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                 $result = $matches->next();
 146                 $limit = $params['limit'];
 147
 148                 while ( $result ) {
 149                         if ( ++$count > $limit ) {
 150                                 // We've reached the one extra which shows that there are
 151                                 // additional items to be had. Stop here...
 152                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $params['limit'] );
 153                                 break;
 154                         }
 155
 156                         // Silently skip broken and missing titles
 157                         if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 158                                 $result = $matches->next();
 159                                 continue;
 160                         }
 161
 162                         if ( $resultPageSet === null ) {
 163                                 $vals = $this->getSearchResultData( $result, $prop, $terms );
 164                                 if ( $vals ) {
 165                                         // Add item to results and see whether it fits
 166                                         $fit = $apiResult->addValue( [ 'query', $this->getModuleName() ], null, $vals );
 167                                         if ( !$fit ) {
 168                                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $count - 1 );
 169                                                 break;
 170                                         }
 171                                 }
 172                         } else {
 173                                 $titles[] = $result->getTitle();
 174                         }
 175
 176                         $result = $matches->next();
 177                 }
 178
 179                 // Here we assume interwiki results do not count with
 180                 // regular search results. We may want to reconsider this
 181                 // if we ever return a lot of interwiki results or want pagination
 182                 // for them.
 183                 // Interwiki results inside main result set
 184                 $canAddInterwiki = (bool)$params['enablerewrites'] && ( $resultPageSet === null );
 185                 if ( $canAddInterwiki ) {
 186                         $this->addInterwikiResults( $matches, $apiResult, $prop, $terms, 'additional',
 187                                 SearchResultSet::INLINE_RESULTS );
 188                 }
 189
 190                 // Interwiki results outside main result set
 191                 if ( $interwiki && $resultPageSet === null ) {
 192                         $this->addInterwikiResults( $matches, $apiResult, $prop, $terms, 'interwiki',
 193                                 SearchResultSet::SECONDARY_RESULTS );
 194                 }
 195
 196                 if ( $resultPageSet === null ) {
 197                         $apiResult->addIndexedTagName( [
 198                                 'query', $this->getModuleName()
 199                         ], 'p' );
 200                 } else {
 201                         $resultPageSet->setRedirectMergePolicy( function ( $current, $new ) {
 202                                 if ( !isset( $current['index'] ) || $new['index'] < $current['index'] ) {
 203                                         $current['index'] = $new['index'];
 204                                 }
 205                                 return $current;
 206                         } );
 207                         $resultPageSet->populateFromTitles( $titles );
 208                         $offset = $params['offset'] + 1;
 209                         foreach ( $titles as $index => $title ) {
 210                                 $resultPageSet->setGeneratorData( $title, [ 'index' => $index + $offset ] );
 211                         }
 212                 }
 213         }
 214
 215         /**
 216          * Assemble search result data.
 217          * @param SearchResult $result Search result
 218          * @param array        $prop Props to extract (as keys)
 219          * @param array        $terms Terms list
 220          * @return array|null Result data or null if result is broken in some way.
 221          */
 222         private function getSearchResultData( SearchResult $result, $prop, $terms ) {
 223                 // Silently skip broken and missing titles
 224                 if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 225                         return null;
 226                 }
 227
 228                 $vals = [];
 229
 230                 $title = $result->getTitle();
 231                 ApiQueryBase::addTitleInfo( $vals, $title );
 232                 $vals['pageid'] = $title->getArticleID();
 233
 234                 if ( isset( $prop['size'] ) ) {
 235                         $vals['size'] = $result->getByteSize();
 236                 }
 237                 if ( isset( $prop['wordcount'] ) ) {
 238                         $vals['wordcount'] = $result->getWordCount();
 239                 }
 240                 if ( isset( $prop['snippet'] ) ) {
 241                         $vals['snippet'] = $result->getTextSnippet( $terms );
 242                 }
 243                 if ( isset( $prop['timestamp'] ) ) {
 244                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $result->getTimestamp() );
 245                 }
 246                 if ( isset( $prop['titlesnippet'] ) ) {
 247                         $vals['titlesnippet'] = $result->getTitleSnippet();
 248                 }
 249                 if ( isset( $prop['categorysnippet'] ) ) {
 250                         $vals['categorysnippet'] = $result->getCategorySnippet();
 251                 }
 252                 if ( !is_null( $result->getRedirectTitle() ) ) {
 253                         if ( isset( $prop['redirecttitle'] ) ) {
 254                                 $vals['redirecttitle'] = $result->getRedirectTitle()->getPrefixedText();
 255                         }
 256                         if ( isset( $prop['redirectsnippet'] ) ) {
 257                                 $vals['redirectsnippet'] = $result->getRedirectSnippet();
 258                         }
 259                 }
 260                 if ( !is_null( $result->getSectionTitle() ) ) {
 261                         if ( isset( $prop['sectiontitle'] ) ) {
 262                                 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
 263                         }
 264                         if ( isset( $prop['sectionsnippet'] ) ) {
 265                                 $vals['sectionsnippet'] = $result->getSectionSnippet();
 266                         }
 267                 }
 268                 if ( isset( $prop['isfilematch'] ) ) {
 269                         $vals['isfilematch'] = $result->isFileMatch();
 270                 }
 271
 272                 if ( isset( $prop['extensiondata'] ) ) {
 273                         $extra = $result->getExtensionData();
 274                         // Add augmented data to the result. The data would be organized as a map:
 275                         // augmentorName => data
 276                         if ( $extra ) {
 277                                 $vals['extensiondata'] = ApiResult::addMetadataToResultVars( $extra );
 278                         }
 279                 }
 280
 281                 return $vals;
 282         }
 283
 284         /**
 285          * Add interwiki results as a section in query results.
