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 use MediaWiki\MediaWikiServices;
  24
  25 /**
  26  * Query module to perform full text search within wiki titles and content
  27  *
  28  * @ingroup API
  29  */
  30 class ApiQuerySearch extends ApiQueryGeneratorBase {
  31         use SearchApi;
  32
  33         /** @var array list of api allowed params */
  34         private $allowedParams;
  35
  36         public function __construct( ApiQuery $query, $moduleName ) {
  37                 parent::__construct( $query, $moduleName, 'sr' );
  38         }
  39
  40         public function execute() {
  41                 $this->run();
  42         }
  43
  44         public function executeGenerator( $resultPageSet ) {
  45                 $this->run( $resultPageSet );
  46         }
  47
  48         /**
  49          * @param ApiPageSet $resultPageSet
  50          * @return void
  51          */
  52         private function run( $resultPageSet = null ) {
  53                 $params = $this->extractRequestParams();
  54
  55                 // Extract parameters
  56                 $query = $params['search'];
  57                 $what = $params['what'];
  58                 $interwiki = $params['interwiki'];
  59                 $searchInfo = array_flip( $params['info'] );
  60                 $prop = array_flip( $params['prop'] );
  61
  62                 // Create search engine instance and set options
  63                 $search = $this->buildSearchEngine( $params );
  64                 if ( isset( $params['sort'] ) ) {
  65                         $search->setSort( $params['sort'] );
  66                 }
  67                 $search->setFeatureData( 'rewrite', (bool)$params['enablerewrites'] );
  68                 $search->setFeatureData( 'interwiki', (bool)$interwiki );
  69
  70                 $nquery = $search->transformSearchTerm( $query );
  71                 if ( $nquery !== $query ) {
  72                         $query = $nquery;
  73                         wfDeprecated( 'SearchEngine::transformSearchTerm() (overridden by ' .
  74                                 get_class( $search ) . ')', '1.32' );
  75                 }
  76
  77                 $nquery = $search->replacePrefixes( $query );
  78                 if ( $nquery !== $query ) {
  79                         $query = $nquery;
  80                         wfDeprecated( 'SearchEngine::replacePrefixes() (overridden by ' .
  81                                                   get_class( $search ) . ')', '1.32' );
  82                 }
  83                 // Perform the actual search
  84                 if ( $what == 'text' ) {
  85                         $matches = $search->searchText( $query );
  86                 } elseif ( $what == 'title' ) {
  87                         $matches = $search->searchTitle( $query );
  88                 } elseif ( $what == 'nearmatch' ) {
  89                         // near matches must receive the user input as provided, otherwise
  90                         // the near matches within namespaces are lost.
  91                         $matches = $search->getNearMatcher( $this->getConfig() )
  92                                 ->getNearMatchResultSet( $params['search'] );
  93                 } else {
  94                         // We default to title searches; this is a terrible legacy
  95                         // of the way we initially set up the MySQL fulltext-based
  96                         // search engine with separate title and text fields.
  97                         // In the future, the default should be for a combined index.
  98                         $what = 'title';
  99                         $matches = $search->searchTitle( $query );
 100
 101                         // Not all search engines support a separate title search,
 102                         // for instance the Lucene-based engine we use on Wikipedia.
 103                         // In this case, fall back to full-text search (which will
 104                         // include titles in it!)
