includes/api/ApiQueryBacklinksprop.php

   1 <?php
   2 /**
   3  * API module to handle links table back-queries
   4  *
   5  * Created on Aug 19, 2014
   6  *
   7  * Copyright © 2014 Brad Jorsch <bjorsch@wikimedia.org>
   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  * @since 1.24
  26  */
  27
  28 /**
  29  * This implements prop=redirects, prop=linkshere, prop=catmembers,
  30  * prop=transcludedin, and prop=fileusage
  31  *
  32  * @ingroup API
  33  * @since 1.24
  34  */
  35 class ApiQueryBacklinksprop extends ApiQueryGeneratorBase {
  36
  37         // Data for the various modules implemented by this class
  38         private static $settings = [
  39                 'redirects' => [
  40                         'code' => 'rd',
  41                         'prefix' => 'rd',
  42                         'linktable' => 'redirect',
  43                         'props' => [
  44                                 'fragment',
  45                         ],
  46                         'showredirects' => false,
  47                         'show' => [
  48                                 'fragment',
  49                                 '!fragment',
  50                         ],
  51                 ],
  52                 'linkshere' => [
  53                         'code' => 'lh',
  54                         'prefix' => 'pl',
  55                         'linktable' => 'pagelinks',
  56                         'indexes' => [ 'pl_namespace', 'pl_backlinks_namespace' ],
  57                         'from_namespace' => true,
  58                         'showredirects' => true,
  59                 ],
  60                 'transcludedin' => [
  61                         'code' => 'ti',
  62                         'prefix' => 'tl',
  63                         'linktable' => 'templatelinks',
  64                         'indexes' => [ 'tl_namespace', 'tl_backlinks_namespace' ],
  65                         'from_namespace' => true,
  66                         'showredirects' => true,
  67                 ],
  68                 'fileusage' => [
  69                         'code' => 'fu',
  70                         'prefix' => 'il',
  71                         'linktable' => 'imagelinks',
  72                         'indexes' => [ 'il_to', 'il_backlinks_namespace' ],
  73                         'from_namespace' => true,
  74                         'to_namespace' => NS_FILE,
  75                         'exampletitle' => 'File:Example.jpg',
  76                         'showredirects' => true,
  77                 ],
  78         ];
  79
  80         public function __construct( ApiQuery $query, $moduleName ) {
  81                 parent::__construct( $query, $moduleName, self::$settings[$moduleName]['code'] );
  82         }
  83
  84         public function execute() {
  85                 $this->run();
  86         }
  87
  88         public function executeGenerator( $resultPageSet ) {
  89                 $this->run( $resultPageSet );
  90         }
  91
  92         /**
  93          * @param ApiPageSet $resultPageSet
  94          */
  95         private function run( ApiPageSet $resultPageSet = null ) {
  96                 $settings = self::$settings[$this->getModuleName()];
  97
  98                 $db = $this->getDB();
  99                 $params = $this->extractRequestParams();
 100                 $prop = array_flip( $params['prop'] );
 101                 $emptyString = $db->addQuotes( '' );
 102
 103                 $pageSet = $this->getPageSet();
 104                 $titles = $pageSet->getGoodAndMissingTitles();
 105                 $map = $pageSet->getGoodAndMissingTitlesByNamespace();
 106
 107                 // Add in special pages, they can theoretically have backlinks too.
 108                 // (although currently they only do for prop=redirects)
 109                 foreach ( $pageSet->getSpecialTitles() as $id => $title ) {
 110                         $titles[] = $title;
 111                         $map[$title->getNamespace()][$title->getDBkey()] = $id;
 112                 }
 113
 114                 // Determine our fields to query on
 115                 $p = $settings['prefix'];
 116                 $hasNS = !isset( $settings['to_namespace'] );
 117                 if ( $hasNS ) {
 118                         $bl_namespace = "{$p}_namespace";
 119                         $bl_title = "{$p}_title";
 120                 } else {
 121                         $bl_namespace = $settings['to_namespace'];
 122                         $bl_title = "{$p}_to";
 123
 124                         $titles = array_filter( $titles, function ( $t ) use ( $bl_namespace ) {
 125                                 return $t->getNamespace() === $bl_namespace;
 126                         } );
 127                         $map = array_intersect_key( $map, [ $bl_namespace => true ] );
 128                 }
 129                 $bl_from = "{$p}_from";
 130
 131                 if ( !$titles ) {
 132                         return; // nothing to do
 133                 }
 134
 135                 // Figure out what we're sorting by, and add associated WHERE clauses.
