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 use MediaWiki\MediaWikiServices;
  28
  29 /**
  30  * Query module to perform full text search within wiki titles and content
  31  *
  32  * @ingroup API
  33  */
  34 class ApiQuerySearch extends ApiQueryGeneratorBase {
  35         use SearchApi;
  36
  37         /** @var array list of api allowed params */
  38         private $allowedParams;
  39
  40         /**
  41          * When $wgSearchType is null, $wgSearchAlternatives[0] is null. Null isn't
  42          * a valid option for an array for PARAM_TYPE, so we'll use a fake name
  43          * that can't possibly be a class name and describes what the null behavior
  44          * does
  45          */
  46         const BACKEND_NULL_PARAM = 'database-backed';
  47
  48         public function __construct( ApiQuery $query, $moduleName ) {
  49                 parent::__construct( $query, $moduleName, 'sr' );
  50         }
  51
  52         public function execute() {
  53                 $this->run();
  54         }
  55
  56         public function executeGenerator( $resultPageSet ) {
  57                 $this->run( $resultPageSet );
  58         }
  59
  60         /**
  61          * @param ApiPageSet $resultPageSet
  62          * @return void
  63          */
  64         private function run( $resultPageSet = null ) {
  65                 global $wgContLang;
  66                 $params = $this->extractRequestParams();
  67
  68                 if ( isset( $params['backend'] ) && $params['backend'] == self::BACKEND_NULL_PARAM ) {
  69                         unset( $params['backend'] );
  70                 }
  71
  72                 // Extract parameters
  73                 $query = $params['search'];
  74                 $what = $params['what'];
  75                 $interwiki = $params['interwiki'];
  76                 $searchInfo = array_flip( $params['info'] );
  77                 $prop = array_flip( $params['prop'] );
  78
  79                 // Deprecated parameters
  80                 if ( isset( $prop['hasrelated'] ) ) {
  81                         $this->logFeatureUsage( 'action=search&srprop=hasrelated' );
  82                         $this->setWarning( 'srprop=hasrelated has been deprecated' );
  83                 }
  84                 if ( isset( $prop['score'] ) ) {
  85                         $this->logFeatureUsage( 'action=search&srprop=score' );
  86                         $this->setWarning( 'srprop=score has been deprecated' );
  87                 }
  88
  89                 // Create search engine instance and set options
  90                 $search = $this->buildSearchEngine( $params );
  91                 $search->setFeatureData( 'rewrite', (bool)$params['enablerewrites'] );
  92
  93                 $query = $search->transformSearchTerm( $query );
  94                 $query = $search->replacePrefixes( $query );
  95
  96                 // Perform the actual search
  97                 if ( $what == 'text' ) {
  98                         $matches = $search->searchText( $query );
  99                 } elseif ( $what == 'title' ) {
 100                         $matches = $search->searchTitle( $query );
 101                 } elseif ( $what == 'nearmatch' ) {
 102                         // near matches must receive the user input as provided, otherwise
 103                         // the near matches within namespaces are lost.
 104                         $matches = $search->getNearMatcher( $this->getConfig() )
 105                                 ->getNearMatchResultSet( $params['search'] );
 106                 } else {
 107                         // We default to title searches; this is a terrible legacy
 108                         // of the way we initially set up the MySQL fulltext-based
 109                         // search engine with separate title and text fields.
 110                         // In the future, the default should be for a combined index.
 111                         $what = 'title';
 112                         $matches = $search->searchTitle( $query );
 113
 114                         // Not all search engines support a separate title search,
 115                         // for instance the Lucene-based engine we use on Wikipedia.
 116                         // In this case, fall back to full-text search (which will
 117                         // include titles in it!)
