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                 // Create search engine instance and set options
  71                 $search = isset( $params['backend'] ) && $params['backend'] != self::BACKEND_NULL_PARAM ?
  72                         SearchEngine::create( $params['backend'] ) : SearchEngine::create();
  73                 $search->setLimitOffset( $limit + 1, $params['offset'] );
  74                 $search->setNamespaces( $params['namespace'] );
  75
  76                 $query = $search->transformSearchTerm( $query );
  77                 $query = $search->replacePrefixes( $query );
  78
  79                 // Perform the actual search
  80                 if ( $what == 'text' ) {
  81                         $matches = $search->searchText( $query );
  82                 } elseif ( $what == 'title' ) {
  83                         $matches = $search->searchTitle( $query );
  84                 } elseif ( $what == 'nearmatch' ) {
  85                         $matches = SearchEngine::getNearMatchResultSet( $query );
  86                 } else {
  87                         // We default to title searches; this is a terrible legacy
  88                         // of the way we initially set up the MySQL fulltext-based
  89                         // search engine with separate title and text fields.
  90                         // In the future, the default should be for a combined index.
  91                         $what = 'title';
  92                         $matches = $search->searchTitle( $query );
  93
  94                         // Not all search engines support a separate title search,
  95                         // for instance the Lucene-based engine we use on Wikipedia.
  96                         // In this case, fall back to full-text search (which will
  97                         // include titles in it!)
  98                         if ( is_null( $matches ) ) {
  99                                 $what = 'text';
 100                                 $matches = $search->searchText( $query );
 101                         }
 102                 }
 103                 if ( is_null( $matches ) ) {
 104                         $this->dieUsage( "{$what} search is disabled", "search-{$what}-disabled" );
 105                 } elseif ( $matches instanceof Status && !$matches->isGood() ) {
 106                         $this->dieUsage( $matches->getWikiText(), 'search-error' );
 107                 }
 108
 109                 $apiResult = $this->getResult();
 110                 // Add search meta data to result
 111                 if ( isset( $searchInfo['totalhits'] ) ) {
 112                         $totalhits = $matches->getTotalHits();
 113                         if ( $totalhits !== null ) {
 114                                 $apiResult->addValue( array( 'query', 'searchinfo' ),
 115                                         'totalhits', $totalhits );
 116                         }
 117                 }
 118                 if ( isset( $searchInfo['suggestion'] ) && $matches->hasSuggestion() ) {
 119                         $apiResult->addValue( array( 'query', 'searchinfo' ),
 120                                 'suggestion', $matches->getSuggestionQuery() );
 121                 }
 122
 123                 // Add the search results to the result
 124                 $terms = $wgContLang->convertForSearchResult( $matches->termMatches() );
 125                 $titles = array();
 126                 $count = 0;
 127                 $result = $matches->next();
 128
 129                 while ( $result ) {
 130                         if ( ++$count > $limit ) {
 131                                 // We've reached the one extra which shows that there are
 132                                 // additional items to be had. Stop here...
