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->logFeatureUsage( 'action=search&srprop=hasrelated' );
  68                         $this->setWarning( 'srprop=hasrelated has been deprecated' );
  69                 }
  70                 if ( isset( $prop['score'] ) ) {
  71                         $this->logFeatureUsage( 'action=search&srprop=score' );
  72                         $this->setWarning( 'srprop=score has been deprecated' );
  73                 }
  74
  75                 // Create search engine instance and set options
  76                 $search = $this->buildSearchEngine( $params );
  77                 $search->setFeatureData( 'rewrite', (bool)$params['enablerewrites'] );
  78                 $search->setFeatureData( 'interwiki', (bool)$interwiki );
  79
  80                 $query = $search->transformSearchTerm( $query );
  81                 $query = $search->replacePrefixes( $query );
  82
  83                 // Perform the actual search
  84                 if ( $what == 'text' ) {
  85                         $matches = $search->searchText( $query );
  86                 } elseif ( $what == 'title' ) {
  87                         $matches = $search->searchTitle( $query );
  88                 } elseif ( $what == 'nearmatch' ) {
  89                         // near matches must receive the user input as provided, otherwise
  90                         // the near matches within namespaces are lost.
  91                         $matches = $search->getNearMatcher( $this->getConfig() )
  92                                 ->getNearMatchResultSet( $params['search'] );
  93                 } else {
  94                         // We default to title searches; this is a terrible legacy
  95                         // of the way we initially set up the MySQL fulltext-based
  96                         // search engine with separate title and text fields.
  97                         // In the future, the default should be for a combined index.
  98                         $what = 'title';
  99                         $matches = $search->searchTitle( $query );
 100
 101                         // Not all search engines support a separate title search,
 102                         // for instance the Lucene-based engine we use on Wikipedia.
 103                         // In this case, fall back to full-text search (which will
 104                         // include titles in it!)
 105                         if ( is_null( $matches ) ) {
 106                                 $what = 'text';
 107                                 $matches = $search->searchText( $query );
 108                         }
 109                 }
 110
 111                 if ( $matches instanceof Status ) {
 112                         $status = $matches;
 113                         $matches = $status->getValue();
 114                 } else {
 115                         $status = null;
 116                 }
 117
 118                 if ( $status ) {
 119                         if ( $status->isOK() ) {
 120                                 $this->getMain()->getErrorFormatter()->addMessagesFromStatus(
 121                                         $this->getModuleName(),
 122                                         $status
 123                                 );
 124                         } else {
 125                                 $this->dieUsage( $status->getWikiText( false, false, 'en' ), 'search-error' );
 126                         }
 127                 } elseif ( is_null( $matches ) ) {
 128                         $this->dieUsage( "{$what} search is disabled", "search-{$what}-disabled" );
 129                 }
 130
 131                 if ( $resultPageSet === null ) {
 132                         $apiResult = $this->getResult();
 133                         // Add search meta data to result
 134                         if ( isset( $searchInfo['totalhits'] ) ) {
 135                                 $totalhits = $matches->getTotalHits();
 136                                 if ( $totalhits !== null ) {
 137                                         $apiResult->addValue( [ 'query', 'searchinfo' ],
 138                                                 'totalhits', $totalhits );
 139                                 }
 140                         }
 141                         if ( isset( $searchInfo['suggestion'] ) && $matches->hasSuggestion() ) {
 142                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 143                                         'suggestion', $matches->getSuggestionQuery() );
 144                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 145                                         'suggestionsnippet', $matches->getSuggestionSnippet() );
 146                         }
 147                         if ( isset( $searchInfo['rewrittenquery'] ) && $matches->hasRewrittenQuery() ) {
 148                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 149                                         'rewrittenquery', $matches->getQueryAfterRewrite() );
 150                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 151                                         'rewrittenquerysnippet', $matches->getQueryAfterRewriteSnippet() );
 152                         }
 153                 }
 154
 155                 // Add the search results to the result
 156                 $terms = $wgContLang->convertForSearchResult( $matches->termMatches() );
 157                 $titles = [];
 158                 $count = 0;
 159                 $result = $matches->next();
 160                 $limit = $params['limit'];
 161
 162                 while ( $result ) {
 163                         if ( ++$count > $limit ) {
 164                                 // We've reached the one extra which shows that there are
 165                                 // additional items to be had. Stop here...
