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                                         if ( $params['startsortkey'] !== null ) {
 162                                                 $this->logFeatureUsage( 'list=categorymembers&cmstartsortkey' );
 163                                         }
 164                                         $startsortkey = $params['startsortkey'];
 165                                 }
 166                                 if ( $params['endsortkeyprefix'] !== null ) {
 167                                         $endsortkey = Collation::singleton()->getSortkey( $params['endsortkeyprefix'] );
 168                                 } elseif ( $params['endhexsortkey'] !== null ) {
 169                                         if ( !$this->validateHexSortkey( $params['endhexsortkey'] ) ) {
 170                                                 $this->dieUsage( 'The endhexsortkey provided is not valid', 'bad_endhexsortkey' );
 171                                         }
 172                                         $endsortkey = pack( 'H*', $params['endhexsortkey'] );
 173                                 } else {
 174                                         if ( $params['endsortkey'] !== null ) {
 175                                                 $this->logFeatureUsage( 'list=categorymembers&cmendsortkey' );
 176                                         }
 177                                         $endsortkey = $params['endsortkey'];
 178                                 }
 179
 180                                 // The below produces ORDER BY cl_sortkey, cl_from, possibly with DESC added to each of them
 181                                 $this->addWhereRange( 'cl_sortkey',
 182                                         $dir,
 183                                         $startsortkey,
 184                                         $endsortkey );
 185                                 $this->addWhereRange( 'cl_from', $dir, null, null );
 186                         }
 187                         $this->addOption( 'USE INDEX', 'cl_sortkey' );
 188                 }
 189
 190                 $this->addWhere( 'cl_from=page_id' );
 191
 192                 $limit = $params['limit'];
 193                 $this->addOption( 'LIMIT', $limit + 1 );
 194
 195                 if ( $params['sort'] == 'sortkey' ) {
 196                         // Run a separate SELECT query for each value of cl_type.
 197                         // This is needed because cl_type is an enum, and MySQL has
 198                         // inconsistencies between ORDER BY cl_type and
 199                         // WHERE cl_type >= 'foo' making proper paging impossible
 200                         // and unindexed.
 201                         $rows = array();
 202                         $first = true;
 203                         foreach ( $queryTypes as $type ) {
 204                                 $extraConds = array( 'cl_type' => $type );
 205                                 if ( $first && $contWhere ) {
 206                                         // Continuation condition. Only added to the
 207                                         // first query, otherwise we'll skip things
 208                                         $extraConds[] = $contWhere;
 209                                 }
 210                                 $res = $this->select( __METHOD__, array( 'where' => $extraConds ) );
 211                                 $rows = array_merge( $rows, iterator_to_array( $res ) );
 212                                 if ( count( $rows ) >= $limit + 1 ) {
 213                                         break;
 214                                 }
 215                                 $first = false;
 216                         }
 217                 } else {
 218                         // Sorting by timestamp
 219                         // No need to worry about per-type queries because we
 220                         // aren't sorting or filtering by type anyway
 221                         $res = $this->select( __METHOD__ );
 222                         $rows = iterator_to_array( $res );
 223                 }
 224
 225                 $result = $this->getResult();
 226                 $count = 0;
 227                 foreach ( $rows as $row ) {
 228                         if ( ++$count > $limit ) {
 229                                 // We've reached the one extra which shows that there are
 230                                 // additional pages to be had. Stop here...
 231                                 // @todo Security issue - if the user has no right to view next
 232                                 // title, it will still be shown
 233                                 if ( $params['sort'] == 'timestamp' ) {
 234                                         $this->setContinueEnumParameter( 'continue', "$row->cl_timestamp|$row->cl_from" );
 235                                 } else {
 236                                         $sortkey = bin2hex( $row->cl_sortkey );
 237                                         $this->setContinueEnumParameter( 'continue',
 238                                                 "{$row->cl_type}|$sortkey|{$row->cl_from}"
 239                                         );
 240                                 }
 241                                 break;
 242                         }
 243
 244                         // Since domas won't tell anyone what he told long ago, apply
 245                         // cmnamespace here. This means the query may return 0 actual
 246                         // results, but on the other hand it could save returning 5000
 247                         // useless results to the client. ~~~~
 248                         if ( count( $miser_ns ) && !in_array( $row->page_namespace, $miser_ns ) ) {
 249                                 continue;
 250                         }
 251
 252                         if ( is_null( $resultPageSet ) ) {
 253                                 $vals = array(
 254                                         ApiResult::META_TYPE => 'assoc',
 255                                 );
 256                                 if ( $fld_ids ) {
 257                                         $vals['pageid'] = intval( $row->page_id );
 258                                 }
 259                                 if ( $fld_title ) {
 260                                         $title = Title::makeTitle( $row->page_namespace, $row->page_title );
 261                                         ApiQueryBase::addTitleInfo( $vals, $title );
 262                                 }
 263                                 if ( $fld_sortkey ) {
 264                                         $vals['sortkey'] = bin2hex( $row->cl_sortkey );
 265                                 }
 266                                 if ( $fld_sortkeyprefix ) {
 267                                         $vals['sortkeyprefix'] = $row->cl_sortkey_prefix;
 268                                 }
 269                                 if ( $fld_type ) {
