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