 118                         if ( is_null( $matches ) ) {
 119                                 $what = 'text';
 120                                 $matches = $search->searchText( $query );
 121                         }
 122                 }
 123                 if ( is_null( $matches ) ) {
 124                         $this->dieUsage( "{$what} search is disabled", "search-{$what}-disabled" );
 125                 } elseif ( $matches instanceof Status && !$matches->isGood() ) {
 126                         $this->dieUsage( $matches->getWikiText( false, false, 'en' ), 'search-error' );
 127                 }
 128
 129                 if ( $resultPageSet === null ) {
 130                         $apiResult = $this->getResult();
 131                         // Add search meta data to result
 132                         if ( isset( $searchInfo['totalhits'] ) ) {
 133                                 $totalhits = $matches->getTotalHits();
 134                                 if ( $totalhits !== null ) {
 135                                         $apiResult->addValue( [ 'query', 'searchinfo' ],
 136                                                 'totalhits', $totalhits );
 137                                 }
 138                         }
 139                         if ( isset( $searchInfo['suggestion'] ) && $matches->hasSuggestion() ) {
 140                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 141                                         'suggestion', $matches->getSuggestionQuery() );
 142                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 143                                         'suggestionsnippet', $matches->getSuggestionSnippet() );
 144                         }
 145                         if ( isset( $searchInfo['rewrittenquery'] ) && $matches->hasRewrittenQuery() ) {
 146                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 147                                         'rewrittenquery', $matches->getQueryAfterRewrite() );
 148                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 149                                         'rewrittenquerysnippet', $matches->getQueryAfterRewriteSnippet() );
 150                         }
 151                 }
 152
 153                 // Add the search results to the result
 154                 $terms = $wgContLang->convertForSearchResult( $matches->termMatches() );
 155                 $titles = [];
 156                 $count = 0;
 157                 $result = $matches->next();
 158                 $limit = $params['limit'];
 159
 160                 while ( $result ) {
 161                         if ( ++$count > $limit ) {
 162                                 // We've reached the one extra which shows that there are
 163                                 // additional items to be had. Stop here...
 164                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $params['limit'] );
 165                                 break;
 166                         }
 167
 168                         // Silently skip broken and missing titles
 169                         if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 170                                 $result = $matches->next();
 171                                 continue;
 172                         }
 173
 174                         $title = $result->getTitle();
 175                         if ( $resultPageSet === null ) {
 176                                 $vals = [];
 177                                 ApiQueryBase::addTitleInfo( $vals, $title );
 178
 179                                 if ( isset( $prop['snippet'] ) ) {
 180                                         $vals['snippet'] = $result->getTextSnippet( $terms );
 181                                 }
 182                                 if ( isset( $prop['size'] ) ) {
 183                                         $vals['size'] = $result->getByteSize();
 184                                 }
 185                                 if ( isset( $prop['wordcount'] ) ) {
 186                                         $vals['wordcount'] = $result->getWordCount();
 187                                 }
 188                                 if ( isset( $prop['timestamp'] ) ) {
 189                                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $result->getTimestamp() );
 190                                 }
 191                                 if ( isset( $prop['titlesnippet'] ) ) {
 192                                         $vals['titlesnippet'] = $result->getTitleSnippet();
 193                                 }
 194                                 if ( isset( $prop['categorysnippet'] ) ) {
 195                                         $vals['categorysnippet'] = $result->getCategorySnippet();
 196                                 }
 197                                 if ( !is_null( $result->getRedirectTitle() ) ) {
 198                                         if ( isset( $prop['redirecttitle'] ) ) {
 199                                                 $vals['redirecttitle'] = $result->getRedirectTitle()->getPrefixedText();
 200                                         }
 201                                         if ( isset( $prop['redirectsnippet'] ) ) {
 202                                                 $vals['redirectsnippet'] = $result->getRedirectSnippet();
 203                                         }
 204                                 }
 205                                 if ( !is_null( $result->getSectionTitle() ) ) {
 206                                         if ( isset( $prop['sectiontitle'] ) ) {
 207                                                 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
 208                                         }
 209                                         if ( isset( $prop['sectionsnippet'] ) ) {
