includes/api/ApiQuerySearch.php

   1 <?php
   2 /**
   3  *
   4  *
   5  * Created on July 30, 2007
   6  *
   7  * Copyright © 2007 Yuri Astrakhan "<Firstname><Lastname>@gmail.com"
   8  *
   9  * This program is free software; you can redistribute it and/or modify
  10  * it under the terms of the GNU General Public License as published by
  11  * the Free Software Foundation; either version 2 of the License, or
  12  * (at your option) any later version.
  13  *
  14  * This program is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  17  * GNU General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU General Public License along
  20  * with this program; if not, write to the Free Software Foundation, Inc.,
  21  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  22  * http://www.gnu.org/copyleft/gpl.html
  23  *
  24  * @file
  25  */
  26
  27 /**
  28  * Query module to perform full text search within wiki titles and content
  29  *
  30  * @ingroup API
  31  */
  32 class ApiQuerySearch extends ApiQueryGeneratorBase {
  33         use SearchApi;
  34
  35         /** @var array list of api allowed params */
  36         private $allowedParams;
  37
  38         public function __construct( ApiQuery $query, $moduleName ) {
  39                 parent::__construct( $query, $moduleName, 'sr' );
  40         }
  41
  42         public function execute() {
  43                 $this->run();
  44         }
  45
  46         public function executeGenerator( $resultPageSet ) {
  47                 $this->run( $resultPageSet );
  48         }
  49
  50         /**
  51          * @param ApiPageSet $resultPageSet
  52          * @return void
  53          */
  54         private function run( $resultPageSet = null ) {
  55                 global $wgContLang;
  56                 $params = $this->extractRequestParams();
  57
  58                 // Extract parameters
  59                 $query = $params['search'];
  60                 $what = $params['what'];
  61                 $interwiki = $params['interwiki'];
  62                 $searchInfo = array_flip( $params['info'] );
  63                 $prop = array_flip( $params['prop'] );
  64
  65                 // Create search engine instance and set options
  66                 $search = $this->buildSearchEngine( $params );
  67                 $search->setFeatureData( 'rewrite', (bool)$params['enablerewrites'] );
  68                 $search->setFeatureData( 'interwiki', (bool)$interwiki );
  69
  70                 $query = $search->transformSearchTerm( $query );
  71                 $query = $search->replacePrefixes( $query );
  72
  73                 // Perform the actual search
  74                 if ( $what == 'text' ) {
  75                         $matches = $search->searchText( $query );
  76                 } elseif ( $what == 'title' ) {
  77                         $matches = $search->searchTitle( $query );
  78                 } elseif ( $what == 'nearmatch' ) {
  79                         // near matches must receive the user input as provided, otherwise
  80                         // the near matches within namespaces are lost.
  81                         $matches = $search->getNearMatcher( $this->getConfig() )
  82                                 ->getNearMatchResultSet( $params['search'] );
  83                 } else {
  84                         // We default to title searches; this is a terrible legacy
  85                         // of the way we initially set up the MySQL fulltext-based
  86                         // search engine with separate title and text fields.
  87                         // In the future, the default should be for a combined index.
  88                         $what = 'title';
  89                         $matches = $search->searchTitle( $query );
  90
  91                         // Not all search engines support a separate title search,
  92                         // for instance the Lucene-based engine we use on Wikipedia.
  93                         // In this case, fall back to full-text search (which will
  94                         // include titles in it!)
  95                         if ( is_null( $matches ) ) {
  96                                 $what = 'text';
  97                                 $matches = $search->searchText( $query );
  98                         }
  99                 }
 100
 101                 if ( $matches instanceof Status ) {
 102                         $status = $matches;
 103                         $matches = $status->getValue();
 104                 } else {
 105                         $status = null;
 106                 }
 107
 108                 if ( $status ) {
 109                         if ( $status->isOK() ) {
 110                                 $this->getMain()->getErrorFormatter()->addMessagesFromStatus(
 111                                         $this->getModuleName(),
 112                                         $status
 113                                 );
 114                         } else {
 115                                 $this->dieStatus( $status );
 116                         }
 117                 } elseif ( is_null( $matches ) ) {
 118                         $this->dieWithError( [ 'apierror-searchdisabled', $what ], "search-{$what}-disabled" );
 119                 }
 120
 121                 if ( $resultPageSet === null ) {
 122                         $apiResult = $this->getResult();
 123                         // Add search meta data to result
 124                         if ( isset( $searchInfo['totalhits'] ) ) {
 125                                 $totalhits = $matches->getTotalHits();
 126                                 if ( $totalhits !== null ) {
 127                                         $apiResult->addValue( [ 'query', 'searchinfo' ],
 128                                                 'totalhits', $totalhits );
 129                                 }
 130                         }
 131                         if ( isset( $searchInfo['suggestion'] ) && $matches->hasSuggestion() ) {
 132                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 133                                         'suggestion', $matches->getSuggestionQuery() );
 134                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 135                                         'suggestionsnippet', $matches->getSuggestionSnippet() );
 136                         }
 137                         if ( isset( $searchInfo['rewrittenquery'] ) && $matches->hasRewrittenQuery() ) {
 138                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 139                                         'rewrittenquery', $matches->getQueryAfterRewrite() );
 140                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 141                                         'rewrittenquerysnippet', $matches->getQueryAfterRewriteSnippet() );
 142                         }
 143                 }
 144
 145                 // Add the search results to the result
 146                 $terms = $wgContLang->convertForSearchResult( $matches->termMatches() );
 147                 $titles = [];
 148                 $count = 0;
 149                 $result = $matches->next();
 150                 $limit = $params['limit'];
 151
 152                 while ( $result ) {
 153                         if ( ++$count > $limit ) {
 154                                 // We've reached the one extra which shows that there are
 155                                 // additional items to be had. Stop here...
