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