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->addParserOutputNoText( $p_result );
 256
 257                         if ( isset( $prop['headitems'] ) ) {
 258                                 $headItems = $this->formatHeadItems( $p_result->getHeadItems() );
 259
 260                                 $userSkin = $wgUser->getSkin();
 261                                 $userSkin->setupUserCss( $out );
 262                                 $css = $this->formatCss( $out->buildCssLinksArray() );
 263
 264                                 $scripts = array( $out->getHeadScripts( $userSkin ) );
 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'], $out->headElement( $userSkin ) );
 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                 $sk = $wgUser->getSkin();
 378                 return $sk->getCategories();
 379         }
 380
 381         /**
 382          * @deprecated No modern skin generates langlinks this way, please use langlinks 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' => null,
 486                         'redirects' => false,
 487                         'oldid' => null,
 488                         'prop' => array(
 489                                 ApiBase::PARAM_DFLT => 'text|langlinks|categories|links|templates|images|externallinks|sections|revid|displaytitle',
 490                                 ApiBase::PARAM_ISMULTI => true,
 491                                 ApiBase::PARAM_TYPE => array(
 492                                         'text',
 493                                         'langlinks',
 494                                         'languageshtml',
 495                                         'categories',
 496                                         'categorieshtml',
 497                                         'links',
 498                                         'templates',
 499                                         'images',
 500                                         'externallinks',
 501                                         'sections',
 502                                         'revid',
 503                                         'displaytitle',
 504                                         'headitems',
 505                                         'headhtml',
 506                                         'iwlinks',
 507                                         'wikitext',
 508                                 )
 509                         ),
 510                         'pst' => false,
 511                         'onlypst' => false,
 512                         'uselang' => null,
 513                         'section' => null,
 514                         'disablepp' => false,
 515                 );
 516         }
 517
 518         public function getParamDescription() {
 519                 $p = $this->getModulePrefix();
 520                 return array(
 521                         'text' => 'Wikitext to parse',
 522                         'summary' => 'Summary to parse',
 523                         'redirects' => "If the {$p}page parameter is set to a redirect, resolve it",
 524                         'title' => 'Title of page the text belongs to',
 525                         'page' => "Parse the content of this page. Cannot be used together with {$p}text and {$p}title",
 526                         'pageid' => "Parse the content of this page. Overrides {$p}page",
 527                         'oldid' => "Parse the content of this revision. Overrides {$p}page and {$p}pageid",
 528                         'prop' => array(
 529                                 'Which pieces of information to get',
 530                                 ' text           - Gives the parsed text of the wikitext',
 531                                 ' langlinks      - Gives the language links in the parsed wikitext',
 532                                 ' categories     - Gives the categories in the parsed wikitext',
 533                                 ' categorieshtml - Gives the HTML version of the categories',
 534                                 ' languageshtml  - Gives the HTML version of the language links',
 535                                 ' links          - Gives the internal links in the parsed wikitext',
 536                                 ' templates      - Gives the templates in the parsed wikitext',
 537                                 ' images         - Gives the images in the parsed wikitext',
 538                                 ' externallinks  - Gives the external links in the parsed wikitext',
 539                                 ' sections       - Gives the sections in the parsed wikitext',
 540                                 ' revid          - Adds the revision ID of the parsed page',
 541                                 ' displaytitle   - Adds the title of the parsed wikitext',
 542                                 ' headitems      - Gives items to put in the <head> of the page',
 543                                 ' headhtml       - Gives parsed <head> of the page',
 544                                 ' iwlinks        - Gives interwiki links in the parsed wikitext',
 545                                 ' wikitext       - Gives the original wikitext that was parsed',
 546                                 'NOTE: Section tree is only generated if there are more than 4 sections, or if the __TOC__ keyword is present'
 547                         ),
 548                         'pst' => array(
 549                                 'Do a pre-save transform on the input before parsing it',
 550                                 'Ignored if page, pageid or oldid is used'
 551                         ),
 552                         'onlypst' => array(
 553                                 'Do a pre-save transform (PST) on the input, but don\'t parse it',
 554                                 'Returns the same wikitext, after a PST has been applied. Ignored if page, pageid or oldid is used'
 555                         ),
 556                         'uselang' => 'Which language to parse the request in',
 557                         'section' => 'Only retrieve the content of this section number',
 558                         'disablepp' => 'Disable the PP Report from the parser output',
 559                 );
 560         }
 561
 562         public function getDescription() {
 563                 return 'This module parses wikitext and returns parser output';
 564         }
 565
 566         public function getPossibleErrors() {
 567                 return array_merge( parent::getPossibleErrors(), array(
 568                         array( 'code' => 'params', 'info' => 'The page parameter cannot be used together with the text and title parameters' ),
 569                         array( 'code' => 'missingrev', 'info' => 'There is no revision ID oldid' ),
 570                         array( 'code' => 'permissiondenied', 'info' => 'You don\'t have permission to view deleted revisions' ),
 571                         array( 'code' => 'missingtitle', 'info' => 'The page you specified doesn\'t exist' ),
 572                         array( 'code' => 'nosuchsection', 'info' => 'There is no section sectionnumber in page' ),
 573                         array( 'nosuchpageid' ),
 574                 ) );
 575         }
 576
 577         protected function getExamples() {
 578                 return array(
 579                         'api.php?action=parse&text={{Project:Sandbox}}'
 580                 );
 581         }
 582
 583         public function getVersion() {
 584                 return __CLASS__ . ': $Id$';
 585         }
 586 }