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
  86                 $query = $search->transformSearchTerm( $query );
  87                 $query = $search->replacePrefixes( $query );
  88
  89                 // Perform the actual search
  90                 if ( $what == 'text' ) {
  91                         $matches = $search->searchText( $query );
  92                 } elseif ( $what == 'title' ) {
  93                         $matches = $search->searchTitle( $query );
  94                 } elseif ( $what == 'nearmatch' ) {
  95                         // near matches must receive the user input as provided, otherwise
  96                         // the near matches within namespaces are lost.
  97                         $matches = SearchEngine::getNearMatchResultSet( $params['search'] );
  98                 } else {
  99                         // We default to title searches; this is a terrible legacy
 100                         // of the way we initially set up the MySQL fulltext-based
 101                         // search engine with separate title and text fields.
 102                         // In the future, the default should be for a combined index.
 103                         $what = 'title';
 104                         $matches = $search->searchTitle( $query );
 105
 106                         // Not all search engines support a separate title search,
 107                         // for instance the Lucene-based engine we use on Wikipedia.
 108                         // In this case, fall back to full-text search (which will
 109                         // include titles in it!)
 110                         if ( is_null( $matches ) ) {
 111                                 $what = 'text';
 112                                 $matches = $search->searchText( $query );
 113                         }
 114                 }
 115                 if ( is_null( $matches ) ) {
 116                         $this->dieUsage( "{$what} search is disabled", "search-{$what}-disabled" );
 117                 } elseif ( $matches instanceof Status && !$matches->isGood() ) {
 118                         $this->dieUsage( $matches->getWikiText(), 'search-error' );
 119                 }
 120
 121                 if ( $resultPageSet === null ) {
 122                         $apiResult = $this->getResult();
 123                         // Add search meta data to result
 124                         if ( isset( $searchInfo['totalhits'] ) ) {
 125                                 $totalhits = $matches->getTotalHits();
 126                                 if ( $totalhits !== null ) {
 127                                         $apiResult->addValue( array( 'query', 'searchinfo' ),
 128                                                 'totalhits', $totalhits );
 129                                 }
 130                         }
 131                         if ( isset( $searchInfo['suggestion'] ) && $matches->hasSuggestion() ) {
 132                                 $apiResult->addValue( array( 'query', 'searchinfo' ),
 133                                         'suggestion', $matches->getSuggestionQuery() );
 134                                 $apiResult->addValue( array( 'query', 'searchinfo' ),
 135                                         'suggestionsnippet', $matches->getSuggestionSnippet() );
 136                         }
 137                 }
 138
 139                 // Add the search results to the result
 140                 $terms = $wgContLang->convertForSearchResult( $matches->termMatches() );
 141                 $titles = array();
 142                 $count = 0;
 143                 $result = $matches->next();
 144
 145                 while ( $result ) {
 146                         if ( ++$count > $limit ) {
 147                                 // We've reached the one extra which shows that there are
 148                                 // additional items to be had. Stop here...
