3 * Copyright © 2007 Yuri Astrakhan "<Firstname><Lastname>@gmail.com"
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
15 * You should have received a copy of the GNU General Public License along
16 * with this program; if not, write to the Free Software Foundation, Inc.,
17 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
18 * http://www.gnu.org/copyleft/gpl.html
23 namespace MediaWiki\Api
;
27 use MediaWiki\Search\TitleMatcher
;
28 use MediaWiki\Status\Status
;
30 use SearchEngineConfig
;
31 use SearchEngineFactory
;
33 use Wikimedia\ParamValidator\ParamValidator
;
34 use Wikimedia\ParamValidator\TypeDef\EnumDef
;
37 * Query module to perform full text search within wiki titles and content
41 class ApiQuerySearch
extends ApiQueryGeneratorBase
{
42 use \MediaWiki\Api\SearchApi
;
44 private TitleMatcher
$titleMatcher;
46 public function __construct(
49 SearchEngineConfig
$searchEngineConfig,
50 SearchEngineFactory
$searchEngineFactory,
51 TitleMatcher
$titleMatcher
53 parent
::__construct( $query, $moduleName, 'sr' );
54 // Services also needed in SearchApi trait
55 $this->searchEngineConfig
= $searchEngineConfig;
56 $this->searchEngineFactory
= $searchEngineFactory;
57 $this->titleMatcher
= $titleMatcher;
60 public function execute() {
64 public function executeGenerator( $resultPageSet ) {
65 $this->run( $resultPageSet );
69 * @param ApiPageSet|null $resultPageSet
72 private function run( $resultPageSet = null ) {
73 $params = $this->extractRequestParams();
76 $query = $params['search'];
77 $what = $params['what'];
78 $interwiki = $params['interwiki'];
79 $searchInfo = array_fill_keys( $params['info'], true );
80 $prop = array_fill_keys( $params['prop'], true );
82 // Create search engine instance and set options
83 $search = $this->buildSearchEngine( $params );
84 if ( isset( $params['sort'] ) ) {
85 $search->setSort( $params['sort'] );
87 $search->setFeatureData( 'rewrite', (bool)$params['enablerewrites'] );
88 $search->setFeatureData( 'interwiki', (bool)$interwiki );
89 // Hint to some SearchEngines about what snippets we would like returned
90 $search->setFeatureData( 'snippets', $this->decideSnippets( $prop ) );
92 $nquery = $search->replacePrefixes( $query );
93 if ( $nquery !== $query ) {
95 wfDeprecatedMsg( 'SearchEngine::replacePrefixes() is overridden by ' .
96 get_class( $search ) . ', this was deprecated in MediaWiki 1.32',
99 // Perform the actual search
100 if ( $what == 'text' ) {
101 $matches = $search->searchText( $query );
102 } elseif ( $what == 'title' ) {
103 $matches = $search->searchTitle( $query );
104 } elseif ( $what == 'nearmatch' ) {
105 // near matches must receive the user input as provided, otherwise
106 // the near matches within namespaces are lost.
107 $matches = $this->titleMatcher
->getNearMatchResultSet( $params['search'] );
109 // We default to title searches; this is a terrible legacy
110 // of the way we initially set up the MySQL fulltext-based
111 // search engine with separate title and text fields.
112 // In the future, the default should be for a combined index.
114 $matches = $search->searchTitle( $query );
116 // Not all search engines support a separate title search,
117 // for instance the Lucene-based engine we use on Wikipedia.
118 // In this case, fall back to full-text search (which will
119 // include titles in it!)
120 if ( $matches === null ) {
122 $matches = $search->searchText( $query );
126 if ( $matches instanceof Status
) {
128 $matches = $status->getValue();
134 if ( $status->isOK() ) {
135 $this->getMain()->getErrorFormatter()->addMessagesFromStatus(
136 $this->getModuleName(),
140 $this->dieStatus( $status );
142 } elseif ( $matches === null ) {
143 $this->dieWithError( [ 'apierror-searchdisabled', $what ], "search-{$what}-disabled" );
146 $apiResult = $this->getResult();
147 // Add search meta data to result
148 if ( isset( $searchInfo['totalhits'] ) ) {
149 $totalhits = $matches->getTotalHits();
150 if ( $totalhits !== null ) {
151 $apiResult->addValue( [ 'query', 'searchinfo' ],
152 'totalhits', $totalhits );
153 if ( $matches->isApproximateTotalHits() ) {
154 $apiResult->addValue( [ 'query', 'searchinfo' ],
155 'approximate_totalhits', $matches->isApproximateTotalHits() );
159 if ( isset( $searchInfo['suggestion'] ) && $matches->hasSuggestion() ) {
160 $apiResult->addValue( [ 'query', 'searchinfo' ],
161 'suggestion', $matches->getSuggestionQuery() );
162 $apiResult->addValue( [ 'query', 'searchinfo' ],
163 'suggestionsnippet', HtmlArmor
::getHtml( $matches->getSuggestionSnippet() ) );
165 if ( isset( $searchInfo['rewrittenquery'] ) && $matches->hasRewrittenQuery() ) {
166 $apiResult->addValue( [ 'query', 'searchinfo' ],
167 'rewrittenquery', $matches->getQueryAfterRewrite() );
168 $apiResult->addValue( [ 'query', 'searchinfo' ],
169 'rewrittenquerysnippet', HtmlArmor
::getHtml( $matches->getQueryAfterRewriteSnippet() ) );
176 if ( $matches->hasMoreResults() ) {
177 $this->setContinueEnumParameter( 'offset', $params['offset'] +
$params['limit'] );
180 foreach ( $matches as $result ) {
182 // Silently skip broken and missing titles
183 if ( $result->isBrokenTitle() ||
$result->isMissingRevision() ) {
187 $vals = $this->getSearchResultData( $result, $prop );
189 if ( $resultPageSet === null ) {
191 // Add item to results and see whether it fits
192 $fit = $apiResult->addValue( [ 'query', $this->getModuleName() ], null, $vals );
194 $this->setContinueEnumParameter( 'offset', $params['offset'] +
$count - 1 );
199 $titles[] = $result->getTitle();
200 $data[] = $vals ?
