includes/api/ApiOpenSearch.php

   1 <?php
   2 /**
   3  * Created on Oct 13, 2006
   4  *
   5  * Copyright © 2006 Yuri Astrakhan "<Firstname><Lastname>@gmail.com"
   6  * Copyright © 2008 Brion Vibber <brion@wikimedia.org>
   7  * Copyright © 2014 Brad Jorsch <bjorsch@wikimedia.org>
   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 use MediaWiki\MediaWikiServices;
  28
  29 /**
  30  * @ingroup API
  31  */
  32 class ApiOpenSearch extends ApiBase {
  33
  34         private $format = null;
  35         private $fm = null;
  36
  37         /**
  38          * Get the output format
  39          *
  40          * @return string
  41          */
  42         protected function getFormat() {
  43                 if ( $this->format === null ) {
  44                         $params = $this->extractRequestParams();
  45                         $format = $params['format'];
  46
  47                         $allowedParams = $this->getAllowedParams();
  48                         if ( !in_array( $format, $allowedParams['format'][ApiBase::PARAM_TYPE] ) ) {
  49                                 $format = $allowedParams['format'][ApiBase::PARAM_DFLT];
  50                         }
  51
  52                         if ( substr( $format, -2 ) === 'fm' ) {
  53                                 $this->format = substr( $format, 0, -2 );
  54                                 $this->fm = 'fm';
  55                         } else {
  56                                 $this->format = $format;
  57                                 $this->fm = '';
  58                         }
  59                 }
  60                 return $this->format;
  61         }
  62
  63         public function getCustomPrinter() {
  64                 switch ( $this->getFormat() ) {
  65                         case 'json':
  66                                 return new ApiOpenSearchFormatJson(
  67                                         $this->getMain(), $this->fm, $this->getParameter( 'warningsaserror' )
  68                                 );
  69
  70                         case 'xml':
  71                                 $printer = $this->getMain()->createPrinterByName( 'xml' . $this->fm );
  72                                 $printer->setRootElement( 'SearchSuggestion' );
  73                                 return $printer;
  74
  75                         default:
  76                                 ApiBase::dieDebug( __METHOD__, "Unsupported format '{$this->getFormat()}'" );
  77                 }
  78         }
  79
  80         public function execute() {
  81                 $params = $this->extractRequestParams();
  82                 $search = $params['search'];
  83                 $limit = $params['limit'];
  84                 $namespaces = $params['namespace'];
  85                 $suggest = $params['suggest'];
  86
  87                 if ( $params['redirects'] === null ) {
  88                         // Backwards compatibility, don't resolve for JSON.
  89                         $resolveRedir = $this->getFormat() !== 'json';
  90                 } else {
  91                         $resolveRedir = $params['redirects'] === 'resolve';
  92                 }
  93
  94                 $results = [];
  95
  96                 if ( !$suggest || $this->getConfig()->get( 'EnableOpenSearchSuggest' ) ) {
  97                         // Open search results may be stored for a very long time
  98                         $this->getMain()->setCacheMaxAge( $this->getConfig()->get( 'SearchSuggestCacheExpiry' ) );
  99                         $this->getMain()->setCacheMode( 'public' );
 100                         $this->search( $search, $limit, $namespaces, $resolveRedir, $results );
 101
 102                         // Allow hooks to populate extracts and images
 103                         Hooks::run( 'ApiOpenSearchSuggest', [ &$results ] );
 104
 105                         // Trim extracts, if necessary
 106                         $length = $this->getConfig()->get( 'OpenSearchDescriptionLength' );
 107                         foreach ( $results as &$r ) {
 108                                 if ( is_string( $r['extract'] ) && !$r['extract trimmed'] ) {
 109                                         $r['extract'] = self::trimExtract( $r['extract'], $length );
 110                                 }
 111                         }
 112                 }
 113
 114                 // Populate result object
 115                 $this->populateResult( $search, $results );
 116         }
 117
 118         /**
 119          * Perform the search
 120          *
 121          * @param string $search Text to search
 122          * @param int $limit Maximum items to return
 123          * @param array $namespaces Namespaces to search
 124          * @param bool $resolveRedir Whether to resolve redirects
 125          * @param array &$results Put results here. Keys have to be integers.