 149                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $params['limit'] );
 150                                 break;
 151                         }
 152
 153                         // Silently skip broken and missing titles
 154                         if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 155                                 $result = $matches->next();
 156                                 continue;
 157                         }
 158
 159                         $title = $result->getTitle();
 160                         if ( $resultPageSet === null ) {
 161                                 $vals = array();
 162                                 ApiQueryBase::addTitleInfo( $vals, $title );
 163
 164                                 if ( isset( $prop['snippet'] ) ) {
 165                                         $vals['snippet'] = $result->getTextSnippet( $terms );
 166                                 }
 167                                 if ( isset( $prop['size'] ) ) {
 168                                         $vals['size'] = $result->getByteSize();
 169                                 }
 170                                 if ( isset( $prop['wordcount'] ) ) {
 171                                         $vals['wordcount'] = $result->getWordCount();
 172                                 }
 173                                 if ( isset( $prop['timestamp'] ) ) {
 174                                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $result->getTimestamp() );
 175                                 }
 176                                 if ( isset( $prop['titlesnippet'] ) ) {
 177                                         $vals['titlesnippet'] = $result->getTitleSnippet();
 178                                 }
 179                                 if ( isset( $prop['categorysnippet'] ) ) {
 180                                         $vals['categorysnippet'] = $result->getCategorySnippet();
 181                                 }
 182                                 if ( !is_null( $result->getRedirectTitle() ) ) {
 183                                         if ( isset( $prop['redirecttitle'] ) ) {
 184                                                 $vals['redirecttitle'] = $result->getRedirectTitle()->getPrefixedText();
 185                                         }
 186                                         if ( isset( $prop['redirectsnippet'] ) ) {
 187                                                 $vals['redirectsnippet'] = $result->getRedirectSnippet();
 188                                         }
 189                                 }
 190                                 if ( !is_null( $result->getSectionTitle() ) ) {
 191                                         if ( isset( $prop['sectiontitle'] ) ) {
 192                                                 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
 193                                         }
 194                                         if ( isset( $prop['sectionsnippet'] ) ) {
 195                                                 $vals['sectionsnippet'] = $result->getSectionSnippet();
 196                                         }
 197                                 }
 198                                 if ( isset( $prop['isfilematch'] ) ) {
 199                                         $vals['isfilematch'] = $result->isFileMatch();
 200                                 }
 201
 202                                 // Add item to results and see whether it fits
 203                                 $fit = $apiResult->addValue( array( 'query', $this->getModuleName() ),
 204                                         null, $vals );
 205                                 if ( !$fit ) {
 206                                         $this->setContinueEnumParameter( 'offset', $params['offset'] + $count - 1 );
 207                                         break;
 208                                 }
 209                         } else {
 210                                 $titles[] = $title;
 211                         }
 212
 213                         $result = $matches->next();
 214                 }
 215
 216                 $hasInterwikiResults = false;
 217                 $totalhits = null;
 218                 if ( $interwiki && $resultPageSet === null && $matches->hasInterwikiResults() ) {
 219                         foreach ( $matches->getInterwikiResults() as $matches ) {
 220                                 $matches = $matches->getInterwikiResults();
 221                                 $hasInterwikiResults = true;
 222
 223                                 // Include number of results if requested
 224                                 if ( $resultPageSet === null && isset( $searchInfo['totalhits'] ) ) {
 225                                         $totalhits += $matches->getTotalHits();
 226                                 }
 227
 228                                 $result = $matches->next();
 229                                 while ( $result ) {
 230                                         $title = $result->getTitle();
 231
 232                                         if ( $resultPageSet === null ) {
 233                                                 $vals = array(
 234                                                         'namespace' => $result->getInterwikiNamespaceText(),
 235                                                         'title' => $title->getText(),
 236                                                         'url' => $title->getFullUrl(),
 237                                                 );
 238
 239                                                 // Add item to results and see whether it fits
 240                                                 $fit = $apiResult->addValue(
 241                                                         array( 'query', 'interwiki' . $this->getModuleName(), $result->getInterwikiPrefix() ),
 242                                                         null,
 243                                                         $vals
 244                                                 );
 245
 246                                                 if ( !$fit ) {
 247                                                         // We hit the limit. We can't really provide any meaningful
 248                                                         // pagination info so just bail out
 249                                                         break;
 250                                                 }
