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 = isset( $data['query']['redirects'] )
 138                                                         ? $data['query']['redirects'] : array();
 139                                                 $to = $page;
 140                                                 foreach ( (array)$redirValues as $r ) {
 141                                                         $to = $r['to'];
 142                                                 }
 143                                         } else {
 144                                                 $to = $page;
 145                                         }
 146                                         $titleObj = Title::newFromText( $to );
 147                                         if ( !$titleObj || !$titleObj->exists() ) {
 148                                                 $this->dieUsage( "The page you specified doesn't exist", 'missingtitle' );
 149                                         }
 150                                 }
 151                                 $wgTitle = $titleObj;
 152
 153                                 $articleObj = new Article( $titleObj, 0 );
 154                                 if ( isset( $prop['revid'] ) ) {
 155                                         $oldid = $articleObj->getRevIdFetched();
 156                                 }
 157
 158                                 $p_result = $this->getParsedSectionOrText( $articleObj, $titleObj, $popts, $pageid,
 159                                          isset( $prop['wikitext'] ) ) ;
 160                         }
 161
 162                 } else { // Not $oldid, $pageid, $page. Hence based on $text
 163
 164                         $this->text = $text;
 165                         $titleObj = Title::newFromText( $title );
 166                         if ( !$titleObj ) {
 167                                 $titleObj = Title::newFromText( 'API' );
 168                         }
 169                         $wgTitle = $titleObj;
 170
 171                         if ( $this->section !== false ) {
 172                                 $this->text = $this->getSectionText( $this->text, $titleObj->getText() );
 173                         }
 174
 175                         if ( $params['pst'] || $params['onlypst'] ) {
 176                                 $this->pstText = $wgParser->preSaveTransform( $this->text, $titleObj, $wgUser, $popts );
 177                         }
 178                         if ( $params['onlypst'] ) {
 179                                 // Build a result and bail out
 180                                 $result_array['text'] = array();
 181                                 $this->getResult()->setContent( $result_array['text'], $this->pstText );
 182                                 if ( isset( $prop['wikitext'] ) ) {
 183                                         $result_array['wikitext'] = array();
 184                                         $this->getResult()->setContent( $result_array['wikitext'], $this->text );
 185                                 }
 186                                 $this->getResult()->addValue( null, $this->getModuleName(), $result_array );
 187                                 return;
 188                         }
 189                         $p_result = $wgParser->parse( $params['pst'] ? $this->pstText : $this->text, $titleObj, $popts );
 190                 }
 191
 192                 // Return result
 193                 $result = $this->getResult();
 194                 $result_array = array();
 195
 196                 $result_array['title'] = $titleObj->getPrefixedText();
 197
 198                 if ( !is_null( $oldid ) ) {
 199                         $result_array['revid'] = intval( $oldid );
 200                 }
 201
 202                 if ( $params['redirects'] && !is_null( $redirValues ) ) {
 203                         $result_array['redirects'] = $redirValues;
 204                 }
 205
 206                 if ( isset( $prop['text'] ) ) {
 207                         $result_array['text'] = array();
 208                         $result->setContent( $result_array['text'], $p_result->getText() );
 209                 }
 210
 211                 if ( !is_null( $params['summary'] ) ) {
 212                         $result_array['parsedsummary'] = array();
 213                         $result->setContent( $result_array['parsedsummary'], $wgUser->getSkin()->formatComment( $params['summary'], $titleObj ) );
 214                 }
 215
 216                 if ( isset( $prop['langlinks'] ) ) {
 217                         $result_array['langlinks'] = $this->formatLangLinks( $p_result->getLanguageLinks() );
 218                 }
 219                 if ( isset( $prop['languageshtml'] ) ) {
 220                         $languagesHtml = $this->languagesHtml( $p_result->getLanguageLinks() );
 221                         $result_array['languageshtml'] = array();
 222                         $result->setContent( $result_array['languageshtml'], $languagesHtml );
 223                 }
 224                 if ( isset( $prop['categories'] ) ) {
 225                         $result_array['categories'] = $this->formatCategoryLinks( $p_result->getCategories() );
 226                 }
 227                 if ( isset( $prop['categorieshtml'] ) ) {
 228                         $categoriesHtml = $this->categoriesHtml( $p_result->getCategories() );
 229                         $result_array['categorieshtml'] = array();
 230                         $result->setContent( $result_array['categorieshtml'], $categoriesHtml );
 231                 }
 232                 if ( isset( $prop['links'] ) ) {
 233                         $result_array['links'] = $this->formatLinks( $p_result->getLinks() );
 234                 }
 235                 if ( isset( $prop['templates'] ) ) {
 236                         $result_array['templates'] = $this->formatLinks( $p_result->getTemplates() );
 237                 }
 238                 if ( isset( $prop['images'] ) ) {
 239                         $result_array['images'] = array_keys( $p_result->getImages() );
 240                 }
 241                 if ( isset( $prop['externallinks'] ) ) {
 242                         $result_array['externallinks'] = array_keys( $p_result->getExternalLinks() );
 243                 }
 244                 if ( isset( $prop['sections'] ) ) {
 245                         $result_array['sections'] = $p_result->getSections();
 246                 }
 247
 248                 if ( isset( $prop['displaytitle'] ) ) {
 249                         $result_array['displaytitle'] = $p_result->getDisplayTitle() ?
