includes/api/ApiQueryCategoryMembers.php

   1 <?php
   2 /**
   3  * Copyright © 2006 Yuri Astrakhan "<Firstname><Lastname>@gmail.com"
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  */
  22
  23 /**
  24  * A query module to enumerate pages that belong to a category.
  25  *
  26  * @ingroup API
  27  */
  28 class ApiQueryCategoryMembers extends ApiQueryGeneratorBase {
  29
  30         public function __construct( ApiQuery $query, $moduleName ) {
  31                 parent::__construct( $query, $moduleName, 'cm' );
  32         }
  33
  34         public function execute() {
  35                 $this->run();
  36         }
  37
  38         public function getCacheMode( $params ) {
  39                 return 'public';
  40         }
  41
  42         public function executeGenerator( $resultPageSet ) {
  43                 $this->run( $resultPageSet );
  44         }
  45
  46         /**
  47          * @param string $hexSortkey
  48          * @return bool
  49          */
  50         private function validateHexSortkey( $hexSortkey ) {
  51                 // A hex sortkey has an unbound number of 2 letter pairs
  52                 return (bool)preg_match( '/^(?:[a-fA-F0-9]{2})*$/D', $hexSortkey );
  53         }
  54
  55         /**
  56          * @param ApiPageSet $resultPageSet
  57          * @return void
  58          */
  59         private function run( $resultPageSet = null ) {
  60                 $params = $this->extractRequestParams();
  61
  62                 $categoryTitle = $this->getTitleOrPageId( $params )->getTitle();
  63                 if ( $categoryTitle->getNamespace() != NS_CATEGORY ) {
  64                         $this->dieWithError( 'apierror-invalidcategory' );
  65                 }
  66
  67                 $prop = array_flip( $params['prop'] );
  68                 $fld_ids = isset( $prop['ids'] );
  69                 $fld_title = isset( $prop['title'] );
  70                 $fld_sortkey = isset( $prop['sortkey'] );
  71                 $fld_sortkeyprefix = isset( $prop['sortkeyprefix'] );
  72                 $fld_timestamp = isset( $prop['timestamp'] );
  73                 $fld_type = isset( $prop['type'] );
  74
  75                 if ( is_null( $resultPageSet ) ) {
  76                         $this->addFields( [ 'cl_from', 'cl_sortkey', 'cl_type', 'page_namespace', 'page_title' ] );
  77                         $this->addFieldsIf( 'page_id', $fld_ids );
  78                         $this->addFieldsIf( 'cl_sortkey_prefix', $fld_sortkeyprefix );
  79                 } else {
  80                         $this->addFields( $resultPageSet->getPageTableFields() ); // will include page_ id, ns, title
  81                         $this->addFields( [ 'cl_from', 'cl_sortkey', 'cl_type' ] );
  82                 }
  83
  84                 $this->addFieldsIf( 'cl_timestamp', $fld_timestamp || $params['sort'] == 'timestamp' );
  85
  86                 $this->addTables( [ 'page', 'categorylinks' ] ); // must be in this order for 'USE INDEX'
  87
  88                 $this->addWhereFld( 'cl_to', $categoryTitle->getDBkey() );
  89                 $queryTypes = $params['type'];
  90                 $contWhere = false;
  91
  92                 // Scanning large datasets for rare categories sucks, and I already told
  93                 // how to have efficient subcategory access :-) ~~~~ (oh well, domas)
  94                 $miser_ns = [];
  95                 if ( $this->getConfig()->get( 'MiserMode' ) ) {
  96                         $miser_ns = $params['namespace'] ?: [];
  97                 } else {
  98                         $this->addWhereFld( 'page_namespace', $params['namespace'] );
  99                 }
 100
 101                 $dir = in_array( $params['dir'], [ 'asc', 'ascending', 'newer' ] ) ? 'newer' : 'older';
 102
 103                 if ( $params['sort'] == 'timestamp' ) {
 104                         $this->addTimestampWhereRange( 'cl_timestamp',
 105                                 $dir,
 106                                 $params['start'],
 107                                 $params['end'] );
 108                         // Include in ORDER BY for uniqueness
 109                         $this->addWhereRange( 'cl_from', $dir, null, null );
 110
 111                         if ( !is_null( $params['continue'] ) ) {
 112                                 $cont = explode( '|', $params['continue'] );
 113                                 $this->dieContinueUsageIf( count( $cont ) != 2 );
 114                                 $op = ( $dir === 'newer' ? '>' : '<' );
 115                                 $db = $this->getDB();
 116                                 $continueTimestamp = $db->addQuotes( $db->timestamp( $cont[0] ) );
 117                                 $continueFrom = (int)$cont[1];
 118                                 $this->dieContinueUsageIf( $continueFrom != $cont[1] );
 119                                 $this->addWhere( "cl_timestamp $op $continueTimestamp OR " .
