includes/api/ApiQueryCategoryMembers.php

   1 <?php
   2 /**
   3  *
   4  *
   5  * Created on June 14, 2007
   6  *
   7  * Copyright © 2006 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  * A query module to enumerate pages that belong to a category.
  29  *
  30  * @ingroup API
  31  */
  32 class ApiQueryCategoryMembers extends ApiQueryGeneratorBase {
  33
  34         public function __construct( ApiQuery $query, $moduleName ) {
  35                 parent::__construct( $query, $moduleName, 'cm' );
  36         }
  37
  38         public function execute() {
  39                 $this->run();
  40         }
  41
  42         public function getCacheMode( $params ) {
  43                 return 'public';
  44         }
  45
  46         public function executeGenerator( $resultPageSet ) {
  47                 $this->run( $resultPageSet );
  48         }
  49
  50         /**
  51          * @param string $hexSortkey
  52          * @return bool
  53          */
  54         private function validateHexSortkey( $hexSortkey ) {
  55                 // A hex sortkey has an unbound number of 2 letter pairs
  56                 return preg_match( '/^(?:[a-fA-F0-9]{2})*$/', $hexSortkey );
  57         }
  58
  59         /**
  60          * @param ApiPageSet $resultPageSet
  61          * @return void
  62          */
  63         private function run( $resultPageSet = null ) {
  64                 $params = $this->extractRequestParams();
  65
  66                 $categoryTitle = $this->getTitleOrPageId( $params )->getTitle();
  67                 if ( $categoryTitle->getNamespace() != NS_CATEGORY ) {
  68                         $this->dieUsage( 'The category name you entered is not valid', 'invalidcategory' );
  69                 }
  70
  71                 $prop = array_flip( $params['prop'] );
  72                 $fld_ids = isset( $prop['ids'] );
  73                 $fld_title = isset( $prop['title'] );
  74                 $fld_sortkey = isset( $prop['sortkey'] );
  75                 $fld_sortkeyprefix = isset( $prop['sortkeyprefix'] );
  76                 $fld_timestamp = isset( $prop['timestamp'] );
  77                 $fld_type = isset( $prop['type'] );
  78
  79                 if ( is_null( $resultPageSet ) ) {
  80                         $this->addFields( array( 'cl_from', 'cl_sortkey', 'cl_type', 'page_namespace', 'page_title' ) );
  81                         $this->addFieldsIf( 'page_id', $fld_ids );
  82                         $this->addFieldsIf( 'cl_sortkey_prefix', $fld_sortkeyprefix );
  83                 } else {
  84                         $this->addFields( $resultPageSet->getPageTableFields() ); // will include page_ id, ns, title
  85                         $this->addFields( array( 'cl_from', 'cl_sortkey', 'cl_type' ) );
  86                 }
  87
  88                 $this->addFieldsIf( 'cl_timestamp', $fld_timestamp || $params['sort'] == 'timestamp' );
  89
  90                 $this->addTables( array( 'page', 'categorylinks' ) ); // must be in this order for 'USE INDEX'
  91
  92                 $this->addWhereFld( 'cl_to', $categoryTitle->getDBkey() );
  93                 $queryTypes = $params['type'];
  94                 $contWhere = false;
  95
  96                 // Scanning large datasets for rare categories sucks, and I already told
  97                 // how to have efficient subcategory access :-) ~~~~ (oh well, domas)
  98                 $miser_ns = array();
  99                 if ( $this->getConfig()->get( 'MiserMode' ) ) {
 100                         $miser_ns = $params['namespace'];
 101                 } else {
 102                         $this->addWhereFld( 'page_namespace', $params['namespace'] );
 103                 }
 104
 105                 $dir = in_array( $params['dir'], array( 'asc', 'ascending', 'newer' ) ) ? 'newer' : 'older';
 106
 107                 if ( $params['sort'] == 'timestamp' ) {
 108                         $this->addTimestampWhereRange( 'cl_timestamp',
 109                                 $dir,
 110                                 $params['start'],
 111                                 $params['end'] );
 112                         // Include in ORDER BY for uniqueness
 113                         $this->addWhereRange( 'cl_from', $dir, null, null );
 114
 115                         if ( !is_null( $params['continue'] ) ) {
 116                                 $cont = explode( '|', $params['continue'] );
 117                                 $this->dieContinueUsageIf( count( $cont ) != 2 );
 118                                 $op = ( $dir === 'newer' ? '>' : '<' );
 119                                 $db = $this->getDB();
 120                                 $continueTimestamp = $db->addQuotes( $db->timestamp( $cont[0] ) );
 121                                 $continueFrom = (int)$cont[1];
 122                                 $this->dieContinueUsageIf( $continueFrom != $cont[1] );
 123                                 $this->addWhere( "cl_timestamp $op $continueTimestamp OR " .
