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
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' );
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
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 ) ) {
64 'The page parameter cannot be used together with the text and title parameters',
69 $prop = array_flip( $params['prop'] );
71 if ( isset( $params['section'] ) ) {
72 $this->section
= $params['section'];
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;
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 );
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();
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'] ) {
124 if ( !is_null( $pageid ) ) {
125 $reqParams['pageids'] = $pageid;
127 $reqParams['titles'] = $page;
129 $req = new FauxRequest( $reqParams );
130 $main = new ApiMain( $req );
132 $data = $main->getResultData();
133 $redirValues = isset( $data['query']['redirects'] )
134 ?
$data['query']['redirects']
137 foreach ( (array)$redirValues as $r ) {
140 $pageParams = array( 'title' => $to );
141 } elseif ( !is_null( $pageid ) ) {
142 $pageParams = array( 'pageid' => $pageid );
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 );
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'] ) ) {
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()
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." );
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 );
226 // Not cached (save or load)
227 if ( $params['pst'] ) {
228 $p_result = $this->pstContent
->getParserOutput( $titleObj, null, $popts );
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 ) );
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();
384 $xml = $dom->__toString();
386 $result_array['parsetree'] = array();
387 ApiResult
::setContent( $result_array['parsetree'], $xml );
390 $result_mapping = array(
393 'categories' => 'cl',
397 'externallinks' => 'el',
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 ) {
423 $popts = $pageObj->makeParserOptions( $this->getContext() );
424 $popts->enableLimitReport( !$params['disablepp'] );
425 $popts->setIsPreview( $params['preview'] ||
$params['sectionpreview'] );
426 $popts->setIsSectionPreview( $params['sectionpreview'] );
427 $popts->setEditSection( !$params['disableeditsection'] );
433 * @param WikiPage $page
434 * @param ParserOptions $popts
436 * @param bool $getWikitext
437 * @return ParserOutput
439 private function getParsedContent( WikiPage
$page, $popts, $pageId = null, $getWikitext = false ) {
440 $this->content
= $page->getContent( Revision
::RAW
); //XXX: really raw?
442 if ( $this->section
!== false && $this->content
!== null ) {
443 $this->content
= $this->getSectionContent(
445 !is_null( $pageId ) ?
'page id ' . $pageId : $page->getTitle()->getPrefixedText()
448 // Not cached (save or load)
449 return $this->content
->getParserOutput( $page->getTitle(), null, $popts );
452 // Try the parser cache first
453 // getParserOutput will save to Parser cache if able
454 $pout = $page->getParserOutput( $popts );
456 $this->dieUsage( "There is no revision ID {$page->getLatest()}", 'missingrev' );
458 if ( $getWikitext ) {
459 $this->content
= $page->getContent( Revision
::RAW
);
466 * @param Content $content
467 * @param string $what Identifies the content in error messages, e.g. page title.
468 * @return Content|bool
470 private function getSectionContent( Content
$content, $what ) {
471 // Not cached (save or load)
472 $section = $content->getSection( $this->section
);
473 if ( $section === false ) {
474 $this->dieUsage( "There is no section {$this->section} in " . $what, 'nosuchsection' );
476 if ( $section === null ) {
477 $this->dieUsage( "Sections are not supported by " . $what, 'nosuchsection' );
484 private function formatLangLinks( $links ) {
486 foreach ( $links as $link ) {
488 $bits = explode( ':', $link, 2 );
489 $title = Title
::newFromText( $link );
491 $entry['lang'] = $bits[0];
493 $entry['url'] = wfExpandUrl( $title->getFullURL(), PROTO_CURRENT
);
494 // localised language name in 'uselang' language
495 $entry['langname'] = Language
::fetchLanguageName(
496 $title->getInterwiki(),
497 $this->getLanguage()->getCode()
500 // native language name
