ApiPageSet.php

Go to the documentation of this file.

<?php
class ApiPageSet extends ApiQueryBase {

        private $mAllPages; // [ns][dbkey] => page_id or negative when missing
        private $mTitles, $mGoodTitles, $mMissingTitles, $mInvalidTitles;
        private $mMissingPageIDs, $mRedirectTitles, $mSpecialTitles;
        private $mNormalizedTitles, $mInterwikiTitles;
        private $mResolveRedirects, $mPendingRedirectIDs;
        private $mConvertTitles, $mConvertedTitles;
        private $mGoodRevIDs, $mMissingRevIDs;
        private $mFakePageId;

        private $mRequestedPageFields;

        public function __construct( $query, $resolveRedirects = false, $convertTitles = false ) {
                parent::__construct( $query, 'query' );

                $this->mAllPages = array();
                $this->mTitles = array();
                $this->mGoodTitles = array();
                $this->mMissingTitles = array();
                $this->mInvalidTitles = array();
                $this->mMissingPageIDs = array();
                $this->mRedirectTitles = array();
                $this->mNormalizedTitles = array();
                $this->mInterwikiTitles = array();
                $this->mGoodRevIDs = array();
                $this->mMissingRevIDs = array();
                $this->mSpecialTitles = array();

                $this->mRequestedPageFields = array();
                $this->mResolveRedirects = $resolveRedirects;
                if ( $resolveRedirects ) {
                        $this->mPendingRedirectIDs = array();
                }

                $this->mConvertTitles = $convertTitles;
                $this->mConvertedTitles = array();

                $this->mFakePageId = - 1;
        }

        public function isResolvingRedirects() {
                return $this->mResolveRedirects;
        }

        public function requestField( $fieldName ) {
                $this->mRequestedPageFields[$fieldName] = null;
        }

        public function getCustomField( $fieldName ) {
                return $this->mRequestedPageFields[$fieldName];
        }

        public function getPageTableFields() {
                // Ensure we get minimum required fields
                // DON'T change this order
                $pageFlds = array(
                        'page_namespace' => null,
                        'page_title' => null,
                        'page_id' => null,
                );

                if ( $this->mResolveRedirects ) {
                        $pageFlds['page_is_redirect'] = null;
                }

                // only store non-default fields
                $this->mRequestedPageFields = array_diff_key( $this->mRequestedPageFields, $pageFlds );

                $pageFlds = array_merge( $pageFlds, $this->mRequestedPageFields );
                return array_keys( $pageFlds );
        }

        public function getAllTitlesByNamespace() {
                return $this->mAllPages;
        }

        public function getTitles() {
                return $this->mTitles;
        }

        public function getTitleCount() {
                return count( $this->mTitles );
        }

        public function getGoodTitles() {
                return $this->mGoodTitles;
        }

        public function getGoodTitleCount() {
                return count( $this->mGoodTitles );
        }

        public function getMissingTitles() {
                return $this->mMissingTitles;
        }

        public function getInvalidTitles() {
                return $this->mInvalidTitles;
        }

        public function getMissingPageIDs() {
                return $this->mMissingPageIDs;
        }

        public function getRedirectTitles() {
                return $this->mRedirectTitles;
        }

        public function getNormalizedTitles() {
                return $this->mNormalizedTitles;
        }

        public function getConvertedTitles() {
                return $this->mConvertedTitles;
        }

        public function getInterwikiTitles() {
                return $this->mInterwikiTitles;
        }

        public function getRevisionIDs() {
                return $this->mGoodRevIDs;
        }

        public function getMissingRevisionIDs() {
                return $this->mMissingRevIDs;
        }

        public function getSpecialTitles() {
                return $this->mSpecialTitles;
        }

        public function getRevisionCount() {
                return count( $this->getRevisionIDs() );
        }

        public function execute() {
                $this->profileIn();
                $params = $this->extractRequestParams();

                // Only one of the titles/pageids/revids is allowed at the same time
                $dataSource = null;
                if ( isset( $params['titles'] ) ) {
                        $dataSource = 'titles';
                }
                if ( isset( $params['pageids'] ) ) {
                        if ( isset( $dataSource ) ) {
                                $this->dieUsage( "Cannot use 'pageids' at the same time as '$dataSource'", 'multisource' );
                        }
                        $dataSource = 'pageids';
                }
                if ( isset( $params['revids'] ) ) {
                        if ( isset( $dataSource ) ) {
                                $this->dieUsage( "Cannot use 'revids' at the same time as '$dataSource'", 'multisource' );
                        }
                        $dataSource = 'revids';
                }

