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                 if ( is_null( $matches ) ) {
 111                         $this->dieUsage( "{$what} search is disabled", "search-{$what}-disabled" );
 112                 } elseif ( $matches instanceof Status && !$matches->isGood() ) {
 113                         $this->dieUsage( $matches->getWikiText( false, false, 'en' ), 'search-error' );
 114                 }
 115
 116                 if ( $resultPageSet === null ) {
 117                         $apiResult = $this->getResult();
 118                         // Add search meta data to result
 119                         if ( isset( $searchInfo['totalhits'] ) ) {
 120                                 $totalhits = $matches->getTotalHits();
 121                                 if ( $totalhits !== null ) {
 122                                         $apiResult->addValue( [ 'query', 'searchinfo' ],
 123                                                 'totalhits', $totalhits );
 124                                 }
 125                         }
 126                         if ( isset( $searchInfo['suggestion'] ) && $matches->hasSuggestion() ) {
 127                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 128                                         'suggestion', $matches->getSuggestionQuery() );
 129                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 130                                         'suggestionsnippet', $matches->getSuggestionSnippet() );
 131                         }
 132                         if ( isset( $searchInfo['rewrittenquery'] ) && $matches->hasRewrittenQuery() ) {
 133                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 134                                         'rewrittenquery', $matches->getQueryAfterRewrite() );
 135                                 $apiResult->addValue( [ 'query', 'searchinfo' ],
 136                                         'rewrittenquerysnippet', $matches->getQueryAfterRewriteSnippet() );
 137                         }
 138                 }
 139
 140                 // Add the search results to the result
 141                 $terms = $wgContLang->convertForSearchResult( $matches->termMatches() );
 142                 $titles = [];
 143                 $count = 0;
 144                 $result = $matches->next();
 145                 $limit = $params['limit'];
 146
 147                 while ( $result ) {
 148                         if ( ++$count > $limit ) {
 149                                 // We've reached the one extra which shows that there are
 150                                 // additional items to be had. Stop here...
 151                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $params['limit'] );
 152                                 break;
 153                         }
 154
 155                         // Silently skip broken and missing titles
 156                         if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 157                                 $result = $matches->next();
 158                                 continue;
 159                         }
 160
 161                         $title = $result->getTitle();
 162                         if ( $resultPageSet === null ) {
 163                                 $vals = [];
 164                                 ApiQueryBase::addTitleInfo( $vals, $title );
 165
 166                                 if ( isset( $prop['snippet'] ) ) {
 167                                         $vals['snippet'] = $result->getTextSnippet( $terms );
 168                                 }
 169                                 if ( isset( $prop['size'] ) ) {
 170                                         $vals['size'] = $result->getByteSize();
 171                                 }
 172                                 if ( isset( $prop['wordcount'] ) ) {
 173                                         $vals['wordcount'] = $result->getWordCount();
 174                                 }
 175                                 if ( isset( $prop['timestamp'] ) ) {
 176                                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $result->getTimestamp() );
 177                                 }
 178                                 if ( isset( $prop['titlesnippet'] ) ) {
 179                                         $vals['titlesnippet'] = $result->getTitleSnippet();
 180                                 }
 181                                 if ( isset( $prop['categorysnippet'] ) ) {
 182                                         $vals['categorysnippet'] = $result->getCategorySnippet();
 183                                 }
 184                                 if ( !is_null( $result->getRedirectTitle() ) ) {
 185                                         if ( isset( $prop['redirecttitle'] ) ) {
 186                                                 $vals['redirecttitle'] = $result->getRedirectTitle()->getPrefixedText();
 187                                         }
 188                                         if ( isset( $prop['redirectsnippet'] ) ) {
 189                                                 $vals['redirectsnippet'] = $result->getRedirectSnippet();
 190                                         }
 191                                 }
 192                                 if ( !is_null( $result->getSectionTitle() ) ) {
 193                                         if ( isset( $prop['sectiontitle'] ) ) {
 194                                                 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
 195                                         }
 196                                         if ( isset( $prop['sectionsnippet'] ) ) {
 197                                                 $vals['sectionsnippet'] = $result->getSectionSnippet();
 198                                         }
 199                                 }
 200                                 if ( isset( $prop['isfilematch'] ) ) {
 201                                         $vals['isfilematch'] = $result->isFileMatch();
 202                                 }
 203
 204                                 // Add item to results and see whether it fits
 205                                 $fit = $apiResult->addValue( [ 'query', $this->getModuleName() ],
 206                                         null, $vals );
 207                                 if ( !$fit ) {
 208                                         $this->setContinueEnumParameter( 'offset', $params['offset'] + $count - 1 );
 209                                         break;
 210                                 }
 211                         } else {
 212                                 $titles[] = $title;
 213                         }
 214
 215                         $result = $matches->next();
 216                 }
 217
 218                 $hasInterwikiResults = false;
 219                 $totalhits = null;
 220                 if ( $interwiki && $resultPageSet === null && $matches->hasInterwikiResults() ) {
 221                         foreach ( $matches->getInterwikiResults() as $interwikiMatches ) {
 222                                 $hasInterwikiResults = true;
 223
 224                                 // Include number of results if requested
 225                                 if ( $resultPageSet === null && isset( $searchInfo['totalhits'] ) ) {
 226                                         $totalhits += $interwikiMatches->getTotalHits();