 120                                         "(cl_timestamp = $continueTimestamp AND " .
 121                                         "cl_from $op= $continueFrom)"
 122                                 );
 123                         }
 124
 125                         $this->addOption( 'USE INDEX', 'cl_timestamp' );
 126                 } else {
 127                         if ( $params['continue'] ) {
 128                                 $cont = explode( '|', $params['continue'], 3 );
 129                                 $this->dieContinueUsageIf( count( $cont ) != 3 );
 130
 131                                 // Remove the types to skip from $queryTypes
 132                                 $contTypeIndex = array_search( $cont[0], $queryTypes );
 133                                 $queryTypes = array_slice( $queryTypes, $contTypeIndex );
 134
 135                                 // Add a WHERE clause for sortkey and from
 136                                 $this->dieContinueUsageIf( !$this->validateHexSortkey( $cont[1] ) );
 137                                 $escSortkey = $this->getDB()->addQuotes( hex2bin( $cont[1] ) );
 138                                 $from = intval( $cont[2] );
 139                                 $op = $dir == 'newer' ? '>' : '<';
 140                                 // $contWhere is used further down
 141                                 $contWhere = "cl_sortkey $op $escSortkey OR " .
 142                                         "(cl_sortkey = $escSortkey AND " .
 143                                         "cl_from $op= $from)";
 144                                 // The below produces ORDER BY cl_sortkey, cl_from, possibly with DESC added to each of them
 145                                 $this->addWhereRange( 'cl_sortkey', $dir, null, null );
 146                                 $this->addWhereRange( 'cl_from', $dir, null, null );
 147                         } else {
 148                                 if ( $params['startsortkeyprefix'] !== null ) {
 149                                         $startsortkey = Collation::singleton()->getSortKey( $params['startsortkeyprefix'] );
 150                                 } elseif ( $params['starthexsortkey'] !== null ) {
 151                                         if ( !$this->validateHexSortkey( $params['starthexsortkey'] ) ) {
 152                                                 $encParamName = $this->encodeParamName( 'starthexsortkey' );
 153                                                 $this->dieWithError( [ 'apierror-badparameter', $encParamName ], "badvalue_$encParamName" );
 154                                         }
 155                                         $startsortkey = hex2bin( $params['starthexsortkey'] );
 156                                 } else {
 157                                         $startsortkey = $params['startsortkey'];
 158                                 }
 159                                 if ( $params['endsortkeyprefix'] !== null ) {
 160                                         $endsortkey = Collation::singleton()->getSortKey( $params['endsortkeyprefix'] );
 161                                 } elseif ( $params['endhexsortkey'] !== null ) {
 162                                         if ( !$this->validateHexSortkey( $params['endhexsortkey'] ) ) {
 163                                                 $encParamName = $this->encodeParamName( 'endhexsortkey' );
 164                                                 $this->dieWithError( [ 'apierror-badparameter', $encParamName ], "badvalue_$encParamName" );
 165                                         }
 166                                         $endsortkey = hex2bin( $params['endhexsortkey'] );
 167                                 } else {
 168                                         $endsortkey = $params['endsortkey'];
 169                                 }
 170
 171                                 // The below produces ORDER BY cl_sortkey, cl_from, possibly with DESC added to each of them
 172                                 $this->addWhereRange( 'cl_sortkey',
 173                                         $dir,
 174                                         $startsortkey,
 175                                         $endsortkey );
 176                                 $this->addWhereRange( 'cl_from', $dir, null, null );
 177                         }
 178                         $this->addOption( 'USE INDEX', 'cl_sortkey' );
 179                 }
 180
 181                 $this->addWhere( 'cl_from=page_id' );
 182
 183                 $limit = $params['limit'];
 184                 $this->addOption( 'LIMIT', $limit + 1 );
 185
 186                 if ( $params['sort'] == 'sortkey' ) {
 187                         // Run a separate SELECT query for each value of cl_type.