 136                 // MySQL's query planner screws up if we include a field in ORDER BY
 137                 // when it's constant in WHERE, so we have to test that for each field.
 138                 $sortby = [];
 139                 if ( $hasNS && count( $map ) > 1 ) {
 140                         $sortby[$bl_namespace] = 'ns';
 141                 }
 142                 $theTitle = null;
 143                 foreach ( $map as $nsTitles ) {
 144                         reset( $nsTitles );
 145                         $key = key( $nsTitles );
 146                         if ( $theTitle === null ) {
 147                                 $theTitle = $key;
 148                         }
 149                         if ( count( $nsTitles ) > 1 || $key !== $theTitle ) {
 150                                 $sortby[$bl_title] = 'title';
 151                                 break;
 152                         }
 153                 }
 154                 $miser_ns = null;
 155                 if ( $params['namespace'] !== null ) {
 156                         if ( empty( $settings['from_namespace'] ) ) {
 157                                 if ( $this->getConfig()->get( 'MiserMode' ) ) {
 158                                         $miser_ns = $params['namespace'];
 159                                 } else {
 160                                         $this->addWhereFld( 'page_namespace', $params['namespace'] );
 161                                 }
 162                         } else {
 163                                 $this->addWhereFld( "{$p}_from_namespace", $params['namespace'] );
 164                                 if ( !empty( $settings['from_namespace'] ) && count( $params['namespace'] ) > 1 ) {
 165                                         $sortby["{$p}_from_namespace"] = 'int';
 166                                 }
 167                         }
 168                 }
 169                 $sortby[$bl_from] = 'int';
 170
 171                 // Now use the $sortby to figure out the continuation
 172                 if ( !is_null( $params['continue'] ) ) {
 173                         $cont = explode( '|', $params['continue'] );
 174                         $this->dieContinueUsageIf( count( $cont ) != count( $sortby ) );
 175                         $where = '';
 176                         $i = count( $sortby ) - 1;
 177                         foreach ( array_reverse( $sortby, true ) as $field => $type ) {
 178                                 $v = $cont[$i];
 179                                 switch ( $type ) {
 180                                         case 'ns':
 181                                         case 'int':
 182                                                 $v = (int)$v;
 183                                                 $this->dieContinueUsageIf( $v != $cont[$i] );
 184                                                 break;
 185                                         default:
 186                                                 $v = $db->addQuotes( $v );
 187                                                 break;
 188                                 }
 189
 190                                 if ( $where === '' ) {
 191                                         $where = "$field >= $v";
 192                                 } else {
 193                                         $where = "$field > $v OR ($field = $v AND ($where))";
 194                                 }
 195
 196                                 $i--;
 197                         }
 198                         $this->addWhere( $where );
 199                 }
 200
 201                 // Populate the rest of the query
 202                 $this->addTables( [ $settings['linktable'], 'page' ] );
 203                 $this->addWhere( "$bl_from = page_id" );
 204
 205                 if ( $this->getModuleName() === 'redirects' ) {
 206                         $this->addWhere( "rd_interwiki = $emptyString OR rd_interwiki IS NULL" );
 207                 }
 208
 209                 $this->addFields( array_keys( $sortby ) );
 210                 $this->addFields( [ 'bl_namespace' => $bl_namespace, 'bl_title' => $bl_title ] );
 211                 if ( is_null( $resultPageSet ) ) {
 212                         $fld_pageid = isset( $prop['pageid'] );
 213                         $fld_title = isset( $prop['title'] );
 214                         $fld_redirect = isset( $prop['redirect'] );
 215
 216                         $this->addFieldsIf( 'page_id', $fld_pageid );
 217                         $this->addFieldsIf( [ 'page_title', 'page_namespace' ], $fld_title );
 218                         $this->addFieldsIf( 'page_is_redirect', $fld_redirect );
 219
 220                         // prop=redirects
 221                         $fld_fragment = isset( $prop['fragment'] );
 222                         $this->addFieldsIf( 'rd_fragment', $fld_fragment );
 223                 } else {
 224                         $this->addFields( $resultPageSet->getPageTableFields() );
 225                 }
 226
 227                 $this->addFieldsIf( 'page_namespace', $miser_ns !== null );
 228
 229                 if ( $hasNS ) {
 230                         // Can't use LinkBatch because it throws away Special titles.
 231                         // And we already have the needed data structure anyway.
