Merge "Typo fixes and non-code tweaks"
[mediawiki.git] / includes / api / ApiParse.php
blobff91b92a48e6025bd21ab49b2600cb88820380a6
1 <?php
2 /**
3 * Created on Dec 01, 2007
5 * Copyright © 2007 Yuri Astrakhan "<Firstname><Lastname>@gmail.com"
7 * This program is free software; you can redistribute it and/or modify
8 * it under the terms of the GNU General Public License as published by
9 * the Free Software Foundation; either version 2 of the License, or
10 * (at your option) any later version.
12 * This program is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 * GNU General Public License for more details.
17 * You should have received a copy of the GNU General Public License along
18 * with this program; if not, write to the Free Software Foundation, Inc.,
19 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
20 * http://www.gnu.org/copyleft/gpl.html
22 * @file
25 /**
26 * @ingroup API
28 class ApiParse extends ApiBase {
30 /** @var string $section */
31 private $section = null;
33 /** @var Content $content */
34 private $content = null;
36 /** @var Content $pstContent */
37 private $pstContent = null;
39 public function execute() {
40 // The data is hot but user-dependent, like page views, so we set vary cookies
41 $this->getMain()->setCacheMode( 'anon-public-user-private' );
43 // Get parameters
44 $params = $this->extractRequestParams();
45 $text = $params['text'];
46 $title = $params['title'];
47 if ( $title === null ) {
48 $titleProvided = false;
49 // A title is needed for parsing, so arbitrarily choose one
50 $title = 'API';
51 } else {
52 $titleProvided = true;
55 $page = $params['page'];
56 $pageid = $params['pageid'];
57 $oldid = $params['oldid'];
59 $model = $params['contentmodel'];
60 $format = $params['contentformat'];
62 if ( !is_null( $page ) && ( !is_null( $text ) || $titleProvided ) ) {
63 $this->dieUsage(
64 'The page parameter cannot be used together with the text and title parameters',
65 'params'
69 $prop = array_flip( $params['prop'] );
71 if ( isset( $params['section'] ) ) {
72 $this->section = $params['section'];
73 } else {
74 $this->section = false;
77 // The parser needs $wgTitle to be set, apparently the
78 // $title parameter in Parser::parse isn't enough *sigh*
79 // TODO: Does this still need $wgTitle?
80 global $wgParser, $wgTitle;
82 $redirValues = null;
84 // Return result
85 $result = $this->getResult();
87 if ( !is_null( $oldid ) || !is_null( $pageid ) || !is_null( $page ) ) {
88 if ( !is_null( $oldid ) ) {
89 // Don't use the parser cache
90 $rev = Revision::newFromId( $oldid );
91 if ( !$rev ) {
92 $this->dieUsage( "There is no revision ID $oldid", 'missingrev' );
94 if ( !$rev->userCan( Revision::DELETED_TEXT, $this->getUser() ) ) {
95 $this->dieUsage( "You don't have permission to view deleted revisions", 'permissiondenied' );
98 $titleObj = $rev->getTitle();
99 $wgTitle = $titleObj;
100 $pageObj = WikiPage::factory( $titleObj );
101 $popts = $this->makeParserOptions( $pageObj, $params );
103 // If for some reason the "oldid" is actually the current revision, it may be cached
104 if ( $rev->isCurrent() ) {
105 // May get from/save to parser cache
106 $p_result = $this->getParsedContent( $pageObj, $popts,
107 $pageid, isset( $prop['wikitext'] ) );
108 } else { // This is an old revision, so get the text differently
109 $this->content = $rev->getContent( Revision::FOR_THIS_USER, $this->getUser() );
111 if ( $this->section !== false ) {
112 $this->content = $this->getSectionContent( $this->content, 'r' . $rev->getId() );
115 // Should we save old revision parses to the parser cache?