 270                                         $vals['type'] = $row->cl_type;
 271                                 }
 272                                 if ( $fld_timestamp ) {
 273                                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $row->cl_timestamp );
 274                                 }
 275                                 $fit = $result->addValue( array( 'query', $this->getModuleName() ),
 276                                         null, $vals );
 277                                 if ( !$fit ) {
 278                                         if ( $params['sort'] == 'timestamp' ) {
 279                                                 $this->setContinueEnumParameter( 'continue', "$row->cl_timestamp|$row->cl_from" );
 280                                         } else {
 281                                                 $sortkey = bin2hex( $row->cl_sortkey );
 282                                                 $this->setContinueEnumParameter( 'continue',
 283                                                         "{$row->cl_type}|$sortkey|{$row->cl_from}"
 284                                                 );
 285                                         }
 286                                         break;
 287                                 }
 288                         } else {
 289                                 $resultPageSet->processDbRow( $row );
 290                         }
 291                 }
 292
 293                 if ( is_null( $resultPageSet ) ) {
 294                         $result->addIndexedTagName(
 295                                 array( 'query', $this->getModuleName() ), 'cm' );
 296                 }
 297         }
 298
 299         public function getAllowedParams() {
 300                 $ret = array(
 301                         'title' => array(
 302                                 ApiBase::PARAM_TYPE => 'string',
 303                         ),
 304                         'pageid' => array(
 305                                 ApiBase::PARAM_TYPE => 'integer'
 306                         ),
 307                         'prop' => array(
 308                                 ApiBase::PARAM_DFLT => 'ids|title',
 309                                 ApiBase::PARAM_ISMULTI => true,
 310                                 ApiBase::PARAM_TYPE => array(
 311                                         'ids',
 312                                         'title',
 313                                         'sortkey',
 314                                         'sortkeyprefix',
 315                                         'type',
 316                                         'timestamp',
 317                                 ),
 318                                 ApiBase::PARAM_HELP_MSG_PER_VALUE => array(),
 319                         ),
 320                         'namespace' => array(
 321                                 ApiBase::PARAM_ISMULTI => true,
 322                                 ApiBase::PARAM_TYPE => 'namespace',
 323                         ),
 324                         'type' => array(
 325                                 ApiBase::PARAM_ISMULTI => true,
 326                                 ApiBase::PARAM_DFLT => 'page|subcat|file',
 327                                 ApiBase::PARAM_TYPE => array(
 328                                         'page',
 329                                         'subcat',
 330                                         'file'
 331                                 )
 332                         ),
 333                         'continue' => array(
 334                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-continue',
 335                         ),
 336                         'limit' => array(
 337                                 ApiBase::PARAM_TYPE => 'limit',
 338                                 ApiBase::PARAM_DFLT => 10,
 339                                 ApiBase::PARAM_MIN => 1,
 340                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_BIG1,
 341                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_BIG2
 342                         ),
 343                         'sort' => array(
 344                                 ApiBase::PARAM_DFLT => 'sortkey',
 345                                 ApiBase::PARAM_TYPE => array(
 346                                         'sortkey',
 347                                         'timestamp'
 348                                 )
 349                         ),
 350                         'dir' => array(
 351                                 ApiBase::PARAM_DFLT => 'ascending',
 352                                 ApiBase::PARAM_TYPE => array(
 353                                         'asc',
 354                                         'desc',
 355                                         // Normalising with other modules
 356                                         'ascending',
 357                                         'descending',
 358                                         'newer',
 359                                         'older',
 360                                 )
 361                         ),
 362                         'start' => array(
 363                                 ApiBase::PARAM_TYPE => 'timestamp'
 364                         ),
 365                         'end' => array(
 366                                 ApiBase::PARAM_TYPE => 'timestamp'
 367                         ),
 368                         'starthexsortkey' => null,
 369                         'endhexsortkey' => null,
 370                         'startsortkeyprefix' => null,
 371                         'endsortkeyprefix' => null,
 372                         'startsortkey' => array(
 373                                 ApiBase::PARAM_DEPRECATED => true,
 374                         ),
 375                         'endsortkey' => array(
 376                                 ApiBase::PARAM_DEPRECATED => true,
 377                         ),
 378                 );
 379
 380                 if ( $this->getConfig()->get( 'MiserMode' ) ) {
 381                         $ret['namespace'][ApiBase::PARAM_HELP_MSG_APPEND] = array(
 382                                 'api-help-param-limited-in-miser-mode',
 383                         );
 384                 }
 385
 386                 return $ret;
 387         }
 388
 389         protected function getExamplesMessages() {
 390                 return array(
 391                         'action=query&list=categorymembers&cmtitle=Category:Physics'
 392                                 => 'apihelp-query+categorymembers-example-simple',
 393                         'action=query&generator=categorymembers&gcmtitle=Category:Physics&prop=info'
 394                                 => 'apihelp-query+categorymembers-example-generator',
 395                 );
 396         }
 397
 398         public function getHelpUrls() {
 399                 return 'https://www.mediawiki.org/wiki/API:Categorymembers';
 400         }
 401 }