 188                         // This is needed because cl_type is an enum, and MySQL has
 189                         // inconsistencies between ORDER BY cl_type and
 190                         // WHERE cl_type >= 'foo' making proper paging impossible
 191                         // and unindexed.
 192                         $rows = [];
 193                         $first = true;
 194                         foreach ( $queryTypes as $type ) {
 195                                 $extraConds = [ 'cl_type' => $type ];
 196                                 if ( $first && $contWhere ) {
 197                                         // Continuation condition. Only added to the
 198                                         // first query, otherwise we'll skip things
 199                                         $extraConds[] = $contWhere;
 200                                 }
 201                                 $res = $this->select( __METHOD__, [ 'where' => $extraConds ] );
 202                                 $rows = array_merge( $rows, iterator_to_array( $res ) );
 203                                 if ( count( $rows ) >= $limit + 1 ) {
 204                                         break;
 205                                 }
 206                                 $first = false;
 207                         }
 208                 } else {
 209                         // Sorting by timestamp
 210                         // No need to worry about per-type queries because we
 211                         // aren't sorting or filtering by type anyway
 212                         $res = $this->select( __METHOD__ );
 213                         $rows = iterator_to_array( $res );
 214                 }
 215
 216                 $result = $this->getResult();
 217                 $count = 0;
 218                 foreach ( $rows as $row ) {
 219                         if ( ++$count > $limit ) {
 220                                 // We've reached the one extra which shows that there are
 221                                 // additional pages to be had. Stop here...
 222                                 // @todo Security issue - if the user has no right to view next
 223                                 // title, it will still be shown
 224                                 if ( $params['sort'] == 'timestamp' ) {
 225                                         $this->setContinueEnumParameter( 'continue', "$row->cl_timestamp|$row->cl_from" );
 226                                 } else {
 227                                         $sortkey = bin2hex( $row->cl_sortkey );
 228                                         $this->setContinueEnumParameter( 'continue',
 229                                                 "{$row->cl_type}|$sortkey|{$row->cl_from}"
 230                                         );
 231                                 }
 232                                 break;
 233                         }
 234
 235                         // Since domas won't tell anyone what he told long ago, apply
 236                         // cmnamespace here. This means the query may return 0 actual
 237                         // results, but on the other hand it could save returning 5000
 238                         // useless results to the client. ~~~~
 239                         if ( count( $miser_ns ) && !in_array( $row->page_namespace, $miser_ns ) ) {
 240                                 continue;
 241                         }
 242
 243                         if ( is_null( $resultPageSet ) ) {
 244                                 $vals = [
 245                                         ApiResult::META_TYPE => 'assoc',
 246                                 ];
 247                                 if ( $fld_ids ) {
 248                                         $vals['pageid'] = intval( $row->page_id );
 249                                 }
 250                                 if ( $fld_title ) {
 251                                         $title = Title::makeTitle( $row->page_namespace, $row->page_title );
 252                                         ApiQueryBase::addTitleInfo( $vals, $title );
 253                                 }
 254                                 if ( $fld_sortkey ) {
 255                                         $vals['sortkey'] = bin2hex( $row->cl_sortkey );
 256                                 }
 257                                 if ( $fld_sortkeyprefix ) {
 258                                         $vals['sortkeyprefix'] = $row->cl_sortkey_prefix;
 259                                 }
 260                                 if ( $fld_type ) {
 261                                         $vals['type'] = $row->cl_type;
 262                                 }
 263                                 if ( $fld_timestamp ) {
 264                                         $vals['timestamp'] = wfTimestamp( TS_ISO_8601, $row->cl_timestamp );
 265                                 }
 266                                 $fit = $result->addValue( [ 'query', $this->getModuleName() ],
 267                                         null, $vals );
 268                                 if ( !$fit ) {
 269                                         if ( $params['sort'] == 'timestamp' ) {
 270                                                 $this->setContinueEnumParameter( 'continue', "$row->cl_timestamp|$row->cl_from" );
 271                                         } else {
 272                                                 $sortkey = bin2hex( $row->cl_sortkey );
 273                                                 $this->setContinueEnumParameter( 'continue',
 274                                                         "{$row->cl_type}|$sortkey|{$row->cl_from}"
 275                                                 );