116 $p_result = $this->content->getParserOutput( $titleObj, $rev->getId(), $popts );
118 } else { // Not $oldid, but $pageid or $page
119 if ( $params['redirects'] ) {
120 $reqParams = array(
121 'action' => 'query',
122 'redirects' => '',
124 if ( !is_null( $pageid ) ) {
125 $reqParams['pageids'] = $pageid;
126 } else { // $page
127 $reqParams['titles'] = $page;
129 $req = new FauxRequest( $reqParams );
130 $main = new ApiMain( $req );
131 $main->execute();
132 $data = $main->getResultData();
133 $redirValues = isset( $data['query']['redirects'] )
134 ? $data['query']['redirects']
135 : array();
136 $to = $page;
137 foreach ( (array)$redirValues as $r ) {
138 $to = $r['to'];
140 $pageParams = array( 'title' => $to );
141 } elseif ( !is_null( $pageid ) ) {
142 $pageParams = array( 'pageid' => $pageid );
143 } else { // $page
144 $pageParams = array( 'title' => $page );
147 $pageObj = $this->getTitleOrPageId( $pageParams, 'fromdb' );
148 $titleObj = $pageObj->getTitle();
149 if ( !$titleObj || !$titleObj->exists() ) {
150 $this->dieUsage( "The page you specified doesn't exist", 'missingtitle' );
152 $wgTitle = $titleObj;
154 if ( isset( $prop['revid'] ) ) {
155 $oldid = $pageObj->getLatest();
158 $popts = $this->makeParserOptions( $pageObj, $params );
160 // Potentially cached
161 $p_result = $this->getParsedContent( $pageObj, $popts, $pageid,
162 isset( $prop['wikitext'] ) );
164 } else { // Not $oldid, $pageid, $page. Hence based on $text
165 $titleObj = Title::newFromText( $title );
166 if ( !$titleObj || $titleObj->isExternal() ) {
167 $this->dieUsageMsg( array( 'invalidtitle', $title ) );
169 $wgTitle = $titleObj;
170 if ( $titleObj->canExist() ) {
171 $pageObj = WikiPage::factory( $titleObj );
172 } else {
173 // Do like MediaWiki::initializeArticle()
174 $article = Article::newFromTitle( $titleObj, $this->getContext() );
175 $pageObj = $article->getPage();
178 $popts = $this->makeParserOptions( $pageObj, $params );
179 $textProvided = !is_null( $text );
181 if ( !$textProvided ) {
182 if ( $titleProvided && ( $prop || $params['generatexml'] ) ) {
183 $this->setWarning(
184 "'title' used without 'text', and parsed page properties were requested " .
185 "(did you mean to use 'page' instead of 'title'?)"
188 // Prevent warning from ContentHandler::makeContent()
189 $text = '';
192 // If we are parsing text, do not use the content model of the default
193 // API title, but default to wikitext to keep BC.
194 if ( $textProvided && !$titleProvided && is_null( $model ) ) {
195 $model = CONTENT_MODEL_WIKITEXT;
196 $this->setWarning( "No 'title' or 'contentmodel' was given, assuming $model." );
199 try {
200 $this->content = ContentHandler::makeContent( $text, $titleObj, $model, $format );
201 } catch ( MWContentSerializationException $ex ) {
202 $this->dieUsage( $ex->getMessage(), 'parseerror' );
205 if ( $this->section !== false ) {
206 $this->content = $this->getSectionContent( $this->content, $titleObj->getPrefixedText() );
209 if ( $params['pst'] || $params['onlypst'] ) {
210 $this->pstContent = $this->content->preSaveTransform( $titleObj, $this->getUser(), $popts );
212 if ( $params['onlypst'] ) {
213 // Build a result and bail out
214 $result_array = array();
215 $result_array['text'] = array();
216 ApiResult::setContent( $result_array['text'], $this->pstContent->serialize( $format ) );
217 if ( isset( $prop['wikitext'] ) ) {
218 $result_array['wikitext'] = array();
219 ApiResult::setContent( $result_array['wikitext'], $this->content->serialize( $format ) );
221 $result->addValue( null, $this->getModuleName(), $result_array );
223 return;
226 // Not cached (save or load)
227 if ( $params['pst'] ) {
228 $p_result = $this->pstContent->getParserOutput( $titleObj, null, $popts );
229 } else {
230 $p_result = $this->content->getParserOutput( $titleObj, null, $popts );
234 $result_array = array();
236 $result_array['title'] = $titleObj->getPrefixedText();
238 if ( !is_null( $oldid ) ) {
239 $result_array['revid'] = intval( $oldid );
242 if ( $params['redirects'] && !is_null( $redirValues ) ) {
243 $result_array['redirects'] = $redirValues;
246 if ( $params['disabletoc'] ) {
247 $p_result->setTOCEnabled( false );
250 if ( isset( $prop['text'] ) ) {
251 $result_array['text'] = array();
252 ApiResult::setContent( $result_array['text'], $p_result->getText() );
255 if ( !is_null( $params['summary'] ) ) {
256 $result_array['parsedsummary'] = array();
257 ApiResult::setContent(
258 $result_array['parsedsummary'],
259 Linker::formatComment( $params['summary'], $titleObj )
263 if ( isset( $prop['langlinks'] ) ) {
264 $langlinks = $p_result->getLanguageLinks();
266 if ( $params['effectivelanglinks'] ) {
267 // Link flags are ignored for now, but may in the future be
268 // included in the result.
