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                 $nquery = $search->transformSearchTerm( $query );
  70                 if ( $nquery !== $query ) {
  71                         $query = $nquery;
  72                         wfDeprecated( 'SearchEngine::transformSearchTerm() (overridden by ' .
  73                                 get_class( $search ) . ')', '1.32' );
  74                 }
  75
  76                 $nquery = $search->replacePrefixes( $query );
  77                 if ( $nquery !== $query ) {
  78                         $query = $nquery;
  79                         wfDeprecated( 'SearchEngine::replacePrefixes() (overridden by ' .
  80                                                   get_class( $search ) . ')', '1.32' );
  81                 }
  82                 // Perform the actual search
  83                 if ( $what == 'text' ) {
  84                         $matches = $search->searchText( $query );
  85                 } elseif ( $what == 'title' ) {
  86                         $matches = $search->searchTitle( $query );
  87                 } elseif ( $what == 'nearmatch' ) {
  88                         // near matches must receive the user input as provided, otherwise
  89                         // the near matches within namespaces are lost.
  90                         $matches = $search->getNearMatcher( $this->getConfig() )
  91                                 ->getNearMatchResultSet( $params['search'] );
  92                 } else {
  93                         // We default to title searches; this is a terrible legacy
  94                         // of the way we initially set up the MySQL fulltext-based
  95                         // search engine with separate title and text fields.
  96                         // In the future, the default should be for a combined index.
  97                         $what = 'title';
  98                         $matches = $search->searchTitle( $query );
  99
 100                         // Not all search engines support a separate title search,
 101                         // for instance the Lucene-based engine we use on Wikipedia.
 102                         // In this case, fall back to full-text search (which will
 103                         // include titles in it!)
 104                         if ( is_null( $matches ) ) {
 105                                 $what = 'text';
 106                                 $matches = $search->searchText( $query );
 107                         }
 108                 }
 109
 110                 if ( $matches instanceof Status ) {
 111                         $status = $matches;
 112                         $matches = $status->getValue();
 113                 } else {
 114                         $status = null;
 115                 }
 116
 117                 if ( $status ) {
 118                         if ( $status->isOK() ) {
 119                                 $this->getMain()->getErrorFormatter()->addMessagesFromStatus(
 120                                         $this->getModuleName(),
 121                                         $status
 122                                 );
 123                         } else {
 124                                 $this->dieStatus( $status );
 125                         }
 126                 } elseif ( is_null( $matches ) ) {
 127                         $this->dieWithError( [ 'apierror-searchdisabled', $what ], "search-{$what}-disabled" );
 128                 }
 129
 130                 if ( $resultPageSet === null ) {
 131                         $apiResult = $this->getResult();
 132                         // Add search meta data to result
 133                         if ( isset( $searchInfo['totalhits'] ) ) {
 134                                 $totalhits = $matches->getTotalHits();
 135                                 if ( $totalhits !== null ) {
 136                                         $apiResult->addValue( [ 'query', 'searchinfo' ],
 137                                                 'totalhits', $totalhits );
 138                                 }
 139                         }
 140                         if ( isset( $searchInfo['suggestion'] ) && $matches->hasSuggestion() ) {
 141                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 142                                         'suggestion', $matches->getSuggestionQuery() );
 143                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 144                                         'suggestionsnippet', $matches->getSuggestionSnippet() );
 145                         }
 146                         if ( isset( $searchInfo['rewrittenquery'] ) && $matches->hasRewrittenQuery() ) {
 147                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 148                                         'rewrittenquery', $matches->getQueryAfterRewrite() );
 149                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 150                                         'rewrittenquerysnippet', $matches->getQueryAfterRewriteSnippet() );
 151                         }
 152                 }
 153
 154                 // Add the search results to the result
 155                 $terms = $wgContLang->convertForSearchResult( $matches->termMatches() );
 156                 $titles = [];
 157                 $count = 0;
 158                 $limit = $params['limit'];
 159
 160                 if ( $matches->hasMoreResults() ) {
 161                         $this->setContinueEnumParameter( 'offset', $params['offset'] + $params['limit'] );
 162                 }
 163
 164                 foreach ( $matches as $result ) {
 165                         $count++;
 166                         // Silently skip broken and missing titles
 167                         if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 168                                 continue;
 169                         }
 170
 171                         if ( $resultPageSet === null ) {
 172                                 $vals = $this->getSearchResultData( $result, $prop, $terms );
 173                                 if ( $vals ) {
 174                                         // Add item to results and see whether it fits
 175                                         $fit = $apiResult->addValue( [ 'query', $this->getModuleName() ], null, $vals );
 176                                         if ( !$fit ) {
 177                                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $count - 1 );
 178                                                 break;
 179                                         }
 180                                 }
 181                         } else {
 182                                 $titles[] = $result->getTitle();
 183                         }
 184                 }
 185
 186                 // Here we assume interwiki results do not count with
 187                 // regular search results. We may want to reconsider this
 188                 // if we ever return a lot of interwiki results or want pagination
 189                 // for them.
