avoid getNativeData()
[lhc/web/wiklou.git] / includes / api / ApiParse.php
1 <?php
2 /**
3 * Created on Dec 01, 2007
4 *
5 * Copyright © 2007 Yuri Astrakhan <Firstname><Lastname>@gmail.com
6 *
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.
11 *
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.
16 *
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
21 *
22 * @file
23 */
24
25 /**
26 * @ingroup API
27 */
28 class ApiParse extends ApiBase {
29 private $section, $text, $pstText = null;
30
31 public function __construct( $main, $action ) {
32 parent::__construct( $main, $action );
33 }
34
35 public function execute() {
36 // The data is hot but user-dependent, like page views, so we set vary cookies
37 $this->getMain()->setCacheMode( 'anon-public-user-private' );
38
39 // Get parameters
40 $params = $this->extractRequestParams();
41 $text = $params['text'];
42 $title = $params['title'];
43 $page = $params['page'];
44 $pageid = $params['pageid'];
45 $oldid = $params['oldid'];
46
47 if ( !is_null( $page ) && ( !is_null( $text ) || $title != 'API' ) ) {
48 $this->dieUsage( 'The page parameter cannot be used together with the text and title parameters', 'params' );
49 }
50
51 $prop = array_flip( $params['prop'] );
52
53 if ( isset( $params['section'] ) ) {
54 $this->section = $params['section'];
55 } else {
56 $this->section = false;
57 }
58
59 // The parser needs $wgTitle to be set, apparently the
60 // $title parameter in Parser::parse isn't enough *sigh*
61 // TODO: Does this still need $wgTitle?
62 global $wgParser, $wgTitle, $wgLang;
63
64 // Currently unnecessary, code to act as a safeguard against any change in current behaviour of uselang breaks
65 $oldLang = null;
66 if ( isset( $params['uselang'] ) && $params['uselang'] != $wgLang->getCode() ) {
67 $oldLang = $wgLang; // Backup wgLang
68 $wgLang = Language::factory( $params['uselang'] );
69 }
70
71 $popts = ParserOptions::newFromContext( $this->getContext() );
72 $popts->setTidy( true );
73 $popts->enableLimitReport( !$params['disablepp'] );
74
75 $redirValues = null;
76
77 // Return result
78 $result = $this->getResult();
79
80 if ( !is_null( $oldid ) || !is_null( $pageid ) || !is_null( $page ) ) {
81 if ( !is_null( $oldid ) ) {
82 // Don't use the parser cache
83 $rev = Revision::newFromID( $oldid );
84 if ( !$rev ) {
85 $this->dieUsage( "There is no revision ID $oldid", 'missingrev' );
86 }
87 if ( !$rev->userCan( Revision::DELETED_TEXT, $this->getUser() ) ) {
88 $this->dieUsage( "You don't have permission to view deleted revisions", 'permissiondenied' );
89 }
90
91 $titleObj = $rev->getTitle();
92
93 $wgTitle = $titleObj;
94
95 // If for some reason the "oldid" is actually the current revision, it may be cached
96 if ( $titleObj->getLatestRevID() === intval( $oldid ) ) {
97 // May get from/save to parser cache
98 $p_result = $this->getParsedSectionOrText( $titleObj, $popts, $pageid,
99 isset( $prop['wikitext'] ) ) ;
100 } else { // This is an old revision, so get the text differently
101 $this->text = $rev->getText( Revision::FOR_THIS_USER, $this->getUser() );
102
103 if ( $this->section !== false ) {
104 $this->text = $this->getSectionText( $this->text, 'r' . $rev->getId() );
105 }
106
107 // Should we save old revision parses to the parser cache?