269 $linkFlags = array();
270 Hooks::run( 'LanguageLinks', array( $titleObj, &$langlinks, &$linkFlags ) );
272 } else {
273 $langlinks = false;
276 if ( isset( $prop['langlinks'] ) ) {
277 $result_array['langlinks'] = $this->formatLangLinks( $langlinks );
279 if ( isset( $prop['categories'] ) ) {
280 $result_array['categories'] = $this->formatCategoryLinks( $p_result->getCategories() );
282 if ( isset( $prop['categorieshtml'] ) ) {
283 $categoriesHtml = $this->categoriesHtml( $p_result->getCategories() );
284 $result_array['categorieshtml'] = array();
285 ApiResult::setContent( $result_array['categorieshtml'], $categoriesHtml );
287 if ( isset( $prop['links'] ) ) {
288 $result_array['links'] = $this->formatLinks( $p_result->getLinks() );
290 if ( isset( $prop['templates'] ) ) {
291 $result_array['templates'] = $this->formatLinks( $p_result->getTemplates() );
293 if ( isset( $prop['images'] ) ) {
294 $result_array['images'] = array_keys( $p_result->getImages() );
296 if ( isset( $prop['externallinks'] ) ) {
297 $result_array['externallinks'] = array_keys( $p_result->getExternalLinks() );
299 if ( isset( $prop['sections'] ) ) {
300 $result_array['sections'] = $p_result->getSections();
303 if ( isset( $prop['displaytitle'] ) ) {
304 $result_array['displaytitle'] = $p_result->getDisplayTitle() ?
305 $p_result->getDisplayTitle() :
306 $titleObj->getPrefixedText();
309 if ( isset( $prop['headitems'] ) || isset( $prop['headhtml'] ) ) {
310 $context = $this->getContext();
311 $context->setTitle( $titleObj );
312 $context->getOutput()->addParserOutputMetadata( $p_result );
314 if ( isset( $prop['headitems'] ) ) {
315 $headItems = $this->formatHeadItems( $p_result->getHeadItems() );
317 $css = $this->formatCss( $context->getOutput()->buildCssLinksArray() );
319 $scripts = array( $context->getOutput()->getHeadScripts() );
321 $result_array['headitems'] = array_merge( $headItems, $css, $scripts );
324 if ( isset( $prop['headhtml'] ) ) {
325 $result_array['headhtml'] = array();
326 ApiResult::setContent(
327 $result_array['headhtml'],
328 $context->getOutput()->headElement( $context->getSkin() )
333 if ( isset( $prop['modules'] ) ) {
334 $result_array['modules'] = array_values( array_unique( $p_result->getModules() ) );
335 $result_array['modulescripts'] = array_values( array_unique( $p_result->getModuleScripts() ) );
336 $result_array['modulestyles'] = array_values( array_unique( $p_result->getModuleStyles() ) );
337 $result_array['modulemessages'] = array_values( array_unique( $p_result->getModuleMessages() ) );
340 if ( isset( $prop['indicators'] ) ) {
341 foreach ( $p_result->getIndicators() as $name => $content ) {
342 $indicator = array( 'name' => $name );
343 ApiResult::setContent( $indicator, $content );
344 $result_array['indicators'][] = $indicator;
348 if ( isset( $prop['iwlinks'] ) ) {
349 $result_array['iwlinks'] = $this->formatIWLinks( $p_result->getInterwikiLinks() );
352 if ( isset( $prop['wikitext'] ) ) {
