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