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 ( !is_null( $result->getScore() ) && isset( $prop['score'] ) ) {
 161                                         $vals['score'] = $result->getScore();
 162                                 }
 163                                 if ( isset( $prop['titlesnippet'] ) ) {
 164                                         $vals['titlesnippet'] = $result->getTitleSnippet( $terms );
 165                                 }
 166                                 if ( !is_null( $result->getRedirectTitle() ) ) {
 167                                         if ( isset( $prop['redirecttitle'] ) ) {
 168                                                 $vals['redirecttitle'] = $result->getRedirectTitle();
 169                                         }
 170                                         if ( isset( $prop['redirectsnippet'] ) ) {
 171                                                 $vals['redirectsnippet'] = $result->getRedirectSnippet( $terms );
 172                                         }
 173                                 }
 174                                 if ( !is_null( $result->getSectionTitle() ) ) {
 175                                         if ( isset( $prop['sectiontitle'] ) ) {
 176                                                 $vals['sectiontitle'] = $result->getSectionTitle()->getFragment();
 177                                         }
 178                                         if ( isset( $prop['sectionsnippet'] ) ) {
 179                                                 $vals['sectionsnippet'] = $result->getSectionSnippet();
 180                                         }
 181                                 }
 182                                 if ( isset( $prop['hasrelated'] ) && $result->hasRelated() ) {
 183                                         $vals['hasrelated'] = '';
 184                                 }
 185
 186                                 // Add item to results and see whether it fits
 187                                 $fit = $apiResult->addValue( array( 'query', $this->getModuleName() ),
 188                                         null, $vals );
 189                                 if ( !$fit ) {
 190                                         $this->setContinueEnumParameter( 'offset', $params['offset'] + $count - 1 );
 191                                         break;
 192                                 }
 193                         } else {
 194                                 $titles[] = $title;
 195                         }
 196
 197                         $result = $matches->next();
 198                 }
 199
 200                 $hasInterwikiResults = false;
 201                 if ( $interwiki && $resultPageSet === null && $matches->hasInterwikiResults() ) {
 202                         $matches = $matches->getInterwikiResults();
 203                         $iwprefixes = array();
 204                         $hasInterwikiResults = true;
 205
 206                         // Include number of results if requested
 207                         if ( isset( $searchInfo['totalhits'] ) ) {
 208                                 $totalhits = $matches->getTotalHits();
 209                                 if ( $totalhits !== null ) {
 210                                         $apiResult->addValue( array( 'query', 'interwikisearchinfo' ),
 211                                                 'totalhits', $totalhits );
 212                                 }
 213                         }
 214
 215                         $result = $matches->next();
 216                         while ( $result ) {
 217                                 $title = $result->getTitle();
 218                                 $vals = array(
 219                                         'namespace' => $result->getInterwikiNamespaceText(),
 220                                         'title' => $title->getText(),
 221                                         'url' => $title->getFullUrl(),
 222                                 );
 223
 224                                 // Add item to results and see whether it fits
 225                                 $fit = $apiResult->addValue(
 226                                         array( 'query', 'interwiki' . $this->getModuleName(), $result->getInterwikiPrefix()  ),
 227                                         null,
 228                                         $vals
 229                                 );
 230
 231                                 if ( !$fit ) {
 232                                         // We hit the limit. We can't really provide any meaningful
 233                                         // pagination info so just bail out
 234                                         break;
 235                                 }
 236
 237                                 $result = $matches->next();
 238                         }
 239                 }
 240
 241                 if ( is_null( $resultPageSet ) ) {
 242                         $apiResult->setIndexedTagName_internal( array(
 243                                 'query', $this->getModuleName()
 244                         ), 'p' );
 245                         if ( $hasInterwikiResults ) {
 246                                 $apiResult->setIndexedTagName_internal( array(
 247                                         'query', 'interwiki' . $this->getModuleName()
 248                                 ), 'p' );
 249                         }
 250                 } else {
 251                         $resultPageSet->populateFromTitles( $titles );
 252                 }
 253         }
 254
 255         public function getCacheMode( $params ) {
 256                 return 'public';
 257         }
 258
 259         public function getAllowedParams() {
 260                 global $wgSearchType;
 261
 262                 $params = array(
 263                         'search' => array(
 264                                 ApiBase::PARAM_TYPE => 'string',
 265                                 ApiBase::PARAM_REQUIRED => true
 266                         ),
 267                         'namespace' => array(
 268                                 ApiBase::PARAM_DFLT => NS_MAIN,
 269                                 ApiBase::PARAM_TYPE => 'namespace',
 270                                 ApiBase::PARAM_ISMULTI => true,
 271                         ),
 272                         'what' => array(
 273                                 ApiBase::PARAM_DFLT => null,
 274                                 ApiBase::PARAM_TYPE => array(
 275                                         'title',
 276                                         'text',
 277                                         'nearmatch',
 278                                 )
 279                         ),
 280                         'info' => array(
 281                                 ApiBase::PARAM_DFLT => 'totalhits|suggestion',
 282                                 ApiBase::PARAM_TYPE => array(
 283                                         'totalhits',
 284                                         'suggestion',
 285                                 ),
 286                                 ApiBase::PARAM_ISMULTI => true,
 287                         ),
 288                         'prop' => array(
 289                                 ApiBase::PARAM_DFLT => 'size|wordcount|timestamp|snippet',
 290                                 ApiBase::PARAM_TYPE => array(
 291                                         'size',
 292                                         'wordcount',
 293                                         'timestamp',
 294                                         'score',
 295                                         'snippet',
 296                                         'titlesnippet',
 297                                         'redirecttitle',
