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( $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 $resultPageSet ApiPageSet
  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                 $search->showRedirects = $params['redirects'];
  76
  77                 $query = $search->transformSearchTerm( $query );
  78                 $query = $search->replacePrefixes( $query );
  79
  80                 // Perform the actual search
  81                 if ( $what == 'text' ) {
  82                         $matches = $search->searchText( $query );
  83                 } elseif ( $what == 'title' ) {
  84                         $matches = $search->searchTitle( $query );
  85                 } elseif ( $what == 'nearmatch' ) {
  86                         $matches = SearchEngine::getNearMatchResultSet( $query );
  87                 } else {
  88                         // We default to title searches; this is a terrible legacy
  89                         // of the way we initially set up the MySQL fulltext-based
  90                         // search engine with separate title and text fields.
  91                         // In the future, the default should be for a combined index.
  92                         $what = 'title';
  93                         $matches = $search->searchTitle( $query );
  94
  95                         // Not all search engines support a separate title search,
  96                         // for instance the Lucene-based engine we use on Wikipedia.
  97                         // In this case, fall back to full-text search (which will
  98                         // include titles in it!)
  99                         if ( is_null( $matches ) ) {
 100                                 $what = 'text';
 101                                 $matches = $search->searchText( $query );
 102                         }
 103                 }
 104                 if ( is_null( $matches ) ) {
 105                         $this->dieUsage( "{$what} search is disabled", "search-{$what}-disabled" );
 106                 } elseif ( $matches instanceof Status && !$matches->isGood() ) {
 107                         $this->dieUsage( $matches->getWikiText(), 'search-error' );
 108                 }
 109
 110                 $apiResult = $this->getResult();
 111                 // Add search meta data to result
 112                 if ( isset( $searchInfo['totalhits'] ) ) {
 113                         $totalhits = $matches->getTotalHits();
 114                         if ( $totalhits !== null ) {
 115                                 $apiResult->addValue( array( 'query', 'searchinfo' ),
 116                                         'totalhits', $totalhits );
 117                         }
 118                 }
 119                 if ( isset( $searchInfo['suggestion'] ) && $matches->hasSuggestion() ) {
 120                         $apiResult->addValue( array( 'query', 'searchinfo' ),
 121                                 'suggestion', $matches->getSuggestionQuery() );
 122                 }
 123
 124                 // Add the search results to the result
 125                 $terms = $wgContLang->convertForSearchResult( $matches->termMatches() );
 126                 $titles = array();
 127                 $count = 0;
 128                 $result = $matches->next();
 129
 130                 while ( $result ) {
 131                         if ( ++$count > $limit ) {
 132                                 // We've reached the one extra which shows that there are
 133                                 // additional items to be had. Stop here...
 134                                 $this->setContinueEnumParameter( 'offset', $params['offset'] + $params['limit'] );
 135                                 break;
 136                         }
 137
 138                         // Silently skip broken and missing titles
 139                         if ( $result->isBrokenTitle() || $result->isMissingRevision() ) {
 140                                 $result = $matches->next();
 141                                 continue;
 142                         }
 143
 144                         $title = $result->getTitle();
 145                         if ( is_null( $resultPageSet ) ) {
 146                                 $vals = array();
 147                                 ApiQueryBase::addTitleInfo( $vals, $title );
 148
 149                                 if ( isset( $prop['snippet'] ) ) {
 150                                         $vals['snippet'] = $result->getTextSnippet( $terms );
 151                                 }
 152                                 if ( isset( $prop['size'] ) ) {
 153                                         $vals['size'] = $result->getByteSize();
 154                                 }
 155                                 if ( isset( $prop['wordcount'] ) ) {
 156                                         $vals['wordcount'] = $result->getWordCount();
 157                                 }
 158                                 if ( isset( $prop['timestamp'] ) ) {
 159                                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $result->getTimestamp() );
 160                                 }
 161                                 if ( !is_null( $result->getScore() ) && isset( $prop['score'] ) ) {
 162                                         $vals['score'] = $result->getScore();
 163                                 }
 164                                 if ( isset( $prop['titlesnippet'] ) ) {
 165                                         $vals['titlesnippet'] = $result->getTitleSnippet( $terms );
 166                                 }
 167                                 if ( !is_null( $result->getRedirectTitle() ) ) {
 168                                         if ( isset( $prop['redirecttitle'] ) ) {
 169                                                 $vals['redirecttitle'] = $result->getRedirectTitle();
 170                                         }
 171                                         if ( isset( $prop['redirectsnippet'] ) ) {
 172                                                 $vals['redirectsnippet'] = $result->getRedirectSnippet( $terms );
 173                                         }
 174                                 }
 175                                 if ( !is_null( $result->getSectionTitle() ) ) {
 176                                         if ( isset( $prop['sectiontitle'] ) ) {
 177                                                 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
 178                                         }
 179                                         if ( isset( $prop['sectionsnippet'] ) ) {
 180                                                 $vals['sectionsnippet'] = $result->getSectionSnippet();
 181                                         }
 182                                 }
 183                                 if ( isset( $prop['hasrelated'] ) && $result->hasRelated() ) {
