includes/api/ApiQueryAllPages.php

   1 <?php
   2 /**
   3  *
   4  *
   5  * Created on Sep 25, 2006
   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  * Query module to enumerate all available pages.
  29  *
  30  * @ingroup API
  31  */
  32 class ApiQueryAllPages extends ApiQueryGeneratorBase {
  33
  34         public function __construct( $query, $moduleName ) {
  35                 parent::__construct( $query, $moduleName, 'ap' );
  36         }
  37
  38         public function execute() {
  39                 $this->run();
  40         }
  41
  42         public function getCacheMode( $params ) {
  43                 return 'public';
  44         }
  45
  46         /**
  47          * @param $resultPageSet ApiPageSet
  48          * @return void
  49          */
  50         public function executeGenerator( $resultPageSet ) {
  51                 if ( $resultPageSet->isResolvingRedirects() ) {
  52                         $this->dieUsage( 'Use "gapfilterredir=nonredirects" option instead of "redirects" when using allpages as a generator', 'params' );
  53                 }
  54
  55                 $this->run( $resultPageSet );
  56         }
  57
  58         /**
  59          * @param $resultPageSet ApiPageSet
  60          * @return void
  61          */
  62         private function run( $resultPageSet = null ) {
  63                 $db = $this->getDB();
  64
  65                 $params = $this->extractRequestParams();
  66
  67                 // Page filters
  68                 $this->addTables( 'page' );
  69
  70                 if ( !is_null( $params['continue'] ) ) {
  71                         $cont = explode( '|', $params['continue'] );
  72                         if ( count( $cont ) != 1 ) {
  73                                 $this->dieUsage( "Invalid continue param. You should pass the " .
  74                                         "original value returned by the previous query", "_badcontinue" );
  75                         }
  76                         $op = $params['dir'] == 'descending' ? '<' : '>';
  77                         $cont_from = $db->addQuotes( $cont[0] );
  78                         $this->addWhere( "page_title $op= $cont_from" );
  79                 }
  80
  81                 if ( $params['filterredir'] == 'redirects' ) {
  82                         $this->addWhereFld( 'page_is_redirect', 1 );
  83                 } elseif ( $params['filterredir'] == 'nonredirects' ) {
  84                         $this->addWhereFld( 'page_is_redirect', 0 );
  85                 }
  86
  87                 $this->addWhereFld( 'page_namespace', $params['namespace'] );
  88                 $dir = ( $params['dir'] == 'descending' ? 'older' : 'newer' );
  89                 $from = ( is_null( $params['from'] ) ? null : $this->titlePartToKey( $params['from'] ) );
  90                 $to = ( is_null( $params['to'] ) ? null : $this->titlePartToKey( $params['to'] ) );
  91                 $this->addWhereRange( 'page_title', $dir, $from, $to );
  92
  93                 if ( isset( $params['prefix'] ) ) {
  94                         $this->addWhere( 'page_title' . $db->buildLike( $this->titlePartToKey( $params['prefix'] ), $db->anyString() ) );
  95                 }
  96
  97                 if ( is_null( $resultPageSet ) ) {
  98                         $selectFields = array(
  99                                 'page_namespace',
 100                                 'page_title',
 101                                 'page_id'
 102                         );
 103                 } else {
 104                         $selectFields = $resultPageSet->getPageTableFields();
 105                 }
 106
 107                 $this->addFields( $selectFields );
 108                 $forceNameTitleIndex = true;
 109                 if ( isset( $params['minsize'] ) ) {
 110                         $this->addWhere( 'page_len>=' . intval( $params['minsize'] ) );
 111                         $forceNameTitleIndex = false;
 112                 }
 113
 114                 if ( isset( $params['maxsize'] ) ) {
 115                         $this->addWhere( 'page_len<=' . intval( $params['maxsize'] ) );
 116                         $forceNameTitleIndex = false;
 117                 }
 118
 119                 // Page protection filtering
 120                 if ( count( $params['prtype'] ) || $params['prexpiry'] != 'all' ) {
 121                         $this->addTables( 'page_restrictions' );
 122                         $this->addWhere( 'page_id=pr_page' );
 123                         $this->addWhere( 'pr_expiry>' . $db->addQuotes( $db->timestamp() ) );
 124
 125                         if ( count( $params['prtype'] ) ) {
 126                                 $this->addWhereFld( 'pr_type', $params['prtype'] );
 127
 128                                 if ( isset( $params['prlevel'] ) ) {
 129                                         // Remove the empty string and '*' from the prlevel array
 130                                         $prlevel = array_diff( $params['prlevel'], array( '', '*' ) );
 131
 132                                         if ( count( $prlevel ) ) {
 133                                                 $this->addWhereFld( 'pr_level', $prlevel );
 134                                         }
 135                                 }
 136                                 if ( $params['prfiltercascade'] == 'cascading' ) {
 137                                         $this->addWhereFld( 'pr_cascade', 1 );
 138                                 } elseif ( $params['prfiltercascade'] == 'noncascading' ) {
 139                                         $this->addWhereFld( 'pr_cascade', 0 );
 140                                 }
 141
 142                                 $this->addOption( 'DISTINCT' );
 143                         }
 144                         $forceNameTitleIndex = false;
 145
 146                         if ( $params['prexpiry'] == 'indefinite' ) {
 147                                 $this->addWhere( "pr_expiry = {$db->addQuotes( $db->getInfinity() )} OR pr_expiry IS NULL" );
