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