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                 $apiResult = $this->getResult();
 120                 // Add search meta data to result
 121                 if ( isset( $searchInfo['totalhits'] ) ) {
 122                         $totalhits = $matches->getTotalHits();
 123                         if ( $totalhits !== null ) {
 124                                 $apiResult->addValue( array( 'query', 'searchinfo' ),
 125                                         'totalhits', $totalhits );
 126                         }
 127                 }
 128                 if ( isset( $searchInfo['suggestion'] ) && $matches->hasSuggestion() ) {
 129                         $apiResult->addValue( array( 'query', 'searchinfo' ),
 130                                 'suggestion', $matches->getSuggestionQuery() );
 131                 }
 132
 133                 // Add the search results to the result
 134                 $terms = $wgContLang->convertForSearchResult( $matches->termMatches() );
 135                 $titles = array();
 136                 $count = 0;
 137                 $result = $matches->next();
 138
 139                 while ( $result ) {
 140                         if ( ++$count > $limit ) {
 141                                 // We've reached the one extra which shows that there are
 142                                 // additional items to be had. Stop here...
 143                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $params['limit'] );
 144                                 break;
 145                         }
 146
 147                         // Silently skip broken and missing titles
 148                         if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 149                                 $result = $matches->next();
 150                                 continue;
 151                         }
 152
 153                         $title = $result->getTitle();
 154                         if ( is_null( $resultPageSet ) ) {
 155                                 $vals = array();
 156                                 ApiQueryBase::addTitleInfo( $vals, $title );
 157
 158                                 if ( isset( $prop['snippet'] ) ) {
 159                                         $vals['snippet'] = $result->getTextSnippet( $terms );
 160                                 }
 161                                 if ( isset( $prop['size'] ) ) {
 162                                         $vals['size'] = $result->getByteSize();
 163                                 }
 164                                 if ( isset( $prop['wordcount'] ) ) {
 165                                         $vals['wordcount'] = $result->getWordCount();
 166                                 }
 167                                 if ( isset( $prop['timestamp'] ) ) {
 168                                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $result->getTimestamp() );
 169                                 }
 170                                 if ( isset( $prop['titlesnippet'] ) ) {
 171                                         $vals['titlesnippet'] = $result->getTitleSnippet();
 172                                 }
 173                                 if ( !is_null( $result->getRedirectTitle() ) ) {
 174                                         if ( isset( $prop['redirecttitle'] ) ) {
 175                                                 $vals['redirecttitle'] = $result->getRedirectTitle();
 176                                         }
 177                                         if ( isset( $prop['redirectsnippet'] ) ) {
 178                                                 $vals['redirectsnippet'] = $result->getRedirectSnippet( $terms );
 179                                         }
 180                                 }
 181                                 if ( !is_null( $result->getSectionTitle() ) ) {
 182                                         if ( isset( $prop['sectiontitle'] ) ) {
 183                                                 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
 184                                         }
 185                                         if ( isset( $prop['sectionsnippet'] ) ) {
 186                                                 $vals['sectionsnippet'] = $result->getSectionSnippet();
 187                                         }
 188                                 }
 189
 190                                 // Add item to results and see whether it fits
 191                                 $fit = $apiResult->addValue( array( 'query', $this->getModuleName() ),
 192                                         null, $vals );
 193                                 if ( !$fit ) {
 194                                         $this->setContinueEnumParameter( 'offset', $params['offset'] + $count - 1 );
 195                                         break;
 196                                 }
 197                         } else {
 198                                 $titles[] = $title;
 199                         }
 200
 201                         $result = $matches->next();
 202                 }
 203
 204                 $hasInterwikiResults = false;
 205                 if ( $interwiki && $resultPageSet === null && $matches->hasInterwikiResults() ) {
 206                         $matches = $matches->getInterwikiResults();
 207                         $hasInterwikiResults = true;
 208
 209                         // Include number of results if requested
 210                         if ( isset( $searchInfo['totalhits'] ) ) {
 211                                 $totalhits = $matches->getTotalHits();
 212                                 if ( $totalhits !== null ) {
 213                                         $apiResult->addValue( array( 'query', 'interwikisearchinfo' ),
 214                                                 'totalhits', $totalhits );
 215                                 }
 216                         }
 217
 218                         $result = $matches->next();
 219                         while ( $result ) {
 220                                 $title = $result->getTitle();
 221                                 $vals = array(
 222                                         'namespace' => $result->getInterwikiNamespaceText(),
 223                                         'title' => $title->getText(),
 224                                         'url' => $title->getFullUrl(),
 225                                 );
 226
 227                                 // Add item to results and see whether it fits
 228                                 $fit = $apiResult->addValue(
 229                                         array( 'query', 'interwiki' . $this->getModuleName(), $result->getInterwikiPrefix()  ),
 230                                         null,
 231                                         $vals
 232                                 );
 233
 234                                 if ( !$fit ) {
 235                                         // We hit the limit. We can't really provide any meaningful
 236                                         // pagination info so just bail out
 237                                         break;
 238                                 }
 239
 240                                 $result = $matches->next();
 241                         }