 105                         if ( is_null( $matches ) ) {
 106                                 $what = 'text';
 107                                 $matches = $search->searchText( $query );
 108                         }
 109                 }
 110
 111                 if ( $matches instanceof Status ) {
 112                         $status = $matches;
 113                         $matches = $status->getValue();
 114                 } else {
 115                         $status = null;
 116                 }
 117
 118                 if ( $status ) {
 119                         if ( $status->isOK() ) {
 120                                 $this->getMain()->getErrorFormatter()->addMessagesFromStatus(
 121                                         $this->getModuleName(),
 122                                         $status
 123                                 );
 124                         } else {
 125                                 $this->dieStatus( $status );
 126                         }
 127                 } elseif ( is_null( $matches ) ) {
 128                         $this->dieWithError( [ 'apierror-searchdisabled', $what ], "search-{$what}-disabled" );
 129                 }
 130
 131                 if ( $resultPageSet === null ) {
 132                         $apiResult = $this->getResult();
 133                         // Add search meta data to result
 134                         if ( isset( $searchInfo['totalhits'] ) ) {
 135                                 $totalhits = $matches->getTotalHits();
 136                                 if ( $totalhits !== null ) {
 137                                         $apiResult->addValue( [ 'query', 'searchinfo' ],
 138                                                 'totalhits', $totalhits );
 139                                 }
 140                         }
 141                         if ( isset( $searchInfo['suggestion'] ) && $matches->hasSuggestion() ) {
 142                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 143                                         'suggestion', $matches->getSuggestionQuery() );
 144                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 145                                         'suggestionsnippet', $matches->getSuggestionSnippet() );
 146                         }
 147                         if ( isset( $searchInfo['rewrittenquery'] ) && $matches->hasRewrittenQuery() ) {
 148                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 149                                         'rewrittenquery', $matches->getQueryAfterRewrite() );
 150                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 151                                         'rewrittenquerysnippet', $matches->getQueryAfterRewriteSnippet() );
 152                         }
 153                 }
 154
 155                 // Add the search results to the result
 156                 $terms = MediaWikiServices::getInstance()->getContentLanguage()->
 157                         convertForSearchResult( $matches->termMatches() );
 158                 $titles = [];
 159                 $count = 0;
 160
 161                 if ( $matches->hasMoreResults() ) {
 162                         $this->setContinueEnumParameter( 'offset', $params['offset'] + $params['limit'] );
 163                 }
 164
 165                 foreach ( $matches as $result ) {
 166                         $count++;
 167                         // Silently skip broken and missing titles
 168                         if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 169                                 continue;
 170                         }
 171
 172                         if ( $resultPageSet === null ) {
 173                                 $vals = $this->getSearchResultData( $result, $prop, $terms );
 174                                 if ( $vals ) {
 175                                         // Add item to results and see whether it fits
 176                                         $fit = $apiResult->addValue( [ 'query', $this->getModuleName() ], null, $vals );
 177                                         if ( !$fit ) {
 178                                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $count - 1 );
 179                                                 break;
 180                                         }
 181                                 }
 182                         } else {
 183                                 $titles[] = $result->getTitle();
 184                         }
 185                 }
 186
 187                 // Here we assume interwiki results do not count with
 188                 // regular search results. We may want to reconsider this
 189                 // if we ever return a lot of interwiki results or want pagination
 190                 // for them.
 191                 // Interwiki results inside main result set
 192                 $canAddInterwiki = (bool)$params['enablerewrites'] && ( $resultPageSet === null );
 193                 if ( $canAddInterwiki ) {
 194                         $this->addInterwikiResults( $matches, $apiResult, $prop, $terms, 'additional',
 195                                 SearchResultSet::INLINE_RESULTS );
 196                 }
 197
 198                 // Interwiki results outside main result set
 199                 if ( $interwiki && $resultPageSet === null ) {
 200                         $this->addInterwikiResults( $matches, $apiResult, $prop, $terms, 'interwiki',
 201                                 SearchResultSet::SECONDARY_RESULTS );
 202                 }
 203
 204                 if ( $resultPageSet === null ) {
 205                         $apiResult->addIndexedTagName( [
 206                                 'query', $this->getModuleName()
 207                         ], 'p' );
 208                 } else {
 209                         $resultPageSet->setRedirectMergePolicy( function ( $current, $new ) {
 210                                 if ( !isset( $current['index'] ) || $new['index'] < $current['index'] ) {
 211                                         $current['index'] = $new['index'];
 212                                 }
 213                                 return $current;
 214                         } );
 215                         $resultPageSet->populateFromTitles( $titles );
 216                         $offset = $params['offset'] + 1;
 217                         foreach ( $titles as $index => $title ) {
 218                                 $resultPageSet->setGeneratorData( $title, [ 'index' => $index + $offset ] );
 219                         }
 220                 }
 221         }
 222
 223         /**
 224          * Assemble search result data.
 225          * @param SearchResult $result Search result
 226          * @param array        $prop Props to extract (as keys)
 227          * @param array        $terms Terms list
 228          * @return array|null Result data or null if result is broken in some way.