 124                                         "(cl_timestamp = $continueTimestamp AND " .
 125                                         "cl_from $op= $continueFrom)"
 126                                 );
 127                         }
 128
 129                         $this->addOption( 'USE INDEX', 'cl_timestamp' );
 130                 } else {
 131                         if ( $params['continue'] ) {
 132                                 $cont = explode( '|', $params['continue'], 3 );
 133                                 $this->dieContinueUsageIf( count( $cont ) != 3 );
 134
 135                                 // Remove the types to skip from $queryTypes
 136                                 $contTypeIndex = array_search( $cont[0], $queryTypes );
 137                                 $queryTypes = array_slice( $queryTypes, $contTypeIndex );
 138
 139                                 // Add a WHERE clause for sortkey and from
 140                                 $this->dieContinueUsageIf( !$this->validateHexSortkey( $cont[1] ) );
 141                                 // pack( "H*", $foo ) is used to convert hex back to binary
 142                                 $escSortkey = $this->getDB()->addQuotes( pack( 'H*', $cont[1] ) );
 143                                 $from = intval( $cont[2] );
 144                                 $op = $dir == 'newer' ? '>' : '<';
 145                                 // $contWhere is used further down
 146                                 $contWhere = "cl_sortkey $op $escSortkey OR " .
 147                                         "(cl_sortkey = $escSortkey AND " .
 148                                         "cl_from $op= $from)";
 149                                 // The below produces ORDER BY cl_sortkey, cl_from, possibly with DESC added to each of them
 150                                 $this->addWhereRange( 'cl_sortkey', $dir, null, null );
 151                                 $this->addWhereRange( 'cl_from', $dir, null, null );
 152                         } else {
 153                                 if ( $params['startsortkeyprefix'] !== null ) {
 154                                         $startsortkey = Collation::singleton()->getSortkey( $params['startsortkeyprefix'] );
 155                                 } elseif ( $params['starthexsortkey'] !== null ) {
 156                                         if ( !$this->validateHexSortkey( $params['starthexsortkey'] ) ) {
 157                                                 $this->dieUsage( 'The starthexsortkey provided is not valid', 'bad_starthexsortkey' );
 158                                         }
 159                                         $startsortkey = pack( 'H*', $params['starthexsortkey'] );
 160                                 } else {
 161                                         $this->logFeatureUsage( 'list=categorymembers&cmstartsortkey' );
 162                                         $startsortkey = $params['startsortkey'];
 163                                 }
 164                                 if ( $params['endsortkeyprefix'] !== null ) {
 165                                         $endsortkey = Collation::singleton()->getSortkey( $params['endsortkeyprefix'] );
 166                                 } elseif ( $params['endhexsortkey'] !== null ) {
 167                                         if ( !$this->validateHexSortkey( $params['endhexsortkey'] ) ) {
 168                                                 $this->dieUsage( 'The endhexsortkey provided is not valid', 'bad_endhexsortkey' );
 169                                         }
 170                                         $endsortkey = pack( 'H*', $params['endhexsortkey'] );
 171                                 } else {
 172                                         $this->logFeatureUsage( 'list=categorymembers&cmendsortkey' );
 173                                         $endsortkey = $params['endsortkey'];
 174                                 }
 175
 176                                 // The below produces ORDER BY cl_sortkey, cl_from, possibly with DESC added to each of them
 177                                 $this->addWhereRange( 'cl_sortkey',
 178                                         $dir,
 179                                         $startsortkey,
 180                                         $endsortkey );
 181                                 $this->addWhereRange( 'cl_from', $dir, null, null );
 182                         }
 183                         $this->addOption( 'USE INDEX', 'cl_sortkey' );
 184                 }
 185
 186                 $this->addWhere( 'cl_from=page_id' );
 187
 188                 $limit = $params['limit'];
 189                 $this->addOption( 'LIMIT', $limit + 1 );
 190
 191                 if ( $params['sort'] == 'sortkey' ) {
 192                         // Run a separate SELECT query for each value of cl_type.
