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