 229          */
 230         private function getSearchResultData( SearchResult $result, $prop, $terms ) {
 231                 // Silently skip broken and missing titles
 232                 if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 233                         return null;
 234                 }
 235
 236                 $vals = [];
 237
 238                 $title = $result->getTitle();
 239                 ApiQueryBase::addTitleInfo( $vals, $title );
 240                 $vals['pageid'] = $title->getArticleID();
 241
 242                 if ( isset( $prop['size'] ) ) {
 243                         $vals['size'] = $result->getByteSize();
 244                 }
 245                 if ( isset( $prop['wordcount'] ) ) {
 246                         $vals['wordcount'] = $result->getWordCount();
 247                 }
 248                 if ( isset( $prop['snippet'] ) ) {
 249                         $vals['snippet'] = $result->getTextSnippet( $terms );
 250                 }
 251                 if ( isset( $prop['timestamp'] ) ) {
 252                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $result->getTimestamp() );
 253                 }
 254                 if ( isset( $prop['titlesnippet'] ) ) {
 255                         $vals['titlesnippet'] = $result->getTitleSnippet();
 256                 }
 257                 if ( isset( $prop['categorysnippet'] ) ) {
 258                         $vals['categorysnippet'] = $result->getCategorySnippet();
 259                 }
 260                 if ( !is_null( $result->getRedirectTitle() ) ) {
 261                         if ( isset( $prop['redirecttitle'] ) ) {
 262                                 $vals['redirecttitle'] = $result->getRedirectTitle()->getPrefixedText();
 263                         }
 264                         if ( isset( $prop['redirectsnippet'] ) ) {
 265                                 $vals['redirectsnippet'] = $result->getRedirectSnippet();
 266                         }
 267                 }
 268                 if ( !is_null( $result->getSectionTitle() ) ) {
 269                         if ( isset( $prop['sectiontitle'] ) ) {
 270                                 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
 271                         }
 272                         if ( isset( $prop['sectionsnippet'] ) ) {
 273                                 $vals['sectionsnippet'] = $result->getSectionSnippet();
 274                         }
 275                 }
 276                 if ( isset( $prop['isfilematch'] ) ) {
 277                         $vals['isfilematch'] = $result->isFileMatch();
 278                 }
 279
 280                 if ( isset( $prop['extensiondata'] ) ) {
 281                         $extra = $result->getExtensionData();
 282                         // Add augmented data to the result. The data would be organized as a map:
 283                         // augmentorName => data
 284                         if ( $extra ) {
 285                                 $vals['extensiondata'] = ApiResult::addMetadataToResultVars( $extra );
 286                         }
 287                 }
 288
 289                 return $vals;
 290         }
 291
 292         /**
 293          * Add interwiki results as a section in query results.
 294          * @param SearchResultSet $matches
 295          * @param ApiResult       $apiResult
 296          * @param array           $prop Props to extract (as keys)
 297          * @param array           $terms Terms list
 298          * @param string          $section Section name where results would go
 299          * @param int             $type Interwiki result type
 300          * @return int|null Number of total hits in the data or null if none was produced
 301          */
 302         private function addInterwikiResults(
 303                 SearchResultSet $matches, ApiResult $apiResult, $prop,
 304                 $terms, $section, $type
 305         ) {
 306                 $totalhits = null;
 307                 if ( $matches->hasInterwikiResults( $type ) ) {
 308                         foreach ( $matches->getInterwikiResults( $type ) as $interwikiMatches ) {
 309                                 // Include number of results if requested
 310                                 $totalhits += $interwikiMatches->getTotalHits();
 311
 312                                 foreach ( $interwikiMatches as $result ) {
 313                                         $title = $result->getTitle();
 314                                         $vals = $this->getSearchResultData( $result, $prop, $terms );
 315
 316                                         $vals['namespace'] = $result->getInterwikiNamespaceText();
 317                                         $vals['title'] = $title->getText();
 318                                         $vals['url'] = $title->getFullURL();
 319
 320                                         // Add item to results and see whether it fits
 321                                         $fit = $apiResult->addValue( [
 322                                                         'query',
 323                                                         $section . $this->getModuleName(),
 324                                                         $result->getInterwikiPrefix()
 325                                                 ], null, $vals );
 326
 327                                         if ( !$fit ) {
 328                                                 // We hit the limit. We can't really provide any meaningful
 329                                                 // pagination info so just bail out