 193                         // This is needed because cl_type is an enum, and MySQL has
 194                         // inconsistencies between ORDER BY cl_type and
 195                         // WHERE cl_type >= 'foo' making proper paging impossible
 196                         // and unindexed.
 197                         $rows = array();
 198                         $first = true;
 199                         foreach ( $queryTypes as $type ) {
 200                                 $extraConds = array( 'cl_type' => $type );
 201                                 if ( $first && $contWhere ) {
 202                                         // Continuation condition. Only added to the
 203                                         // first query, otherwise we'll skip things
 204                                         $extraConds[] = $contWhere;
 205                                 }
 206                                 $res = $this->select( __METHOD__, array( 'where' => $extraConds ) );
 207                                 $rows = array_merge( $rows, iterator_to_array( $res ) );
 208                                 if ( count( $rows ) >= $limit + 1 ) {
 209                                         break;
 210                                 }
 211                                 $first = false;
 212                         }
 213                 } else {
 214                         // Sorting by timestamp
 215                         // No need to worry about per-type queries because we
 216                         // aren't sorting or filtering by type anyway
 217                         $res = $this->select( __METHOD__ );
 218                         $rows = iterator_to_array( $res );
 219                 }
 220
 221                 $result = $this->getResult();
 222                 $count = 0;
 223                 foreach ( $rows as $row ) {
 224                         if ( ++$count > $limit ) {
 225                                 // We've reached the one extra which shows that there are
 226                                 // additional pages to be had. Stop here...
 227                                 // @todo Security issue - if the user has no right to view next
 228                                 // title, it will still be shown
 229                                 if ( $params['sort'] == 'timestamp' ) {
 230                                         $this->setContinueEnumParameter( 'continue', "$row->cl_timestamp|$row->cl_from" );
 231                                 } else {
 232                                         $sortkey = bin2hex( $row->cl_sortkey );
 233                                         $this->setContinueEnumParameter( 'continue',
 234                                                 "{$row->cl_type}|$sortkey|{$row->cl_from}"
 235                                         );
 236                                 }
 237                                 break;
 238                         }
 239
 240                         // Since domas won't tell anyone what he told long ago, apply
 241                         // cmnamespace here. This means the query may return 0 actual
 242                         // results, but on the other hand it could save returning 5000
 243                         // useless results to the client. ~~~~
 244                         if ( count( $miser_ns ) && !in_array( $row->page_namespace, $miser_ns ) ) {
 245                                 continue;
 246                         }
 247
 248                         if ( is_null( $resultPageSet ) ) {
 249                                 $vals = array(
 250                                         ApiResult::META_TYPE => 'assoc',
 251                                 );
 252                                 if ( $fld_ids ) {
 253                                         $vals['pageid'] = intval( $row->page_id );
 254                                 }
 255                                 if ( $fld_title ) {
 256                                         $title = Title::makeTitle( $row->page_namespace, $row->page_title );
 257                                         ApiQueryBase::addTitleInfo( $vals, $title );
 258                                 }
 259                                 if ( $fld_sortkey ) {
 260                                         $vals['sortkey'] = bin2hex( $row->cl_sortkey );
 261                                 }
 262                                 if ( $fld_sortkeyprefix ) {
 263                                         $vals['sortkeyprefix'] = $row->cl_sortkey_prefix;
 264                                 }
 265                                 if ( $fld_type ) {
 266                                         $vals['type'] = $row->cl_type;
 267                                 }
 268                                 if ( $fld_timestamp ) {
 269                                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $row->cl_timestamp );
 270                                 }
 271                                 $fit = $result->addValue( array( 'query', $this->getModuleName() ),
 272                                         null, $vals );
 273                                 if ( !$fit ) {
 274                                         if ( $params['sort'] == 'timestamp' ) {
 275                                                 $this->setContinueEnumParameter( 'continue', "$row->cl_timestamp|$row->cl_from" );
 276                                         } else {
 277                                                 $sortkey = bin2hex( $row->cl_sortkey );
 278                                                 $this->setContinueEnumParameter( 'continue',
 279                                                         "{$row->cl_type}|$sortkey|{$row->cl_from}"
 280                                                 );