108 $p_result = $wgParser->parse( $this->text, $titleObj, $popts );
109 }
110 } else { // Not $oldid, but $pageid or $page
111 if ( $params['redirects'] ) {
112 $reqParams = array(
113 'action' => 'query',
114 'redirects' => '',
115 );
116 if ( !is_null ( $pageid ) ) {
117 $reqParams['pageids'] = $pageid;
118 } else { // $page
119 $reqParams['titles'] = $page;
120 }
121 $req = new FauxRequest( $reqParams );
122 $main = new ApiMain( $req );
123 $main->execute();
124 $data = $main->getResultData();
125 $redirValues = isset( $data['query']['redirects'] )
126 ? $data['query']['redirects']
127 : array();
128 $to = $page;
129 foreach ( (array)$redirValues as $r ) {
130 $to = $r['to'];
131 }
132 $titleObj = Title::newFromText( $to );
133 } else {
134 if ( !is_null ( $pageid ) ) {
135 $reqParams['pageids'] = $pageid;
136 $titleObj = Title::newFromID( $pageid );
137 } else { // $page
138 $to = $page;
139 $titleObj = Title::newFromText( $to );
140 }
141 }
142 if ( !is_null ( $pageid ) ) {
143 if ( !$titleObj ) {
144 // Still throw nosuchpageid error if pageid was provided
145 $this->dieUsageMsg( array( 'nosuchpageid', $pageid ) );
146 }
147 } elseif ( !$titleObj || !$titleObj->exists() ) {
148 $this->dieUsage( "The page you specified doesn't exist", 'missingtitle' );
149 }
150 $wgTitle = $titleObj;
151
152 if ( isset( $prop['revid'] ) ) {
153 $oldid = $titleObj->getLatestRevID();
154 }
155
156 // Potentially cached
157 $p_result = $this->getParsedSectionOrText( $titleObj, $popts, $pageid,
158 isset( $prop['wikitext'] ) ) ;
159 }
160 } else { // Not $oldid, $pageid, $page. Hence based on $text
161
162 if ( is_null( $text ) ) {
163 $this->dieUsage( 'The text parameter should be passed with the title parameter. Should you be using the "page" parameter instead?', 'params' );
164 }
165 $this->text = $text;
166 $titleObj = Title::newFromText( $title );
167 if ( !$titleObj ) {
168 $this->dieUsageMsg( array( 'invalidtitle', $title ) );
169 }
170 $wgTitle = $titleObj;
171
172 if ( $this->section !== false ) {
173 $this->text = $this->getSectionText( $this->text, $titleObj->getText() );
174 }
175
176 if ( $params['pst'] || $params['onlypst'] ) {
177 $this->pstText = $wgParser->preSaveTransform( $this->text, $titleObj, $this->getUser(), $popts );
178 }
179 if ( $params['onlypst'] ) {
180 // Build a result and bail out
181 $result_array = array();
182 $result_array['text'] = array();
183 $result->setContent( $result_array['text'], $this->pstText );
184 if ( isset( $prop['wikitext'] ) ) {
185 $result_array['wikitext'] = array();
186 $result->setContent( $result_array['wikitext'], $this->text );
187 }
188 $result->addValue( null, $this->getModuleName(), $result_array );
189 return;
190 }
191 // Not cached (save or load)
192 $p_result = $wgParser->parse( $params['pst'] ? $this->pstText : $this->text, $titleObj, $popts );
193 }
194
195 $result_array = array();
196
197 $result_array['title'] = $titleObj->getPrefixedText();
198
199 if ( !is_null( $oldid ) ) {
200 $result_array['revid'] = intval( $oldid );
201 }
202
203 if ( $params['redirects'] && !is_null( $redirValues ) ) {
204 $result_array['redirects'] = $redirValues;
205 }
206
207 if ( isset( $prop['text'] ) ) {
208 $result_array['text'] = array();
209 $result->setContent( $result_array['text'], $p_result->getText() );
210 }
211
212 if ( !is_null( $params['summary'] ) ) {
213 $result_array['parsedsummary'] = array();
214 $result->setContent( $result_array['parsedsummary'], Linker::formatComment( $params['summary'], $titleObj ) );
215 }
216
217 if ( isset( $prop['langlinks'] ) ) {
218 $result_array['langlinks'] = $this->formatLangLinks( $p_result->getLanguageLinks() );
219 }
220 if ( isset( $prop['languageshtml'] ) ) {
221 $languagesHtml = $this->languagesHtml( $p_result->getLanguageLinks() );
222 $result_array['languageshtml'] = array();
223 $result->setContent( $result_array['languageshtml'], $languagesHtml );
224 }
225 if ( isset( $prop['categories'] ) ) {
226 $result_array['categories'] = $this->formatCategoryLinks( $p_result->getCategories() );
227 }
228 if ( isset( $prop['categorieshtml'] ) ) {
229 $categoriesHtml = $this->categoriesHtml( $p_result->getCategories() );
230 $result_array['categorieshtml'] = array();
231 $result->setContent( $result_array['categorieshtml'], $categoriesHtml );
232 }
233 if ( isset( $prop['links'] ) ) {
234 $result_array['links'] = $this->formatLinks( $p_result->getLinks() );
235 }
236 if ( isset( $prop['templates'] ) ) {
237 $result_array['templates'] = $this->formatLinks( $p_result->getTemplates() );
238 }
239 if ( isset( $prop['images'] ) ) {
240 $result_array['images'] = array_keys( $p_result->getImages() );
241 }
242 if ( isset( $prop['externallinks'] ) ) {
243 $result_array['externallinks'] = array_keys( $p_result->getExternalLinks() );
244 }
245 if ( isset( $prop['sections'] ) ) {
246 $result_array['sections'] = $p_result->getSections();
247 }
248
249 if ( isset( $prop['displaytitle'] ) ) {
250 $result_array['displaytitle'] = $p_result->getDisplayTitle() ?