353 $result_array['wikitext'] = array();
354 ApiResult::setContent( $result_array['wikitext'], $this->content->serialize( $format ) );
355 if ( !is_null( $this->pstContent ) ) {
356 $result_array['psttext'] = array();
357 ApiResult::setContent( $result_array['psttext'], $this->pstContent->serialize( $format ) );
360 if ( isset( $prop['properties'] ) ) {
361 $result_array['properties'] = $this->formatProperties( $p_result->getProperties() );
364 if ( isset( $prop['limitreportdata'] ) ) {
365 $result_array['limitreportdata'] =
366 $this->formatLimitReportData( $p_result->getLimitReportData() );
368 if ( isset( $prop['limitreporthtml'] ) ) {
369 $limitreportHtml = EditPage::getPreviewLimitReport( $p_result );
370 $result_array['limitreporthtml'] = array();
371 ApiResult::setContent( $result_array['limitreporthtml'], $limitreportHtml );
374 if ( $params['generatexml'] ) {
375 if ( $this->content->getModel() != CONTENT_MODEL_WIKITEXT ) {
376 $this->dieUsage( "generatexml is only supported for wikitext content", "notwikitext" );
379 $wgParser->startExternalParse( $titleObj, $popts, Parser::OT_PREPROCESS );
380 $dom = $wgParser->preprocessToDom( $this->content->getNativeData() );
381 if ( is_callable( array( $dom, 'saveXML' ) ) ) {
382 $xml = $dom->saveXML();
383 } else {
384 $xml = $dom->__toString();
386 $result_array['parsetree'] = array();
387 ApiResult::setContent( $result_array['parsetree'], $xml );
390 $result_mapping = array(
391 'redirects' => 'r',
392 'langlinks' => 'll',
393 'categories' => 'cl',
394 'links' => 'pl',
395 'templates' => 'tl',
396 'images' => 'img',
397 'externallinks' => 'el',
398 'iwlinks' => 'iw',
399 'sections' => 's',
400 'headitems' => 'hi',
401 'modules' => 'm',
402 'indicators' => 'ind',
403 'modulescripts' => 'm',
404 'modulestyles' => 'm',
405 'modulemessages' => 'm',
406 'properties' => 'pp',
407 'limitreportdata' => 'lr',
409 $this->setIndexedTagNames( $result_array, $result_mapping );
410 $result->addValue( null, $this->getModuleName(), $result_array );
414 * Constructs a ParserOptions object
416 * @param WikiPage $pageObj
417 * @param array $params
419 * @return ParserOptions
421 protected function makeParserOptions( WikiPage $pageObj, array $params ) {
422 wfProfileIn( __METHOD__ );
424 $popts = $pageObj->makeParserOptions( $this->getContext() );
425 $popts->enableLimitReport( !$params['disablepp'] );
426 $popts->setIsPreview( $params['preview'] || $params['sectionpreview'] );
427 $popts->setIsSectionPreview( $params['sectionpreview'] );
428 $popts->setEditSection( !$params['disableeditsection'] );
430 wfProfileOut( __METHOD__ );
432 return $popts;
436 * @param WikiPage $page
437 * @param ParserOptions $popts
438 * @param int $pageId
439 * @param bool $getWikitext
440 * @return ParserOutput
442 private function getParsedContent( WikiPage $page, $popts, $pageId = null, $getWikitext = false ) {
443 $this->content = $page->getContent( Revision::RAW ); //XXX: really raw?