 184                                         $vals['hasrelated'] = '';
 185                                 }
 186
 187                                 // Add item to results and see whether it fits
 188                                 $fit = $apiResult->addValue( array( 'query', $this->getModuleName() ),
 189                                         null, $vals );
 190                                 if ( !$fit ) {
 191                                         $this->setContinueEnumParameter( 'offset', $params['offset'] + $count - 1 );
 192                                         break;
 193                                 }
 194                         } else {
 195                                 $titles[] = $title;
 196                         }
 197
 198                         $result = $matches->next();
 199                 }
 200
 201                 $hasInterwikiResults = false;
 202                 if ( $interwiki && $resultPageSet === null && $matches->hasInterwikiResults() ) {
 203                         $matches = $matches->getInterwikiResults();
 204                         $iwprefixes = array();
 205                         $hasInterwikiResults = true;
 206
 207                         // Include number of results if requested
 208                         if ( isset( $searchInfo['totalhits'] ) ) {
 209                                 $totalhits = $matches->getTotalHits();
 210                                 if ( $totalhits !== null ) {
 211                                         $apiResult->addValue( array( 'query', 'interwikisearchinfo' ),
 212                                                 'totalhits', $totalhits );
 213                                 }
 214                         }
 215
 216                         $result = $matches->next();
 217                         while ( $result ) {
 218                                 $title = $result->getTitle();
 219                                 $vals = array(
 220                                         'namespace' => $result->getInterwikiNamespaceText(),
 221                                         'title' => $title->getText(),
 222                                         'url' => $title->getFullUrl(),
 223                                 );
 224
 225                                 // Add item to results and see whether it fits
 226                                 $fit = $apiResult->addValue( array( 'query', 'interwiki' . $this->getModuleName(), $result->getInterwikiPrefix()  ),
 227                                         null, $vals );
 228                                 if ( !$fit ) {
 229                                         // We hit the limit. We can't really provide any meaningful
 230                                         // pagination info so just bail out
 231                                         break;
 232                                 }
 233
 234                                 $result = $matches->next();
 235                         }
 236                 }
 237
 238                 if ( is_null( $resultPageSet ) ) {
 239                         $apiResult->setIndexedTagName_internal( array(
 240                                 'query', $this->getModuleName()
 241                         ), 'p' );
 242                         if ( $hasInterwikiResults ) {
 243                                 $apiResult->setIndexedTagName_internal( array(
 244                                         'query', 'interwiki' . $this->getModuleName()
 245                                 ), 'p' );
 246                         }
 247                 } else {
 248                         $resultPageSet->populateFromTitles( $titles );
 249                 }
 250         }
 251
 252         public function getCacheMode( $params ) {
 253                 return 'public';
 254         }
 255
 256         public function getAllowedParams() {
 257                 global $wgSearchType;
 258
 259                 $params = array(
 260                         'search' => array(
 261                                 ApiBase::PARAM_TYPE => 'string',
 262                                 ApiBase::PARAM_REQUIRED => true
 263                         ),
 264                         'namespace' => array(
 265                                 ApiBase::PARAM_DFLT => NS_MAIN,
 266                                 ApiBase::PARAM_TYPE => 'namespace',
 267                                 ApiBase::PARAM_ISMULTI => true,
 268                         ),
 269                         'what' => array(
 270                                 ApiBase::PARAM_DFLT => null,
 271                                 ApiBase::PARAM_TYPE => array(
 272                                         'title',
 273                                         'text',
 274                                         'nearmatch',
 275                                 )
 276                         ),
 277                         'info' => array(
 278                                 ApiBase::PARAM_DFLT => 'totalhits|suggestion',
 279                                 ApiBase::PARAM_TYPE => array(
 280                                         'totalhits',
 281                                         'suggestion',
 282                                 ),
 283                                 ApiBase::PARAM_ISMULTI => true,
 284                         ),
 285                         'prop' => array(
 286                                 ApiBase::PARAM_DFLT => 'size|wordcount|timestamp|snippet',
 287                                 ApiBase::PARAM_TYPE => array(
 288                                         'size',
 289                                         'wordcount',
 290                                         'timestamp',
 291                                         'score',
 292                                         'snippet',
 293                                         'titlesnippet',
 294                                         'redirecttitle',
 295                                         'redirectsnippet',
 296                                         'sectiontitle',
 297                                         'sectionsnippet',
 298                                         'hasrelated',
 299                                 ),
 300                                 ApiBase::PARAM_ISMULTI => true,
 301                         ),
 302                         'redirects' => false,
 303                         'offset' => 0,
 304                         'limit' => array(
 305                                 ApiBase::PARAM_DFLT => 10,
 306                                 ApiBase::PARAM_TYPE => 'limit',
 307                                 ApiBase::PARAM_MIN => 1,
 308                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_SML1,
 309                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_SML2
 310                         ),
 311                         'interwiki' => false,
 312                 );
 313