 251                                         } else {
 252                                                 $titles[] = $title;
 253                                         }
 254
 255                                         $result = $matches->next();
 256                                 }
 257                         }
 258                         if ( $totalhits !== null ) {
 259                                 $apiResult->addValue( array( 'query', 'interwikisearchinfo' ),
 260                                         'totalhits', $totalhits );
 261                         }
 262                 }
 263
 264                 if ( $resultPageSet === null ) {
 265                         $apiResult->addIndexedTagName( array(
 266                                 'query', $this->getModuleName()
 267                         ), 'p' );
 268                         if ( $hasInterwikiResults ) {
 269                                 $apiResult->addIndexedTagName( array(
 270                                         'query', 'interwiki' . $this->getModuleName()
 271                                 ), 'p' );
 272                         }
 273                 } else {
 274                         $resultPageSet->populateFromTitles( $titles );
 275                         $offset = $params['offset'] + 1;
 276                         foreach ( $titles as $index => $title ) {
 277                                 $resultPageSet->setGeneratorData( $title, array( 'index' => $index + $offset ) );
 278                         }
 279                 }
 280         }
 281
 282         public function getCacheMode( $params ) {
 283                 return 'public';
 284         }
 285
 286         public function getAllowedParams() {
 287                 $params = array(
 288                         'search' => array(
 289                                 ApiBase::PARAM_TYPE => 'string',
 290                                 ApiBase::PARAM_REQUIRED => true
 291                         ),
 292                         'namespace' => array(
 293                                 ApiBase::PARAM_DFLT => NS_MAIN,
 294                                 ApiBase::PARAM_TYPE => 'namespace',
 295                                 ApiBase::PARAM_ISMULTI => true,
 296                         ),
 297                         'what' => array(
 298                                 ApiBase::PARAM_DFLT => null,
 299                                 ApiBase::PARAM_TYPE => array(
 300                                         'title',
 301                                         'text',
 302                                         'nearmatch',
 303                                 )
 304                         ),
 305                         'info' => array(
 306                                 ApiBase::PARAM_DFLT => 'totalhits|suggestion',
 307                                 ApiBase::PARAM_TYPE => array(
 308                                         'totalhits',
 309                                         'suggestion',
 310                                 ),
 311                                 ApiBase::PARAM_ISMULTI => true,
 312                         ),
 313                         'prop' => array(
 314                                 ApiBase::PARAM_DFLT => 'size|wordcount|timestamp|snippet',
 315                                 ApiBase::PARAM_TYPE => array(
 316                                         'size',
 317                                         'wordcount',
 318                                         'timestamp',
 319                                         'snippet',
 320                                         'titlesnippet',
 321                                         'redirecttitle',
 322                                         'redirectsnippet',
 323                                         'sectiontitle',
 324                                         'sectionsnippet',
 325                                         'isfilematch',
 326                                         'categorysnippet',
 327                                         'score', // deprecated
 328                                         'hasrelated', // deprecated
 329                                 ),
 330                                 ApiBase::PARAM_ISMULTI => true,
 331                                 ApiBase::PARAM_HELP_MSG_PER_VALUE => array(),
 332                         ),
 333                         'offset' => array(
 334                                 ApiBase::PARAM_DFLT => 0,
 335                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-continue',
 336                         ),
 337                         'limit' => array(
 338                                 ApiBase::PARAM_DFLT => 10,
 339                                 ApiBase::PARAM_TYPE => 'limit',
 340                                 ApiBase::PARAM_MIN => 1,
 341                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_SML1,
 342                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_SML2
 343                         ),
 344                         'interwiki' => false,
 345                 );
 346
 347                 $alternatives = SearchEngine::getSearchTypes();
 348                 if ( count( $alternatives ) > 1 ) {
 349                         if ( $alternatives[0] === null ) {
 350                                 $alternatives[0] = self::BACKEND_NULL_PARAM;
 351                         }
 352                         $params['backend'] = array(
 353                                 ApiBase::PARAM_DFLT => $this->getConfig()->get( 'SearchType' ),
 354                                 ApiBase::PARAM_TYPE => $alternatives,
 355                         );
 356                 }
 357
 358                 return $params;
 359         }
 360
 361         protected function getExamplesMessages() {
 362                 return array(
 363                         'action=query&list=search&srsearch=meaning'
 364                                 => 'apihelp-query+search-example-simple',
 365                         'action=query&list=search&srwhat=text&srsearch=meaning'
 366                                 => 'apihelp-query+search-example-text',
 367                         'action=query&generator=search&gsrsearch=meaning&prop=info'
 368                                 => 'apihelp-query+search-example-generator',
 369                 );
 370         }
 371
 372         public function getHelpUrls() {
 373                 return 'https://www.mediawiki.org/wiki/API:Search';
 374         }
 375 }