 242                 }
 243
 244                 if ( is_null( $resultPageSet ) ) {
 245                         $apiResult->setIndexedTagName_internal( array(
 246                                 'query', $this->getModuleName()
 247                         ), 'p' );
 248                         if ( $hasInterwikiResults ) {
 249                                 $apiResult->setIndexedTagName_internal( array(
 250                                         'query', 'interwiki' . $this->getModuleName()
 251                                 ), 'p' );
 252                         }
 253                 } else {
 254                         $resultPageSet->populateFromTitles( $titles );
 255                 }
 256         }
 257
 258         public function getCacheMode( $params ) {
 259                 return 'public';
 260         }
 261
 262         public function getAllowedParams() {
 263                 $params = array(
 264                         'search' => array(
 265                                 ApiBase::PARAM_TYPE => 'string',
 266                                 ApiBase::PARAM_REQUIRED => true
 267                         ),
 268                         'namespace' => array(
 269                                 ApiBase::PARAM_DFLT => NS_MAIN,
 270                                 ApiBase::PARAM_TYPE => 'namespace',
 271                                 ApiBase::PARAM_ISMULTI => true,
 272                         ),
 273                         'what' => array(
 274                                 ApiBase::PARAM_DFLT => null,
 275                                 ApiBase::PARAM_TYPE => array(
 276                                         'title',
 277                                         'text',
 278                                         'nearmatch',
 279                                 )
 280                         ),
 281                         'info' => array(
 282                                 ApiBase::PARAM_DFLT => 'totalhits|suggestion',
 283                                 ApiBase::PARAM_TYPE => array(
 284                                         'totalhits',
 285                                         'suggestion',
 286                                 ),
 287                                 ApiBase::PARAM_ISMULTI => true,
 288                         ),
 289                         'prop' => array(
 290                                 ApiBase::PARAM_DFLT => 'size|wordcount|timestamp|snippet',
 291                                 ApiBase::PARAM_TYPE => array(
 292                                         'size',
 293                                         'wordcount',
 294                                         'timestamp',
 295                                         'score',
 296                                         'snippet',
 297                                         'titlesnippet',
 298                                         'redirecttitle',
 299                                         'redirectsnippet',
 300                                         'sectiontitle',
 301                                         'sectionsnippet',
 302                                         'hasrelated',
 303                                 ),
 304                                 ApiBase::PARAM_ISMULTI => true,
 305                         ),
 306                         'offset' => 0,
 307                         'limit' => array(
 308                                 ApiBase::PARAM_DFLT => 10,
 309                                 ApiBase::PARAM_TYPE => 'limit',
 310                                 ApiBase::PARAM_MIN => 1,
 311                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_SML1,
 312                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_SML2
 313                         ),
 314                         'interwiki' => false,
 315                 );
 316
 317                 $alternatives = SearchEngine::getSearchTypes();
 318                 if ( count( $alternatives ) > 1 ) {
 319                         if ( $alternatives[0] === null ) {
 320                                 $alternatives[0] = self::BACKEND_NULL_PARAM;
 321                         }
 322                         $params['backend'] = array(
 323                                 ApiBase::PARAM_DFLT => $this->getConfig()->get( 'SearchType' ),
 324                                 ApiBase::PARAM_TYPE => $alternatives,
 325                         );
 326                 }
 327
 328                 return $params;
 329         }
 330
 331         public function getParamDescription() {
 332                 $descriptions = array(
 333                         'search' => 'Search for all page titles (or content) that has this value',
 334                         'namespace' => 'The namespace(s) to enumerate',
 335                         'what' => 'Search inside the text or titles',
 336                         'info' => 'What metadata to return',
 337                         'prop' => array(
 338                                 'What properties to return',
 339                                 ' size             - Adds the size of the page in bytes',
 340                                 ' wordcount        - Adds the word count of the page',
 341                                 ' timestamp        - Adds the timestamp of when the page was last edited',
 342                                 ' score            - DEPRECATED and IGNORED',
 343                                 ' snippet          - Adds a parsed snippet of the page',
 344                                 ' titlesnippet     - Adds a parsed snippet of the page title',
 345                                 ' redirectsnippet  - Adds a parsed snippet of the redirect title',
 346                                 ' redirecttitle    - Adds the title of the matching redirect',
 347                                 ' sectionsnippet   - Adds a parsed snippet of the matching section title',
 348                                 ' sectiontitle     - Adds the title of the matching section',
 349                                 ' hasrelated       - DEPRECATED and IGNORED',
 350                         ),
 351                         'offset' => 'Use this value to continue paging (return by query)',
 352                         'limit' => 'How many total pages to return',
 353                         'interwiki' => 'Include interwiki results in the search, if available'
 354                 );
 355
 356                 if ( count( SearchEngine::getSearchTypes() ) > 1 ) {
 357                         $descriptions['backend'] = 'Which search backend to use, if not the default';
 358                 }
 359
 360                 return $descriptions;
 361         }
 362
 363         public function getDescription() {
 364                 return 'Perform a full text search.';
 365         }
 366
 367         public function getExamples() {
 368                 return array(
 369                         'api.php?action=query&list=search&srsearch=meaning',
 370                         'api.php?action=query&list=search&srwhat=text&srsearch=meaning',
 371                         'api.php?action=query&generator=search&gsrsearch=meaning&prop=info',
 372                 );
 373         }
 374
 375         public function getHelpUrls() {
 376                 return 'https://www.mediawiki.org/wiki/API:Search';
 377         }
 378 }