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