includes/api/ApiQuerySearch.php

   1 <?php
   2 /**
   3  * Copyright © 2007 Yuri Astrakhan "<Firstname><Lastname>@gmail.com"
   4  *
   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.
   9  *
  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.
  14  *
  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
  19  *
  20  * @file
  21  */
  22
  23 /**
  24  * Query module to perform full text search within wiki titles and content
  25  *
  26  * @ingroup API
  27  */
  28 class ApiQuerySearch extends ApiQueryGeneratorBase {
  29         use SearchApi;
  30
  31         /** @var array list of api allowed params */
  32         private $allowedParams;
  33
  34         public function __construct( ApiQuery $query, $moduleName ) {
  35                 parent::__construct( $query, $moduleName, 'sr' );
  36         }
  37
  38         public function execute() {
  39                 $this->run();
  40         }
  41
  42         public function executeGenerator( $resultPageSet ) {
  43                 $this->run( $resultPageSet );
  44         }
  45
  46         /**
  47          * @param ApiPageSet $resultPageSet
  48          * @return void
  49          */
  50         private function run( $resultPageSet = null ) {
  51                 global $wgContLang;
  52                 $params = $this->extractRequestParams();
  53
  54                 // Extract parameters
  55                 $query = $params['search'];
  56                 $what = $params['what'];
  57                 $interwiki = $params['interwiki'];
  58                 $searchInfo = array_flip( $params['info'] );
  59                 $prop = array_flip( $params['prop'] );
  60
  61                 // Create search engine instance and set options
  62                 $search = $this->buildSearchEngine( $params );
  63                 if ( isset( $params['sort'] ) ) {
  64                         $search->setSort( $params['sort'] );
  65                 }
  66                 $search->setFeatureData( 'rewrite', (bool)$params['enablerewrites'] );
  67                 $search->setFeatureData( 'interwiki', (bool)$interwiki );
  68
  69                 $query = $search->transformSearchTerm( $query );
  70                 $query = $search->replacePrefixes( $query );
  71
  72                 // Perform the actual search
  73                 if ( $what == 'text' ) {
  74                         $matches = $search->searchText( $query );
  75                 } elseif ( $what == 'title' ) {
  76                         $matches = $search->searchTitle( $query );
  77                 } elseif ( $what == 'nearmatch' ) {
  78                         // near matches must receive the user input as provided, otherwise
  79                         // the near matches within namespaces are lost.
  80                         $matches = $search->getNearMatcher( $this->getConfig() )
  81                                 ->getNearMatchResultSet( $params['search'] );
  82                 } else {
  83                         // We default to title searches; this is a terrible legacy
  84                         // of the way we initially set up the MySQL fulltext-based
  85                         // search engine with separate title and text fields.
  86                         // In the future, the default should be for a combined index.
  87                         $what = 'title';
  88                         $matches = $search->searchTitle( $query );
  89
  90                         // Not all search engines support a separate title search,
  91                         // for instance the Lucene-based engine we use on Wikipedia.
  92                         // In this case, fall back to full-text search (which will
  93                         // include titles in it!)
