $result = array();
 347                 foreach ( $links as $link ) {
 348                         $entry = array();
 349                         $bits = explode( ':', $link, 2 );
 350                         $title = Title::newFromText( $link );
 351
 352                         $entry['lang'] = $bits[0];
 353                         if ( $title ) {
 354                                 $entry['url'] = $title->getFullURL();
 355                         }
 356                         $this->getResult()->setContent( $entry, $bits[1] );
 357                         $result[] = $entry;
 358                 }
 359                 return $result;
 360         }
 361
 362         private function formatCategoryLinks( $links ) {
 363                 $result = array();
 364                 foreach ( $links as $link => $sortkey ) {
 365                         $entry = array();
 366                         $entry['sortkey'] = $sortkey;
 367                         $this->getResult()->setContent( $entry, $link );
 368                         $result[] = $entry;
 369                 }
 370                 return $result;
 371         }
 372
 373         private function categoriesHtml( $categories ) {
 374                 global $wgOut, $wgUser;
 375                 $wgOut->addCategoryLinks( $categories );
 376                 return $wgUser->getSkin()->getCategories( $wgOut );
 377         }
 378
 379         /**
 380          * @deprecated No modern skin generates langlinks this way, please use langlinks data to generate your own html
 381          */
 382         private function languagesHtml( $languages ) {
 383                 global $wgContLang, $wgHideInterlanguageLinks;
 384
 385                 if ( $wgHideInterlanguageLinks || count( $languages ) == 0 ) {
 386                         return '';
 387                 }
 388
 389                 $s = htmlspecialchars( wfMsg( 'otherlanguages' ) . wfMsg( 'colon-separator' ) );
 390
 391                 $langs = array();
 392                 foreach ( $languages as $l ) {
 393                         $nt = Title::newFromText( $l );
 394                         $text = $wgContLang->getLanguageName( $nt->getInterwiki() );
 395
 396                         $langs[] = Html::element( 'a',
 397                                 array( 'href' => $nt->getFullURL(), 'title' => $nt->getText(), 'class' => "external" ),
 398                                 $text == '' ? $l : $text );
 399                 }
 400
 401                 $s .= implode( htmlspecialchars( wfMsgExt( 'pipe-separator', 'escapenoentities' ) ), $langs );
 402
 403                 if ( $wgContLang->isRTL() ) {
 404                         $s = Html::rawElement( 'span', array( 'dir' => "LTR" ), $s );
 405                 }
 406
 407                 return $s;
 408         }
 409
 410         private function formatLinks( $links ) {
 411                 $result = array();
 412                 foreach ( $links as $ns => $nslinks ) {
 413                         foreach ( $nslinks as $title => $id ) {
 414                                 $entry = array();
 415                                 $entry['ns'] = $ns;
 416                                 $this->getResult()->setContent( $entry, Title::makeTitle( $ns, $title )->getFullText() );
 417                                 if ( $id != 0 ) {
 418                                         $entry['exists'] = '';
 419                                 }
 420                                 $result[] = $entry;
 421                         }
 422                 }
 423                 return $result;
 424         }
 425
 426         private function formatIWLinks( $iw ) {
 427                 $result = array();
 428                 foreach ( $iw as $prefix => $titles ) {
 429                         foreach ( array_keys( $titles ) as $title ) {
 430                                 $entry = array();
 431                                 $entry['prefix'] = $prefix;
 432
 433                                 $title = Title::newFromText( "{$prefix}:{$title}" );
 434                                 if ( $title ) {
 435                                         $entry['url'] = $title->getFullURL();
 436                                 }
 437
 438                                 $this->getResult()->setContent( $entry, $title->getFullText() );
 439                                 $result[] = $entry;
 440                         }
 441                 }
 442                 return $result;
 443         }
 444
 445         private function formatHeadItems( $headItems ) {
 446                 $result = array();
 447                 foreach ( $headItems as $tag => $content ) {
 448                         $entry = array();
 449                         $entry['tag'] = $tag;
 450                         $this->getResult()->setContent( $entry, $content );
 451                         $result[] = $entry;
 452                 }
 453                 return $result;
 454         }
 455
 456         private function formatCss( $css ) {
 457                 $result = array();
 458                 foreach ( $css as $file => $link ) {
 459                         $entry = array();
 460                         $entry['file'] = $file;
 461                         $this->getResult()->setContent( $entry, $link );
 462                         $result[] = $entry;
 463                 }
 464                 return $result;
 465         }
 466
 467         private function setIndexedTagNames( &$array, $mapping ) {
 468                 foreach ( $mapping as $key => $name ) {
 469                         if ( isset( $array[$key] ) ) {
 470                                 $this->getResult()->setIndexedTagName( $array[$key], $name );
 471                         }
 472                 }
 473         }
 474
 475         public function getAllowedParams() {
 476                 return array(
 477                         'title' => array(
 478                                 ApiBase::PARAM_DFLT => 'API',
 479                         ),
 480                         'text' => null,
 481                         'summary' => null,
 482                         'page' => null,
 483                         'pageid' => array(
 484                                 ApiBase::PARAM_TYPE => 'integer',
 485                         ),
 486                         'redirects' => false,
 487                         'oldid' => array(
 488                                 ApiBase::PARAM_TYPE => 'integer',
 489                         ),