                switch ( $dataSource ) {
                        case 'titles':
                                $this->initFromTitles( $params['titles'] );
                                break;
                        case 'pageids':
                                $this->initFromPageIds( $params['pageids'] );
                                break;
                        case 'revids':
                                if ( $this->mResolveRedirects ) {
                                        $this->setWarning( 'Redirect resolution cannot be used together with the revids= parameter. ' .
                                        'Any redirects the revids= point to have not been resolved.' );
                                }
                                $this->mResolveRedirects = false;
                                $this->initFromRevIDs( $params['revids'] );
                                break;
                        default:
                                // Do nothing - some queries do not need any of the data sources.
                                break;
                }
                $this->profileOut();
        }

        public function populateFromTitles( $titles ) {
                $this->profileIn();
                $this->initFromTitles( $titles );
                $this->profileOut();
        }

        public function populateFromPageIDs( $pageIDs ) {
                $this->profileIn();
                $this->initFromPageIds( $pageIDs );
                $this->profileOut();
        }

        public function populateFromQueryResult( $db, $queryResult ) {
                $this->profileIn();
                $this->initFromQueryResult( $queryResult );
                $this->profileOut();
        }

        public function populateFromRevisionIDs( $revIDs ) {
                $this->profileIn();
                $this->initFromRevIDs( $revIDs );
                $this->profileOut();
        }

        public function processDbRow( $row ) {
                // Store Title object in various data structures
                $title = Title::newFromRow( $row );

                $pageId = intval( $row->page_id );
                $this->mAllPages[$row->page_namespace][$row->page_title] = $pageId;
                $this->mTitles[] = $title;

                if ( $this->mResolveRedirects && $row->page_is_redirect == '1' ) {
                        $this->mPendingRedirectIDs[$pageId] = $title;
                } else {
                        $this->mGoodTitles[$pageId] = $title;
                }

                foreach ( $this->mRequestedPageFields as $fieldName => &$fieldValues ) {
                        $fieldValues[$pageId] = $row-> $fieldName;
                }
        }

        public function finishPageSetGeneration() {
                $this->profileIn();
                $this->resolvePendingRedirects();
                $this->profileOut();
        }

        private function initFromTitles( $titles ) {
                // Get validated and normalized title objects
                $linkBatch = $this->processTitlesArray( $titles );
                if ( $linkBatch->isEmpty() ) {
                        return;
                }

                $db = $this->getDB();
                $set = $linkBatch->constructSet( 'page', $db );

                // Get pageIDs data from the `page` table
                $this->profileDBIn();
                $res = $db->select( 'page', $this->getPageTableFields(), $set,
                                        __METHOD__ );
                $this->profileDBOut();

                // Hack: get the ns:titles stored in array(ns => array(titles)) format
                $this->initFromQueryResult( $res, $linkBatch->data, true ); // process Titles

                // Resolve any found redirects
                $this->resolvePendingRedirects();
        }

        private function initFromPageIds( $pageids ) {
                if ( !count( $pageids ) ) {
                        return;
                }

                $pageids = array_map( 'intval', $pageids ); // paranoia
                $remaining = array_flip( $pageids );

                $pageids = self::getPositiveIntegers( $pageids );

                $res = null;
                if ( count( $pageids ) ) {
                        $set = array(
                                'page_id' => $pageids
                        );
                        $db = $this->getDB();

                        // Get pageIDs data from the `page` table
                        $this->profileDBIn();
                        $res = $db->select( 'page', $this->getPageTableFields(), $set,
                                                __METHOD__ );
                        $this->profileDBOut();
                }

                $this->initFromQueryResult( $res, $remaining, false );  // process PageIDs

                // Resolve any found redirects
                $this->resolvePendingRedirects();
        }

        private function initFromQueryResult( $res, &$remaining = null, $processTitles = null ) {
                if ( !is_null( $remaining ) && is_null( $processTitles ) ) {
                        ApiBase::dieDebug( __METHOD__, 'Missing $processTitles parameter when $remaining is provided' );
                }

                $usernames = array();
                if ( $res ) {
                        foreach ( $res as $row ) {
                                $pageId = intval( $row->page_id );

                                // Remove found page from the list of remaining items
                                if ( isset( $remaining ) ) {
                                        if ( $processTitles ) {
                                                unset( $remaining[$row->page_namespace][$row->page_title] );
                                        } else {
                                                unset( $remaining[$pageId] );
                                        }
                                }

                                // Store any extra fields requested by modules
                                $this->processDbRow( $row );