251 $p_result->getDisplayTitle() :
252 $titleObj->getPrefixedText();
253 }
254
255 if ( isset( $prop['headitems'] ) || isset( $prop['headhtml'] ) ) {
256 $context = $this->getContext();
257 $context->setTitle( $titleObj );
258 $context->getOutput()->addParserOutputNoText( $p_result );
259
260 if ( isset( $prop['headitems'] ) ) {
261 $headItems = $this->formatHeadItems( $p_result->getHeadItems() );
262
263 $css = $this->formatCss( $context->getOutput()->buildCssLinksArray() );
264
265 $scripts = array( $context->getOutput()->getHeadScripts() );
266
267 $result_array['headitems'] = array_merge( $headItems, $css, $scripts );
268 }
269
270 if ( isset( $prop['headhtml'] ) ) {
271 $result_array['headhtml'] = array();
272 $result->setContent( $result_array['headhtml'], $context->getOutput()->headElement( $context->getSkin() ) );
273 }
274 }
275
276 if ( isset( $prop['iwlinks'] ) ) {
277 $result_array['iwlinks'] = $this->formatIWLinks( $p_result->getInterwikiLinks() );
278 }
279
280 if ( isset( $prop['wikitext'] ) ) {
281 $result_array['wikitext'] = array();
282 $result->setContent( $result_array['wikitext'], $this->text );
283 if ( !is_null( $this->pstText ) ) {
284 $result_array['psttext'] = array();
285 $result->setContent( $result_array['psttext'], $this->pstText );
286 }
287 }
288
289 $result_mapping = array(
290 'redirects' => 'r',
291 'langlinks' => 'll',
292 'categories' => 'cl',
293 'links' => 'pl',
294 'templates' => 'tl',
295 'images' => 'img',
296 'externallinks' => 'el',
297 'iwlinks' => 'iw',
298 'sections' => 's',
299 'headitems' => 'hi',
300 );
301 $this->setIndexedTagNames( $result_array, $result_mapping );
302 $result->addValue( null, $this->getModuleName(), $result_array );
303
304 if ( !is_null( $oldLang ) ) {
305 $wgLang = $oldLang; // Reset $wgLang to $oldLang
306 }
307 }
308
309 /**
310 * @param $titleObj Title
311 * @param $popts ParserOptions
312 * @param $pageId Int
313 * @param $getWikitext Bool
314 * @return ParserOutput
315 */
316 private function getParsedSectionOrText( $titleObj, $popts, $pageId = null, $getWikitext = false ) {
317 global $wgParser;
318
319 $page = WikiPage::factory( $titleObj );
320
321 if ( $this->section !== false ) { #FIXME: get section Content, get parser output, ...
322 $this->text = $this->getSectionText( $page->getRawText(), !is_null( $pageId )
323 ? 'page id ' . $pageId : $titleObj->getText() ); #FIXME: get section...
324
325 // Not cached (save or load)
326 return $wgParser->parse( $this->text, $titleObj, $popts );
327 } else {
328 // Try the parser cache first
329 // getParserOutput will save to Parser cache if able
330 $pout = $page->getParserOutput( $popts );
331 if ( $getWikitext ) {
332 $this->content = $page->getContent( Revision::RAW ); #FIXME: use $this->content everywhere
333 $this->text = ContentHandler::getContentText( $this->content ); #FIXME: serialize, get format from params; or use object structure in result?