  94                         if ( is_null( $matches ) ) {
  95                                 $what = 'text';
  96                                 $matches = $search->searchText( $query );
  97                         }
  98                 }
  99
 100                 if ( $matches instanceof Status ) {
 101                         $status = $matches;
 102                         $matches = $status->getValue();
 103                 } else {
 104                         $status = null;
 105                 }
 106
 107                 if ( $status ) {
 108                         if ( $status->isOK() ) {
 109                                 $this->getMain()->getErrorFormatter()->addMessagesFromStatus(
 110                                         $this->getModuleName(),
 111                                         $status
 112                                 );
 113                         } else {
 114                                 $this->dieStatus( $status );
 115                         }
 116                 } elseif ( is_null( $matches ) ) {
 117                         $this->dieWithError( [ 'apierror-searchdisabled', $what ], "search-{$what}-disabled" );
 118                 }
 119
 120                 if ( $resultPageSet === null ) {
 121                         $apiResult = $this->getResult();
 122                         // Add search meta data to result
 123                         if ( isset( $searchInfo['totalhits'] ) ) {
 124                                 $totalhits = $matches->getTotalHits();
 125                                 if ( $totalhits !== null ) {
 126                                         $apiResult->addValue( [ 'query', 'searchinfo' ],
 127                                                 'totalhits', $totalhits );
 128                                 }
 129                         }
 130                         if ( isset( $searchInfo['suggestion'] ) && $matches->hasSuggestion() ) {
 131                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 132                                         'suggestion', $matches->getSuggestionQuery() );
 133                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 134                                         'suggestionsnippet', $matches->getSuggestionSnippet() );
 135                         }
 136                         if ( isset( $searchInfo['rewrittenquery'] ) && $matches->hasRewrittenQuery() ) {
 137                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 138                                         'rewrittenquery', $matches->getQueryAfterRewrite() );
 139                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 140                                         'rewrittenquerysnippet', $matches->getQueryAfterRewriteSnippet() );
 141                         }
 142                 }
 143
 144                 // Add the search results to the result
 145                 $terms = $wgContLang->convertForSearchResult( $matches->termMatches() );
 146                 $titles = [];
 147                 $count = 0;
 148                 $limit = $params['limit'];
 149
 150                 if ( $matches->hasMoreResults() ) {
 151                         $this->setContinueEnumParameter( 'offset', $params['offset'] + $params['limit'] );
 152                 }
 153
 154                 foreach ( $matches as $result ) {
 155                         $count++;
 156                         // Silently skip broken and missing titles
 157                         if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 158                                 continue;
 159                         }
 160
 161                         if ( $resultPageSet === null ) {
 162                                 $vals = $this->getSearchResultData( $result, $prop, $terms );
 163                                 if ( $vals ) {
 164                                         // Add item to results and see whether it fits
 165                                         $fit = $apiResult->addValue( [ 'query', $this->getModuleName() ], null, $vals );
 166                                         if ( !$fit ) {
 167                                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $count - 1 );
 168                                                 break;
 169                                         }
 170                                 }
 171                         } else {
 172                                 $titles[] = $result->getTitle();
 173                         }
 174                 }
 175
 176                 // Here we assume interwiki results do not count with
 177                 // regular search results. We may want to reconsider this
 178                 // if we ever return a lot of interwiki results or want pagination
 179                 // for them.
 180                 // Interwiki results inside main result set
 181                 $canAddInterwiki = (bool)$params['enablerewrites'] && ( $resultPageSet === null );
 182                 if ( $canAddInterwiki ) {
 183                         $this->addInterwikiResults( $matches, $apiResult, $prop, $terms, 'additional',
 184                                 SearchResultSet::INLINE_RESULTS );
 185                 }
 186
 187                 // Interwiki results outside main result set
 188                 if ( $interwiki && $resultPageSet === null ) {
 189                         $this->addInterwikiResults( $matches, $apiResult, $prop, $terms, 'interwiki',
 190                                 SearchResultSet::SECONDARY_RESULTS );
 191                 }
 192
 193                 if ( $resultPageSet === null ) {
 194                         $apiResult->addIndexedTagName( [
 195                                 'query', $this->getModuleName()
 196                         ], 'p' );
 197                 } else {
 198                         $resultPageSet->setRedirectMergePolicy( function ( $current, $new ) {
 199                                 if ( !isset( $current['index'] ) || $new['index'] < $current['index'] ) {
 200                                         $current['index'] = $new['index'];
 201                                 }
 202                                 return $current;
 203                         } );
 204                         $resultPageSet->populateFromTitles( $titles );
 205                         $offset = $params['offset'] + 1;
 206                         foreach ( $titles as $index => $title ) {
 207                                 $resultPageSet->setGeneratorData( $title, [ 'index' => $index + $offset ] );
 208                         }
 209                 }
 210         }
 211
 212         /**
 213          * Assemble search result data.
 214          * @param SearchResult $result Search result
 215          * @param array        $prop Props to extract (as keys)
 216          * @param array        $terms Terms list
 217          * @return array|null Result data or null if result is broken in some way.
