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