334 }
335 return $pout;
336 }
337 }
338
339 private function getSectionText( $text, $what ) { #FIXME: replace with Content::getSection
340 global $wgParser;
341 // Not cached (save or load)
342 $text = $wgParser->getSection( $text, $this->section, false );
343 if ( $text === false ) {
344 $this->dieUsage( "There is no section {$this->section} in " . $what, 'nosuchsection' );
345 }
346 return $text;
347 }
348
349 private function formatLangLinks( $links ) {
350 $result = array();
351 foreach ( $links as $link ) {
352 $entry = array();
353 $bits = explode( ':', $link, 2 );
354 $title = Title::newFromText( $link );
355
356 $entry['lang'] = $bits[0];
357 if ( $title ) {
358 $entry['url'] = wfExpandUrl( $title->getFullURL(), PROTO_CURRENT );
359 }
360 $this->getResult()->setContent( $entry, $bits[1] );
361 $result[] = $entry;
362 }
363 return $result;
364 }
365
366 private function formatCategoryLinks( $links ) {
367 $result = array();
368 foreach ( $links as $link => $sortkey ) {
369 $entry = array();
370 $entry['sortkey'] = $sortkey;
371 $this->getResult()->setContent( $entry, $link );
372 $result[] = $entry;
373 }
374 return $result;
375 }
376
377 private function categoriesHtml( $categories ) {
378 $context = $this->getContext();
379 $context->getOutput()->addCategoryLinks( $categories );
380 return $context->getSkin()->getCategories();
381 }
382
383 /**
384 * @deprecated since 1.18 No modern skin generates language links this way, please use language links
385 * data to generate your own HTML.
386 * @param $languages array
387 * @return string
388 */
389 private function languagesHtml( $languages ) {
390 wfDeprecated( __METHOD__, '1.18' );
391
392 global $wgContLang, $wgHideInterlanguageLinks;
393
394 if ( $wgHideInterlanguageLinks || count( $languages ) == 0 ) {
395 return '';
396 }
397
398 $s = htmlspecialchars( wfMsg( 'otherlanguages' ) . wfMsg( 'colon-separator' ) );
399
400 $langs = array();
401 foreach ( $languages as $l ) {
402 $nt = Title::newFromText( $l );
403 $text = $wgContLang->getLanguageName( $nt->getInterwiki() );
404
405 $langs[] = Html::element( 'a',
406 array( 'href' => $nt->getFullURL(), 'title' => $nt->getText(), 'class' => "external" ),
407 $text == '' ? $l : $text );
408 }
409
410 $s .= implode( htmlspecialchars( wfMsgExt( 'pipe-separator', 'escapenoentities' ) ), $langs );
411
412 if ( $wgContLang->isRTL() ) {
413 $s = Html::rawElement( 'span', array( 'dir' => "LTR" ), $s );
414 }
415
416 return $s;
417 }
418
419 private function formatLinks( $links ) {
420 $result = array();
421 foreach ( $links as $ns => $nslinks ) {
422 foreach ( $nslinks as $title => $id ) {
423 $entry = array();
424 $entry['ns'] = $ns;
425 $this->getResult()->setContent( $entry, Title::makeTitle( $ns, $title )->getFullText() );
426 if ( $id != 0 ) {
427 $entry['exists'] = '';
428 }
429 $result[] = $entry;
430 }
431 }
432 return $result;
433 }
434
435 private function formatIWLinks( $iw ) {
436 $result = array();
437 foreach ( $iw as $prefix => $titles ) {
438 foreach ( array_keys( $titles ) as $title ) {
439 $entry = array();
440 $entry['prefix'] = $prefix;
441
442 $title = Title::newFromText( "{$prefix}:{$title}" );
443 if ( $title ) {
444 $entry['url'] = wfExpandUrl( $title->getFullURL(), PROTO_CURRENT );
445 }
446
447 $this->getResult()->setContent( $entry, $title->getFullText() );
448 $result[] = $entry;
449 }
450 }
451 return $result;
452 }
453
454 private function formatHeadItems( $headItems ) {
455 $result = array();
456 foreach ( $headItems as $tag => $content ) {
457 $entry = array();
458 $entry['tag'] = $tag;
459 $this->getResult()->setContent( $entry, $content );
460 $result[] = $entry;
461 }
462 return $result;
463 }
464
465 private function formatCss( $css ) {
466 $result = array();
467 foreach ( $css as $file => $link ) {
468 $entry = array();
469 $entry['file'] = $file;
470 $this->getResult()->setContent( $entry, $link );
471 $result[] = $entry;
472 }
473 return $result;
474 }
475
476 private function setIndexedTagNames( &$array, $mapping ) {
477 foreach ( $mapping as $key => $name ) {
478 if ( isset( $array[$key] ) ) {
479 $this->getResult()->setIndexedTagName( $array[$key], $name );
480 }
481 }
482 }
483
484 public function getAllowedParams() {
485 return array(
486 'title' => array(
487 ApiBase::PARAM_DFLT => 'API',
488 ),
