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