 210                                                 $vals['sectionsnippet'] = $result->getSectionSnippet();
 211                                         }
 212                                 }
 213                                 if ( isset( $prop['isfilematch'] ) ) {
 214                                         $vals['isfilematch'] = $result->isFileMatch();
 215                                 }
 216
 217                                 // Add item to results and see whether it fits
 218                                 $fit = $apiResult->addValue( [ 'query', $this->getModuleName() ],
 219                                         null, $vals );
 220                                 if ( !$fit ) {
 221                                         $this->setContinueEnumParameter( 'offset', $params['offset'] + $count - 1 );
 222                                         break;
 223                                 }
 224                         } else {
 225                                 $titles[] = $title;
 226                         }
 227
 228                         $result = $matches->next();
 229                 }
 230
 231                 $hasInterwikiResults = false;
 232                 $totalhits = null;
 233                 if ( $interwiki && $resultPageSet === null && $matches->hasInterwikiResults() ) {
 234                         foreach ( $matches->getInterwikiResults() as $interwikiMatches ) {
 235                                 $hasInterwikiResults = true;
 236
 237                                 // Include number of results if requested
 238                                 if ( $resultPageSet === null && isset( $searchInfo['totalhits'] ) ) {
 239                                         $totalhits += $interwikiMatches->getTotalHits();
 240                                 }
 241
 242                                 $result = $interwikiMatches->next();
 243                                 while ( $result ) {
 244                                         $title = $result->getTitle();
 245
 246                                         if ( $resultPageSet === null ) {
 247                                                 $vals = [
 248                                                         'namespace' => $result->getInterwikiNamespaceText(),
 249                                                         'title' => $title->getText(),
 250                                                         'url' => $title->getFullUrl(),
 251                                                 ];
 252
 253                                                 // Add item to results and see whether it fits
 254                                                 $fit = $apiResult->addValue(
 255                                                         [ 'query', 'interwiki' . $this->getModuleName(), $result->getInterwikiPrefix() ],
 256                                                         null,
 257                                                         $vals
 258                                                 );
 259
 260                                                 if ( !$fit ) {
 261                                                         // We hit the limit. We can't really provide any meaningful
 262                                                         // pagination info so just bail out
 263                                                         break;
 264                                                 }
 265                                         } else {
 266                                                 $titles[] = $title;
 267                                         }
 268
 269                                         $result = $interwikiMatches->next();
 270                                 }
 271                         }
 272                         if ( $totalhits !== null ) {
 273                                 $apiResult->addValue( [ 'query', 'interwikisearchinfo' ],
 274                                         'totalhits', $totalhits );
 275                         }
 276                 }
 277
 278                 if ( $resultPageSet === null ) {
 279                         $apiResult->addIndexedTagName( [
 280                                 'query', $this->getModuleName()
 281                         ], 'p' );
 282                         if ( $hasInterwikiResults ) {
 283                                 $apiResult->addIndexedTagName( [
 284                                         'query', 'interwiki' . $this->getModuleName()
 285                                 ], 'p' );
 286                         }
 287                 } else {
 288                         $resultPageSet->setRedirectMergePolicy( function ( $current, $new ) {
 289                                 if ( !isset( $current['index'] ) || $new['index'] < $current['index'] ) {
 290                                         $current['index'] = $new['index'];
 291                                 }
 292                                 return $current;
 293                         } );
 294                         $resultPageSet->populateFromTitles( $titles );
 295                         $offset = $params['offset'] + 1;
 296                         foreach ( $titles as $index => $title ) {
 297                                 $resultPageSet->setGeneratorData( $title, [ 'index' => $index + $offset ] );
 298                         }
 299                 }
 300         }
 301
 302         public function getCacheMode( $params ) {
 303                 return 'public';
 304         }
 305
 306         public function getAllowedParams() {
 307                 if ( $this->allowedParams !== null ) {
 308                         return $this->allowedParams;
 309                 }
 310
 311                 $this->allowedParams = [
 312                         'search' => [
 313                                 ApiBase::PARAM_TYPE => 'string',