 298                                         'redirectsnippet',
 299                                         'sectiontitle',
 300                                         'sectionsnippet',
 301                                         'hasrelated',
 302                                 ),
 303                                 ApiBase::PARAM_ISMULTI => true,
 304                         ),
 305                         'offset' => 0,
 306                         'limit' => array(
 307                                 ApiBase::PARAM_DFLT => 10,
 308                                 ApiBase::PARAM_TYPE => 'limit',
 309                                 ApiBase::PARAM_MIN => 1,
 310                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_SML1,
 311                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_SML2
 312                         ),
 313                         'interwiki' => false,
 314                 );
 315
 316                 $alternatives = SearchEngine::getSearchTypes();
 317                 if ( count( $alternatives ) > 1 ) {
 318                         if ( $alternatives[0] === null ) {
 319                                 $alternatives[0] = self::BACKEND_NULL_PARAM;
 320                         }
 321                         $params['backend'] = array(
 322                                 ApiBase::PARAM_DFLT => $wgSearchType,
 323                                 ApiBase::PARAM_TYPE => $alternatives,
 324                         );
 325                 }
 326
 327                 return $params;
 328         }
 329
 330         public function getParamDescription() {
 331                 $descriptions = array(
 332                         'search' => 'Search for all page titles (or content) that has this value',
 333                         'namespace' => 'The namespace(s) to enumerate',
 334                         'what' => 'Search inside the text or titles',
 335                         'info' => 'What metadata to return',
 336                         'prop' => array(
 337                                 'What properties to return',
 338                                 ' size             - Adds the size of the page in bytes',
 339                                 ' wordcount        - Adds the word count of the page',
 340                                 ' timestamp        - Adds the timestamp of when the page was last edited',
 341                                 ' score            - Adds the score (if any) from the search engine',
 342                                 ' snippet          - Adds a parsed snippet of the page',
 343                                 ' titlesnippet     - Adds a parsed snippet of the page title',
 344                                 ' redirectsnippet  - Adds a parsed snippet of the redirect title',
 345                                 ' redirecttitle    - Adds the title of the matching redirect',
 346                                 ' sectionsnippet   - Adds a parsed snippet of the matching section title',
 347                                 ' sectiontitle     - Adds the title of the matching section',
 348                                 ' hasrelated       - Indicates whether a related search is available',
 349                         ),
 350                         'offset' => 'Use this value to continue paging (return by query)',
 351                         'limit' => 'How many total pages to return',
 352                         'interwiki' => 'Include interwiki results in the search, if available'
 353                 );
 354
 355                 if ( count( SearchEngine::getSearchTypes() ) > 1 ) {
 356                         $descriptions['backend'] = 'Which search backend to use, if not the default';
 357                 }
 358
 359                 return $descriptions;
 360         }
 361
 362         public function getResultProperties() {
 363                 return array(
 364                         '' => array(
 365                                 'ns' => 'namespace',
 366                                 'title' => 'string'
 367                         ),
 368                         'snippet' => array(
 369                                 'snippet' => 'string'
 370                         ),
 371                         'size' => array(
 372                                 'size' => 'integer'
 373                         ),
 374                         'wordcount' => array(
 375                                 'wordcount' => 'integer'
 376                         ),
 377                         'timestamp' => array(
 378                                 'timestamp' => 'timestamp'
 379                         ),
 380                         'score' => array(
 381                                 'score' => array(
 382                                         ApiBase::PROP_TYPE => 'string',
 383                                         ApiBase::PROP_NULLABLE => true
 384                                 )
 385                         ),
 386                         'titlesnippet' => array(
 387                                 'titlesnippet' => 'string'
 388                         ),
 389                         'redirecttitle' => array(
 390                                 'redirecttitle' => array(
 391                                         ApiBase::PROP_TYPE => 'string',
 392                                         ApiBase::PROP_NULLABLE => true
 393                                 )
 394                         ),
 395                         'redirectsnippet' => array(
 396                                 'redirectsnippet' => array(
 397                                         ApiBase::PROP_TYPE => 'string',
 398                                         ApiBase::PROP_NULLABLE => true
 399                                 )
 400                         ),
 401                         'sectiontitle' => array(
 402                                 'sectiontitle' => array(
 403                                         ApiBase::PROP_TYPE => 'string',
 404                                         ApiBase::PROP_NULLABLE => true
 405                                 )
 406                         ),
 407                         'sectionsnippet' => array(
 408                                 'sectionsnippet' => array(
 409                                         ApiBase::PROP_TYPE => 'string',
 410                                         ApiBase::PROP_NULLABLE => true
 411                                 )
 412                         ),
 413                         'hasrelated' => array(
 414                                 'hasrelated' => 'boolean'
 415                         )
 416                 );
 417         }
 418
 419         public function getDescription() {
 420                 return 'Perform a full text search.';
 421         }
 422
 423         public function getPossibleErrors() {
 424                 return array_merge( parent::getPossibleErrors(), array(
 425                         array( 'code' => 'search-text-disabled', 'info' => 'text search is disabled' ),
 426                         array( 'code' => 'search-title-disabled', 'info' => 'title search is disabled' ),
 427                         array( 'code' => 'search-error', 'info' => 'search error has occurred' ),
 428                 ) );
 429         }
 430
 431         public function getExamples() {
 432                 return array(
 433                         'api.php?action=query&list=search&srsearch=meaning',
 434                         'api.php?action=query&list=search&srwhat=text&srsearch=meaning',
 435                         'api.php?action=query&generator=search&gsrsearch=meaning&prop=info',
 436                 );
 437         }
 438
 439         public function getHelpUrls() {
 440                 return 'https://www.mediawiki.org/wiki/API:Search';
 441         }
 442 }