 330                                                 break;
 331                                         }
 332                                 }
 333                         }
 334                         if ( $totalhits !== null ) {
 335                                 $apiResult->addValue( [ 'query', $section . 'searchinfo' ], 'totalhits', $totalhits );
 336                                 $apiResult->addIndexedTagName( [
 337                                         'query', $section . $this->getModuleName()
 338                                 ], 'p' );
 339                         }
 340                 }
 341                 return $totalhits;
 342         }
 343
 344         public function getCacheMode( $params ) {
 345                 return 'public';
 346         }
 347
 348         public function getAllowedParams() {
 349                 if ( $this->allowedParams !== null ) {
 350                         return $this->allowedParams;
 351                 }
 352
 353                 $this->allowedParams = $this->buildCommonApiParams() + [
 354                         'what' => [
 355                                 ApiBase::PARAM_TYPE => [
 356                                         'title',
 357                                         'text',
 358                                         'nearmatch',
 359                                 ]
 360                         ],
 361                         'info' => [
 362                                 ApiBase::PARAM_DFLT => 'totalhits|suggestion|rewrittenquery',
 363                                 ApiBase::PARAM_TYPE => [
 364                                         'totalhits',
 365                                         'suggestion',
 366                                         'rewrittenquery',
 367                                 ],
 368                                 ApiBase::PARAM_ISMULTI => true,
 369                         ],
 370                         'prop' => [
 371                                 ApiBase::PARAM_DFLT => 'size|wordcount|timestamp|snippet',
 372                                 ApiBase::PARAM_TYPE => [
 373                                         'size',
 374                                         'wordcount',
 375                                         'timestamp',
 376                                         'snippet',
 377                                         'titlesnippet',
 378                                         'redirecttitle',
 379                                         'redirectsnippet',
 380                                         'sectiontitle',
 381                                         'sectionsnippet',
 382                                         'isfilematch',
 383                                         'categorysnippet',
 384                                         'score', // deprecated
 385                                         'hasrelated', // deprecated
 386                                         'extensiondata',
 387                                 ],
 388                                 ApiBase::PARAM_ISMULTI => true,
 389                                 ApiBase::PARAM_HELP_MSG_PER_VALUE => [],
 390                                 ApiBase::PARAM_DEPRECATED_VALUES => [
 391                                         'score' => true,
 392                                         'hasrelated' => true
 393                                 ],
 394                         ],
 395                         'interwiki' => false,
 396                         'enablerewrites' => false,
 397                 ];
 398
 399                 // If we have more than one engine the list of available sorts is
 400                 // difficult to represent. For now don't expose it.
 401                 $services = MediaWiki\MediaWikiServices::getInstance();
 402                 $alternatives = $services
 403                         ->getSearchEngineConfig()
 404                         ->getSearchTypes();
 405                 if ( count( $alternatives ) == 1 ) {
 406                         $this->allowedParams['sort'] = [
 407                                 ApiBase::PARAM_DFLT => 'relevance',
 408                                 ApiBase::PARAM_TYPE => $services
 409                                         ->newSearchEngine()
 410                                         ->getValidSorts(),
 411                         ];
 412                 }
 413
 414                 return $this->allowedParams;
 415         }
 416
 417         public function getSearchProfileParams() {
 418                 return [
 419                         'qiprofile' => [
 420                                 'profile-type' => SearchEngine::FT_QUERY_INDEP_PROFILE_TYPE,
 421                                 'help-message' => 'apihelp-query+search-param-qiprofile',
 422                         ],
 423                 ];
 424         }
 425
 426         protected function getExamplesMessages() {
 427                 return [
 428                         'action=query&list=search&srsearch=meaning'
 429                                 => 'apihelp-query+search-example-simple',
 430                         'action=query&list=search&srwhat=text&srsearch=meaning'
 431                                 => 'apihelp-query+search-example-text',
 432                         'action=query&generator=search&gsrsearch=meaning&prop=info'
 433                                 => 'apihelp-query+search-example-generator',
 434                 ];
 435         }
 436
 437         public function getHelpUrls() {
 438                 return 'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Search';
 439         }
 440 }