501 $entry['autonym'] = Language
::fetchLanguageName( $title->getInterwiki() );
503 ApiResult
::setContent( $entry, $bits[1] );
510 private function formatCategoryLinks( $links ) {
517 // Fetch hiddencat property
519 $lb->setArray( array( NS_CATEGORY
=> $links ) );
520 $db = $this->getDB();
521 $res = $db->select( array( 'page', 'page_props' ),
522 array( 'page_title', 'pp_propname' ),
523 $lb->constructSet( 'page', $db ),
526 array( 'page_props' => array(
527 'LEFT JOIN', array( 'pp_propname' => 'hiddencat', 'pp_page = page_id' )
530 $hiddencats = array();
531 foreach ( $res as $row ) {
532 $hiddencats[$row->page_title
] = isset( $row->pp_propname
);
535 foreach ( $links as $link => $sortkey ) {
537 $entry['sortkey'] = $sortkey;
538 ApiResult
::setContent( $entry, $link );
539 if ( !isset( $hiddencats[$link] ) ) {
540 $entry['missing'] = '';
541 } elseif ( $hiddencats[$link] ) {
542 $entry['hidden'] = '';
550 private function categoriesHtml( $categories ) {
551 $context = $this->getContext();
552 $context->getOutput()->addCategoryLinks( $categories );
554 return $context->getSkin()->getCategories();
557 private function formatLinks( $links ) {
559 foreach ( $links as $ns => $nslinks ) {
560 foreach ( $nslinks as $title => $id ) {
563 ApiResult
::setContent( $entry, Title
::makeTitle( $ns, $title )->getFullText() );
565 $entry['exists'] = '';
574 private function formatIWLinks( $iw ) {
576 foreach ( $iw as $prefix => $titles ) {
577 foreach ( array_keys( $titles ) as $title ) {
579 $entry['prefix'] = $prefix;
581 $title = Title
::newFromText( "{$prefix}:{$title}" );
583 $entry['url'] = wfExpandUrl( $title->getFullURL(), PROTO_CURRENT
);
586 ApiResult
::setContent( $entry, $title->getFullText() );
594 private function formatHeadItems( $headItems ) {
596 foreach ( $headItems as $tag => $content ) {
598 $entry['tag'] = $tag;
599 ApiResult
::setContent( $entry, $content );
606 private function formatProperties( $properties ) {
608 foreach ( $properties as $name => $value ) {
610 $entry['name'] = $name;
611 ApiResult
::setContent( $entry, $value );
618 private function formatCss( $css ) {
620 foreach ( $css as $file => $link ) {
622 $entry['file'] = $file;
623 ApiResult
::setContent( $entry, $link );
630 private function formatLimitReportData( $limitReportData ) {
632 $apiResult = $this->getResult();
634 foreach ( $limitReportData as $name => $value ) {
636 $entry['name'] = $name;
637 if ( !is_array( $value ) ) {
638 $value = array( $value );
640 $apiResult->setIndexedTagName( $value, 'param' );
641 $apiResult->setIndexedTagName_recursive( $value, 'param' );
642 $entry = array_merge( $entry, $value );
649 private function setIndexedTagNames( &$array, $mapping ) {
650 foreach ( $mapping as $key => $name ) {
651 if ( isset( $array[$key] ) ) {
652 $this->getResult()->setIndexedTagName( $array[$key], $name );
657 public function getAllowedParams() {
664 ApiBase
::PARAM_TYPE
=> 'integer',
666 'redirects' => false,
668 ApiBase
::PARAM_TYPE
=> 'integer',
671 ApiBase
::PARAM_DFLT
=> 'text|langlinks|categories|links|templates|' .
672 'images|externallinks|sections|revid|displaytitle|iwlinks|properties',
673 ApiBase
::PARAM_ISMULTI
=> true,
674 ApiBase
::PARAM_TYPE
=> array(
699 'effectivelanglinks' => false,
701 'disablepp' => false,
702 'disableeditsection' => false,
703 'generatexml' => array(
704 ApiBase
::PARAM_DFLT
=> false,
705 ApiBase
::PARAM_HELP_MSG
=> array(
706 'apihelp-parse-param-generatexml', CONTENT_MODEL_WIKITEXT
710 'sectionpreview' => false,
711 'disabletoc' => false,
712 'contentformat' => array(
713 ApiBase
::PARAM_TYPE
=> ContentHandler
::getAllContentFormats(),
715 'contentmodel' => array(
716 ApiBase
::PARAM_TYPE
=> ContentHandler
::getContentModels(),
721 protected function getExamplesMessages() {
723 'action=parse&page=Project:Sandbox'
724 => 'apihelp-parse-example-page',
725 'action=parse&text={{Project:Sandbox}}&contentmodel=wikitext'
726 => 'apihelp-parse-example-text',
727 'action=parse&text={{PAGENAME}}&title=Test'
728 => 'apihelp-parse-example-texttitle',
729 'action=parse&summary=Some+[[link]]&prop='
730 => 'apihelp-parse-example-summary',
734 public function getHelpUrls() {
735 return 'https://www.mediawiki.org/wiki/API:Parsing_wikitext#parse';