 218          */
 219         private function getSearchResultData( SearchResult $result, $prop, $terms ) {
 220                 // Silently skip broken and missing titles
 221                 if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 222                         return null;
 223                 }
 224
 225                 $vals = [];
 226
 227                 $title = $result->getTitle();
 228                 ApiQueryBase::addTitleInfo( $vals, $title );
 229                 $vals['pageid'] = $title->getArticleID();
 230
 231                 if ( isset( $prop['size'] ) ) {
 232                         $vals['size'] = $result->getByteSize();
 233                 }
 234                 if ( isset( $prop['wordcount'] ) ) {
 235                         $vals['wordcount'] = $result->getWordCount();
 236                 }
 237                 if ( isset( $prop['snippet'] ) ) {
 238                         $vals['snippet'] = $result->getTextSnippet( $terms );
 239                 }
 240                 if ( isset( $prop['timestamp'] ) ) {
 241                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $result->getTimestamp() );
 242                 }
 243                 if ( isset( $prop['titlesnippet'] ) ) {
 244                         $vals['titlesnippet'] = $result->getTitleSnippet();
 245                 }
 246                 if ( isset( $prop['categorysnippet'] ) ) {
 247                         $vals['categorysnippet'] = $result->getCategorySnippet();
 248                 }
 249                 if ( !is_null( $result->getRedirectTitle() ) ) {
 250                         if ( isset( $prop['redirecttitle'] ) ) {
 251                                 $vals['redirecttitle'] = $result->getRedirectTitle()->getPrefixedText();
 252                         }
 253                         if ( isset( $prop['redirectsnippet'] ) ) {
 254                                 $vals['redirectsnippet'] = $result->getRedirectSnippet();
 255                         }
 256                 }
 257                 if ( !is_null( $result->getSectionTitle() ) ) {
 258                         if ( isset( $prop['sectiontitle'] ) ) {
 259                                 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
 260                         }
 261                         if ( isset( $prop['sectionsnippet'] ) ) {
 262                                 $vals['sectionsnippet'] = $result->getSectionSnippet();
 263                         }
 264                 }
 265                 if ( isset( $prop['isfilematch'] ) ) {
 266                         $vals['isfilematch'] = $result->isFileMatch();
 267                 }
 268
 269                 if ( isset( $prop['extensiondata'] ) ) {
 270                         $extra = $result->getExtensionData();
 271                         // Add augmented data to the result. The data would be organized as a map:
 272                         // augmentorName => data
 273                         if ( $extra ) {
 274                                 $vals['extensiondata'] = ApiResult::addMetadataToResultVars( $extra );
 275                         }
 276                 }
 277
 278                 return $vals;
 279         }
 280
 281         /**
 282          * Add interwiki results as a section in query results.
 283          * @param SearchResultSet $matches
 284          * @param ApiResult       $apiResult
 285          * @param array           $prop Props to extract (as keys)
 286          * @param array           $terms Terms list
 287          * @param string          $section Section name where results would go
 288          * @param int             $type Interwiki result type
 289          * @return int|null Number of total hits in the data or null if none was produced
 290          */
 291         private function addInterwikiResults(
 292                 SearchResultSet $matches, ApiResult $apiResult, $prop,
 293                 $terms, $section, $type
 294         ) {
 295                 $totalhits = null;
 296                 if ( $matches->hasInterwikiResults( $type ) ) {
 297                         foreach ( $matches->getInterwikiResults( $type ) as $interwikiMatches ) {
 298                                 // Include number of results if requested
 299                                 $totalhits += $interwikiMatches->getTotalHits();
 300
 301                                 foreach ( $interwikiMatches as $result ) {
 302                                         $title = $result->getTitle();
 303                                         $vals = $this->getSearchResultData( $result, $prop, $terms );
 304
 305                                         $vals['namespace'] = $result->getInterwikiNamespaceText();
 306                                         $vals['title'] = $title->getText();
 307                                         $vals['url'] = $title->getFullURL();
 308
 309                                         // Add item to results and see whether it fits
 310                                         $fit = $apiResult->addValue( [
 311                                                         'query',
 312                                                         $section . $this->getModuleName(),
 313                                                         $result->getInterwikiPrefix()
 314                                                 ], null, $vals );
 315
 316                                         if ( !$fit ) {
 317                                                 // We hit the limit. We can't really provide any meaningful
 318                                                 // pagination info so just bail out