 156                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $params['limit'] );
 157                                 break;
 158                         }
 159
 160                         // Silently skip broken and missing titles
 161                         if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 162                                 $result = $matches->next();
 163                                 continue;
 164                         }
 165
 166                         if ( $resultPageSet === null ) {
 167                                 $vals = $this->getSearchResultData( $result, $prop, $terms );
 168                                 if ( $vals ) {
 169                                         // Add item to results and see whether it fits
 170                                         $fit = $apiResult->addValue( [ 'query', $this->getModuleName() ], null, $vals );
 171                                         if ( !$fit ) {
 172                                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $count - 1 );
 173                                                 break;
 174                                         }
 175                                 }
 176                         } else {
 177                                 $titles[] = $result->getTitle();
 178                         }
 179
 180                         $result = $matches->next();
 181                 }
 182
 183                 // Here we assume interwiki results do not count with
 184                 // regular search results. We may want to reconsider this
 185                 // if we ever return a lot of interwiki results or want pagination
 186                 // for them.
 187                 // Interwiki results inside main result set
 188                 $canAddInterwiki = (bool)$params['enablerewrites'] && ( $resultPageSet === null );
 189                 if ( $canAddInterwiki ) {
 190                         $this->addInterwikiResults( $matches, $apiResult, $prop, $terms, 'additional',
 191                                 SearchResultSet::INLINE_RESULTS );
 192                 }
 193
 194                 // Interwiki results outside main result set
 195                 if ( $interwiki && $resultPageSet === null ) {
 196                         $this->addInterwikiResults( $matches, $apiResult, $prop, $terms, 'interwiki',
 197                                 SearchResultSet::SECONDARY_RESULTS );
 198                 }
 199
 200                 if ( $resultPageSet === null ) {
 201                         $apiResult->addIndexedTagName( [
 202                                 'query', $this->getModuleName()
 203                         ], 'p' );
 204                 } else {
 205                         $resultPageSet->setRedirectMergePolicy( function ( $current, $new ) {
 206                                 if ( !isset( $current['index'] ) || $new['index'] < $current['index'] ) {
 207                                         $current['index'] = $new['index'];
 208                                 }
 209                                 return $current;
 210                         } );
 211                         $resultPageSet->populateFromTitles( $titles );
 212                         $offset = $params['offset'] + 1;
 213                         foreach ( $titles as $index => $title ) {
 214                                 $resultPageSet->setGeneratorData( $title, [ 'index' => $index + $offset ] );
 215                         }
 216                 }
 217         }
 218
 219         /**
 220          * Assemble search result data.
 221          * @param SearchResult $result Search result
 222          * @param array        $prop Props to extract (as keys)
 223          * @param array        $terms Terms list
 224          * @return array|null Result data or null if result is broken in some way.