 148                         } elseif ( $params['prexpiry'] == 'definite' ) {
 149                                 $this->addWhere( "pr_expiry != {$db->addQuotes( $db->getInfinity() )}" );
 150                         }
 151
 152                 } elseif ( isset( $params['prlevel'] ) ) {
 153                         $this->dieUsage( 'prlevel may not be used without prtype', 'params' );
 154                 }
 155
 156                 if ( $params['filterlanglinks'] == 'withoutlanglinks' ) {
 157                         $this->addTables( 'langlinks' );
 158                         $this->addJoinConds( array( 'langlinks' => array( 'LEFT JOIN', 'page_id=ll_from' ) ) );
 159                         $this->addWhere( 'll_from IS NULL' );
 160                         $forceNameTitleIndex = false;
 161                 } elseif ( $params['filterlanglinks'] == 'withlanglinks' ) {
 162                         $this->addTables( 'langlinks' );
 163                         $this->addWhere( 'page_id=ll_from' );
 164                         $this->addOption( 'STRAIGHT_JOIN' );
 165                         // We have to GROUP BY all selected fields to stop
 166                         // PostgreSQL from whining
 167                         $this->addOption( 'GROUP BY', $selectFields );
 168                         $forceNameTitleIndex = false;
 169                 }
 170
 171                 if ( $forceNameTitleIndex ) {
 172                         $this->addOption( 'USE INDEX', 'name_title' );
 173                 }
 174
 175                 $limit = $params['limit'];
 176                 $this->addOption( 'LIMIT', $limit + 1 );
 177                 $res = $this->select( __METHOD__ );
 178
 179                 //Get gender information
 180                 if( MWNamespace::hasGenderDistinction( $params['namespace'] ) ) {
 181                         $users = array();
 182                         foreach ( $res as $row ) {
 183                                 $users[] = $row->page_title;
 184                         }
 185                         GenderCache::singleton()->doQuery( $users, __METHOD__ );
 186                         $res->rewind(); //reset
 187                 }
 188
 189                 $count = 0;
 190                 $result = $this->getResult();
 191                 foreach ( $res as $row ) {
 192                         if ( ++ $count > $limit ) {
 193                                 // We've reached the one extra which shows that there are additional pages to be had. Stop here...
 194                                 $this->setContinueEnumParameter( 'continue', $row->page_title );
 195                                 break;
 196                         }
 197
 198                         if ( is_null( $resultPageSet ) ) {
 199                                 $title = Title::makeTitle( $row->page_namespace, $row->page_title );
 200                                 $vals = array(
 201                                         'pageid' => intval( $row->page_id ),
 202                                         'ns' => intval( $title->getNamespace() ),
 203                                         'title' => $title->getPrefixedText()
 204                                 );
 205                                 $fit = $result->addValue( array( 'query', $this->getModuleName() ), null, $vals );
 206                                 if ( !$fit ) {
 207                                         $this->setContinueEnumParameter( 'continue', $row->page_title );
 208                                         break;
 209                                 }
 210                         } else {
 211                                 $resultPageSet->processDbRow( $row );
 212                         }
 213                 }
 214
 215                 if ( is_null( $resultPageSet ) ) {
 216                         $result->setIndexedTagName_internal( array( 'query', $this->getModuleName() ), 'p' );
 217                 }
 218         }
 219
 220         public function getAllowedParams() {
 221                 global $wgRestrictionLevels;
 222
 223                 return array(
 224                         'from' => null,
 225                         'continue' => null,
 226                         'to' => null,
 227                         'prefix' => null,
 228                         'namespace' => array(
 229                                 ApiBase::PARAM_DFLT => 0,
 230                                 ApiBase::PARAM_TYPE => 'namespace',
 231                         ),
 232                         'filterredir' => array(
 233                                 ApiBase::PARAM_DFLT => 'all',
 234                                 ApiBase::PARAM_TYPE => array(
 235                                         'all',
 236                                         'redirects',
 237                                         'nonredirects'
 238                                 )
 239                         ),
 240                         'minsize' => array(
 241                                 ApiBase::PARAM_TYPE => 'integer',
 242                         ),
 243                         'maxsize' => array(
 244                                 ApiBase::PARAM_TYPE => 'integer',
 245                         ),
 246                         'prtype' => array(
 247                                 ApiBase::PARAM_TYPE => Title::getFilteredRestrictionTypes( true ),
 248                                 ApiBase::PARAM_ISMULTI => true
 249                         ),
 250                         'prlevel' => array(
 251                                 ApiBase::PARAM_TYPE => $wgRestrictionLevels,
 252                                 ApiBase::PARAM_ISMULTI => true
 253                         ),
 254                         'prfiltercascade' => array(
 255                                 ApiBase::PARAM_DFLT => 'all',
 256                                 ApiBase::PARAM_TYPE => array(
 257                                         'cascading',
 258                                         'noncascading',