 250                                                         $p_result->getDisplayTitle() :
 251                                                         $titleObj->getPrefixedText();
 252                 }
 253
 254                 if ( isset( $prop['headitems'] ) || isset( $prop['headhtml'] ) ) {
 255                         $context = new RequestContext;
 256                         $context->output->addParserOutputNoText( $p_result );
 257
 258                         if ( isset( $prop['headitems'] ) ) {
 259                                 $headItems = $this->formatHeadItems( $p_result->getHeadItems() );
 260
 261                                 $context->skin->setupUserCss( $context->output );
 262                                 $css = $this->formatCss( $context->output->buildCssLinksArray() );
 263
 264                                 $scripts = array( $context->output->getHeadScripts( $context->skin ) );
 265
 266                                 $result_array['headitems'] = array_merge( $headItems, $css, $scripts );
 267                         }
 268
 269                         if ( isset( $prop['headhtml'] ) ) {
 270                                 $result_array['headhtml'] = array();
 271                                 $result->setContent( $result_array['headhtml'], $context->output->headElement( $context->skin ) );
 272                         }
 273                 }
 274
 275                 if ( isset( $prop['iwlinks'] ) ) {
 276                         $result_array['iwlinks'] = $this->formatIWLinks( $p_result->getInterwikiLinks() );
 277                 }
 278
 279                 if ( isset( $prop['wikitext'] ) ) {
 280                         $result_array['wikitext'] = array();
 281                         $result->setContent( $result_array['wikitext'], $this->text );
 282                         if ( !is_null( $this->pstText ) ) {
 283                                 $result_array['psttext'] = array();
 284                                 $result->setContent( $result_array['psttext'], $this->pstText );
 285                         }
 286                 }
 287
 288                 $result_mapping = array(
 289                         'redirects' => 'r',
 290                         'langlinks' => 'll',
 291                         'categories' => 'cl',
 292                         'links' => 'pl',
 293                         'templates' => 'tl',
 294                         'images' => 'img',
 295                         'externallinks' => 'el',
 296                         'iwlinks' => 'iw',
 297                         'sections' => 's',
 298                         'headitems' => 'hi',
 299                 );
 300                 $this->setIndexedTagNames( $result_array, $result_mapping );
 301                 $result->addValue( null, $this->getModuleName(), $result_array );
 302
 303                 if ( !is_null( $oldLang ) ) {
 304                         $wgLang = $oldLang; // Reset $wgLang to $oldLang
 305                 }
 306         }
 307
 308         /**
 309          * @param  $articleObj Article
 310          * @param  $titleObj Title
 311          * @param  $popts ParserOptions
 312          * @param  $pageId Int
 313          * @param  $getWikitext Bool
 314          * @return ParserOutput
 315          */
 316         private function getParsedSectionOrText( $articleObj, $titleObj, $popts, $pageId = null, $getWikitext = false ) {
 317                 if ( $this->section !== false ) {
 318                         global $wgParser;
 319
 320                         $this->text = $this->getSectionText( $articleObj->getRawText(), !is_null ( $pageId )
 321                                         ? 'page id ' . $pageId : $titleObj->getText() );
 322
 323                         return $wgParser->parse( $this->text, $titleObj, $popts );
 324                 } else {
 325                         // Try the parser cache first
 326                         $pout = $articleObj->getParserOutput();
 327                         if ( $getWikitext ) {
 328                                 $rev = Revision::newFromTitle( $titleObj );
 329                                 if ( $rev ) {
 330                                         $this->text = $rev->getText();
 331                                 }
 332                         }
 333                         return $pout;
 334                 }
 335         }
 336
 337         private function getSectionText( $text, $what ) {
 338                 global $wgParser;
 339                 $text = $wgParser->getSection( $text, $this->section, false );
 340                 if ( $text === false ) {
 341                         $this->dieUsage( "There is no section {$this->section} in " . $what, 'nosuchsection' );
 342                 }
 343                 return $text;
 344         }
 345
 346         private function formatLangLinks( $links ) {
 347                 $result = array();
 348                 foreach ( $links as $link ) {
 349                         $entry = array();
 350                         $bits = explode( ':', $link, 2 );
 351                         $title = Title::newFromText( $link );
 352
 353                         $entry['lang'] = $bits[0];