 225          */
 226         private function getSearchResultData( SearchResult $result, $prop, $terms ) {
 227                 // Silently skip broken and missing titles
 228                 if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 229                         return null;
 230                 }
 231
 232                 $vals = [];
 233
 234                 $title = $result->getTitle();
 235                 ApiQueryBase::addTitleInfo( $vals, $title );
 236                 $vals['pageid'] = $title->getArticleID();
 237
 238                 if ( isset( $prop['size'] ) ) {
 239                         $vals['size'] = $result->getByteSize();
 240                 }
 241                 if ( isset( $prop['wordcount'] ) ) {
 242                         $vals['wordcount'] = $result->getWordCount();
 243                 }
 244                 if ( isset( $prop['snippet'] ) ) {
 245                         $vals['snippet'] = $result->getTextSnippet( $terms );
 246                 }
 247                 if ( isset( $prop['timestamp'] ) ) {
 248                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $result->getTimestamp() );
 249                 }
 250                 if ( isset( $prop['titlesnippet'] ) ) {
 251                         $vals['titlesnippet'] = $result->getTitleSnippet();
 252                 }
 253                 if ( isset( $prop['categorysnippet'] ) ) {
 254                         $vals['categorysnippet'] = $result->getCategorySnippet();
 255                 }
 256                 if ( !is_null( $result->getRedirectTitle() ) ) {
 257                         if ( isset( $prop['redirecttitle'] ) ) {
 258                                 $vals['redirecttitle'] = $result->getRedirectTitle()->getPrefixedText();
 259                         }
 260                         if ( isset( $prop['redirectsnippet'] ) ) {
 261                                 $vals['redirectsnippet'] = $result->getRedirectSnippet();
 262                         }
 263                 }
 264                 if ( !is_null( $result->getSectionTitle() ) ) {
 265                         if ( isset( $prop['sectiontitle'] ) ) {
 266                                 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
 267                         }
 268                         if ( isset( $prop['sectionsnippet'] ) ) {
 269                                 $vals['sectionsnippet'] = $result->getSectionSnippet();
 270                         }
 271                 }
 272                 if ( isset( $prop['isfilematch'] ) ) {
 273                         $vals['isfilematch'] = $result->isFileMatch();
 274                 }
 275                 return $vals;
 276         }
 277
 278         /**
 279          * Add interwiki results as a section in query results.
 280          * @param SearchResultSet $matches
 281          * @param ApiResult       $apiResult
 282          * @param array           $prop Props to extract (as keys)
 283          * @param array           $terms Terms list
 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                 SearchResultSet $matches, ApiResult $apiResult, $prop,
 290                 $terms, $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                                 $result = $interwikiMatches->next();
 299                                 while ( $result ) {
 300                                         $title = $result->getTitle();
 301                                         $vals = $this->getSearchResultData( $result, $prop, $terms );
 302
 303                                         $vals['namespace'] = $result->getInterwikiNamespaceText();
 304                                         $vals['title'] = $title->getText();
 305                                         $vals['url'] = $title->getFullURL();
 306
 307                                         // Add item to results and see whether it fits
 308                                         $fit = $apiResult->addValue( [
 309                                                         'query',
 310                                                         $section . $this->getModuleName(),
 311                                                         $result->getInterwikiPrefix()
 312                                                 ], null, $vals );
 313
 314                                         if ( !$fit ) {
 315                                                 // We hit the limit. We can't really provide any meaningful
 316                                                 // pagination info so just bail out
 317                                                 break;
 318                                         }
 319
 320                                         $result = $interwikiMatches->next();
 321                                 }
 322                         }
 323                         if ( $totalhits !== null ) {
 324                                 $apiResult->addValue( [ 'query', $section . 'searchinfo' ], 'totalhits', $totalhits );
 325                                 $apiResult->addIndexedTagName( [
 326                                         'query', $section . $this->getModuleName()
 327                                 ], 'p' );
 328                         }
 329                 }
 330                 return $totalhits;
 331         }
 332
 333         public function getCacheMode( $params ) {
 334                 return 'public';
 335         }
 336
 337         public function getAllowedParams() {
 338                 if ( $this->allowedParams !== null ) {
 339                         return $this->allowedParams;
 340                 }
 341
 342                 $this->allowedParams = $this->buildCommonApiParams() + [
 343                         'what' => [
 344                                 ApiBase::PARAM_TYPE => [
 345                                         'title',
 346                                         'text',
 347                                         'nearmatch',
 348                                 ]
 349                         ],
 350                         'info' => [
 351                                 ApiBase::PARAM_DFLT => 'totalhits|suggestion|rewrittenquery',
 352                                 ApiBase::PARAM_TYPE => [
 353                                         'totalhits',
 354                                         'suggestion',
 355                                         'rewrittenquery',
 356                                 ],
 357                                 ApiBase::PARAM_ISMULTI => true,
 358                         ],
 359                         'prop' => [
 360                                 ApiBase::PARAM_DFLT => 'size|wordcount|timestamp|snippet',
 361                                 ApiBase::PARAM_TYPE => [
 362                                         'size',
 363                                         'wordcount',
 364                                         'timestamp',
 365                                         'snippet',
 366                                         'titlesnippet',
 367                                         'redirecttitle',
 368                                         'redirectsnippet',
 369                                         'sectiontitle',
 370                                         'sectionsnippet',
 371                                         'isfilematch',
 372                                         'categorysnippet',
 373                                         'score', // deprecated
 374                                         'hasrelated', // deprecated
 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 }