includes/api/ApiQuerySearch.php

   1 <?php
   2 /**
   3  * Copyright © 2007 Yuri Astrakhan "<Firstname><Lastname>@gmail.com"
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  */
  22
  23 /**
  24  * Query module to perform full text search within wiki titles and content
  25  *
  26  * @ingroup API
  27  */
  28 class ApiQuerySearch extends ApiQueryGeneratorBase {
  29         use SearchApi;
  30
  31         /** @var array list of api allowed params */
  32         private $allowedParams;
  33
  34         public function __construct( ApiQuery $query, $moduleName ) {
  35                 parent::__construct( $query, $moduleName, 'sr' );
  36         }
  37
  38         public function execute() {
  39                 $this->run();
  40         }
  41
  42         public function executeGenerator( $resultPageSet ) {
  43                 $this->run( $resultPageSet );
  44         }
  45
  46         /**
  47          * @param ApiPageSet $resultPageSet
  48          * @return void
  49          */
  50         private function run( $resultPageSet = null ) {
  51                 global $wgContLang;
  52                 $params = $this->extractRequestParams();
  53
  54                 // Extract parameters
  55                 $query = $params['search'];
  56                 $what = $params['what'];
  57                 $interwiki = $params['interwiki'];
  58                 $searchInfo = array_flip( $params['info'] );
  59                 $prop = array_flip( $params['prop'] );
  60
  61                 // Create search engine instance and set options
  62                 $search = $this->buildSearchEngine( $params );
  63                 $search->setFeatureData( 'rewrite', (bool)$params['enablerewrites'] );
  64                 $search->setFeatureData( 'interwiki', (bool)$interwiki );
  65
  66                 $query = $search->transformSearchTerm( $query );
  67                 $query = $search->replacePrefixes( $query );
  68
  69                 // Perform the actual search
  70                 if ( $what == 'text' ) {
  71                         $matches = $search->searchText( $query );
  72                 } elseif ( $what == 'title' ) {
  73                         $matches = $search->searchTitle( $query );
  74                 } elseif ( $what == 'nearmatch' ) {
  75                         // near matches must receive the user input as provided, otherwise
  76                         // the near matches within namespaces are lost.
  77                         $matches = $search->getNearMatcher( $this->getConfig() )
  78                                 ->getNearMatchResultSet( $params['search'] );
  79                 } else {
  80                         // We default to title searches; this is a terrible legacy
  81                         // of the way we initially set up the MySQL fulltext-based
  82                         // search engine with separate title and text fields.
  83                         // In the future, the default should be for a combined index.
  84                         $what = 'title';
  85                         $matches = $search->searchTitle( $query );
  86
  87                         // Not all search engines support a separate title search,
  88                         // for instance the Lucene-based engine we use on Wikipedia.
  89                         // In this case, fall back to full-text search (which will
  90                         // include titles in it!)
  91                         if ( is_null( $matches ) ) {
  92                                 $what = 'text';
  93                                 $matches = $search->searchText( $query );
  94                         }
  95                 }
  96
  97                 if ( $matches instanceof Status ) {
  98                         $status = $matches;
  99                         $matches = $status->getValue();
 100                 } else {
 101                         $status = null;
 102                 }
 103
 104                 if ( $status ) {
 105                         if ( $status->isOK() ) {
 106                                 $this->getMain()->getErrorFormatter()->addMessagesFromStatus(
 107                                         $this->getModuleName(),
 108                                         $status
 109                                 );
 110                         } else {
 111                                 $this->dieStatus( $status );
 112                         }
 113                 } elseif ( is_null( $matches ) ) {
 114                         $this->dieWithError( [ 'apierror-searchdisabled', $what ], "search-{$what}-disabled" );
 115                 }
 116
 117                 if ( $resultPageSet === null ) {
 118                         $apiResult = $this->getResult();
 119                         // Add search meta data to result
 120                         if ( isset( $searchInfo['totalhits'] ) ) {
 121                                 $totalhits = $matches->getTotalHits();
 122                                 if ( $totalhits !== null ) {
 123                                         $apiResult->addValue( [ 'query', 'searchinfo' ],
 124                                                 'totalhits', $totalhits );
 125                                 }
 126                         }
 127                         if ( isset( $searchInfo['suggestion'] ) && $matches->hasSuggestion() ) {
 128                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 129                                         'suggestion', $matches->getSuggestionQuery() );
 130                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 131                                         'suggestionsnippet', $matches->getSuggestionSnippet() );
 132                         }
 133                         if ( isset( $searchInfo['rewrittenquery'] ) && $matches->hasRewrittenQuery() ) {
 134                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 135                                         'rewrittenquery', $matches->getQueryAfterRewrite() );
 136                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 137                                         'rewrittenquerysnippet', $matches->getQueryAfterRewriteSnippet() );
 138                         }
 139                 }
 140
 141                 // Add the search results to the result
 142                 $terms = $wgContLang->convertForSearchResult( $matches->termMatches() );
 143                 $titles = [];
 144                 $count = 0;
 145                 $limit = $params['limit'];
 146
 147                 foreach ( $matches as $result ) {
 148                         if ( ++$count > $limit ) {
 149                                 // We've reached the one extra which shows that there are
 150                                 // additional items to be had. Stop here...