445 if ( $this->section !== false && $this->content !== null ) {
446 $this->content = $this->getSectionContent(
447 $this->content,
448 !is_null( $pageId ) ? 'page id ' . $pageId : $page->getTitle()->getPrefixedText()
451 // Not cached (save or load)
452 return $this->content->getParserOutput( $page->getTitle(), null, $popts );
455 // Try the parser cache first
456 // getParserOutput will save to Parser cache if able
457 $pout = $page->getParserOutput( $popts );
458 if ( !$pout ) {
459 $this->dieUsage( "There is no revision ID {$page->getLatest()}", 'missingrev' );
461 if ( $getWikitext ) {
462 $this->content = $page->getContent( Revision::RAW );
465 return $pout;
469 * @param Content $content
470 * @param string $what Identifies the content in error messages, e.g. page title.
471 * @return Content|bool
473 private function getSectionContent( Content $content, $what ) {
474 // Not cached (save or load)
475 $section = $content->getSection( $this->section );
476 if ( $section === false ) {
477 $this->dieUsage( "There is no section {$this->section} in " . $what, 'nosuchsection' );
479 if ( $section === null ) {
480 $this->dieUsage( "Sections are not supported by " . $what, 'nosuchsection' );
481 $section = false;
484 return $section;
487 private function formatLangLinks( $links ) {
488 $result = array();
489 foreach ( $links as $link ) {
490 $entry = array();
491 $bits = explode( ':', $link, 2 );
492 $title = Title::newFromText( $link );
494 $entry['lang'] = $bits[0];
495 if ( $title ) {
496 $entry['url'] = wfExpandUrl( $title->getFullURL(), PROTO_CURRENT );
497 // localised language name in 'uselang' language
498 $entry['langname'] = Language::fetchLanguageName(
499 $title->getInterwiki(),
500 $this->getLanguage()->getCode()
503 // native language name
504 $entry['autonym'] = Language::fetchLanguageName( $title->getInterwiki() );
506 ApiResult::setContent( $entry, $bits[1] );
507 $result[] = $entry;
510 return $result;
513 private function formatCategoryLinks( $links ) {
514 $result = array();
516 if ( !$links ) {
517 return $result;
520 // Fetch hiddencat property
521 $lb = new LinkBatch;
522 $lb->setArray( array( NS_CATEGORY => $links ) );
523 $db = $this->getDB();
524 $res = $db->select( array( 'page', 'page_props' ),
525 array( 'page_title', 'pp_propname' ),
526 $lb->constructSet( 'page', $db ),
527 __METHOD__,
528 array(),
529 array( 'page_props' => array(
530 'LEFT JOIN', array( 'pp_propname' => 'hiddencat', 'pp_page = page_id' )
533 $hiddencats = array();
534 foreach ( $res as $row ) {
535 $hiddencats[$row->page_title] = isset( $row->pp_propname );
538 foreach ( $links as $link => $sortkey ) {
539 $entry = array();
540 $entry['sortkey'] = $sortkey;
541 ApiResult::setContent( $entry, $link );
542 if ( !isset( $hiddencats[$link] ) ) {
543 $entry['missing'] = '';
544 } elseif ( $hiddencats[$link] ) {
545 $entry['hidden'] = '';
547 $result[] = $entry;
550 return $result;
553 private function categoriesHtml( $categories ) {
554 $context = $this->getContext();
555 $context->getOutput()->addCategoryLinks( $categories );
557 return $context->getSkin()->getCategories();
560 private function formatLinks( $links ) {
561 $result = array();
562 foreach ( $links as $ns => $nslinks ) {
563 foreach ( $nslinks as $title => $id ) {
564 $entry = array();
565 $entry['ns'] = $ns;
566 ApiResult::setContent( $entry, Title::makeTitle( $ns, $title )->getFullText() );
567 if ( $id != 0 ) {
568 $entry['exists'] = '';
570 $result[] = $entry;
574 return $result;
577 private function formatIWLinks( $iw ) {
578 $result = array();
579 foreach ( $iw as $prefix => $titles ) {
580 foreach ( array_keys( $titles ) as $title ) {
581 $entry = array();
582 $entry['prefix'] = $prefix;