 319                                                 break;
 320                                         }
 321                                 }
 322                         }
 323                         if ( $totalhits !== null ) {
 324                                 $apiResult->addValue( [ 'query', $section . 'searchinfo' ], 'totalhits', $totalhits );
 325                                 $apiResult->addIndexedTagName( [
 326                                         'query', $section . $this->getModuleName()
 327                                 ], 'p' );
 328                         }
 329                 }
 330                 return $totalhits;
 331         }
 332
 333         public function getCacheMode( $params ) {
 334                 return 'public';
 335         }
 336
 337         public function getAllowedParams() {
 338                 if ( $this->allowedParams !== null ) {
 339                         return $this->allowedParams;
 340                 }
 341
 342                 $this->allowedParams = $this->buildCommonApiParams() + [
 343                         'what' => [
 344                                 ApiBase::PARAM_TYPE => [
 345                                         'title',
 346                                         'text',
 347                                         'nearmatch',
 348                                 ]
 349                         ],
 350                         'info' => [
 351                                 ApiBase::PARAM_DFLT => 'totalhits|suggestion|rewrittenquery',
 352                                 ApiBase::PARAM_TYPE => [
 353                                         'totalhits',
 354                                         'suggestion',
 355                                         'rewrittenquery',
 356                                 ],
 357                                 ApiBase::PARAM_ISMULTI => true,
 358                         ],
 359                         'prop' => [
 360                                 ApiBase::PARAM_DFLT => 'size|wordcount|timestamp|snippet',
 361                                 ApiBase::PARAM_TYPE => [
 362                                         'size',
 363                                         'wordcount',
 364                                         'timestamp',
 365                                         'snippet',
 366                                         'titlesnippet',
 367                                         'redirecttitle',
 368                                         'redirectsnippet',
 369                                         'sectiontitle',
 370                                         'sectionsnippet',
 371                                         'isfilematch',
 372                                         'categorysnippet',
 373                                         'score', // deprecated
 374                                         'hasrelated', // deprecated
 375                                         'extensiondata',
 376                                 ],
 377                                 ApiBase::PARAM_ISMULTI => true,
 378                                 ApiBase::PARAM_HELP_MSG_PER_VALUE => [],
 379                                 ApiBase::PARAM_DEPRECATED_VALUES => [
 380                                         'score' => true,
 381                                         'hasrelated' => true
 382                                 ],
 383                         ],
 384                         'interwiki' => false,
 385                         'enablerewrites' => false,
 386                 ];
 387
 388                 // If we have more than one engine the list of available sorts is
 389                 // difficult to represent. For now don't expose it.
 390                 $alternatives = MediaWiki\MediaWikiServices::getInstance()
 391                         ->getSearchEngineConfig()
 392                         ->getSearchTypes();
 393                 if ( count( $alternatives ) == 1 ) {
 394                         $this->allowedParams['sort'] = [
 395                                 ApiBase::PARAM_DFLT => 'relevance',
 396                                 ApiBase::PARAM_TYPE => MediaWiki\MediaWikiServices::getInstance()
 397                                         ->newSearchEngine()
 398                                         ->getValidSorts(),
 399                         ];
 400                 }
 401
 402                 return $this->allowedParams;
 403         }
 404
 405         public function getSearchProfileParams() {
 406                 return [
 407                         'qiprofile' => [
 408                                 'profile-type' => SearchEngine::FT_QUERY_INDEP_PROFILE_TYPE,
 409                                 'help-message' => 'apihelp-query+search-param-qiprofile',
 410                         ],
 411                 ];
 412         }
 413
 414         protected function getExamplesMessages() {
 415                 return [
 416                         'action=query&list=search&srsearch=meaning'
 417                                 => 'apihelp-query+search-example-simple',
 418                         'action=query&list=search&srwhat=text&srsearch=meaning'
 419                                 => 'apihelp-query+search-example-text',
 420                         'action=query&generator=search&gsrsearch=meaning&prop=info'
 421                                 => 'apihelp-query+search-example-generator',
 422                 ];
 423         }
 424
 425         public function getHelpUrls() {
 426                 return 'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Search';
 427         }
 428 }