 354                         if ( $title ) {
 355                                 $entry['url'] = $title->getFullURL();
 356                         }
 357                         $this->getResult()->setContent( $entry, $bits[1] );
 358                         $result[] = $entry;
 359                 }
 360                 return $result;
 361         }
 362
 363         private function formatCategoryLinks( $links ) {
 364                 $result = array();
 365                 foreach ( $links as $link => $sortkey ) {
 366                         $entry = array();
 367                         $entry['sortkey'] = $sortkey;
 368                         $this->getResult()->setContent( $entry, $link );
 369                         $result[] = $entry;
 370                 }
 371                 return $result;
 372         }
 373
 374         private function categoriesHtml( $categories ) {
 375                 global $wgOut, $wgUser;
 376                 $wgOut->addCategoryLinks( $categories );
 377                 return $wgUser->getSkin()->getCategories();
 378         }
 379
 380         /**
 381          * @deprecated No modern skin generates langlinks this way, please use langlinks data to generate your own html
 382          */
 383         private function languagesHtml( $languages ) {
 384                 global $wgContLang, $wgHideInterlanguageLinks;
 385
 386                 if ( $wgHideInterlanguageLinks || count( $languages ) == 0 ) {
 387                         return '';
 388                 }
 389
 390                 $s = htmlspecialchars( wfMsg( 'otherlanguages' ) . wfMsg( 'colon-separator' ) );
 391
 392                 $langs = array();
 393                 foreach ( $languages as $l ) {
 394                         $nt = Title::newFromText( $l );
 395                         $text = $wgContLang->getLanguageName( $nt->getInterwiki() );
 396
 397                         $langs[] = Html::element( 'a',
 398                                 array( 'href' => $nt->getFullURL(), 'title' => $nt->getText(), 'class' => "external" ),
 399                                 $text == '' ? $l : $text );
 400                 }
 401
 402                 $s .= implode( htmlspecialchars( wfMsgExt( 'pipe-separator', 'escapenoentities' ) ), $langs );
 403
 404                 if ( $wgContLang->isRTL() ) {
 405                         $s = Html::rawElement( 'span', array( 'dir' => "LTR" ), $s );
 406                 }
 407
 408                 return $s;
 409         }
 410
 411         private function formatLinks( $links ) {
 412                 $result = array();
 413                 foreach ( $links as $ns => $nslinks ) {
 414                         foreach ( $nslinks as $title => $id ) {
 415                                 $entry = array();
 416                                 $entry['ns'] = $ns;
 417                                 $this->getResult()->setContent( $entry, Title::makeTitle( $ns, $title )->getFullText() );
 418                                 if ( $id != 0 ) {
 419                                         $entry['exists'] = '';
 420                                 }
 421                                 $result[] = $entry;
 422                         }
 423                 }
 424                 return $result;
 425         }
 426
 427         private function formatIWLinks( $iw ) {
 428                 $result = array();
 429                 foreach ( $iw as $prefix => $titles ) {
 430                         foreach ( array_keys( $titles ) as $title ) {
 431                                 $entry = array();
 432                                 $entry['prefix'] = $prefix;
 433
 434                                 $title = Title::newFromText( "{$prefix}:{$title}" );
 435                                 if ( $title ) {
 436                                         $entry['url'] = $title->getFullURL();
 437                                 }
 438
 439                                 $this->getResult()->setContent( $entry, $title->getFullText() );
 440                                 $result[] = $entry;
 441                         }
 442                 }
 443                 return $result;
 444         }
 445
 446         private function formatHeadItems( $headItems ) {
 447                 $result = array();
 448                 foreach ( $headItems as $tag => $content ) {
 449                         $entry = array();
 450                         $entry['tag'] = $tag;
 451                         $this->getResult()->setContent( $entry, $content );
 452                         $result[] = $entry;
 453                 }
 454                 return $result;
 455         }
 456
 457         private function formatCss( $css ) {
 458                 $result = array();
 459                 foreach ( $css as $file => $link ) {
 460                         $entry = array();
 461                         $entry['file'] = $file;
 462                         $this->getResult()->setContent( $entry, $link );
 463                         $result[] = $entry;
 464                 }
 465                 return $result;
 466         }
 467
 468         private function setIndexedTagNames( &$array, $mapping ) {