 166                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $params['limit'] );
 167                                 break;
 168                         }
 169
 170                         // Silently skip broken and missing titles
 171                         if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 172                                 return null;
 173                         }
 174
 175                         if ( $resultPageSet === null ) {
 176                                 $vals = $this->getSearchResultData( $result, $prop, $terms );
 177                                 if ( $vals ) {
 178                                         // Add item to results and see whether it fits
 179                                         $fit = $apiResult->addValue( [ 'query', $this->getModuleName() ], null, $vals );
 180                                         if ( !$fit ) {
 181                                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $count - 1 );
 182                                                 break;
 183                                         }
 184                                 }
 185                         } else {
 186                                 $titles[] = $result->getTitle();
 187                         }
 188
 189                         $result = $matches->next();
 190                 }
 191
 192                 // Here we assume interwiki results do not count with
 193                 // regular search results. We may want to reconsider this
 194                 // if we ever return a lot of interwiki results or want pagination
 195                 // for them.
 196                 // Interwiki results inside main result set
 197                 $canAddInterwiki = (bool)$params['enablerewrites'] && ( $resultPageSet === null );
 198                 if ( $canAddInterwiki ) {
 199                         $this->addInterwikiResults( $matches, $apiResult, $prop, $terms, 'additional',
 200                                 SearchResultSet::INLINE_RESULTS );
 201                 }
 202
 203                 // Interwiki results outside main result set
 204                 if ( $interwiki && $resultPageSet === null ) {
 205                         $this->addInterwikiResults( $matches, $apiResult, $prop, $terms, 'interwiki',
 206                                 SearchResultSet::SECONDARY_RESULTS );
 207                 }
 208
 209                 if ( $resultPageSet === null ) {
 210                         $apiResult->addIndexedTagName( [
 211                                 'query', $this->getModuleName()
 212                         ], 'p' );
 213                 } else {
 214                         $resultPageSet->setRedirectMergePolicy( function ( $current, $new ) {
 215                                 if ( !isset( $current['index'] ) || $new['index'] < $current['index'] ) {
 216                                         $current['index'] = $new['index'];
 217                                 }
 218                                 return $current;
 219                         } );
 220                         $resultPageSet->populateFromTitles( $titles );
 221                         $offset = $params['offset'] + 1;
 222                         foreach ( $titles as $index => $title ) {
 223                                 $resultPageSet->setGeneratorData( $title, [ 'index' => $index + $offset ] );
 224                         }
 225                 }
 226         }
 227
 228         /**
 229          * Assemble search result data.
 230          * @param SearchResult $result Search result
 231          * @param array        $prop Props to extract (as keys)
 232          * @param array        $terms Terms list
 233          * @return array|null Result data or null if result is broken in some way.
 234          */
 235         private function getSearchResultData( SearchResult $result, $prop, $terms ) {
 236                 // Silently skip broken and missing titles
 237                 if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 238                         return null;
 239                 }
 240
 241                 $vals = [];
 242
 243                 $title = $result->getTitle();
 244                 ApiQueryBase::addTitleInfo( $vals, $title );
 245
 246                 if ( isset( $prop['size'] ) ) {
 247                         $vals['size'] = $result->getByteSize();
 248                 }
 249                 if ( isset( $prop['wordcount'] ) ) {
 250                         $vals['wordcount'] = $result->getWordCount();
 251                 }
 252                 if ( isset( $prop['snippet'] ) ) {
 253                         $vals['snippet'] = $result->getTextSnippet( $terms );
 254                 }
 255                 if ( isset( $prop['timestamp'] ) ) {
 256                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $result->getTimestamp() );
 257                 }
 258                 if ( isset( $prop['titlesnippet'] ) ) {
 259                         $vals['titlesnippet'] = $result->getTitleSnippet();
 260                 }
 261                 if ( isset( $prop['categorysnippet'] ) ) {
 262                         $vals['categorysnippet'] = $result->getCategorySnippet();
 263                 }
 264                 if ( !is_null( $result->getRedirectTitle() ) ) {
 265                         if ( isset( $prop['redirecttitle'] ) ) {
 266                                 $vals['redirecttitle'] = $result->getRedirectTitle()->getPrefixedText();
 267                         }
 268                         if ( isset( $prop['redirectsnippet'] ) ) {
 269                                 $vals['redirectsnippet'] = $result->getRedirectSnippet();
 270                         }
 271                 }
 272                 if ( !is_null( $result->getSectionTitle() ) ) {
 273                         if ( isset( $prop['sectiontitle'] ) ) {
 274                                 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
 275                         }
 276                         if ( isset( $prop['sectionsnippet'] ) ) {
 277                                 $vals['sectionsnippet'] = $result->getSectionSnippet();
 278                         }
 279                 }
 280                 if ( isset( $prop['isfilematch'] ) ) {
 281                         $vals['isfilematch'] = $result->isFileMatch();
 282                 }
 283                 return $vals;
 284         }
 285
 286         /**
 287          * Add interwiki results as a section in query results.