 227                                 }
 228
 229                                 $result = $interwikiMatches->next();
 230                                 while ( $result ) {
 231                                         $title = $result->getTitle();
 232
 233                                         if ( $resultPageSet === null ) {
 234                                                 $vals = [
 235                                                         'namespace' => $result->getInterwikiNamespaceText(),
 236                                                         'title' => $title->getText(),
 237                                                         'url' => $title->getFullURL(),
 238                                                 ];
 239
 240                                                 // Add item to results and see whether it fits
 241                                                 $fit = $apiResult->addValue(
 242                                                         [ 'query', 'interwiki' . $this->getModuleName(), $result->getInterwikiPrefix() ],
 243                                                         null,
 244                                                         $vals
 245                                                 );
 246
 247                                                 if ( !$fit ) {
 248                                                         // We hit the limit. We can't really provide any meaningful
 249                                                         // pagination info so just bail out
 250                                                         break;
 251                                                 }
 252                                         } else {
 253                                                 $titles[] = $title;
 254                                         }
 255
 256                                         $result = $interwikiMatches->next();
 257                                 }
 258                         }
 259                         if ( $totalhits !== null ) {
 260                                 $apiResult->addValue( [ 'query', 'interwikisearchinfo' ],
 261                                         'totalhits', $totalhits );
 262                         }
 263                 }
 264
 265                 if ( $resultPageSet === null ) {
 266                         $apiResult->addIndexedTagName( [
 267                                 'query', $this->getModuleName()
 268                         ], 'p' );
 269                         if ( $hasInterwikiResults ) {
 270                                 $apiResult->addIndexedTagName( [
 271                                         'query', 'interwiki' . $this->getModuleName()
 272                                 ], 'p' );
 273                         }
 274                 } else {
 275                         $resultPageSet->setRedirectMergePolicy( function ( $current, $new ) {
 276                                 if ( !isset( $current['index'] ) || $new['index'] < $current['index'] ) {
 277                                         $current['index'] = $new['index'];
 278                                 }
 279                                 return $current;
 280                         } );
 281                         $resultPageSet->populateFromTitles( $titles );
 282                         $offset = $params['offset'] + 1;
 283                         foreach ( $titles as $index => $title ) {
 284                                 $resultPageSet->setGeneratorData( $title, [ 'index' => $index + $offset ] );
 285                         }
 286                 }
 287         }
 288
 289         public function getCacheMode( $params ) {
 290                 return 'public';
 291         }
 292
 293         public function getAllowedParams() {
 294                 if ( $this->allowedParams !== null ) {
 295                         return $this->allowedParams;
 296                 }
 297
 298                 $this->allowedParams = $this->buildCommonApiParams() + [
 299                         'what' => [
 300                                 ApiBase::PARAM_TYPE => [
 301                                         'title',
 302                                         'text',
 303                                         'nearmatch',
 304                                 ]
 305                         ],
 306                         'info' => [
 307                                 ApiBase::PARAM_DFLT => 'totalhits|suggestion|rewrittenquery',
 308                                 ApiBase::PARAM_TYPE => [
 309                                         'totalhits',
 310                                         'suggestion',
 311                                         'rewrittenquery',
 312                                 ],
 313                                 ApiBase::PARAM_ISMULTI => true,
 314                         ],
 315                         'prop' => [
 316                                 ApiBase::PARAM_DFLT => 'size|wordcount|timestamp|snippet',
 317                                 ApiBase::PARAM_TYPE => [
 318                                         'size',
 319                                         'wordcount',
 320                                         'timestamp',
 321                                         'snippet',
 322                                         'titlesnippet',
 323                                         'redirecttitle',
 324                                         'redirectsnippet',
 325                                         'sectiontitle',
 326                                         'sectionsnippet',
 327                                         'isfilematch',
 328                                         'categorysnippet',
 329                                         'score', // deprecated
 330                                         'hasrelated', // deprecated
 331                                 ],
 332                                 ApiBase::PARAM_ISMULTI => true,
 333                                 ApiBase::PARAM_HELP_MSG_PER_VALUE => [],
 334                         ],
 335                         'interwiki' => false,
 336                         'enablerewrites' => false,
 337                 ];
 338
 339                 return $this->allowedParams;
 340         }
 341
 342         public function getSearchProfileParams() {
 343                 return [
 344                         'qiprofile' => [
 345                                 'profile-type' => SearchEngine::FT_QUERY_INDEP_PROFILE_TYPE,
 346                                 'help-message' => 'apihelp-query+search-param-qiprofile',
 347                         ],
 348                 ];
 349         }
 350
 351         protected function getExamplesMessages() {
 352                 return [
 353                         'action=query&list=search&srsearch=meaning'
 354                                 => 'apihelp-query+search-example-simple',
 355                         'action=query&list=search&srwhat=text&srsearch=meaning'
 356                                 => 'apihelp-query+search-example-text',
 357                         'action=query&generator=search&gsrsearch=meaning&prop=info'
 358                                 => 'apihelp-query+search-example-generator',
 359                 ];
 360         }
 361
 362         public function getHelpUrls() {
 363                 return 'https://www.mediawiki.org/wiki/API:Search';
 364         }
 365 }