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