 286          * @param SearchResultSet $matches
 287          * @param ApiResult       $apiResult
 288          * @param array           $prop Props to extract (as keys)
 289          * @param array           $terms Terms list
 290          * @param string          $section Section name where results would go
 291          * @param int             $type Interwiki result type
 292          * @return int|null Number of total hits in the data or null if none was produced
 293          */
 294         private function addInterwikiResults(
 295                 SearchResultSet $matches, ApiResult $apiResult, $prop,
 296                 $terms, $section, $type
 297         ) {
 298                 $totalhits = null;
 299                 if ( $matches->hasInterwikiResults( $type ) ) {
 300                         foreach ( $matches->getInterwikiResults( $type ) as $interwikiMatches ) {
 301                                 // Include number of results if requested
 302                                 $totalhits += $interwikiMatches->getTotalHits();
 303
 304                                 $result = $interwikiMatches->next();
 305                                 while ( $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                                         $result = $interwikiMatches->next();
 327                                 }
 328                         }
 329                         if ( $totalhits !== null ) {
 330                                 $apiResult->addValue( [ 'query', $section . 'searchinfo' ], 'totalhits', $totalhits );
 331                                 $apiResult->addIndexedTagName( [
 332                                         'query', $section . $this->getModuleName()
 333                                 ], 'p' );
 334                         }
 335                 }
 336                 return $totalhits;
 337         }
 338
 339         public function getCacheMode( $params ) {
 340                 return 'public';
 341         }
 342
 343         public function getAllowedParams() {
 344                 if ( $this->allowedParams !== null ) {
 345                         return $this->allowedParams;
 346                 }
 347
 348                 $this->allowedParams = $this->buildCommonApiParams() + [
 349                         'what' => [
 350                                 ApiBase::PARAM_TYPE => [
 351                                         'title',
 352                                         'text',
 353                                         'nearmatch',
 354                                 ]
 355                         ],
 356                         'info' => [
 357                                 ApiBase::PARAM_DFLT => 'totalhits|suggestion|rewrittenquery',
 358                                 ApiBase::PARAM_TYPE => [
 359                                         'totalhits',
 360                                         'suggestion',
 361                                         'rewrittenquery',
 362                                 ],
 363                                 ApiBase::PARAM_ISMULTI => true,
 364                         ],
 365                         'prop' => [
 366                                 ApiBase::PARAM_DFLT => 'size|wordcount|timestamp|snippet',
 367                                 ApiBase::PARAM_TYPE => [
 368                                         'size',
 369                                         'wordcount',
 370                                         'timestamp',
 371                                         'snippet',
 372                                         'titlesnippet',
 373                                         'redirecttitle',
 374                                         'redirectsnippet',
 375                                         'sectiontitle',
 376                                         'sectionsnippet',
 377                                         'isfilematch',
 378                                         'categorysnippet',
 379                                         'score', // deprecated
 380                                         'hasrelated', // deprecated
 381                                         'extensiondata',
 382                                 ],
 383                                 ApiBase::PARAM_ISMULTI => true,
 384                                 ApiBase::PARAM_HELP_MSG_PER_VALUE => [],
 385                                 ApiBase::PARAM_DEPRECATED_VALUES => [
 386                                         'score' => true,
 387                                         'hasrelated' => true
 388                                 ],
 389                         ],
 390                         'interwiki' => false,
 391                         'enablerewrites' => false,
 392                 ];
 393
 394                 return $this->allowedParams;
 395         }
 396
 397         public function getSearchProfileParams() {
 398                 return [
 399                         'qiprofile' => [
 400                                 'profile-type' => SearchEngine::FT_QUERY_INDEP_PROFILE_TYPE,
 401                                 'help-message' => 'apihelp-query+search-param-qiprofile',
 402                         ],
 403                 ];
 404         }
 405
 406         protected function getExamplesMessages() {
 407                 return [
 408                         'action=query&list=search&srsearch=meaning'
 409                                 => 'apihelp-query+search-example-simple',
 410                         'action=query&list=search&srwhat=text&srsearch=meaning'
 411                                 => 'apihelp-query+search-example-text',
 412                         'action=query&generator=search&gsrsearch=meaning&prop=info'
 413                                 => 'apihelp-query+search-example-generator',
 414                 ];
 415         }
 416
 417         public function getHelpUrls() {
 418                 return 'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Search';
 419         }
 420 }