 490                         'prop' => array(
 491                                 ApiBase::PARAM_DFLT => 'text|langlinks|categories|links|templates|images|externallinks|sections|revid|displaytitle',
 492                                 ApiBase::PARAM_ISMULTI => true,
 493                                 ApiBase::PARAM_TYPE => array(
 494                                         'text',
 495                                         'langlinks',
 496                                         'languageshtml',
 497                                         'categories',
 498                                         'categorieshtml',
 499                                         'links',
 500                                         'templates',
 501                                         'images',
 502                                         'externallinks',
 503                                         'sections',
 504                                         'revid',
 505                                         'displaytitle',
 506                                         'headitems',
 507                                         'headhtml',
 508                                         'iwlinks',
 509                                         'wikitext',
 510                                 )
 511                         ),
 512                         'pst' => false,
 513                         'onlypst' => false,
 514                         'uselang' => null,
 515                         'section' => null,
 516                         'disablepp' => false,
 517                 );
 518         }
 519
 520         public function getParamDescription() {
 521                 $p = $this->getModulePrefix();
 522                 return array(
 523                         'text' => 'Wikitext to parse',
 524                         'summary' => 'Summary to parse',
 525                         'redirects' => "If the {$p}page parameter is set to a redirect, resolve it",
 526                         'title' => 'Title of page the text belongs to',
 527                         'page' => "Parse the content of this page. Cannot be used together with {$p}text and {$p}title",
 528                         'pageid' => "Parse the content of this page. Overrides {$p}page",
 529                         'oldid' => "Parse the content of this revision. Overrides {$p}page and {$p}pageid",
 530                         'prop' => array(
 531                                 'Which pieces of information to get',
 532                                 ' text           - Gives the parsed text of the wikitext',
 533                                 ' langlinks      - Gives the language links in the parsed wikitext',
 534                                 ' categories     - Gives the categories in the parsed wikitext',
 535                                 ' categorieshtml - Gives the HTML version of the categories',
 536                                 ' languageshtml  - Gives the HTML version of the language links',
 537                                 ' links          - Gives the internal links in the parsed wikitext',
 538                                 ' templates      - Gives the templates in the parsed wikitext',
 539                                 ' images         - Gives the images in the parsed wikitext',
 540                                 ' externallinks  - Gives the external links in the parsed wikitext',
 541                                 ' sections       - Gives the sections in the parsed wikitext',
 542                                 ' revid          - Adds the revision ID of the parsed page',
 543                                 ' displaytitle   - Adds the title of the parsed wikitext',
 544                                 ' headitems      - Gives items to put in the <head> of the page',
 545                                 ' headhtml       - Gives parsed <head> of the page',
 546                                 ' iwlinks        - Gives interwiki links in the parsed wikitext',
 547                                 ' wikitext       - Gives the original wikitext that was parsed',
 548                                 'NOTE: Section tree is only generated if there are more than 4 sections, or if the __TOC__ keyword is present'
 549                         ),
 550                         'pst' => array(
 551                                 'Do a pre-save transform on the input before parsing it',
 552                                 'Ignored if page, pageid or oldid is used'
 553                         ),
 554                         'onlypst' => array(
 555                                 'Do a pre-save transform (PST) on the input, but don\'t parse it',
 556                                 'Returns the same wikitext, after a PST has been applied. Ignored if page, pageid or oldid is used'
 557                         ),
 558                         'uselang' => 'Which language to parse the request in',
 559                         'section' => 'Only retrieve the content of this section number',
 560                         'disablepp' => 'Disable the PP Report from the parser output',
 561                 );
 562         }
 563
 564         public function getDescription() {
 565                 return 'This module parses wikitext and returns parser output';
 566         }
 567
 568         public function getPossibleErrors() {
 569                 return array_merge( parent::getPossibleErrors(), array(
 570                         array( 'code' => 'params', 'info' => 'The page parameter cannot be used together with the text and title parameters' ),
 571                         array( 'code' => 'missingrev', 'info' => 'There is no revision ID oldid' ),
 572                         array( 'code' => 'permissiondenied', 'info' => 'You don\'t have permission to view deleted revisions' ),
 573                         array( 'code' => 'missingtitle', 'info' => 'The page you specified doesn\'t exist' ),
 574                         array( 'code' => 'nosuchsection', 'info' => 'There is no section sectionnumber in page' ),
 575                         array( 'nosuchpageid' ),
 576                 ) );
 577         }
 578
 579         protected function getExamples() {
 580                 return array(
 581                         'api.php?action=parse&text={{Project:Sandbox}}'
 582                 );
 583         }
 584
 585         public function getVersion() {
 586                 return __CLASS__ . ': $Id$';
 587         }
 588 }