 126          */
 127         protected function search( $search, $limit, $namespaces, $resolveRedir, &$results ) {
 128                 $searchEngine = MediaWikiServices::getInstance()->newSearchEngine();
 129                 $searchEngine->setLimitOffset( $limit );
 130                 $searchEngine->setNamespaces( $namespaces );
 131                 $titles = $searchEngine->extractTitles( $searchEngine->completionSearchWithVariants( $search ) );
 132
 133                 if ( !$titles ) {
 134                         return;
 135                 }
 136
 137                 // Special pages need unique integer ids in the return list, so we just
 138                 // assign them negative numbers because those won't clash with the
 139                 // always positive articleIds that non-special pages get.
 140                 $nextSpecialPageId = -1;
 141
 142                 if ( $resolveRedir ) {
 143                         // Query for redirects
 144                         $redirects = [];
 145                         $lb = new LinkBatch( $titles );
 146                         if ( !$lb->isEmpty() ) {
 147                                 $db = $this->getDB();
 148                                 $res = $db->select(
 149                                         [ 'page', 'redirect' ],
 150                                         [ 'page_namespace', 'page_title', 'rd_namespace', 'rd_title' ],
 151                                         [
 152                                                 'rd_from = page_id',
 153                                                 'rd_interwiki IS NULL OR rd_interwiki = ' . $db->addQuotes( '' ),
 154                                                 $lb->constructSet( 'page', $db ),
 155                                         ],
 156                                         __METHOD__
 157                                 );
 158                                 foreach ( $res as $row ) {
 159                                         $redirects[$row->page_namespace][$row->page_title] =
 160                                                 [ $row->rd_namespace, $row->rd_title ];
 161                                 }
 162                         }
 163
 164                         // Bypass any redirects
 165                         $seen = [];
 166                         foreach ( $titles as $title ) {
 167                                 $ns = $title->getNamespace();
 168                                 $dbkey = $title->getDBkey();
 169                                 $from = null;
 170                                 if ( isset( $redirects[$ns][$dbkey] ) ) {
 171                                         list( $ns, $dbkey ) = $redirects[$ns][$dbkey];
 172                                         $from = $title;
 173                                         $title = Title::makeTitle( $ns, $dbkey );
 174                                 }
 175                                 if ( !isset( $seen[$ns][$dbkey] ) ) {
 176                                         $seen[$ns][$dbkey] = true;
 177                                         $resultId = $title->getArticleID();
 178                                         if ( $resultId === 0 ) {
 179                                                 $resultId = $nextSpecialPageId;
 180                                                 $nextSpecialPageId -= 1;
 181                                         }
 182                                         $results[$resultId] = [
 183                                                 'title' => $title,
 184                                                 'redirect from' => $from,
 185                                                 'extract' => false,
 186                                                 'extract trimmed' => false,
 187                                                 'image' => false,
 188                                                 'url' => wfExpandUrl( $title->getFullURL(), PROTO_CURRENT ),
 189                                         ];
 190                                 }
 191                         }
 192                 } else {
 193                         foreach ( $titles as $title ) {
 194                                 $resultId = $title->getArticleID();
 195                                 if ( $resultId === 0 ) {
 196                                         $resultId = $nextSpecialPageId;
 197                                         $nextSpecialPageId -= 1;
 198                                 }
 199                                 $results[$resultId] = [
 200                                         'title' => $title,
 201                                         'redirect from' => null,
 202                                         'extract' => false,
 203                                         'extract trimmed' => false,
 204                                         'image' => false,
 205                                         'url' => wfExpandUrl( $title->getFullURL(), PROTO_CURRENT ),
 206                                 ];
 207                         }
 208                 }
 209         }
 210
 211         /**
 212          * @param string $search
 213          * @param array &$results
 214          */
 215         protected function populateResult( $search, &$results ) {
 216                 $result = $this->getResult();
 217
 218                 switch ( $this->getFormat() ) {
