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 ApiPageSet $resultPageSet
  52          * @return void
  53          */
  54         private function run( $resultPageSet = null ) {
  55                 $params = $this->extractRequestParams();
  56
  57                 $categoryTitle = $this->getTitleOrPageId( $params )->getTitle();
  58                 if ( $categoryTitle->getNamespace() != NS_CATEGORY ) {
  59                         $this->dieUsage( 'The category name you entered is not valid', 'invalidcategory' );
  60                 }
  61
  62                 $prop = array_flip( $params['prop'] );
  63                 $fld_ids = isset( $prop['ids'] );
  64                 $fld_title = isset( $prop['title'] );
  65                 $fld_sortkey = isset( $prop['sortkey'] );
  66                 $fld_sortkeyprefix = isset( $prop['sortkeyprefix'] );
  67                 $fld_timestamp = isset( $prop['timestamp'] );
  68                 $fld_type = isset( $prop['type'] );
  69
  70                 if ( is_null( $resultPageSet ) ) {
  71                         $this->addFields( array( 'cl_from', 'cl_sortkey', 'cl_type', 'page_namespace', 'page_title' ) );
  72                         $this->addFieldsIf( 'page_id', $fld_ids );
  73                         $this->addFieldsIf( 'cl_sortkey_prefix', $fld_sortkeyprefix );
  74                 } else {
  75                         $this->addFields( $resultPageSet->getPageTableFields() ); // will include page_ id, ns, title
  76                         $this->addFields( array( 'cl_from', 'cl_sortkey', 'cl_type' ) );
  77                 }
  78
  79                 $this->addFieldsIf( 'cl_timestamp', $fld_timestamp || $params['sort'] == 'timestamp' );
  80
  81                 $this->addTables( array( 'page', 'categorylinks' ) ); // must be in this order for 'USE INDEX'
  82
  83                 $this->addWhereFld( 'cl_to', $categoryTitle->getDBkey() );
  84                 $queryTypes = $params['type'];
  85                 $contWhere = false;
  86
  87                 // Scanning large datasets for rare categories sucks, and I already told
  88                 // how to have efficient subcategory access :-) ~~~~ (oh well, domas)
  89                 $miser_ns = array();
  90                 if ( $this->getConfig()->get( 'MiserMode' ) ) {
  91                         $miser_ns = $params['namespace'];
  92                 } else {
  93                         $this->addWhereFld( 'page_namespace', $params['namespace'] );
  94                 }
  95
  96                 $dir = in_array( $params['dir'], array( 'asc', 'ascending', 'newer' ) ) ? 'newer' : 'older';
  97
  98                 if ( $params['sort'] == 'timestamp' ) {
  99                         $this->addTimestampWhereRange( 'cl_timestamp',
 100                                 $dir,
 101                                 $params['start'],
 102                                 $params['end'] );
 103                         // Include in ORDER BY for uniqueness
 104                         $this->addWhereRange( 'cl_from', $dir, null, null );
 105
 106                         if ( !is_null( $params['continue'] ) ) {
 107                                 $cont = explode( '|', $params['continue'] );
 108                                 $this->dieContinueUsageIf( count( $cont ) != 2 );
 109                                 $op = ( $dir === 'newer' ? '>' : '<' );
 110                                 $db = $this->getDB();
 111                                 $continueTimestamp = $db->addQuotes( $db->timestamp( $cont[0] ) );
 112                                 $continueFrom = (int)$cont[1];
 113                                 $this->dieContinueUsageIf( $continueFrom != $cont[1] );
 114                                 $this->addWhere( "cl_timestamp $op $continueTimestamp OR " .
 115                                         "(cl_timestamp = $continueTimestamp AND " .
 116                                         "cl_from $op= $continueFrom)"
 117                                 );
 118                         }
 119
 120                         $this->addOption( 'USE INDEX', 'cl_timestamp' );
 121                 } else {
 122                         if ( $params['continue'] ) {
 123                                 $cont = explode( '|', $params['continue'], 3 );
 124                                 $this->dieContinueUsageIf( count( $cont ) != 3 );
 125
 126                                 // Remove the types to skip from $queryTypes
 127                                 $contTypeIndex = array_search( $cont[0], $queryTypes );
 128                                 $queryTypes = array_slice( $queryTypes, $contTypeIndex );
 129
 130                                 // Add a WHERE clause for sortkey and from
 131                                 // pack( "H*", $foo ) is used to convert hex back to binary
 132                                 $escSortkey = $this->getDB()->addQuotes( pack( 'H*', $cont[1] ) );
 133                                 $from = intval( $cont[2] );
 134                                 $op = $dir == 'newer' ? '>' : '<';
 135                                 // $contWhere is used further down
 136                                 $contWhere = "cl_sortkey $op $escSortkey OR " .