 133                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $params['limit'] );
 134                                 break;
 135                         }
 136
 137                         // Silently skip broken and missing titles
 138                         if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 139                                 $result = $matches->next();
 140                                 continue;
 141                         }
 142
 143                         $title = $result->getTitle();
 144                         if ( is_null( $resultPageSet ) ) {
 145                                 $vals = array();
 146                                 ApiQueryBase::addTitleInfo( $vals, $title );
 147
 148                                 if ( isset( $prop['snippet'] ) ) {
 149                                         $vals['snippet'] = $result->getTextSnippet( $terms );
 150                                 }
 151                                 if ( isset( $prop['size'] ) ) {
 152                                         $vals['size'] = $result->getByteSize();
 153                                 }
 154                                 if ( isset( $prop['wordcount'] ) ) {
 155                                         $vals['wordcount'] = $result->getWordCount();
 156                                 }
 157                                 if ( isset( $prop['timestamp'] ) ) {
 158                                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $result->getTimestamp() );
 159                                 }
 160                                 if ( isset( $prop['titlesnippet'] ) ) {
 161                                         $vals['titlesnippet'] = $result->getTitleSnippet( $terms );
 162                                 }
 163                                 if ( !is_null( $result->getRedirectTitle() ) ) {
 164                                         if ( isset( $prop['redirecttitle'] ) ) {
 165                                                 $vals['redirecttitle'] = $result->getRedirectTitle();
 166                                         }
 167                                         if ( isset( $prop['redirectsnippet'] ) ) {
 168                                                 $vals['redirectsnippet'] = $result->getRedirectSnippet( $terms );
 169                                         }
 170                                 }
 171                                 if ( !is_null( $result->getSectionTitle() ) ) {
 172                                         if ( isset( $prop['sectiontitle'] ) ) {
 173                                                 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
 174                                         }
 175                                         if ( isset( $prop['sectionsnippet'] ) ) {
 176                                                 $vals['sectionsnippet'] = $result->getSectionSnippet();
 177                                         }
 178                                 }
 179
 180                                 // Add item to results and see whether it fits
 181                                 $fit = $apiResult->addValue( array( 'query', $this->getModuleName() ),
 182                                         null, $vals );
 183                                 if ( !$fit ) {
 184                                         $this->setContinueEnumParameter( 'offset', $params['offset'] + $count - 1 );
 185                                         break;
 186                                 }
 187                         } else {
 188                                 $titles[] = $title;
 189                         }
 190
 191                         $result = $matches->next();
 192                 }
 193
 194                 $hasInterwikiResults = false;
 195                 if ( $interwiki && $resultPageSet === null && $matches->hasInterwikiResults() ) {
 196                         $matches = $matches->getInterwikiResults();
 197                         $hasInterwikiResults = true;
 198
 199                         // Include number of results if requested
 200                         if ( isset( $searchInfo['totalhits'] ) ) {
 201                                 $totalhits = $matches->getTotalHits();
 202                                 if ( $totalhits !== null ) {
 203                                         $apiResult->addValue( array( 'query', 'interwikisearchinfo' ),
 204                                                 'totalhits', $totalhits );
 205                                 }
 206                         }
 207
 208                         $result = $matches->next();
 209                         while ( $result ) {
 210                                 $title = $result->getTitle();
 211                                 $vals = array(
 212                                         'namespace' => $result->getInterwikiNamespaceText(),
 213                                         'title' => $title->getText(),
 214                                         'url' => $title->getFullUrl(),
 215                                 );
 216
 217                                 // Add item to results and see whether it fits
 218                                 $fit = $apiResult->addValue(
 219                                         array( 'query', 'interwiki' . $this->getModuleName(), $result->getInterwikiPrefix()  ),
 220                                         null,
 221                                         $vals
 222                                 );
 223
 224                                 if ( !$fit ) {
 225                                         // We hit the limit. We can't really provide any meaningful
 226                                         // pagination info so just bail out
 227                                         break;
 228                                 }
 229
 230                                 $result = $matches->next();
 231                         }
 232                 }
 233
 234                 if ( is_null( $resultPageSet ) ) {
 235                         $apiResult->setIndexedTagName_internal( array(
 236                                 'query', $this->getModuleName()
 237                         ), 'p' );
 238                         if ( $hasInterwikiResults ) {
 239                                 $apiResult->setIndexedTagName_internal( array(
 240                                         'query', 'interwiki' . $this->getModuleName()
 241                                 ), 'p' );
 242                         }
 243                 } else {
 244                         $resultPageSet->populateFromTitles( $titles );
 245                 }
 246         }
 247