 219                         case 'json':
 220                                 // http://www.opensearch.org/Specifications/OpenSearch/Extensions/Suggestions/1.1
 221                                 $result->addArrayType( null, 'array' );
 222                                 $result->addValue( null, 0, strval( $search ) );
 223                                 $terms = [];
 224                                 $descriptions = [];
 225                                 $urls = [];
 226                                 foreach ( $results as $r ) {
 227                                         $terms[] = $r['title']->getPrefixedText();
 228                                         $descriptions[] = strval( $r['extract'] );
 229                                         $urls[] = $r['url'];
 230                                 }
 231                                 $result->addValue( null, 1, $terms );
 232                                 $result->addValue( null, 2, $descriptions );
 233                                 $result->addValue( null, 3, $urls );
 234                                 break;
 235
 236                         case 'xml':
 237                                 // http://msdn.microsoft.com/en-us/library/cc891508%28v=vs.85%29.aspx
 238                                 $imageKeys = [
 239                                         'source' => true,
 240                                         'alt' => true,
 241                                         'width' => true,
 242                                         'height' => true,
 243                                         'align' => true,
 244                                 ];
 245                                 $items = [];
 246                                 foreach ( $results as $r ) {
 247                                         $item = [
 248                                                 'Text' => $r['title']->getPrefixedText(),
 249                                                 'Url' => $r['url'],
 250                                         ];
 251                                         if ( is_string( $r['extract'] ) && $r['extract'] !== '' ) {
 252                                                 $item['Description'] = $r['extract'];
 253                                         }
 254                                         if ( is_array( $r['image'] ) && isset( $r['image']['source'] ) ) {
 255                                                 $item['Image'] = array_intersect_key( $r['image'], $imageKeys );
 256                                         }
 257                                         ApiResult::setSubelementsList( $item, array_keys( $item ) );
 258                                         $items[] = $item;
 259                                 }
 260                                 ApiResult::setIndexedTagName( $items, 'Item' );
 261                                 $result->addValue( null, 'version', '2.0' );
 262                                 $result->addValue( null, 'xmlns', 'http://opensearch.org/searchsuggest2' );
 263                                 $result->addValue( null, 'Query', strval( $search ) );
 264                                 $result->addSubelementsList( null, 'Query' );
 265                                 $result->addValue( null, 'Section', $items );
 266                                 break;
 267
 268                         default:
 269                                 ApiBase::dieDebug( __METHOD__, "Unsupported format '{$this->getFormat()}'" );
 270                 }
 271         }
 272
 273         public function getAllowedParams() {
 274                 return [
 275                         'search' => null,
 276                         'limit' => [
 277                                 ApiBase::PARAM_DFLT => $this->getConfig()->get( 'OpenSearchDefaultLimit' ),
 278                                 ApiBase::PARAM_TYPE => 'limit',
 279                                 ApiBase::PARAM_MIN => 1,
 280                                 ApiBase::PARAM_MAX => 100,
 281                                 ApiBase::PARAM_MAX2 => 100
 282                         ],
 283                         'namespace' => [
 284                                 ApiBase::PARAM_DFLT => NS_MAIN,
 285                                 ApiBase::PARAM_TYPE => 'namespace',
 286                                 ApiBase::PARAM_ISMULTI => true
 287                         ],
 288                         'suggest' => false,
 289                         'redirects' => [
 290                                 ApiBase::PARAM_TYPE => [ 'return', 'resolve' ],
 291                         ],
 292                         'format' => [
 293                                 ApiBase::PARAM_DFLT => 'json',
 294                                 ApiBase::PARAM_TYPE => [ 'json', 'jsonfm', 'xml', 'xmlfm' ],
 295                         ],
 296                         'warningsaserror' => false,
 297                 ];
 298         }
 299
 300         protected function getExamplesMessages() {
 301                 return [
 302                         'action=opensearch&search=Te'
 303                                 => 'apihelp-opensearch-example-te',
 304                 ];
 305         }
 306
 307         public function getHelpUrls() {
 308                 return 'https://www.mediawiki.org/wiki/API:Opensearch';
 309         }
 310
 311         /**
 312          * Trim an extract to a sensible length.