: [];
204 // Here we assume interwiki results do not count with
205 // regular search results. We may want to reconsider this
206 // if we ever return a lot of interwiki results or want pagination
208 // Interwiki results inside main result set
209 $canAddInterwiki = (bool)$params['enablerewrites'] && ( $resultPageSet === null );
210 if ( $canAddInterwiki ) {
211 $this->addInterwikiResults( $matches, $apiResult, $prop, 'additional',
212 ISearchResultSet
::INLINE_RESULTS
);
215 // Interwiki results outside main result set
216 if ( $interwiki && $resultPageSet === null ) {
217 $this->addInterwikiResults( $matches, $apiResult, $prop, 'interwiki',
218 ISearchResultSet
::SECONDARY_RESULTS
);
221 if ( $resultPageSet === null ) {
222 $apiResult->addIndexedTagName( [
223 'query', $this->getModuleName()
226 $resultPageSet->setRedirectMergePolicy( static function ( $current, $new ) {
227 if ( !isset( $current['index'] ) ||
$new['index'] < $current['index'] ) {
228 $current['index'] = $new['index'];
232 $resultPageSet->populateFromTitles( $titles );
233 $offset = $params['offset'] +
1;
234 foreach ( $titles as $index => $title ) {
235 $resultPageSet->setGeneratorData(
237 $data[ $index ] +
[ 'index' => $index +
$offset ]
244 * Assemble search result data.
245 * @param SearchResult $result Search result
246 * @param array $prop Props to extract (as keys)
247 * @return array|null Result data or null if result is broken in some way.
249 private function getSearchResultData( SearchResult
$result, $prop ) {
250 // Silently skip broken and missing titles
251 if ( $result->isBrokenTitle() ||
$result->isMissingRevision() ) {
257 $title = $result->getTitle();
258 ApiQueryBase
::addTitleInfo( $vals, $title );
259 $vals['pageid'] = $title->getArticleID();
261 if ( isset( $prop['size'] ) ) {
262 $vals['size'] = $result->getByteSize();
264 if ( isset( $prop['wordcount'] ) ) {
265 $vals['wordcount'] = $result->getWordCount();
267 if ( isset( $prop['snippet'] ) ) {
268 $vals['snippet'] = $result->getTextSnippet();
270 if ( isset( $prop['timestamp'] ) ) {
271 $vals['timestamp'] = wfTimestamp( TS_ISO_8601
, $result->getTimestamp() );
273 if ( isset( $prop['titlesnippet'] ) ) {
274 $vals['titlesnippet'] = $result->getTitleSnippet();
276 if ( isset( $prop['categorysnippet'] ) ) {
277 $vals['categorysnippet'] = $result->getCategorySnippet();
279 if ( $result->getRedirectTitle() !== null ) {
280 if ( isset( $prop['redirecttitle'] ) ) {
281 $vals['redirecttitle'] = $result->getRedirectTitle()->getPrefixedText();
283 if ( isset( $prop['redirectsnippet'] ) ) {
284 $vals['redirectsnippet'] = $result->getRedirectSnippet();
287 if ( $result->getSectionTitle() !== null ) {
288 if ( isset( $prop['sectiontitle'] ) ) {
289 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
291 if ( isset( $prop['sectionsnippet'] ) ) {
292 $vals['sectionsnippet'] = $result->getSectionSnippet();
295 if ( isset( $prop['isfilematch'] ) ) {
296 $vals['isfilematch'] = $result->isFileMatch();
299 if ( isset( $prop['extensiondata'] ) ) {
300 $extra = $result->getExtensionData();
301 // Add augmented data to the result. The data would be organized as a map:
302 // augmentorName => data
304 $vals['extensiondata'] = ApiResult
::addMetadataToResultVars( $extra );
312 * Add interwiki results as a section in query results.