 137                                         "(cl_sortkey = $escSortkey AND " .
 138                                         "cl_from $op= $from)";
 139                                 // The below produces ORDER BY cl_sortkey, cl_from, possibly with DESC added to each of them
 140                                 $this->addWhereRange( 'cl_sortkey', $dir, null, null );
 141                                 $this->addWhereRange( 'cl_from', $dir, null, null );
 142                         } else {
 143                                 $startsortkey = $params['startsortkeyprefix'] !== null ?
 144                                         Collation::singleton()->getSortkey( $params['startsortkeyprefix'] ) :
 145                                         $params['startsortkey'];
 146                                 $endsortkey = $params['endsortkeyprefix'] !== null ?
 147                                         Collation::singleton()->getSortkey( $params['endsortkeyprefix'] ) :
 148                                         $params['endsortkey'];
 149
 150                                 // The below produces ORDER BY cl_sortkey, cl_from, possibly with DESC added to each of them
 151                                 $this->addWhereRange( 'cl_sortkey',
 152                                         $dir,
 153                                         $startsortkey,
 154                                         $endsortkey );
 155                                 $this->addWhereRange( 'cl_from', $dir, null, null );
 156                         }
 157                         $this->addOption( 'USE INDEX', 'cl_sortkey' );
 158                 }
 159
 160                 $this->addWhere( 'cl_from=page_id' );
 161
 162                 $limit = $params['limit'];
 163                 $this->addOption( 'LIMIT', $limit + 1 );
 164
 165                 if ( $params['sort'] == 'sortkey' ) {
 166                         // Run a separate SELECT query for each value of cl_type.
 167                         // This is needed because cl_type is an enum, and MySQL has
 168                         // inconsistencies between ORDER BY cl_type and
 169                         // WHERE cl_type >= 'foo' making proper paging impossible
 170                         // and unindexed.
 171                         $rows = array();
 172                         $first = true;
 173                         foreach ( $queryTypes as $type ) {
 174                                 $extraConds = array( 'cl_type' => $type );
 175                                 if ( $first && $contWhere ) {
 176                                         // Continuation condition. Only added to the
 177                                         // first query, otherwise we'll skip things
 178                                         $extraConds[] = $contWhere;
 179                                 }
 180                                 $res = $this->select( __METHOD__, array( 'where' => $extraConds ) );
 181                                 $rows = array_merge( $rows, iterator_to_array( $res ) );
 182                                 if ( count( $rows ) >= $limit + 1 ) {
 183                                         break;
 184                                 }
 185                                 $first = false;
 186                         }
 187                 } else {
 188                         // Sorting by timestamp
 189                         // No need to worry about per-type queries because we
 190                         // aren't sorting or filtering by type anyway
 191                         $res = $this->select( __METHOD__ );
 192                         $rows = iterator_to_array( $res );
 193                 }
 194
 195                 $result = $this->getResult();
 196                 $count = 0;
 197                 foreach ( $rows as $row ) {
 198                         if ( ++$count > $limit ) {
 199                                 // We've reached the one extra which shows that there are
 200                                 // additional pages to be had. Stop here...