 151                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $params['limit'] );
 152                                 break;
 153                         }
 154
 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, $terms );
 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, $terms, 'additional',
 183                                 SearchResultSet::INLINE_RESULTS );
 184                 }
 185
 186                 // Interwiki results outside main result set
 187                 if ( $interwiki && $resultPageSet === null ) {
 188                         $this->addInterwikiResults( $matches, $apiResult, $prop, $terms, 'interwiki',
 189                                 SearchResultSet::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          * @param array        $terms Terms list
 216          * @return array|null Result data or null if result is broken in some way.
 217          */
 218         private function getSearchResultData( SearchResult $result, $prop, $terms ) {
 219                 // Silently skip broken and missing titles
 220                 if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 221                         return null;
 222                 }
 223
 224                 $vals = [];
 225
 226                 $title = $result->getTitle();
 227                 ApiQueryBase::addTitleInfo( $vals, $title );
 228                 $vals['pageid'] = $title->getArticleID();
 229
 230                 if ( isset( $prop['size'] ) ) {
 231                         $vals['size'] = $result->getByteSize();
 232                 }
 233                 if ( isset( $prop['wordcount'] ) ) {
 234                         $vals['wordcount'] = $result->getWordCount();
 235                 }
 236                 if ( isset( $prop['snippet'] ) ) {
 237                         $vals['snippet'] = $result->getTextSnippet( $terms );
 238                 }
 239                 if ( isset( $prop['timestamp'] ) ) {
 240                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $result->getTimestamp() );
 241                 }
 242                 if ( isset( $prop['titlesnippet'] ) ) {
 243                         $vals['titlesnippet'] = $result->getTitleSnippet();
 244                 }
 245                 if ( isset( $prop['categorysnippet'] ) ) {
 246                         $vals['categorysnippet'] = $result->getCategorySnippet();
 247                 }
 248                 if ( !is_null( $result->getRedirectTitle() ) ) {
 249                         if ( isset( $prop['redirecttitle'] ) ) {
 250                                 $vals['redirecttitle'] = $result->getRedirectTitle()->getPrefixedText();
 251                         }
 252                         if ( isset( $prop['redirectsnippet'] ) ) {
 253                                 $vals['redirectsnippet'] = $result->getRedirectSnippet();
 254                         }
 255                 }
 256                 if ( !is_null( $result->getSectionTitle() ) ) {
 257                         if ( isset( $prop['sectiontitle'] ) ) {
 258                                 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
 259                         }
 260                         if ( isset( $prop['sectionsnippet'] ) ) {
 261                                 $vals['sectionsnippet'] = $result->getSectionSnippet();
 262                         }
 263                 }
 264                 if ( isset( $prop['isfilematch'] ) ) {
 265                         $vals['isfilematch'] = $result->isFileMatch();
 266                 }
 267
 268                 if ( isset( $prop['extensiondata'] ) ) {
 269                         $extra = $result->getExtensionData();
 270                         // Add augmented data to the result. The data would be organized as a map:
 271                         // augmentorName => data
 272                         if ( $extra ) {
 273                                 $vals['extensiondata'] = ApiResult::addMetadataToResultVars( $extra );
 274                         }
 275                 }
 276
 277                 return $vals;
 278         }
 279
 280         /**
 281          * Add interwiki results as a section in query results.