 281                                         }
 282                                         break;
 283                                 }
 284                         } else {
 285                                 $resultPageSet->processDbRow( $row );
 286                         }
 287                 }
 288
 289                 if ( is_null( $resultPageSet ) ) {
 290                         $result->addIndexedTagName(
 291                                 array( 'query', $this->getModuleName() ), 'cm' );
 292                 }
 293         }
 294
 295         public function getAllowedParams() {
 296                 $ret = array(
 297                         'title' => array(
 298                                 ApiBase::PARAM_TYPE => 'string',
 299                         ),
 300                         'pageid' => array(
 301                                 ApiBase::PARAM_TYPE => 'integer'
 302                         ),
 303                         'prop' => array(
 304                                 ApiBase::PARAM_DFLT => 'ids|title',
 305                                 ApiBase::PARAM_ISMULTI => true,
 306                                 ApiBase::PARAM_TYPE => array(
 307                                         'ids',
 308                                         'title',
 309                                         'sortkey',
 310                                         'sortkeyprefix',
 311                                         'type',
 312                                         'timestamp',
 313                                 )
 314                         ),
 315                         'namespace' => array(
 316                                 ApiBase::PARAM_ISMULTI => true,
 317                                 ApiBase::PARAM_TYPE => 'namespace',
 318                         ),
 319                         'type' => array(
 320                                 ApiBase::PARAM_ISMULTI => true,
 321                                 ApiBase::PARAM_DFLT => 'page|subcat|file',
 322                                 ApiBase::PARAM_TYPE => array(
 323                                         'page',
 324                                         'subcat',
 325                                         'file'
 326                                 )
 327                         ),
 328                         'continue' => array(
 329                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-continue',
 330                         ),
 331                         'limit' => array(
 332                                 ApiBase::PARAM_TYPE => 'limit',
 333                                 ApiBase::PARAM_DFLT => 10,
 334                                 ApiBase::PARAM_MIN => 1,
 335                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_BIG1,
 336                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_BIG2
 337                         ),
 338                         'sort' => array(
 339                                 ApiBase::PARAM_DFLT => 'sortkey',
 340                                 ApiBase::PARAM_TYPE => array(
 341                                         'sortkey',
 342                                         'timestamp'
 343                                 )
 344                         ),
 345                         'dir' => array(
 346                                 ApiBase::PARAM_DFLT => 'ascending',
 347                                 ApiBase::PARAM_TYPE => array(
 348                                         'asc',
 349                                         'desc',
 350                                         // Normalising with other modules
 351                                         'ascending',
 352                                         'descending',
 353                                         'newer',
 354                                         'older',
 355                                 )
 356                         ),
 357                         'start' => array(
 358                                 ApiBase::PARAM_TYPE => 'timestamp'
 359                         ),
 360                         'end' => array(
 361                                 ApiBase::PARAM_TYPE => 'timestamp'
 362                         ),
 363                         'starthexsortkey' => null,
 364                         'endhexsortkey' => null,
 365                         'startsortkeyprefix' => null,
 366                         'endsortkeyprefix' => null,
 367                         'startsortkey' => array(
 368                                 ApiBase::PARAM_DEPRECATED => true,
 369                         ),
 370                         'endsortkey' => array(
 371                                 ApiBase::PARAM_DEPRECATED => true,
 372                         ),
 373                 );
 374
 375                 if ( $this->getConfig()->get( 'MiserMode' ) ) {
 376                         $ret['namespace'][ApiBase::PARAM_HELP_MSG_APPEND] = array(
 377                                 'api-help-param-limited-in-miser-mode',
 378                         );
 379                 }
 380
 381                 return $ret;
 382         }
 383
 384         protected function getExamplesMessages() {
 385                 return array(
 386                         'action=query&list=categorymembers&cmtitle=Category:Physics'
 387                                 => 'apihelp-query+categorymembers-example-simple',
 388                         'action=query&generator=categorymembers&gcmtitle=Category:Physics&prop=info'
 389                                 => 'apihelp-query+categorymembers-example-generator',
 390                 );
 391         }
 392
 393         public function getHelpUrls() {
 394                 return 'https://www.mediawiki.org/wiki/API:Categorymembers';
 395         }
 396 }