 201                                 // @todo Security issue - if the user has no right to view next
 202                                 // title, it will still be shown
 203                                 if ( $params['sort'] == 'timestamp' ) {
 204                                         $this->setContinueEnumParameter( 'continue', "$row->cl_timestamp|$row->cl_from" );
 205                                 } else {
 206                                         $sortkey = bin2hex( $row->cl_sortkey );
 207                                         $this->setContinueEnumParameter( 'continue',
 208                                                 "{$row->cl_type}|$sortkey|{$row->cl_from}"
 209                                         );
 210                                 }
 211                                 break;
 212                         }
 213
 214                         // Since domas won't tell anyone what he told long ago, apply
 215                         // cmnamespace here. This means the query may return 0 actual
 216                         // results, but on the other hand it could save returning 5000
 217                         // useless results to the client. ~~~~
 218                         if ( count( $miser_ns ) && !in_array( $row->page_namespace, $miser_ns ) ) {
 219                                 continue;
 220                         }
 221
 222                         if ( is_null( $resultPageSet ) ) {
 223                                 $vals = array();
 224                                 if ( $fld_ids ) {
 225                                         $vals['pageid'] = intval( $row->page_id );
 226                                 }
 227                                 if ( $fld_title ) {
 228                                         $title = Title::makeTitle( $row->page_namespace, $row->page_title );
 229                                         ApiQueryBase::addTitleInfo( $vals, $title );
 230                                 }
 231                                 if ( $fld_sortkey ) {
 232                                         $vals['sortkey'] = bin2hex( $row->cl_sortkey );
 233                                 }
 234                                 if ( $fld_sortkeyprefix ) {
 235                                         $vals['sortkeyprefix'] = $row->cl_sortkey_prefix;
 236                                 }
 237                                 if ( $fld_type ) {
 238                                         $vals['type'] = $row->cl_type;
 239                                 }
 240                                 if ( $fld_timestamp ) {
 241                                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $row->cl_timestamp );
 242                                 }
 243                                 $fit = $result->addValue( array( 'query', $this->getModuleName() ),
 244                                         null, $vals );
 245                                 if ( !$fit ) {
 246                                         if ( $params['sort'] == 'timestamp' ) {
 247                                                 $this->setContinueEnumParameter( 'continue', "$row->cl_timestamp|$row->cl_from" );
 248                                         } else {
 249                                                 $sortkey = bin2hex( $row->cl_sortkey );
 250                                                 $this->setContinueEnumParameter( 'continue',
 251                                                         "{$row->cl_type}|$sortkey|{$row->cl_from}"
 252                                                 );
 253                                         }
 254                                         break;
 255                                 }
 256                         } else {
 257                                 $resultPageSet->processDbRow( $row );
 258                         }
 259                 }
 260
 261                 if ( is_null( $resultPageSet ) ) {
 262                         $result->setIndexedTagName_internal(
 263                                 array( 'query', $this->getModuleName() ), 'cm' );
 264                 }
 265         }
 266
 267         public function getAllowedParams() {
 268                 return array(
 269                         'title' => array(
 270                                 ApiBase::PARAM_TYPE => 'string',
 271                         ),
 272                         'pageid' => array(
 273                                 ApiBase::PARAM_TYPE => 'integer'
 274                         ),
 275                         'prop' => array(
 276                                 ApiBase::PARAM_DFLT => 'ids|title',
 277                                 ApiBase::PARAM_ISMULTI => true,
 278                                 ApiBase::PARAM_TYPE => array(
 279                                         'ids',
 280                                         'title',
 281                                         'sortkey',
 282                                         'sortkeyprefix',
 283                                         'type',
 284                                         'timestamp',
 285                                 )
 286                         ),
 287                         'namespace' => array(
 288                                 ApiBase::PARAM_ISMULTI => true,
 289                                 ApiBase::PARAM_TYPE => 'namespace',
 290                         ),
 291                         'type' => array(
 292                                 ApiBase::PARAM_ISMULTI => true,
 293                                 ApiBase::PARAM_DFLT => 'page|subcat|file',
 294                                 ApiBase::PARAM_TYPE => array(
 295                                         'page',
 296                                         'subcat',
 297                                         'file'
 298                                 )
 299                         ),
 300                         'continue' => null,
 301                         'limit' => array(
 302                                 ApiBase::PARAM_TYPE => 'limit',
 303                                 ApiBase::PARAM_DFLT => 10,
 304                                 ApiBase::PARAM_MIN => 1,
 305                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_BIG1,
 306                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_BIG2
 307                         ),
 308                         'sort' => array(
 309                                 ApiBase::PARAM_DFLT => 'sortkey',
 310                                 ApiBase::PARAM_TYPE => array(
 311                                         'sortkey',
 312                                         'timestamp'
 313                                 )
 314                         ),
 315                         'dir' => array(
 316                                 ApiBase::PARAM_DFLT => 'ascending',
 317                                 ApiBase::PARAM_TYPE => array(
 318                                         'asc',
 319                                         'desc',
 320                                         // Normalising with other modules
 321                                         'ascending',
 322                                         'descending',
 323                                         'newer',
 324                                         'older',
 325                                 )
 326                         ),
 327                         'start' => array(
 328                                 ApiBase::PARAM_TYPE => 'timestamp'
 329                         ),
 330                         'end' => array(
 331                                 ApiBase::PARAM_TYPE => 'timestamp'
 332                         ),
 333                         'startsortkey' => null,
 334                         'endsortkey' => null,
 335                         'startsortkeyprefix' => null,
 336                         'endsortkeyprefix' => null,
 337                 );
 338         }
 339
 340         public function getParamDescription() {
 341                 $p = $this->getModulePrefix();
 342                 $desc = array(
 343                         'title' => "Which category to enumerate (required). Must include " .