 259                                         'all'
 260                                 ),
 261                         ),
 262                         'limit' => array(
 263                                 ApiBase::PARAM_DFLT => 10,
 264                                 ApiBase::PARAM_TYPE => 'limit',
 265                                 ApiBase::PARAM_MIN => 1,
 266                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_BIG1,
 267                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_BIG2
 268                         ),
 269                         'dir' => array(
 270                                 ApiBase::PARAM_DFLT => 'ascending',
 271                                 ApiBase::PARAM_TYPE => array(
 272                                         'ascending',
 273                                         'descending'
 274                                 )
 275                         ),
 276                         'filterlanglinks' => array(
 277                                 ApiBase::PARAM_TYPE => array(
 278                                         'withlanglinks',
 279                                         'withoutlanglinks',
 280                                         'all'
 281                                 ),
 282                                 ApiBase::PARAM_DFLT => 'all'
 283                         ),
 284                         'prexpiry' => array(
 285                                 ApiBase::PARAM_TYPE => array(
 286                                         'indefinite',
 287                                         'definite',
 288                                         'all'
 289                                 ),
 290                                 ApiBase::PARAM_DFLT => 'all'
 291                         ),
 292                 );
 293         }
 294
 295         public function getParamDescription() {
 296                 $p = $this->getModulePrefix();
 297                 return array(
 298                         'from' => 'The page title to start enumerating from',
 299                         'continue' => 'When more results are available, use this to continue',
 300                         'to' => 'The page title to stop enumerating at',
 301                         'prefix' => 'Search for all page titles that begin with this value',
 302                         'namespace' => 'The namespace to enumerate',
 303                         'filterredir' => 'Which pages to list',
 304                         'dir' => 'The direction in which to list',
 305                         'minsize' => 'Limit to pages with at least this many bytes',
 306                         'maxsize' => 'Limit to pages with at most this many bytes',
 307                         'prtype' => 'Limit to protected pages only',
 308                         'prlevel' => "The protection level (must be used with {$p}prtype= parameter)",
 309                         'prfiltercascade' => "Filter protections based on cascadingness (ignored when {$p}prtype isn't set)",
 310                         'filterlanglinks' => 'Filter based on whether a page has langlinks',
 311                         'limit' => 'How many total pages to return.',
 312                         'prexpiry' => array(
 313                                 'Which protection expiry to filter the page on',
 314                                 ' indefinite - Get only pages with indefinite protection expiry',
 315                                 ' definite - Get only pages with a definite (specific) protection expiry',
 316                                 ' all - Get pages with any protections expiry'
 317                         ),
 318                 );
 319         }
 320
 321         public function getResultProperties() {
 322                 return array(
 323                         '' => array(
 324                                 'pageid' => 'integer',
 325                                 'ns' => 'namespace',
 326                                 'title' => 'string'
 327                         )
 328                 );
 329         }
 330
 331         public function getDescription() {
 332                 return 'Enumerate all pages sequentially in a given namespace';
 333         }
 334
 335         public function getPossibleErrors() {
 336                 return array_merge( parent::getPossibleErrors(), array(
 337                         array( 'code' => 'params', 'info' => 'Use "gapfilterredir=nonredirects" option instead of "redirects" when using allpages as a generator' ),
 338                         array( 'code' => 'params', 'info' => 'prlevel may not be used without prtype' ),
 339                         array( 'code' => '_badcontinue', 'info' => 'Invalid continue param. You should pass the original value returned by the previous query' ),
 340                 ) );
 341         }
 342
 343         public function getExamples() {
 344                 return array(
 345                         'api.php?action=query&list=allpages&apfrom=B' => array(
 346                                 'Simple Use',
 347                                 'Show a list of pages starting at the letter "B"',
 348                         ),
 349                         'api.php?action=query&generator=allpages&gaplimit=4&gapfrom=T&prop=info' => array(
 350                                 'Using as Generator',
 351                                 'Show info about 4 pages starting at the letter "T"',
 352                         ),
 353                         'api.php?action=query&generator=allpages&gaplimit=2&gapfilterredir=nonredirects&gapfrom=Re&prop=revisions&rvprop=content' => array(
 354                                 'Show content of first 2 non-redirect pages begining at "Re"',
 355                         )
 356                 );
 357         }
 358
 359         public function getHelpUrls() {
 360                 return 'https://www.mediawiki.org/wiki/API:Allpages';
 361         }
 362
 363         public function getVersion() {
 364                 return __CLASS__ . ': $Id$';
 365         }
 366 }