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                 $limit = $params['limit'];
 161
 162                 if ( $matches->hasMoreResults() ) {
 163                         $this->setContinueEnumParameter( 'offset', $params['offset'] + $params['limit'] );
 164                 }
 165
 166                 foreach ( $matches as $result ) {
 167                         $count++;
 168                         // Silently skip broken and missing titles
 169                         if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 170                                 continue;
 171                         }
 172
 173                         if ( $resultPageSet === null ) {
 174                                 $vals = $this->getSearchResultData( $result, $prop, $terms );
 175                                 if ( $vals ) {
 176                                         // Add item to results and see whether it fits
 177                                         $fit = $apiResult->addValue( [ 'query', $this->getModuleName() ], null, $vals );
 178                                         if ( !$fit ) {
 179                                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $count - 1 );
 180                                                 break;
 181                                         }
 182                                 }
 183                         } else {
 184                                 $titles[] = $result->getTitle();
 185                         }
 186                 }
 187
 188                 // Here we assume interwiki results do not count with
 189                 // regular search results. We may want to reconsider this
 190                 // if we ever return a lot of interwiki results or want pagination
 191                 // for them.
 192                 // Interwiki results inside main result set
 193                 $canAddInterwiki = (bool)$params['enablerewrites'] && ( $resultPageSet === null );
 194                 if ( $canAddInterwiki ) {
 195                         $this->addInterwikiResults( $matches, $apiResult, $prop, $terms, 'additional',
 196                                 SearchResultSet::INLINE_RESULTS );
 197                 }
 198
 199                 // Interwiki results outside main result set
 200                 if ( $interwiki && $resultPageSet === null ) {
 201                         $this->addInterwikiResults( $matches, $apiResult, $prop, $terms, 'interwiki',
 202                                 SearchResultSet::SECONDARY_RESULTS );
 203                 }
 204
 205                 if ( $resultPageSet === null ) {
 206                         $apiResult->addIndexedTagName( [
 207                                 'query', $this->getModuleName()
 208                         ], 'p' );
 209                 } else {
 210                         $resultPageSet->setRedirectMergePolicy( function ( $current, $new ) {
 211                                 if ( !isset( $current['index'] ) || $new['index'] < $current['index'] ) {
 212                                         $current['index'] = $new['index'];
 213                                 }
 214                                 return $current;
 215                         } );
 216                         $resultPageSet->populateFromTitles( $titles );
 217                         $offset = $params['offset'] + 1;
 218                         foreach ( $titles as $index => $title ) {
 219                                 $resultPageSet->setGeneratorData( $title, [ 'index' => $index + $offset ] );
 220                         }
 221                 }
 222         }
 223
 224         /**
 225          * Assemble search result data.
 226          * @param SearchResult $result Search result
 227          * @param array        $prop Props to extract (as keys)
 228          * @param array        $terms Terms list
 229          * @return array|null Result data or null if result is broken in some way.
 230          */
 231         private function getSearchResultData( SearchResult $result, $prop, $terms ) {
 232                 // Silently skip broken and missing titles
 233                 if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 234                         return null;
 235                 }
 236
 237                 $vals = [];
 238
 239                 $title = $result->getTitle();
 240                 ApiQueryBase::addTitleInfo( $vals, $title );
 241                 $vals['pageid'] = $title->getArticleID();
 242
 243                 if ( isset( $prop['size'] ) ) {
 244                         $vals['size'] = $result->getByteSize();
 245                 }
 246                 if ( isset( $prop['wordcount'] ) ) {
 247                         $vals['wordcount'] = $result->getWordCount();
 248                 }
 249                 if ( isset( $prop['snippet'] ) ) {
 250                         $vals['snippet'] = $result->getTextSnippet( $terms );
 251                 }
 252                 if ( isset( $prop['timestamp'] ) ) {
 253                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $result->getTimestamp() );
 254                 }
 255                 if ( isset( $prop['titlesnippet'] ) ) {
 256                         $vals['titlesnippet'] = $result->getTitleSnippet();
 257                 }
 258                 if ( isset( $prop['categorysnippet'] ) ) {
 259                         $vals['categorysnippet'] = $result->getCategorySnippet();
 260                 }
 261                 if ( !is_null( $result->getRedirectTitle() ) ) {
 262                         if ( isset( $prop['redirecttitle'] ) ) {
 263                                 $vals['redirecttitle'] = $result->getRedirectTitle()->getPrefixedText();
 264                         }
 265                         if ( isset( $prop['redirectsnippet'] ) ) {
 266                                 $vals['redirectsnippet'] = $result->getRedirectSnippet();
 267                         }
 268                 }
 269                 if ( !is_null( $result->getSectionTitle() ) ) {
 270                         if ( isset( $prop['sectiontitle'] ) ) {
 271                                 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
 272                         }
 273                         if ( isset( $prop['sectionsnippet'] ) ) {
 274                                 $vals['sectionsnippet'] = $result->getSectionSnippet();
 275                         }
 276                 }
 277                 if ( isset( $prop['isfilematch'] ) ) {
 278                         $vals['isfilematch'] = $result->isFileMatch();
 279                 }
 280
 281                 if ( isset( $prop['extensiondata'] ) ) {
 282                         $extra = $result->getExtensionData();
 283                         // Add augmented data to the result. The data would be organized as a map:
 284                         // augmentorName => data
 285                         if ( $extra ) {
 286                                 $vals['extensiondata'] = ApiResult::addMetadataToResultVars( $extra );
 287                         }
 288                 }
 289
 290                 return $vals;
 291         }
 292
 293         /**
 294          * Add interwiki results as a section in query results.