 288          * @param SearchResultSet $matches
 289          * @param ApiResult       $apiResult
 290          * @param array           $prop Props to extract (as keys)
 291          * @param array           $terms Terms list
 292          * @param string          $section Section name where results would go
 293          * @param int             $type Interwiki result type
 294          * @return int|null Number of total hits in the data or null if none was produced
 295          */
 296         private function addInterwikiResults(
 297                 SearchResultSet $matches, ApiResult $apiResult, $prop,
 298                 $terms, $section, $type
 299         ) {
 300                 $totalhits = null;
 301                 if ( $matches->hasInterwikiResults( $type ) ) {
 302                         foreach ( $matches->getInterwikiResults( $type ) as $interwikiMatches ) {
 303                                 // Include number of results if requested
 304                                 $totalhits += $interwikiMatches->getTotalHits();
 305
 306                                 $result = $interwikiMatches->next();
 307                                 while ( $result ) {
 308                                         $title = $result->getTitle();
 309                                         $vals = $this->getSearchResultData( $result, $prop, $terms );
 310
 311                                         $vals['namespace'] = $result->getInterwikiNamespaceText();
 312                                         $vals['title'] = $title->getText();
 313                                         $vals['url'] = $title->getFullURL();
 314
 315                                         // Add item to results and see whether it fits
 316                                         $fit = $apiResult->addValue( [
 317                                                         'query',
 318                                                         $section . $this->getModuleName(),
 319                                                         $result->getInterwikiPrefix()
 320                                                 ], null, $vals );
 321
 322                                         if ( !$fit ) {
 323                                                 // We hit the limit. We can't really provide any meaningful
 324                                                 // pagination info so just bail out
 325                                                 break;
 326                                         }
 327
 328                                         $result = $interwikiMatches->next();
 329                                 }
 330                         }
 331                         if ( $totalhits !== null ) {
 332                                 $apiResult->addValue( [ 'query', $section . 'searchinfo' ], 'totalhits', $totalhits );
 333                                 $apiResult->addIndexedTagName( [
 334                                         'query', $section . $this->getModuleName()
 335                                 ], 'p' );
 336                         }
 337                 }
 338                 return $totalhits;
 339         }
 340
 341         public function getCacheMode( $params ) {
 342                 return 'public';
 343         }
 344
 345         public function getAllowedParams() {
 346                 if ( $this->allowedParams !== null ) {
 347                         return $this->allowedParams;
 348                 }
 349
 350                 $this->allowedParams = $this->buildCommonApiParams() + [
 351                         'what' => [
 352                                 ApiBase::PARAM_TYPE => [
 353                                         'title',
 354                                         'text',
 355                                         'nearmatch',
 356                                 ]
 357                         ],
 358                         'info' => [
 359                                 ApiBase::PARAM_DFLT => 'totalhits|suggestion|rewrittenquery',
 360                                 ApiBase::PARAM_TYPE => [
 361                                         'totalhits',
 362                                         'suggestion',
 363                                         'rewrittenquery',
 364                                 ],
 365                                 ApiBase::PARAM_ISMULTI => true,
 366                         ],
 367                         'prop' => [
 368                                 ApiBase::PARAM_DFLT => 'size|wordcount|timestamp|snippet',
 369                                 ApiBase::PARAM_TYPE => [
 370                                         'size',
 371                                         'wordcount',
 372                                         'timestamp',
 373                                         'snippet',
 374                                         'titlesnippet',
 375                                         'redirecttitle',
 376                                         'redirectsnippet',
 377                                         'sectiontitle',
 378                                         'sectionsnippet',
 379                                         'isfilematch',
 380                                         'categorysnippet',
 381                                         'score', // deprecated
 382                                         'hasrelated', // deprecated
 383                                 ],
 384                                 ApiBase::PARAM_ISMULTI => true,
 385                                 ApiBase::PARAM_HELP_MSG_PER_VALUE => [],
 386                         ],
 387                         'interwiki' => false,
 388                         'enablerewrites' => false,
 389                 ];
 390
 391                 return $this->allowedParams;
 392         }
 393
 394         public function getSearchProfileParams() {
 395                 return [
 396                         'qiprofile' => [
 397                                 'profile-type' => SearchEngine::FT_QUERY_INDEP_PROFILE_TYPE,
 398                                 'help-message' => 'apihelp-query+search-param-qiprofile',
 399                         ],
 400                 ];
 401         }
 402
 403         protected function getExamplesMessages() {
 404                 return [
 405                         'action=query&list=search&srsearch=meaning'
 406                                 => 'apihelp-query+search-example-simple',
 407                         'action=query&list=search&srwhat=text&srsearch=meaning'
 408                                 => 'apihelp-query+search-example-text',
 409                         'action=query&generator=search&gsrsearch=meaning&prop=info'
 410                                 => 'apihelp-query+search-example-generator',
 411                 ];
 412         }
 413
 414         public function getHelpUrls() {
 415                 return 'https://www.mediawiki.org/wiki/API:Search';
 416         }
 417 }