 282          * @param SearchResultSet $matches
 283          * @param ApiResult       $apiResult
 284          * @param array           $prop Props to extract (as keys)
 285          * @param array           $terms Terms list
 286          * @param string          $section Section name where results would go
 287          * @param int             $type Interwiki result type
 288          * @return int|null Number of total hits in the data or null if none was produced
 289          */
 290         private function addInterwikiResults(
 291                 SearchResultSet $matches, ApiResult $apiResult, $prop,
 292                 $terms, $section, $type
 293         ) {
 294                 $totalhits = null;
 295                 if ( $matches->hasInterwikiResults( $type ) ) {
 296                         foreach ( $matches->getInterwikiResults( $type ) as $interwikiMatches ) {
 297                                 // Include number of results if requested
 298                                 $totalhits += $interwikiMatches->getTotalHits();
 299
 300                                 foreach ( $interwikiMatches as $result ) {
 301                                         $title = $result->getTitle();
 302                                         $vals = $this->getSearchResultData( $result, $prop, $terms );
 303
 304                                         $vals['namespace'] = $result->getInterwikiNamespaceText();
 305                                         $vals['title'] = $title->getText();
 306                                         $vals['url'] = $title->getFullURL();
 307
 308                                         // Add item to results and see whether it fits
 309                                         $fit = $apiResult->addValue( [
 310                                                         'query',
 311                                                         $section . $this->getModuleName(),
 312                                                         $result->getInterwikiPrefix()
 313                                                 ], null, $vals );
 314
 315                                         if ( !$fit ) {
 316                                                 // We hit the limit. We can't really provide any meaningful
 317                                                 // pagination info so just bail out
 318                                                 break;
 319                                         }
 320                                 }
 321                         }
 322                         if ( $totalhits !== null ) {
 323                                 $apiResult->addValue( [ 'query', $section . 'searchinfo' ], 'totalhits', $totalhits );
 324                                 $apiResult->addIndexedTagName( [
 325                                         'query', $section . $this->getModuleName()
 326                                 ], 'p' );
 327                         }
 328                 }
 329                 return $totalhits;
 330         }
 331
 332         public function getCacheMode( $params ) {
 333                 return 'public';
 334         }
 335
 336         public function getAllowedParams() {
 337                 if ( $this->allowedParams !== null ) {
 338                         return $this->allowedParams;
 339                 }
 340
 341                 $this->allowedParams = $this->buildCommonApiParams() + [
 342                         'what' => [
 343                                 ApiBase::PARAM_TYPE => [
 344                                         'title',
 345                                         'text',
 346                                         'nearmatch',
 347                                 ]
 348                         ],
 349                         'info' => [
 350                                 ApiBase::PARAM_DFLT => 'totalhits|suggestion|rewrittenquery',
 351                                 ApiBase::PARAM_TYPE => [
 352                                         'totalhits',
 353                                         'suggestion',
 354                                         'rewrittenquery',
 355                                 ],
 356                                 ApiBase::PARAM_ISMULTI => true,
 357                         ],
 358                         'prop' => [
 359                                 ApiBase::PARAM_DFLT => 'size|wordcount|timestamp|snippet',
 360                                 ApiBase::PARAM_TYPE => [
 361                                         'size',
 362                                         'wordcount',
 363                                         'timestamp',
 364                                         'snippet',
 365                                         'titlesnippet',
 366                                         'redirecttitle',
 367                                         'redirectsnippet',
 368                                         'sectiontitle',
 369                                         'sectionsnippet',
 370                                         'isfilematch',
 371                                         'categorysnippet',
 372                                         'score', // deprecated
 373                                         'hasrelated', // deprecated
 374                                         'extensiondata',
 375                                 ],
 376                                 ApiBase::PARAM_ISMULTI => true,
 377                                 ApiBase::PARAM_HELP_MSG_PER_VALUE => [],
 378                                 ApiBase::PARAM_DEPRECATED_VALUES => [
 379                                         'score' => true,
 380                                         'hasrelated' => true
 381                                 ],
 382                         ],
 383                         'interwiki' => false,
 384                         'enablerewrites' => false,
 385                 ];
 386
 387                 return $this->allowedParams;
 388         }
 389
 390         public function getSearchProfileParams() {
 391                 return [
 392                         'qiprofile' => [
 393                                 'profile-type' => SearchEngine::FT_QUERY_INDEP_PROFILE_TYPE,
 394                                 'help-message' => 'apihelp-query+search-param-qiprofile',
 395                         ],
 396                 ];
 397         }
 398
 399         protected function getExamplesMessages() {
 400                 return [
 401                         'action=query&list=search&srsearch=meaning'
 402                                 => 'apihelp-query+search-example-simple',
 403                         'action=query&list=search&srwhat=text&srsearch=meaning'
 404                                 => 'apihelp-query+search-example-text',
 405                         'action=query&generator=search&gsrsearch=meaning&prop=info'
 406                                 => 'apihelp-query+search-example-generator',
 407                 ];
 408         }
 409
 410         public function getHelpUrls() {
 411                 return 'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Search';
 412         }
 413 }