 276                                         }
 277                                         break;
 278                                 }
 279                         } else {
 280                                 $resultPageSet->processDbRow( $row );
 281                         }
 282                 }
 283
 284                 if ( is_null( $resultPageSet ) ) {
 285                         $result->addIndexedTagName(
 286                                 [ 'query', $this->getModuleName() ], 'cm' );
 287                 }
 288         }
 289
 290         public function getAllowedParams() {
 291                 $ret = [
 292                         'title' => [
 293                                 ApiBase::PARAM_TYPE => 'string',
 294                         ],
 295                         'pageid' => [
 296                                 ApiBase::PARAM_TYPE => 'integer'
 297                         ],
 298                         'prop' => [
 299                                 ApiBase::PARAM_DFLT => 'ids|title',
 300                                 ApiBase::PARAM_ISMULTI => true,
 301                                 ApiBase::PARAM_TYPE => [
 302                                         'ids',
 303                                         'title',
 304                                         'sortkey',
 305                                         'sortkeyprefix',
 306                                         'type',
 307                                         'timestamp',
 308                                 ],
 309                                 ApiBase::PARAM_HELP_MSG_PER_VALUE => [],
 310                         ],
 311                         'namespace' => [
 312                                 ApiBase::PARAM_ISMULTI => true,
 313                                 ApiBase::PARAM_TYPE => 'namespace',
 314                         ],
 315                         'type' => [
 316                                 ApiBase::PARAM_ISMULTI => true,
 317                                 ApiBase::PARAM_DFLT => 'page|subcat|file',
 318                                 ApiBase::PARAM_TYPE => [
 319                                         'page',
 320                                         'subcat',
 321                                         'file'
 322                                 ]
 323                         ],
 324                         'continue' => [
 325                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-continue',
 326                         ],
 327                         'limit' => [
 328                                 ApiBase::PARAM_TYPE => 'limit',
 329                                 ApiBase::PARAM_DFLT => 10,
 330                                 ApiBase::PARAM_MIN => 1,
 331                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_BIG1,
 332                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_BIG2
 333                         ],
 334                         'sort' => [
 335                                 ApiBase::PARAM_DFLT => 'sortkey',
 336                                 ApiBase::PARAM_TYPE => [
 337                                         'sortkey',
 338                                         'timestamp'
 339                                 ]
 340                         ],
 341                         'dir' => [
 342                                 ApiBase::PARAM_DFLT => 'ascending',
 343                                 ApiBase::PARAM_TYPE => [
 344                                         'asc',
 345                                         'desc',
 346                                         // Normalising with other modules
 347                                         'ascending',
 348                                         'descending',
 349                                         'newer',
 350                                         'older',
 351                                 ]
 352                         ],
 353                         'start' => [
 354                                 ApiBase::PARAM_TYPE => 'timestamp'
 355                         ],
 356                         'end' => [
 357                                 ApiBase::PARAM_TYPE => 'timestamp'
 358                         ],
 359                         'starthexsortkey' => null,
 360                         'endhexsortkey' => null,
 361                         'startsortkeyprefix' => null,
 362                         'endsortkeyprefix' => null,
 363                         'startsortkey' => [
 364                                 ApiBase::PARAM_DEPRECATED => true,
 365                         ],
 366                         'endsortkey' => [
 367                                 ApiBase::PARAM_DEPRECATED => true,
 368                         ],
 369                 ];
 370
 371                 if ( $this->getConfig()->get( 'MiserMode' ) ) {
 372                         $ret['namespace'][ApiBase::PARAM_HELP_MSG_APPEND] = [
 373                                 'api-help-param-limited-in-miser-mode',
 374                         ];
 375                 }
 376
 377                 return $ret;
 378         }
 379
 380         protected function getExamplesMessages() {
 381                 return [
 382                         'action=query&list=categorymembers&cmtitle=Category:Physics'
 383                                 => 'apihelp-query+categorymembers-example-simple',
 384                         'action=query&generator=categorymembers&gcmtitle=Category:Physics&prop=info'
 385                                 => 'apihelp-query+categorymembers-example-generator',
 386                 ];
 387         }
 388
 389         public function getHelpUrls() {
 390                 return 'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Categorymembers';
 391         }
 392 }