 190                 // Interwiki results inside main result set
 191                 $canAddInterwiki = (bool)$params['enablerewrites'] && ( $resultPageSet === null );
 192                 if ( $canAddInterwiki ) {
 193                         $this->addInterwikiResults( $matches, $apiResult, $prop, $terms, 'additional',
 194                                 SearchResultSet::INLINE_RESULTS );
 195                 }
 196
 197                 // Interwiki results outside main result set
 198                 if ( $interwiki && $resultPageSet === null ) {
 199                         $this->addInterwikiResults( $matches, $apiResult, $prop, $terms, 'interwiki',
 200                                 SearchResultSet::SECONDARY_RESULTS );
 201                 }
 202
 203                 if ( $resultPageSet === null ) {
 204                         $apiResult->addIndexedTagName( [
 205                                 'query', $this->getModuleName()
 206                         ], 'p' );
 207                 } else {
 208                         $resultPageSet->setRedirectMergePolicy( function ( $current, $new ) {
 209                                 if ( !isset( $current['index'] ) || $new['index'] < $current['index'] ) {
 210                                         $current['index'] = $new['index'];
 211                                 }
 212                                 return $current;
 213                         } );
 214                         $resultPageSet->populateFromTitles( $titles );
 215                         $offset = $params['offset'] + 1;
 216                         foreach ( $titles as $index => $title ) {
 217                                 $resultPageSet->setGeneratorData( $title, [ 'index' => $index + $offset ] );
 218                         }
 219                 }
 220         }
 221
 222         /**
 223          * Assemble search result data.
 224          * @param SearchResult $result Search result
 225          * @param array        $prop Props to extract (as keys)
 226          * @param array        $terms Terms list
 227          * @return array|null Result data or null if result is broken in some way.
 228          */
 229         private function getSearchResultData( SearchResult $result, $prop, $terms ) {
 230                 // Silently skip broken and missing titles
 231                 if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 232                         return null;
 233                 }
 234
 235                 $vals = [];
 236
 237                 $title = $result->getTitle();
 238                 ApiQueryBase::addTitleInfo( $vals, $title );
 239                 $vals['pageid'] = $title->getArticleID();
 240
 241                 if ( isset( $prop['size'] ) ) {
 242                         $vals['size'] = $result->getByteSize();
 243                 }
 244                 if ( isset( $prop['wordcount'] ) ) {
 245                         $vals['wordcount'] = $result->getWordCount();
 246                 }
 247                 if ( isset( $prop['snippet'] ) ) {
 248                         $vals['snippet'] = $result->getTextSnippet( $terms );
 249                 }
 250                 if ( isset( $prop['timestamp'] ) ) {
 251                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $result->getTimestamp() );
 252                 }
 253                 if ( isset( $prop['titlesnippet'] ) ) {
 254                         $vals['titlesnippet'] = $result->getTitleSnippet();
 255                 }
 256                 if ( isset( $prop['categorysnippet'] ) ) {
 257                         $vals['categorysnippet'] = $result->getCategorySnippet();
 258                 }
 259                 if ( !is_null( $result->getRedirectTitle() ) ) {
 260                         if ( isset( $prop['redirecttitle'] ) ) {
 261                                 $vals['redirecttitle'] = $result->getRedirectTitle()->getPrefixedText();
 262                         }
 263                         if ( isset( $prop['redirectsnippet'] ) ) {
 264                                 $vals['redirectsnippet'] = $result->getRedirectSnippet();
 265                         }
 266                 }
 267                 if ( !is_null( $result->getSectionTitle() ) ) {
 268                         if ( isset( $prop['sectiontitle'] ) ) {
 269                                 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
 270                         }
 271                         if ( isset( $prop['sectionsnippet'] ) ) {
 272                                 $vals['sectionsnippet'] = $result->getSectionSnippet();
 273                         }
 274                 }
 275                 if ( isset( $prop['isfilematch'] ) ) {
 276                         $vals['isfilematch'] = $result->isFileMatch();
 277                 }
 278
 279                 if ( isset( $prop['extensiondata'] ) ) {
 280                         $extra = $result->getExtensionData();
 281                         // Add augmented data to the result. The data would be organized as a map:
 282                         // augmentorName => data
 283                         if ( $extra ) {
 284                                 $vals['extensiondata'] = ApiResult::addMetadataToResultVars( $extra );
 285                         }
 286                 }
 287
 288                 return $vals;
 289         }
 290
 291         /**
 292          * Add interwiki results as a section in query results.
