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