 232                         $this->addWhere( $db->makeWhereFrom2d( $map, $bl_namespace, $bl_title ) );
 233                 } else {
 234                         $where = [];
 235                         foreach ( $titles as $t ) {
 236                                 if ( $t->getNamespace() == $bl_namespace ) {
 237                                         $where[] = "$bl_title = " . $db->addQuotes( $t->getDBkey() );
 238                                 }
 239                         }
 240                         $this->addWhere( $db->makeList( $where, LIST_OR ) );
 241                 }
 242
 243                 if ( $params['show'] !== null ) {
 244                         // prop=redirects only
 245                         $show = array_flip( $params['show'] );
 246                         if ( isset( $show['fragment'] ) && isset( $show['!fragment'] ) ||
 247                                 isset( $show['redirect'] ) && isset( $show['!redirect'] )
 248                         ) {
 249                                 $this->dieWithError( 'apierror-show' );
 250                         }
 251                         $this->addWhereIf( "rd_fragment != $emptyString", isset( $show['fragment'] ) );
 252                         $this->addWhereIf(
 253                                 "rd_fragment = $emptyString OR rd_fragment IS NULL",
 254                                 isset( $show['!fragment'] )
 255                         );
 256                         $this->addWhereIf( [ 'page_is_redirect' => 1 ], isset( $show['redirect'] ) );
 257                         $this->addWhereIf( [ 'page_is_redirect' => 0 ], isset( $show['!redirect'] ) );
 258                 }
 259
 260                 // Override any ORDER BY from above with what we calculated earlier.
 261                 $this->addOption( 'ORDER BY', array_keys( $sortby ) );
 262
 263                 // MySQL's optimizer chokes if we have too many values in "$bl_title IN
 264                 // (...)" and chooses the wrong index, so specify the correct index to
 265                 // use for the query. See T139056 for details.
 266                 if ( !empty( $settings['indexes'] ) ) {
 267                         list( $idxNoFromNS, $idxWithFromNS ) = $settings['indexes'];
 268                         if ( $params['namespace'] !== null && !empty( $settings['from_namespace'] ) ) {
 269                                 $this->addOption( 'USE INDEX', [ $settings['linktable'] => $idxWithFromNS ] );
 270                         } else {
 271                                 $this->addOption( 'USE INDEX', [ $settings['linktable'] => $idxNoFromNS ] );
 272                         }
 273                 }
 274
 275                 // MySQL (or at least 5.5.5-10.0.23-MariaDB) chooses a really bad query
 276                 // plan if it thinks there will be more matching rows in the linktable
 277                 // than are in page. Use STRAIGHT_JOIN here to force it to use the
 278                 // intended, fast plan. See T145079 for details.
 279                 $this->addOption( 'STRAIGHT_JOIN' );
 280
 281                 $this->addOption( 'LIMIT', $params['limit'] + 1 );
 282
 283                 $res = $this->select( __METHOD__ );
 284
 285                 if ( is_null( $resultPageSet ) ) {
 286                         $count = 0;
 287                         foreach ( $res as $row ) {
 288                                 if ( ++$count > $params['limit'] ) {
 289                                         // We've reached the one extra which shows that
 290                                         // there are additional pages to be had. Stop here...
 291                                         $this->setContinue( $row, $sortby );
 292                                         break;
 293                                 }
 294
 295                                 if ( $miser_ns !== null && !in_array( $row->page_namespace, $miser_ns ) ) {
 296                                         // Miser mode namespace check
 297                                         continue;
 298                                 }
 299
 300                                 // Get the ID of the current page
 301                                 $id = $map[$row->bl_namespace][$row->bl_title];
 302
 303                                 $vals = [];
 304                                 if ( $fld_pageid ) {
 305                                         $vals['pageid'] = (int)$row->page_id;
 306                                 }
 307                                 if ( $fld_title ) {
 308                                         ApiQueryBase::addTitleInfo( $vals,
 309                                                 Title::makeTitle( $row->page_namespace, $row->page_title )
 310                                         );
 311                                 }
 312                                 if ( $fld_fragment && $row->rd_fragment !== null && $row->rd_fragment !== '' ) {
 313                                         $vals['fragment'] = $row->rd_fragment;
 314                                 }
 315                                 if ( $fld_redirect ) {
 316                                         $vals['redirect'] = (bool)$row->page_is_redirect;
 317                                 }
 318                                 $fit = $this->addPageSubItem( $id, $vals );
 319                                 if ( !$fit ) {
 320                                         $this->setContinue( $row, $sortby );
 321                                         break;
 322                                 }
 323                         }
 324                 } else {
 325                         $titles = [];
 326                         $count = 0;
 327                         foreach ( $res as $row ) {
 328                                 if ( ++$count > $params['limit'] ) {
 329                                         // We've reached the one extra which shows that
 330                                         // there are additional pages to be had. Stop here...