 293          * @param SearchResultSet $matches
 294          * @param ApiResult       $apiResult
 295          * @param array           $prop Props to extract (as keys)
 296          * @param array           $terms Terms list
 297          * @param string          $section Section name where results would go
 298          * @param int             $type Interwiki result type
 299          * @return int|null Number of total hits in the data or null if none was produced
 300          */
 301         private function addInterwikiResults(
 302                 SearchResultSet $matches, ApiResult $apiResult, $prop,
 303                 $terms, $section, $type
 304         ) {
 305                 $totalhits = null;
 306                 if ( $matches->hasInterwikiResults( $type ) ) {
 307                         foreach ( $matches->getInterwikiResults( $type ) as $interwikiMatches ) {
 308                                 // Include number of results if requested
 309                                 $totalhits += $interwikiMatches->getTotalHits();
 310
 311                                 foreach ( $interwikiMatches as $result ) {
 312                                         $title = $result->getTitle();
 313                                         $vals = $this->getSearchResultData( $result, $prop, $terms );
 314
 315                                         $vals['namespace'] = $result->getInterwikiNamespaceText();
 316                                         $vals['title'] = $title->getText();
 317                                         $vals['url'] = $title->getFullURL();
 318
 319                                         // Add item to results and see whether it fits
 320                                         $fit = $apiResult->addValue( [
 321                                                         'query',
 322                                                         $section . $this->getModuleName(),
 323                                                         $result->getInterwikiPrefix()
 324                                                 ], null, $vals );
 325
 326                                         if ( !$fit ) {
 327                                                 // We hit the limit. We can't really provide any meaningful
 328                                                 // pagination info so just bail out
 329                                                 break;
 330                                         }
 331                                 }
 332                         }
 333                         if ( $totalhits !== null ) {
 334                                 $apiResult->addValue( [ 'query', $section . 'searchinfo' ], 'totalhits', $totalhits );
 335                                 $apiResult->addIndexedTagName( [
 336                                         'query', $section . $this->getModuleName()
 337                                 ], 'p' );
 338                         }
 339                 }
 340                 return $totalhits;
 341         }
 342
 343         public function getCacheMode( $params ) {
 344                 return 'public';
 345         }
 346
 347         public function getAllowedParams() {
 348                 if ( $this->allowedParams !== null ) {
 349                         return $this->allowedParams;
 350                 }
 351
 352                 $this->allowedParams = $this->buildCommonApiParams() + [
 353                         'what' => [
 354                                 ApiBase::PARAM_TYPE => [
 355                                         'title',
 356                                         'text',
 357                                         'nearmatch',
 358                                 ]
 359                         ],
 360                         'info' => [
 361                                 ApiBase::PARAM_DFLT => 'totalhits|suggestion|rewrittenquery',
 362                                 ApiBase::PARAM_TYPE => [
 363                                         'totalhits',
 364                                         'suggestion',
 365                                         'rewrittenquery',
 366                                 ],
 367                                 ApiBase::PARAM_ISMULTI => true,
 368                         ],
 369                         'prop' => [
 370                                 ApiBase::PARAM_DFLT => 'size|wordcount|timestamp|snippet',
 371                                 ApiBase::PARAM_TYPE => [
 372                                         'size',
 373                                         'wordcount',
 374                                         'timestamp',
 375                                         'snippet',
 376                                         'titlesnippet',
 377                                         'redirecttitle',
 378                                         'redirectsnippet',
 379                                         'sectiontitle',
 380                                         'sectionsnippet',
 381                                         'isfilematch',
 382                                         'categorysnippet',
 383                                         'score', // deprecated
 384                                         'hasrelated', // deprecated
 385                                         'extensiondata',
 386                                 ],
 387                                 ApiBase::PARAM_ISMULTI => true,
 388                                 ApiBase::PARAM_HELP_MSG_PER_VALUE => [],
 389                                 ApiBase::PARAM_DEPRECATED_VALUES => [
 390                                         'score' => true,
 391                                         'hasrelated' => true
 392                                 ],
 393                         ],
 394                         'interwiki' => false,
 395                         'enablerewrites' => false,
 396                 ];
 397
 398                 // If we have more than one engine the list of available sorts is
 399                 // difficult to represent. For now don't expose it.
 400                 $alternatives = MediaWiki\MediaWikiServices::getInstance()
 401                         ->getSearchEngineConfig()
 402                         ->getSearchTypes();
 403                 if ( count( $alternatives ) == 1 ) {
 404                         $this->allowedParams['sort'] = [
 405                                 ApiBase::PARAM_DFLT => 'relevance',
 406                                 ApiBase::PARAM_TYPE => MediaWiki\MediaWikiServices::getInstance()
 407                                         ->newSearchEngine()
 408                                         ->getValidSorts(),
 409                         ];
 410                 }
 411
 412                 return $this->allowedParams;
 413         }
 414
 415         public function getSearchProfileParams() {
 416                 return [
 417                         'qiprofile' => [
 418                                 'profile-type' => SearchEngine::FT_QUERY_INDEP_PROFILE_TYPE,
 419                                 'help-message' => 'apihelp-query+search-param-qiprofile',
 420                         ],
 421                 ];
 422         }
 423
 424         protected function getExamplesMessages() {
 425                 return [
 426                         'action=query&list=search&srsearch=meaning'
 427                                 => 'apihelp-query+search-example-simple',
 428                         'action=query&list=search&srwhat=text&srsearch=meaning'
 429                                 => 'apihelp-query+search-example-text',
 430                         'action=query&generator=search&gsrsearch=meaning&prop=info'
 431                                 => 'apihelp-query+search-example-generator',
 432                 ];
 433         }
 434
 435         public function getHelpUrls() {
 436                 return 'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Search';
 437         }
 438 }