 248         public function getCacheMode( $params ) {
 249                 return 'public';
 250         }
 251
 252         public function getAllowedParams() {
 253                 $params = array(
 254                         'search' => array(
 255                                 ApiBase::PARAM_TYPE => 'string',
 256                                 ApiBase::PARAM_REQUIRED => true
 257                         ),
 258                         'namespace' => array(
 259                                 ApiBase::PARAM_DFLT => NS_MAIN,
 260                                 ApiBase::PARAM_TYPE => 'namespace',
 261                                 ApiBase::PARAM_ISMULTI => true,
 262                         ),
 263                         'what' => array(
 264                                 ApiBase::PARAM_DFLT => null,
 265                                 ApiBase::PARAM_TYPE => array(
 266                                         'title',
 267                                         'text',
 268                                         'nearmatch',
 269                                 )
 270                         ),
 271                         'info' => array(
 272                                 ApiBase::PARAM_DFLT => 'totalhits|suggestion',
 273                                 ApiBase::PARAM_TYPE => array(
 274                                         'totalhits',
 275                                         'suggestion',
 276                                 ),
 277                                 ApiBase::PARAM_ISMULTI => true,
 278                         ),
 279                         'prop' => array(
 280                                 ApiBase::PARAM_DFLT => 'size|wordcount|timestamp|snippet',
 281                                 ApiBase::PARAM_TYPE => array(
 282                                         'size',
 283                                         'wordcount',
 284                                         'timestamp',
 285                                         'score',
 286                                         'snippet',
 287                                         'titlesnippet',
 288                                         'redirecttitle',
 289                                         'redirectsnippet',
 290                                         'sectiontitle',
 291                                         'sectionsnippet',
 292                                         'hasrelated',
 293                                 ),
 294                                 ApiBase::PARAM_ISMULTI => true,
 295                         ),
 296                         'offset' => 0,
 297                         'limit' => array(
 298                                 ApiBase::PARAM_DFLT => 10,
 299                                 ApiBase::PARAM_TYPE => 'limit',
 300                                 ApiBase::PARAM_MIN => 1,
 301                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_SML1,
 302                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_SML2
 303                         ),
 304                         'interwiki' => false,
 305                 );
 306
 307                 $alternatives = SearchEngine::getSearchTypes();
 308                 if ( count( $alternatives ) > 1 ) {
 309                         if ( $alternatives[0] === null ) {
 310                                 $alternatives[0] = self::BACKEND_NULL_PARAM;
 311                         }
 312                         $params['backend'] = array(
 313                                 ApiBase::PARAM_DFLT => $this->getConfig()->get( 'SearchType' ),
 314                                 ApiBase::PARAM_TYPE => $alternatives,
 315                         );
 316                 }
 317
 318                 return $params;
 319         }
 320
 321         public function getParamDescription() {
 322                 $descriptions = array(
 323                         'search' => 'Search for all page titles (or content) that has this value',
 324                         'namespace' => 'The namespace(s) to enumerate',
 325                         'what' => 'Search inside the text or titles',
 326                         'info' => 'What metadata to return',
 327                         'prop' => array(
 328                                 'What properties to return',
 329                                 ' size             - Adds the size of the page in bytes',
 330                                 ' wordcount        - Adds the word count of the page',
 331                                 ' timestamp        - Adds the timestamp of when the page was last edited',
 332                                 ' score            - Adds the score (if any) from the search engine',
 333                                 ' snippet          - Adds a parsed snippet of the page',
 334                                 ' titlesnippet     - Adds a parsed snippet of the page title',
 335                                 ' redirectsnippet  - Adds a parsed snippet of the redirect title',
 336                                 ' redirecttitle    - Adds the title of the matching redirect',
 337                                 ' sectionsnippet   - Adds a parsed snippet of the matching section title',
 338                                 ' sectiontitle     - Adds the title of the matching section',
 339                                 ' hasrelated       - Indicates whether a related search is available',
 340                         ),
 341                         'offset' => 'Use this value to continue paging (return by query)',
 342                         'limit' => 'How many total pages to return',
 343                         'interwiki' => 'Include interwiki results in the search, if available'
 344                 );
 345
 346                 if ( count( SearchEngine::getSearchTypes() ) > 1 ) {
 347                         $descriptions['backend'] = 'Which search backend to use, if not the default';
 348                 }
 349
 350                 return $descriptions;
 351         }
 352
 353         public function getDescription() {
 354                 return 'Perform a full text search.';
 355         }
 356
 357         public function getExamples() {
 358                 return array(
 359                         'api.php?action=query&list=search&srsearch=meaning',
 360                         'api.php?action=query&list=search&srwhat=text&srsearch=meaning',
 361                         'api.php?action=query&generator=search&gsrsearch=meaning&prop=info',
 362                 );
 363         }
 364
 365         public function getHelpUrls() {
 366                 return 'https://www.mediawiki.org/wiki/API:Search';
 367         }
 368 }