                                // Need gender information
                                if( MWNamespace::hasGenderDistinction( $row->page_namespace ) ) {
                                        $usernames[] = $row->page_title;
                                }
                        }
                }

                if ( isset( $remaining ) ) {
                        // Any items left in the $remaining list are added as missing
                        if ( $processTitles ) {
                                // The remaining titles in $remaining are non-existent pages
                                foreach ( $remaining as $ns => $dbkeys ) {
                                        foreach ( array_keys( $dbkeys ) as $dbkey ) {
                                                $title = Title::makeTitle( $ns, $dbkey );
                                                $this->mAllPages[$ns][$dbkey] = $this->mFakePageId;
                                                $this->mMissingTitles[$this->mFakePageId] = $title;
                                                $this->mFakePageId--;
                                                $this->mTitles[] = $title;

                                                // need gender information
                                                if( MWNamespace::hasGenderDistinction( $ns ) ) {
                                                        $usernames[] = $dbkey;
                                                }
                                        }
                                }
                        } else {
                                // The remaining pageids do not exist
                                if ( !$this->mMissingPageIDs ) {
                                        $this->mMissingPageIDs = array_keys( $remaining );
                                } else {
                                        $this->mMissingPageIDs = array_merge( $this->mMissingPageIDs, array_keys( $remaining ) );
                                }
                        }
                }

                // Get gender information
                $genderCache = GenderCache::singleton();
                $genderCache->doQuery( $usernames, __METHOD__ );
        }

        private function initFromRevIDs( $revids ) {
                if ( !count( $revids ) ) {
                        return;
                }

                $revids = array_map( 'intval', $revids ); // paranoia
                $db = $this->getDB();
                $pageids = array();
                $remaining = array_flip( $revids );

                $revids = self::getPositiveIntegers( $revids );

                if ( count( $revids ) ) {
                        $tables = array( 'revision', 'page' );
                        $fields = array( 'rev_id', 'rev_page' );
                        $where = array( 'rev_id' => $revids, 'rev_page = page_id' );

                        // Get pageIDs data from the `page` table
                        $this->profileDBIn();
                        $res = $db->select( $tables, $fields, $where,  __METHOD__ );
                        foreach ( $res as $row ) {
                                $revid = intval( $row->rev_id );
                                $pageid = intval( $row->rev_page );
                                $this->mGoodRevIDs[$revid] = $pageid;
                                $pageids[$pageid] = '';
                                unset( $remaining[$revid] );
                        }
                        $this->profileDBOut();
                }

                $this->mMissingRevIDs = array_keys( $remaining );

                // Populate all the page information
                $this->initFromPageIds( array_keys( $pageids ) );
        }

        private function resolvePendingRedirects() {
                if ( $this->mResolveRedirects ) {
                        $db = $this->getDB();
                        $pageFlds = $this->getPageTableFields();

                        // Repeat until all redirects have been resolved
                        // The infinite loop is prevented by keeping all known pages in $this->mAllPages
                        while ( $this->mPendingRedirectIDs ) {
                                // Resolve redirects by querying the pagelinks table, and repeat the process
                                // Create a new linkBatch object for the next pass
                                $linkBatch = $this->getRedirectTargets();

                                if ( $linkBatch->isEmpty() ) {
                                        break;
                                }

                                $set = $linkBatch->constructSet( 'page', $db );
                                if ( $set === false ) {
                                        break;
                                }

                                // Get pageIDs data from the `page` table
                                $this->profileDBIn();
                                $res = $db->select( 'page', $pageFlds, $set, __METHOD__ );
                                $this->profileDBOut();

                                // Hack: get the ns:titles stored in array(ns => array(titles)) format
                                $this->initFromQueryResult( $res, $linkBatch->data, true );
                        }
                }
        }

        private function getRedirectTargets() {
                $lb = new LinkBatch();
                $db = $this->getDB();

                $this->profileDBIn();
                $res = $db->select(
                        'redirect',
                        array(
                                'rd_from',
                                'rd_namespace',
                                'rd_fragment',
                                'rd_interwiki',
                                'rd_title'
                        ), array( 'rd_from' => array_keys( $this->mPendingRedirectIDs ) ),
                        __METHOD__
                );
                $this->profileDBOut();
                foreach ( $res as $row ) {
                        $rdfrom = intval( $row->rd_from );
                        $from = $this->mPendingRedirectIDs[$rdfrom]->getPrefixedText();
                        $to = Title::makeTitle( $row->rd_namespace, $row->rd_title, $row->rd_fragment, $row->rd_interwiki );
                        unset( $this->mPendingRedirectIDs[$rdfrom] );
                        if ( !isset( $this->mAllPages[$row->rd_namespace][$row->rd_title] ) ) {
                                $lb->add( $row->rd_namespace, $row->rd_title );
                        }
                        $this->mRedirectTitles[$from] = $to;
                }