 313          *
 314          * Adapted from Extension:OpenSearchXml, which adapted it from
 315          * Extension:ActiveAbstract.
 316          *
 317          * @param string $text
 318          * @param int $length Target length; actual result will continue to the end of a sentence.
 319          * @return string
 320          */
 321         public static function trimExtract( $text, $length ) {
 322                 static $regex = null;
 323
 324                 if ( $regex === null ) {
 325                         $endchars = [
 326                                 '([^\d])\.\s', '\!\s', '\?\s', // regular ASCII
 327                                 '。', // full-width ideographic full-stop
 328                                 '．', '！', '？', // double-width roman forms
 329                                 '｡', // half-width ideographic full stop
 330                         ];
 331                         $endgroup = implode( '|', $endchars );
 332                         $end = "(?:$endgroup)";
 333                         $sentence = ".{{$length},}?$end+";
 334                         $regex = "/^($sentence)/u";
 335                 }
 336
 337                 $matches = [];
 338                 if ( preg_match( $regex, $text, $matches ) ) {
 339                         return trim( $matches[1] );
 340                 } else {
 341                         // Just return the first line
 342                         return trim( explode( "\n", $text )[0] );
 343                 }
 344         }
 345
 346         /**
 347          * Fetch the template for a type.
 348          *
 349          * @param string $type MIME type
 350          * @return string
 351          * @throws MWException
 352          */
 353         public static function getOpenSearchTemplate( $type ) {
 354                 $config = MediaWikiServices::getInstance()->getSearchEngineConfig();
 355                 $template = $config->getConfig()->get( 'OpenSearchTemplate' );
 356
 357                 if ( $template && $type === 'application/x-suggestions+json' ) {
 358                         return $template;
 359                 }
 360
 361                 $ns = implode( '|', $config->defaultNamespaces() );
 362                 if ( !$ns ) {
 363                         $ns = '0';
 364                 }
 365
 366                 switch ( $type ) {
 367                         case 'application/x-suggestions+json':
 368                                 return $config->getConfig()->get( 'CanonicalServer' ) . wfScript( 'api' )
 369                                         . '?action=opensearch&search={searchTerms}&namespace=' . $ns;
 370
 371                         case 'application/x-suggestions+xml':
 372                                 return $config->getConfig()->get( 'CanonicalServer' ) . wfScript( 'api' )
 373                                         . '?action=opensearch&format=xml&search={searchTerms}&namespace=' . $ns;
 374
 375                         default:
 376                                 throw new MWException( __METHOD__ . ": Unknown type '$type'" );
 377                 }
 378         }
 379 }
 380
 381 class ApiOpenSearchFormatJson extends ApiFormatJson {
 382         private $warningsAsError = false;
 383
 384         public function __construct( ApiMain $main, $fm, $warningsAsError ) {
 385                 parent::__construct( $main, "json$fm" );
 386                 $this->warningsAsError = $warningsAsError;
 387         }
 388
 389         public function execute() {
 390                 if ( !$this->getResult()->getResultData( 'error' ) ) {
 391                         $result = $this->getResult();
 392
 393                         // Ignore warnings or treat as errors, as requested
 394                         $warnings = $result->removeValue( 'warnings', null );
 395                         if ( $this->warningsAsError && $warnings ) {
 396                                 $this->dieUsage(
 397                                         'Warnings cannot be represented in OpenSearch JSON format', 'warnings', 0,
 398                                         [ 'warnings' => $warnings ]
 399                                 );
 400                         }
 401
 402                         // Ignore any other unexpected keys (e.g. from $wgDebugToolbar)
 403                         $remove = array_keys( array_diff_key(
 404                                 $result->getResultData(),
 405                                 [ 0 => 'search', 1 => 'terms', 2 => 'descriptions', 3 => 'urls' ]
 406                         ) );
 407                         foreach ( $remove as $key ) {
 408                                 $result->removeValue( $key, null );
 409                         }
 410                 }
 411
 412                 parent::execute();
 413         }
 414 }