 344                                 "'Category:' prefix. Cannot be used together with {$p}pageid",
 345                         'pageid' => "Page ID of the category to enumerate. Cannot be used together with {$p}title",
 346                         'prop' => array(
 347                                 'What pieces of information to include',
 348                                 ' ids           - Adds the page ID',
 349                                 ' title         - Adds the title and namespace ID of the page',
 350                                 ' sortkey       - Adds the sortkey used for sorting in the category (hexadecimal string)',
 351                                 ' sortkeyprefix - Adds the sortkey prefix used for sorting in the ' .
 352                                         'category (human-readable part of the sortkey)',
 353                                 ' type          - Adds the type that the page has been categorised as (page, subcat or file)',
 354                                 ' timestamp     - Adds the timestamp of when the page was included',
 355                         ),
 356                         'namespace' => 'Only include pages in these namespaces',
 357                         'type' => "What type of category members to include. Ignored when {$p}sort=timestamp is set",
 358                         'sort' => 'Property to sort by',
 359                         'dir' => 'In which direction to sort',
 360                         'start' => "Timestamp to start listing from. Can only be used with {$p}sort=timestamp",
 361                         'end' => "Timestamp to end listing at. Can only be used with {$p}sort=timestamp",
 362                         'startsortkey' => "Sortkey to start listing from. Must be given in " .
 363                                 "binary format. Can only be used with {$p}sort=sortkey",
 364                         'endsortkey' => "Sortkey to end listing at. Must be given in binary " .
 365                                 "format. Can only be used with {$p}sort=sortkey",
 366                         'startsortkeyprefix' => "Sortkey prefix to start listing from. Can " .
 367                                 "only be used with {$p}sort=sortkey. Overrides {$p}startsortkey",
 368                         'endsortkeyprefix' => "Sortkey prefix to end listing BEFORE (not at, " .
 369                                 "if this value occurs it will not be included!). Can only be used with " .
 370                                 "{$p}sort=sortkey. Overrides {$p}endsortkey",
 371                         'continue' => 'For large categories, give the value returned from previous query',
 372                         'limit' => 'The maximum number of pages to return.',
 373                 );
 374
 375                 if ( $this->getConfig()->get( 'MiserMode' ) ) {
 376                         $desc['namespace'] = array(
 377                                 $desc['namespace'],
 378                                 "NOTE: Due to \$wgMiserMode, using this may result in fewer than \"{$p}limit\" results",
 379                                 'returned before continuing; in extreme cases, zero results may be returned.',
 380                                 "Note that you can use {$p}type=subcat or {$p}type=file instead of {$p}namespace=14 or 6.",
 381                         );
 382                 }
 383
 384                 return $desc;
 385         }
 386
 387         public function getDescription() {
 388                 return 'List all pages in a given category.';
 389         }
 390
 391         public function getExamples() {
 392                 return array(
 393                         'api.php?action=query&list=categorymembers&cmtitle=Category:Physics'
 394                                 => 'Get first 10 pages in [[Category:Physics]]',
 395                         'api.php?action=query&generator=categorymembers&gcmtitle=Category:Physics&prop=info'
 396                                 => 'Get page info about first 10 pages in [[Category:Physics]]',
 397                 );
 398         }
 399
 400         public function getHelpUrls() {
 401                 return 'https://www.mediawiki.org/wiki/API:Categorymembers';
 402         }
 403 }