                if ( $this->mPendingRedirectIDs ) {
                        // We found pages that aren't in the redirect table
                        // Add them
                        foreach ( $this->mPendingRedirectIDs as $id => $title ) {
                                $page = WikiPage::factory( $title );
                                $rt = $page->insertRedirect();
                                if ( !$rt ) {
                                        // What the hell. Let's just ignore this
                                        continue;
                                }
                                $lb->addObj( $rt );
                                $this->mRedirectTitles[$title->getPrefixedText()] = $rt;
                                unset( $this->mPendingRedirectIDs[$id] );
                        }
                }
                return $lb;
        }

        private function processTitlesArray( $titles ) {
                $genderCache = GenderCache::singleton();
                $genderCache->doTitlesArray( $titles, __METHOD__ );

                $linkBatch = new LinkBatch();

                foreach ( $titles as $title ) {
                        $titleObj = is_string( $title ) ? Title::newFromText( $title ) : $title;
                        if ( !$titleObj ) {
                                // Handle invalid titles gracefully
                                $this->mAllpages[0][$title] = $this->mFakePageId;
                                $this->mInvalidTitles[$this->mFakePageId] = $title;
                                $this->mFakePageId--;
                                continue; // There's nothing else we can do
                        }
                        $unconvertedTitle = $titleObj->getPrefixedText();
                        $titleWasConverted = false;
                        $iw = $titleObj->getInterwiki();
                        if ( strval( $iw ) !== '' ) {
                                // This title is an interwiki link.
                                $this->mInterwikiTitles[$titleObj->getPrefixedText()] = $iw;
                        } else {
                                // Variants checking
                                global $wgContLang;
                                if ( $this->mConvertTitles &&
                                                count( $wgContLang->getVariants() ) > 1  &&
                                                !$titleObj->exists() ) {
                                        // Language::findVariantLink will modify titleObj into
                                        // the canonical variant if possible
                                        $wgContLang->findVariantLink( $title, $titleObj );
                                        $titleWasConverted = $unconvertedTitle !== $titleObj->getPrefixedText();
                                }

                                if ( $titleObj->getNamespace() < 0 ) {
                                        // Handle Special and Media pages
                                        $titleObj = $titleObj->fixSpecialName();
                                        $this->mSpecialTitles[$this->mFakePageId] = $titleObj;
                                        $this->mFakePageId--;
                                } else {
                                        // Regular page
                                        $linkBatch->addObj( $titleObj );
                                }
                        }

                        // Make sure we remember the original title that was
                        // given to us. This way the caller can correlate new
                        // titles with the originally requested when e.g. the
                        // namespace is localized or the capitalization is
                        // different
                        if ( $titleWasConverted ) {
                                $this->mConvertedTitles[$title] = $titleObj->getPrefixedText();
                        } elseif ( is_string( $title ) && $title !== $titleObj->getPrefixedText() ) {
                                $this->mNormalizedTitles[$title] = $titleObj->getPrefixedText();
                        }
                }

                return $linkBatch;
        }

        private static function getPositiveIntegers( $array ) {
                // bug 25734 API: possible issue with revids validation
                // It seems with a load of revision rows, MySQL gets upset
                // Remove any < 0 integers, as they can't be valid
                foreach( $array as $i => $int ) {
                        if ( $int < 0 ) {
                                unset( $array[$i] );
                        }
                }

                return $array;
        }

        public function getAllowedParams() {
                return array(
                        'titles' => array(
                                ApiBase::PARAM_ISMULTI => true
                        ),
                        'pageids' => array(
                                ApiBase::PARAM_TYPE => 'integer',
                                ApiBase::PARAM_ISMULTI => true
                        ),
                        'revids' => array(
                                ApiBase::PARAM_TYPE => 'integer',
                                ApiBase::PARAM_ISMULTI => true
                        )
                );
        }

        public function getParamDescription() {
                return array(
                        'titles' => 'A list of titles to work on',
                        'pageids' => 'A list of page IDs to work on',
                        'revids' => 'A list of revision IDs to work on'
                );
        }

        public function getPossibleErrors() {
                return array_merge( parent::getPossibleErrors(), array(
                        array( 'code' => 'multisource', 'info' => "Cannot use 'pageids' at the same time as 'dataSource'" ),
                        array( 'code' => 'multisource', 'info' => "Cannot use 'revids' at the same time as 'dataSource'" ),
                ) );
        }

        public function getVersion() {
                return __CLASS__ . ': $Id$';
        }
}