 314                                 ApiBase::PARAM_REQUIRED => true
 315                         ],
 316                         'namespace' => [
 317                                 ApiBase::PARAM_DFLT => NS_MAIN,
 318                                 ApiBase::PARAM_TYPE => 'namespace',
 319                                 ApiBase::PARAM_ISMULTI => true,
 320                         ],
 321                         'what' => [
 322                                 ApiBase::PARAM_TYPE => [
 323                                         'title',
 324                                         'text',
 325                                         'nearmatch',
 326                                 ]
 327                         ],
 328                         'info' => [
 329                                 ApiBase::PARAM_DFLT => 'totalhits|suggestion|rewrittenquery',
 330                                 ApiBase::PARAM_TYPE => [
 331                                         'totalhits',
 332                                         'suggestion',
 333                                         'rewrittenquery',
 334                                 ],
 335                                 ApiBase::PARAM_ISMULTI => true,
 336                         ],
 337                         'prop' => [
 338                                 ApiBase::PARAM_DFLT => 'size|wordcount|timestamp|snippet',
 339                                 ApiBase::PARAM_TYPE => [
 340                                         'size',
 341                                         'wordcount',
 342                                         'timestamp',
 343                                         'snippet',
 344                                         'titlesnippet',
 345                                         'redirecttitle',
 346                                         'redirectsnippet',
 347                                         'sectiontitle',
 348                                         'sectionsnippet',
 349                                         'isfilematch',
 350                                         'categorysnippet',
 351                                         'score', // deprecated
 352                                         'hasrelated', // deprecated
 353                                 ],
 354                                 ApiBase::PARAM_ISMULTI => true,
 355                                 ApiBase::PARAM_HELP_MSG_PER_VALUE => [],
 356                         ],
 357                         'offset' => [
 358                                 ApiBase::PARAM_DFLT => 0,
 359                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-continue',
 360                         ],
 361                         'limit' => [
 362                                 ApiBase::PARAM_DFLT => 10,
 363                                 ApiBase::PARAM_TYPE => 'limit',
 364                                 ApiBase::PARAM_MIN => 1,
 365                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_BIG1,
 366                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_BIG2
 367                         ],
 368                         'interwiki' => false,
 369                         'enablerewrites' => false,
 370                 ];
 371
 372                 $searchConfig = MediaWikiServices::getInstance()->getSearchEngineConfig();
 373                 $alternatives = $searchConfig->getSearchTypes();
 374                 if ( count( $alternatives ) > 1 ) {
 375                         if ( $alternatives[0] === null ) {
 376                                 $alternatives[0] = self::BACKEND_NULL_PARAM;
 377                         }
 378                         $this->allowedParams['backend'] = [
 379                                 ApiBase::PARAM_DFLT => $searchConfig->getSearchType(),
 380                                 ApiBase::PARAM_TYPE => $alternatives,
 381                         ];
 382                         // @todo: support profile selection when multiple
 383                         // backends are available. The solution could be to
 384                         // merge all possible profiles and let ApiBase
 385                         // subclasses do the check. Making ApiHelp and ApiSandbox
 386                         // comprehensive might be more difficult.
 387                 } else {
 388                         $profileParam = $this->buildProfileApiParam( SearchEngine::FT_QUERY_INDEP_PROFILE_TYPE,
 389                                 'apihelp-query+search-param-qiprofile' );
 390                         if ( $profileParam ) {
 391                                 $this->allowedParams['qiprofile'] = $profileParam;
 392                         }
 393                 }
 394
 395                 return $this->allowedParams;
 396         }
 397
 398         public function getSearchProfileParams() {
 399                 if ( isset( $this->getAllowedParams()['qiprofile'] ) ) {
 400                         return [ SearchEngine::FT_QUERY_INDEP_PROFILE_TYPE => 'qiprofile' ];
 401                 }
 402                 return [];
 403         }
 404
 405         protected function getExamplesMessages() {
 406                 return [
 407                         'action=query&list=search&srsearch=meaning'
 408                                 => 'apihelp-query+search-example-simple',
 409                         'action=query&list=search&srwhat=text&srsearch=meaning'
 410                                 => 'apihelp-query+search-example-text',
 411                         'action=query&generator=search&gsrsearch=meaning&prop=info'
 412                                 => 'apihelp-query+search-example-generator',
 413                 ];
 414         }
 415
 416         public function getHelpUrls() {
 417                 return 'https://www.mediawiki.org/wiki/API:Search';
 418         }
 419 }