 314                 $alternatives = SearchEngine::getSearchTypes();
 315                 if ( count( $alternatives ) > 1 ) {
 316                         if ( $alternatives[0] === null ) {
 317                                 $alternatives[0] = self::BACKEND_NULL_PARAM;
 318                         }
 319                         $params['backend'] = array(
 320                                 ApiBase::PARAM_DFLT => $wgSearchType,
 321                                 ApiBase::PARAM_TYPE => $alternatives,
 322                         );
 323                 }
 324
 325                 return $params;
 326         }
 327
 328         public function getParamDescription() {
 329                 $descriptions = array(
 330                         'search' => 'Search for all page titles (or content) that has this value',
 331                         'namespace' => 'The namespace(s) to enumerate',
 332                         'what' => 'Search inside the text or titles',
 333                         'info' => 'What metadata to return',
 334                         'prop' => array(
 335                                 'What properties to return',
 336                                 ' size             - Adds the size of the page in bytes',
 337                                 ' wordcount        - Adds the word count of the page',
 338                                 ' timestamp        - Adds the timestamp of when the page was last edited',
 339                                 ' score            - Adds the score (if any) from the search engine',
 340                                 ' snippet          - Adds a parsed snippet of the page',
 341                                 ' titlesnippet     - Adds a parsed snippet of the page title',
 342                                 ' redirectsnippet  - Adds a parsed snippet of the redirect title',
 343                                 ' redirecttitle    - Adds the title of the matching redirect',
 344                                 ' sectionsnippet   - Adds a parsed snippet of the matching section title',
 345                                 ' sectiontitle     - Adds the title of the matching section',
 346                                 ' hasrelated       - Indicates whether a related search is available',
 347                         ),
 348                         'redirects' => 'Include redirect pages in the search',
 349                         'offset' => 'Use this value to continue paging (return by query)',
 350                         'limit' => 'How many total pages to return',
 351                         'interwiki' => 'Include interwiki results in the search, if available'
 352                 );
 353
 354                 if ( count( SearchEngine::getSearchTypes() ) > 1 ) {
 355                         $descriptions['backend'] = 'Which search backend to use, if not the default';
 356                 }
 357
 358                 return $descriptions;
 359         }
 360
 361         public function getResultProperties() {
 362                 return array(
 363                         '' => array(
 364                                 'ns' => 'namespace',
 365                                 'title' => 'string'
 366                         ),
 367                         'snippet' => array(
 368                                 'snippet' => 'string'
 369                         ),
 370                         'size' => array(
 371                                 'size' => 'integer'
 372                         ),
 373                         'wordcount' => array(
 374                                 'wordcount' => 'integer'
 375                         ),
 376                         'timestamp' => array(
 377                                 'timestamp' => 'timestamp'
 378                         ),
 379                         'score' => array(
 380                                 'score' => array(
 381                                         ApiBase::PROP_TYPE => 'string',
 382                                         ApiBase::PROP_NULLABLE => true
 383                                 )
 384                         ),
 385                         'titlesnippet' => array(
 386                                 'titlesnippet' => 'string'
 387                         ),
 388                         'redirecttitle' => array(
 389                                 'redirecttitle' => array(
 390                                         ApiBase::PROP_TYPE => 'string',
 391                                         ApiBase::PROP_NULLABLE => true
 392                                 )
 393                         ),
 394                         'redirectsnippet' => array(
 395                                 'redirectsnippet' => array(
 396                                         ApiBase::PROP_TYPE => 'string',
 397                                         ApiBase::PROP_NULLABLE => true
 398                                 )
 399                         ),
 400                         'sectiontitle' => array(
 401                                 'sectiontitle' => array(
 402                                         ApiBase::PROP_TYPE => 'string',
 403                                         ApiBase::PROP_NULLABLE => true
 404                                 )
 405                         ),
 406                         'sectionsnippet' => array(
 407                                 'sectionsnippet' => array(
 408                                         ApiBase::PROP_TYPE => 'string',
 409                                         ApiBase::PROP_NULLABLE => true
 410                                 )
 411                         ),
 412                         'hasrelated' => array(
 413                                 'hasrelated' => 'boolean'
 414                         )
 415                 );
 416         }
 417
 418         public function getDescription() {
 419                 return 'Perform a full text search';
 420         }
 421
 422         public function getPossibleErrors() {
 423                 return array_merge( parent::getPossibleErrors(), array(
 424                         array( 'code' => 'search-text-disabled', 'info' => 'text search is disabled' ),
 425                         array( 'code' => 'search-title-disabled', 'info' => 'title search is disabled' ),
 426                         array( 'code' => 'search-error', 'info' => 'search error has occurred' ),
 427                 ) );
 428         }
 429
 430         public function getExamples() {
 431                 return array(
 432                         'api.php?action=query&list=search&srsearch=meaning',
 433                         'api.php?action=query&list=search&srwhat=text&srsearch=meaning',
 434                         'api.php?action=query&generator=search&gsrsearch=meaning&prop=info',
 435                 );
 436         }
 437
 438         public function getHelpUrls() {
 439                 return 'https://www.mediawiki.org/wiki/API:Search';
 440         }
 441 }