313 * @param ISearchResultSet $matches
314 * @param ApiResult $apiResult
315 * @param array $prop Props to extract (as keys)
316 * @param string $section Section name where results would go
317 * @param int $type Interwiki result type
318 * @return int|null Number of total hits in the data or null if none was produced
320 private function addInterwikiResults(
321 ISearchResultSet
$matches, ApiResult
$apiResult, $prop,
325 $approximateTotalHits = false;
326 if ( $matches->hasInterwikiResults( $type ) ) {
327 foreach ( $matches->getInterwikiResults( $type ) as $interwikiMatches ) {
328 // Include number of results if requested
329 $interwikiTotalHits = $interwikiMatches->getTotalHits();
330 if ( $interwikiTotalHits !== null ) {
331 $totalhits +
= $interwikiTotalHits;
332 $approximateTotalHits = $approximateTotalHits ||
$interwikiMatches->isApproximateTotalHits();
335 foreach ( $interwikiMatches as $result ) {
336 $title = $result->getTitle();
337 $vals = $this->getSearchResultData( $result, $prop );
339 $vals['namespace'] = $result->getInterwikiNamespaceText();
340 $vals['title'] = $title->getText();
341 $vals['url'] = $title->getFullURL();
343 // Add item to results and see whether it fits
344 $fit = $apiResult->addValue( [
346 $section . $this->getModuleName(),
347 $result->getInterwikiPrefix()
351 // We hit the limit. We can't really provide any meaningful
352 // pagination info so just bail out
357 if ( $totalhits !== null ) {
358 $apiResult->addValue( [ 'query', $section . 'searchinfo' ], 'totalhits', $totalhits );
359 if ( $approximateTotalHits ) {
360 $apiResult->addValue( [ 'query', $section . 'searchinfo' ], 'approximate_totalhits', true );
362 $apiResult->addIndexedTagName( [
363 'query', $section . $this->getModuleName()
370 private function decideSnippets( array $prop ): array {
371 // Field names align with definitions in ContentHandler::getFieldsForSearchIndex.
372 // Except `redirect` which isn't explicitly created, but refers to the title of
373 // pages that redirect to the result page.
375 if ( isset( $prop['titlesnippet'] ) ) {
378 // checking snippet and title variants is a bit special cased, but some search
379 // engines generate the title variant from the snippet and thus must have the
380 // snippet requested to provide the title.
381 if ( isset( $prop['redirectsnippet'] ) ||
isset( $prop['redirecttitle'] ) ) {
382 $fields[] = 'redirect';
384 if ( isset( $prop['categorysnippet'] ) ) {
385 $fields[] = 'category';
387 if ( isset( $prop['sectionsnippet'] ) ||
isset( $prop['sectiontitle'] ) ) {
388 $fields[] = 'heading';
393 public function getCacheMode( $params ) {
397 public function getAllowedParams() {
398 $allowedParams = $this->buildCommonApiParams() +
[
400 ParamValidator
::PARAM_TYPE
=> [
407 ParamValidator
::PARAM_DEFAULT
=> 'totalhits|suggestion|rewrittenquery',
408 ParamValidator
::PARAM_TYPE
=> [
413 ParamValidator
::PARAM_ISMULTI
=> true,
416 ParamValidator
::PARAM_DEFAULT
=> 'size|wordcount|timestamp|snippet',
417 ParamValidator
::PARAM_TYPE
=> [
429 'score', // deprecated
430 'hasrelated', // deprecated
433 ParamValidator
::PARAM_ISMULTI
=> true,
434 ApiBase
::PARAM_HELP_MSG_PER_VALUE
=> [],
435 EnumDef
::PARAM_DEPRECATED_VALUES
=> [
440 'interwiki' => false,
441 'enablerewrites' => false,
444 // Generators only add info/properties if explicitly requested. T263841
445 if ( $this->isInGeneratorMode() ) {
446 $allowedParams['prop'][ParamValidator
::PARAM_DEFAULT
] = '';
447 $allowedParams['info'][ParamValidator
::PARAM_DEFAULT
] = '';
450 // If we have more than one engine the list of available sorts is
451 // difficult to represent. For now don't expose it.
452 $alternatives = $this->searchEngineConfig
->getSearchTypes();
453 if ( count( $alternatives ) == 1 ) {
454 $allowedParams['sort'] = [
455 ParamValidator
::PARAM_DEFAULT
=> SearchEngine
::DEFAULT_SORT
,
456 ParamValidator
::PARAM_TYPE
=> $this->searchEngineFactory
->create()->getValidSorts(),
460 return $allowedParams;
463 public function getSearchProfileParams() {
466 'profile-type' => SearchEngine
::FT_QUERY_INDEP_PROFILE_TYPE
,
467 'help-message' => 'apihelp-query+search-param-qiprofile',
472 protected function getExamplesMessages() {
474 'action=query&list=search&srsearch=meaning'
475 => 'apihelp-query+search-example-simple',
476 'action=query&list=search&srwhat=text&srsearch=meaning'
477 => 'apihelp-query+search-example-text',
478 'action=query&generator=search&gsrsearch=meaning&prop=info'
479 => 'apihelp-query+search-example-generator',
483 public function getHelpUrls() {
484 return 'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Search';
488 /** @deprecated class alias since 1.43 */
489 class_alias( ApiQuerySearch
::class, 'ApiQuerySearch' );