 331                                         $this->setContinue( $row, $sortby );
 332                                         break;
 333                                 }
 334                                 $titles[] = Title::makeTitle( $row->page_namespace, $row->page_title );
 335                         }
 336                         $resultPageSet->populateFromTitles( $titles );
 337                 }
 338         }
 339
 340         private function setContinue( $row, $sortby ) {
 341                 $cont = [];
 342                 foreach ( $sortby as $field => $v ) {
 343                         $cont[] = $row->$field;
 344                 }
 345                 $this->setContinueEnumParameter( 'continue', implode( '|', $cont ) );
 346         }
 347
 348         public function getCacheMode( $params ) {
 349                 return 'public';
 350         }
 351
 352         public function getAllowedParams() {
 353                 $settings = self::$settings[$this->getModuleName()];
 354
 355                 $ret = [
 356                         'prop' => [
 357                                 ApiBase::PARAM_TYPE => [
 358                                         'pageid',
 359                                         'title',
 360                                 ],
 361                                 ApiBase::PARAM_ISMULTI => true,
 362                                 ApiBase::PARAM_DFLT => 'pageid|title',
 363                                 ApiBase::PARAM_HELP_MSG_PER_VALUE => [],
 364                         ],
 365                         'namespace' => [
 366                                 ApiBase::PARAM_ISMULTI => true,
 367                                 ApiBase::PARAM_TYPE => 'namespace',
 368                         ],
 369                         'show' => null, // Will be filled/removed below
 370                         'limit' => [
 371                                 ApiBase::PARAM_DFLT => 10,
 372                                 ApiBase::PARAM_TYPE => 'limit',
 373                                 ApiBase::PARAM_MIN => 1,
 374                                 ApiBase::PARAM_MAX => ApiBase::LIMIT_BIG1,
 375                                 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_BIG2
 376                         ],
 377                         'continue' => [
 378                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-continue',
 379                         ],
 380                 ];
 381
 382                 if ( empty( $settings['from_namespace'] ) && $this->getConfig()->get( 'MiserMode' ) ) {
 383                         $ret['namespace'][ApiBase::PARAM_HELP_MSG_APPEND] = [
 384                                 'api-help-param-limited-in-miser-mode',
 385                         ];
 386                 }
 387
 388                 if ( !empty( $settings['showredirects'] ) ) {
 389                         $ret['prop'][ApiBase::PARAM_TYPE][] = 'redirect';
 390                         $ret['prop'][ApiBase::PARAM_DFLT] .= '|redirect';
 391                 }
 392                 if ( isset( $settings['props'] ) ) {
 393                         $ret['prop'][ApiBase::PARAM_TYPE] = array_merge(
 394                                 $ret['prop'][ApiBase::PARAM_TYPE], $settings['props']
 395                         );
 396                 }
 397
 398                 $show = [];
 399                 if ( !empty( $settings['showredirects'] ) ) {
 400                         $show[] = 'redirect';
 401                         $show[] = '!redirect';
 402                 }
 403                 if ( isset( $settings['show'] ) ) {
 404                         $show = array_merge( $show, $settings['show'] );
 405                 }
 406                 if ( $show ) {
 407                         $ret['show'] = [
 408                                 ApiBase::PARAM_TYPE => $show,
 409                                 ApiBase::PARAM_ISMULTI => true,
 410                         ];
 411                 } else {
 412                         unset( $ret['show'] );
 413                 }
 414
 415                 return $ret;
 416         }
 417
 418         protected function getExamplesMessages() {
 419                 $settings = self::$settings[$this->getModuleName()];
 420                 $name = $this->getModuleName();
 421                 $path = $this->getModulePath();
 422                 $title = isset( $settings['exampletitle'] ) ? $settings['exampletitle'] : 'Main Page';
 423                 $etitle = rawurlencode( $title );
 424
 425                 return [
 426                         "action=query&prop={$name}&titles={$etitle}"
 427                                 => "apihelp-$path-example-simple",
 428                         "action=query&generator={$name}&titles={$etitle}&prop=info"
 429                                 => "apihelp-$path-example-generator",
 430                 ];
 431         }
 432
 433         public function getHelpUrls() {
 434                 $name = ucfirst( $this->getModuleName() );
 435                 return "https://www.mediawiki.org/wiki/API:{$name}";
 436         }
 437 }