584 $title = Title::newFromText( "{$prefix}:{$title}" );
585 if ( $title ) {
586 $entry['url'] = wfExpandUrl( $title->getFullURL(), PROTO_CURRENT );
589 ApiResult::setContent( $entry, $title->getFullText() );
590 $result[] = $entry;
594 return $result;
597 private function formatHeadItems( $headItems ) {
598 $result = array();
599 foreach ( $headItems as $tag => $content ) {
600 $entry = array();
601 $entry['tag'] = $tag;
602 ApiResult::setContent( $entry, $content );
603 $result[] = $entry;
606 return $result;
609 private function formatProperties( $properties ) {
610 $result = array();
611 foreach ( $properties as $name => $value ) {
612 $entry = array();
613 $entry['name'] = $name;
614 ApiResult::setContent( $entry, $value );
615 $result[] = $entry;
618 return $result;
621 private function formatCss( $css ) {
622 $result = array();
623 foreach ( $css as $file => $link ) {
624 $entry = array();
625 $entry['file'] = $file;
626 ApiResult::setContent( $entry, $link );
627 $result[] = $entry;
630 return $result;
633 private function formatLimitReportData( $limitReportData ) {
634 $result = array();
635 $apiResult = $this->getResult();
637 foreach ( $limitReportData as $name => $value ) {
638 $entry = array();
639 $entry['name'] = $name;
640 if ( !is_array( $value ) ) {
641 $value = array( $value );
643 $apiResult->setIndexedTagName( $value, 'param' );
644 $apiResult->setIndexedTagName_recursive( $value, 'param' );
645 $entry = array_merge( $entry, $value );
646 $result[] = $entry;
649 return $result;
652 private function setIndexedTagNames( &$array, $mapping ) {
653 foreach ( $mapping as $key => $name ) {
654 if ( isset( $array[$key] ) ) {
655 $this->getResult()->setIndexedTagName( $array[$key], $name );
660 public function getAllowedParams() {
661 return array(
662 'title' => null,
663 'text' => null,
664 'summary' => null,
665 'page' => null,
666 'pageid' => array(
667 ApiBase::PARAM_TYPE => 'integer',
669 'redirects' => false,
670 'oldid' => array(
671 ApiBase::PARAM_TYPE => 'integer',
673 'prop' => array(
674 ApiBase::PARAM_DFLT => 'text|langlinks|categories|links|templates|' .
675 'images|externallinks|sections|revid|displaytitle|iwlinks|properties',
676 ApiBase::PARAM_ISMULTI => true,
677 ApiBase::PARAM_TYPE => array(
678 'text',
679 'langlinks',
680 'categories',
681 'categorieshtml',
682 'links',
683 'templates',
684 'images',
685 'externallinks',
686 'sections',
687 'revid',
688 'displaytitle',
689 'headitems',
690 'headhtml',
691 'modules',
692 'indicators',
693 'iwlinks',
694 'wikitext',
695 'properties',
696 'limitreportdata',
697 'limitreporthtml',
700 'pst' => false,
701 'onlypst' => false,
702 'effectivelanglinks' => false,
703 'section' => null,
704 'disablepp' => false,
705 'disableeditsection' => false,
706 'generatexml' => array(
707 ApiBase::PARAM_DFLT => false,
708 ApiBase::PARAM_HELP_MSG => array(
709 'apihelp-parse-param-generatexml', CONTENT_MODEL_WIKITEXT
712 'preview' => false,
713 'sectionpreview' => false,
714 'disabletoc' => false,
715 'contentformat' => array(
716 ApiBase::PARAM_TYPE => ContentHandler::getAllContentFormats(),
718 'contentmodel' => array(
719 ApiBase::PARAM_TYPE => ContentHandler::getContentModels(),
724 protected function getExamplesMessages() {
725 return array(
726 'action=parse&page=Project:Sandbox'
727 => 'apihelp-parse-example-page',
728 'action=parse&text={{Project:Sandbox}}&contentmodel=wikitext'
729 => 'apihelp-parse-example-text',
730 'action=parse&text={{PAGENAME}}&title=Test'
731 => 'apihelp-parse-example-texttitle',
732 'action=parse&summary=Some+[[link]]&prop='
733 => 'apihelp-parse-example-summary',
737 public function getHelpUrls() {
738 return 'https://www.mediawiki.org/wiki/API:Parsing_wikitext#parse';