 295          * @param SearchResultSet $matches
 296          * @param ApiResult       $apiResult
 297          * @param array           $prop Props to extract (as keys)
 298          * @param array           $terms Terms list
 299          * @param string          $section Section name where results would go
 300          * @param int             $type Interwiki result type
 301          * @return int|null Number of total hits in the data or null if none was produced
 302          */
 303         private function addInterwikiResults(
 304                 SearchResultSet $matches, ApiResult $apiResult, $prop,
 305                 $terms, $section, $type
 306         ) {
 307                 $totalhits = null;
 308                 if ( $matches->hasInterwikiResults( $type ) ) {
 309                         foreach ( $matches->getInterwikiResults( $type ) as $interwikiMatches ) {
 310                                 // Include number of results if requested
 311                                 $totalhits += $interwikiMatches->getTotalHits();
 312
 313                                 foreach ( $interwikiMatches as $result ) {
 314                                         $title = $result->getTitle();
 315                                         $vals = $this->getSearchResultData( $result, $prop, $terms );
 316
 317                                         $vals['namespace'] = $result->getInterwikiNamespaceText();
 318                                         $vals['title'] = $title->getText();
 319                                         $vals['url'] = $title->getFullURL();
 320
 321                                         // Add item to results and see whether it fits
 322                                         $fit = $apiResult->addValue( [
 323                                                         'query',
 324                                                         $section . $this->getModuleName(),
 325                                                         $result->getInterwikiPrefix()
 326                                                 ], null, $vals );
 327
 328                                         if ( !$fit ) {
 329                                                 // We hit the limit. We can't really provide any meaningful
 330                                                 // pagination info so just bail out
 331                                                 break;
 332                                         }
 333                                 }
 334                         }
 335                         if ( $totalhits !== null ) {
 336                                 $apiResult->addValue( [ 'query', $section . 'searchinfo' ], 'totalhits', $totalhits );
 337                                 $apiResult->addIndexedTagName( [
 338                                         'query', $section . $this->getModuleName()
 339                                 ], 'p' );
 340                         }
 341                 }
 342                 return $totalhits;
 343         }
 344
 345         public function getCacheMode( $params ) {
 346                 return 'public';
 347         }
 348
 349         public function getAllowedParams() {
 350                 if ( $this->allowedParams !== null ) {
 351                         return $this->allowedParams;
 352                 }
 353
 354                 $this->allowedParams = $this->buildCommonApiParams() + [
 355                         'what' => [
 356                                 ApiBase::PARAM_TYPE => [
 357                                         'title',
 358                                         'text',
 359                                         'nearmatch',
 360                                 ]
 361                         ],
 362                         'info' => [
 363                                 ApiBase::PARAM_DFLT => 'totalhits|suggestion|rewrittenquery',
 364                                 ApiBase::PARAM_TYPE => [
 365                                         'totalhits',
 366                                         'suggestion',
 367                                         'rewrittenquery',
 368                                 ],
 369                                 ApiBase::PARAM_ISMULTI => true,
 370                         ],
 371                         'prop' => [
 372                                 ApiBase::PARAM_DFLT => 'size|wordcount|timestamp|snippet',
 373                                 ApiBase::PARAM_TYPE => [
 374                                         'size',
 375                                         'wordcount',
 376                                         'timestamp',
 377                                         'snippet',
 378                                         'titlesnippet',
 379                                         'redirecttitle',
 380                                         'redirectsnippet',
 381                                         'sectiontitle',
 382                                         'sectionsnippet',
 383                                         'isfilematch',
 384                                         'categorysnippet',
 385                                         'score', // deprecated
 386                                         'hasrelated', // deprecated
 387                                         'extensiondata',
 388                                 ],
 389                                 ApiBase::PARAM_ISMULTI => true,
 390                                 ApiBase::PARAM_HELP_MSG_PER_VALUE => [],
 391                                 ApiBase::PARAM_DEPRECATED_VALUES => [
 392                                         'score' => true,
 393                                         'hasrelated' => true
 394                                 ],
 395                         ],
 396                         'interwiki' => false,
 397                         'enablerewrites' => false,
 398                 ];
 399
 400                 // If we have more than one engine the list of available sorts is
 401                 // difficult to represent. For now don't expose it.
 402                 $services = MediaWiki\MediaWikiServices::getInstance();
 403                 $alternatives = $services
 404                         ->getSearchEngineConfig()
 405                         ->getSearchTypes();
 406                 if ( count( $alternatives ) == 1 ) {
 407                         $this->allowedParams['sort'] = [
 408                                 ApiBase::PARAM_DFLT => 'relevance',
 409                                 ApiBase::PARAM_TYPE => $services
 410                                         ->newSearchEngine()
 411                                         ->getValidSorts(),
 412                         ];
 413                 }
 414
 415                 return $this->allowedParams;
 416         }
 417
 418         public function getSearchProfileParams() {
 419                 return [
 420                         'qiprofile' => [
 421                                 'profile-type' => SearchEngine::FT_QUERY_INDEP_PROFILE_TYPE,
 422                                 'help-message' => 'apihelp-query+search-param-qiprofile',
 423                         ],
 424                 ];
 425         }
 426
 427         protected function getExamplesMessages() {
 428                 return [
 429                         'action=query&list=search&srsearch=meaning'
 430                                 => 'apihelp-query+search-example-simple',
 431                         'action=query&list=search&srwhat=text&srsearch=meaning'
 432                                 => 'apihelp-query+search-example-text',
 433                         'action=query&generator=search&gsrsearch=meaning&prop=info'
 434                                 => 'apihelp-query+search-example-generator',
 435                 ];
 436         }
 437
 438         public function getHelpUrls() {
 439                 return 'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Search';
 440         }
 441 }