489 'text' => null,
490 'summary' => null,
491 'page' => null,
492 'pageid' => array(
493 ApiBase::PARAM_TYPE => 'integer',
494 ),
495 'redirects' => false,
496 'oldid' => array(
497 ApiBase::PARAM_TYPE => 'integer',
498 ),
499 'prop' => array(
500 ApiBase::PARAM_DFLT => 'text|langlinks|categories|links|templates|images|externallinks|sections|revid|displaytitle',
501 ApiBase::PARAM_ISMULTI => true,
502 ApiBase::PARAM_TYPE => array(
503 'text',
504 'langlinks',
505 'languageshtml',
506 'categories',
507 'categorieshtml',
508 'links',
509 'templates',
510 'images',
511 'externallinks',
512 'sections',
513 'revid',
514 'displaytitle',
515 'headitems',
516 'headhtml',
517 'iwlinks',
518 'wikitext',
519 )
520 ),
521 'pst' => false,
522 'onlypst' => false,
523 'uselang' => null,
524 'section' => null,
525 'disablepp' => false,
526 );
527 }
528
529 public function getParamDescription() {
530 $p = $this->getModulePrefix();
531 return array(
532 'text' => 'Wikitext to parse',
533 'summary' => 'Summary to parse',
534 'redirects' => "If the {$p}page or the {$p}pageid parameter is set to a redirect, resolve it",
535 'title' => 'Title of page the text belongs to',
536 'page' => "Parse the content of this page. Cannot be used together with {$p}text and {$p}title",
537 'pageid' => "Parse the content of this page. Overrides {$p}page",
538 'oldid' => "Parse the content of this revision. Overrides {$p}page and {$p}pageid",
539 'prop' => array(
540 'Which pieces of information to get',
541 ' text - Gives the parsed text of the wikitext',
542 ' langlinks - Gives the language links in the parsed wikitext',
543 ' categories - Gives the categories in the parsed wikitext',
544 ' categorieshtml - Gives the HTML version of the categories',
545 ' languageshtml - Gives the HTML version of the language links',
546 ' links - Gives the internal links in the parsed wikitext',
547 ' templates - Gives the templates in the parsed wikitext',
548 ' images - Gives the images in the parsed wikitext',
549 ' externallinks - Gives the external links in the parsed wikitext',
550 ' sections - Gives the sections in the parsed wikitext',
551 ' revid - Adds the revision ID of the parsed page',
552 ' displaytitle - Adds the title of the parsed wikitext',
553 ' headitems - Gives items to put in the <head> of the page',
554 ' headhtml - Gives parsed <head> of the page',
555 ' iwlinks - Gives interwiki links in the parsed wikitext',
556 ' wikitext - Gives the original wikitext that was parsed',
557 ),
558 'pst' => array(
559 'Do a pre-save transform on the input before parsing it',
560 'Ignored if page, pageid or oldid is used'
561 ),
562 'onlypst' => array(
563 'Do a pre-save transform (PST) on the input, but don\'t parse it',
564 'Returns the same wikitext, after a PST has been applied. Ignored if page, pageid or oldid is used'
565 ),
566 'uselang' => 'Which language to parse the request in',
567 'section' => 'Only retrieve the content of this section number',
568 'disablepp' => 'Disable the PP Report from the parser output',
569 );
570 }
571
572 public function getDescription() {
573 return 'Parses wikitext and returns parser output';
574 }
575
576 public function getPossibleErrors() {
577 return array_merge( parent::getPossibleErrors(), array(
578 array( 'code' => 'params', 'info' => 'The page parameter cannot be used together with the text and title parameters' ),
579 array( 'code' => 'params', 'info' => 'The text parameter should be passed with the title parameter. Should you be using the "page" parameter instead?' ),
580 array( 'code' => 'missingrev', 'info' => 'There is no revision ID oldid' ),
581 array( 'code' => 'permissiondenied', 'info' => 'You don\'t have permission to view deleted revisions' ),
582 array( 'code' => 'missingtitle', 'info' => 'The page you specified doesn\'t exist' ),
583 array( 'code' => 'nosuchsection', 'info' => 'There is no section sectionnumber in page' ),
584 array( 'nosuchpageid' ),
585 array( 'invalidtitle', 'title' ),
586 ) );
587 }
588
589 public function getExamples() {
590 return array(
591 'api.php?action=parse&text={{Project:Sandbox}}'
592 );
593 }
594
595 public function getHelpUrls() {
596 return 'https://www.mediawiki.org/wiki/API:Parsing_wikitext#parse';
597 }
598
599 public function getVersion() {
600 return __CLASS__ . ': $Id$';
601 }
602 }