 469                 foreach ( $mapping as $key => $name ) {
 470                         if ( isset( $array[$key] ) ) {
 471                                 $this->getResult()->setIndexedTagName( $array[$key], $name );
 472                         }
 473                 }
 474         }
 475
 476         public function getAllowedParams() {
 477                 return array(
 478                         'title' => array(
 479                                 ApiBase::PARAM_DFLT => 'API',
 480                         ),
 481                         'text' => null,
 482                         'summary' => null,
 483                         'page' => null,
 484                         'pageid' => array(
 485                                 ApiBase::PARAM_TYPE => 'integer',
 486                         ),
 487                         'redirects' => false,
 488                         'oldid' => array(
 489                                 ApiBase::PARAM_TYPE => 'integer',
 490                         ),
 491                         'prop' => array(
 492                                 ApiBase::PARAM_DFLT => 'text|langlinks|categories|links|templates|images|externallinks|sections|revid|displaytitle',
 493                                 ApiBase::PARAM_ISMULTI => true,
 494                                 ApiBase::PARAM_TYPE => array(
 495                                         'text',
 496                                         'langlinks',
 497                                         'languageshtml',
 498                                         'categories',
 499                                         'categorieshtml',
 500                                         'links',
 501                                         'templates',
 502                                         'images',
 503                                         'externallinks',
 504                                         'sections',
 505                                         'revid',
 506                                         'displaytitle',
 507                                         'headitems',
 508                                         'headhtml',
 509                                         'iwlinks',
 510                                         'wikitext',
 511                                 )
 512                         ),
 513                         'pst' => false,
 514                         'onlypst' => false,
 515                         'uselang' => null,
 516                         'section' => null,
 517                         'disablepp' => false,
 518                 );
 519         }
 520
 521         public function getParamDescription() {
 522                 $p = $this->getModulePrefix();
 523                 return array(
 524                         'text' => 'Wikitext to parse',
 525                         'summary' => 'Summary to parse',
 526                         'redirects' => "If the {$p}page parameter is set to a redirect, resolve it",
 527                         'title' => 'Title of page the text belongs to',
 528                         'page' => "Parse the content of this page. Cannot be used together with {$p}text and {$p}title",
 529                         'pageid' => "Parse the content of this page. Overrides {$p}page",
 530                         'oldid' => "Parse the content of this revision. Overrides {$p}page and {$p}pageid",
 531                         'prop' => array(
 532                                 'Which pieces of information to get',
 533                                 ' text           - Gives the parsed text of the wikitext',
 534                                 ' langlinks      - Gives the language links in the parsed wikitext',
 535                                 ' categories     - Gives the categories in the parsed wikitext',
 536                                 ' categorieshtml - Gives the HTML version of the categories',
 537                                 ' languageshtml  - Gives the HTML version of the language links',
 538                                 ' links          - Gives the internal links in the parsed wikitext',
 539                                 ' templates      - Gives the templates in the parsed wikitext',
 540                                 ' images         - Gives the images in the parsed wikitext',
 541                                 ' externallinks  - Gives the external links in the parsed wikitext',
 542                                 ' sections       - Gives the sections in the parsed wikitext',
 543                                 ' revid          - Adds the revision ID of the parsed page',
 544                                 ' displaytitle   - Adds the title of the parsed wikitext',
 545                                 ' headitems      - Gives items to put in the <head> of the page',
 546                                 ' headhtml       - Gives parsed <head> of the page',
 547                                 ' iwlinks        - Gives interwiki links in the parsed wikitext',
 548                                 ' wikitext       - Gives the original wikitext that was parsed',
 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 '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 }