PageIndexer::getSolrConnectionsByItem() - Code Metrics - Inspection of "Use Guzzle API for HTTP Crawling Requests (#1837)" - TYPO3-Solr/ext-solr - Measure and Improve Code Quality continuously with Scrutinizer

Passed
Push — master ( 02f8b8...776b21 )

by Timo
created 2018-02-15 15:13 UTC
PageIndexer::getSolrConnectionsByItem() B

↳ Parent: PageIndexer
Complexity

Conditions	6
Paths	6
Size

Total Lines	30
Code Lines	15
Duplication

Lines	0
Ratio	0 %
Code Coverage

Tests	0
CRAP Score	42
Importance

Changes
Metric	Value
dl	0
loc	30
ccs	0
cts	22
cp	0
rs	8.439
c	0
b	0
f	0
cc	6
eloc	15
nc	6
nop	1
crap	42
<?php
namespace ApacheSolrForTypo3\Solr\IndexQueue;

/***************************************************************
 *  Copyright notice
 *
 *  (c) 2009-2015 Ingo Renner <[email protected]>
 *  All rights reserved
 *
 *  This script is part of the TYPO3 project. The TYPO3 project is
 *  free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; either version 3 of the License, or
 *  (at your option) any later version.
 *
 *  The GNU General Public License can be found at
 *  http://www.gnu.org/copyleft/gpl.html.
 *
 *  This script is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 *  This copyright notice MUST APPEAR in all copies of the script!
 ***************************************************************/

use ApacheSolrForTypo3\Solr\Access\Rootline;
use ApacheSolrForTypo3\Solr\Access\RootlineElement;
use ApacheSolrForTypo3\Solr\System\Logging\SolrLogManager;
use TYPO3\CMS\Backend\Utility\BackendUtility;
use TYPO3\CMS\Core\Utility\GeneralUtility;

/**
 * A special purpose indexer to index pages.
 *
 * In the case of pages we can't directly index the page records, we need to
 * retrieve the content that belongs to a page from tt_content, too. Also
 * plugins may be included on a page and thus may need to be executed.
 *
 * @author Ingo Renner <[email protected]>
 */
class PageIndexer extends Indexer
{
    /**
     * Indexes an item from the indexing queue.
     *
     * @param Item $item An index queue item
     * @return bool Whether indexing was successful
     */
    public function index(Item $item)
    {
        $this->setLogging($item);

        // check whether we should move on at all
        if (!$this->isPageIndexable($item)) {
            return false;
        }

        $solrConnections = $this->getSolrConnectionsByItem($item);
        foreach ($solrConnections as $systemLanguageUid => $solrConnection) {
            $contentAccessGroups = $this->getAccessGroupsFromContent($item,
                $systemLanguageUid);

            if (empty($contentAccessGroups)) {
                // might be an empty page w/no content elements or some TYPO3 error / bug
                // FIXME logging needed
                continue;
            }

            foreach ($contentAccessGroups as $userGroup) {
                $this->indexPage($item, $systemLanguageUid, $userGroup);
            }
        }

        return true;
    }

    /**
     * Checks whether we can index this page.
     *
     * @param Item $item The page we want to index encapsulated in an index queue item
     * @return bool True if we can index this page, FALSE otherwise
     */
    protected function isPageIndexable(Item $item)
    {

        // TODO do we still need this?
        // shouldn't those be sorted out by the record monitor / garbage collector already?

        $isIndexable = true;
        $record = $item->getRecord();

        if (isset($GLOBALS['TCA']['pages']['ctrl']['enablecolumns']['disabled'])
            && $record[$GLOBALS['TCA']['pages']['ctrl']['enablecolumns']['disabled']]
        ) {
            $isIndexable = false;
        }

        return $isIndexable;
    }

    /**
     * Gets the Solr connections applicable for a page.
     *
     * The connections include the default connection and connections to be used
     * for translations of a page.
     *
     * @param Item $item An index queue item
     * @return array An array of ApacheSolrForTypo3\Solr\System\Solr\SolrConnection connections, the array's keys are the sys_language_uid of the language of the connection
     */
    protected function getSolrConnectionsByItem(Item $item)
    {
        $solrConnections = parent::getSolrConnectionsByItem($item);

        $page = $item->getRecord();
        // may use \TYPO3\CMS\Core\Utility\GeneralUtility::hideIfDefaultLanguage($page['l18n_cfg']) with TYPO3 4.6

        if ($page['l18n_cfg'] & 1) {
            // page is configured to hide the default translation -> remove Solr connection for default language
            unset($solrConnections[0]);
        }

        if (GeneralUtility::hideIfNotTranslated($page['l18n_cfg'])) {
            $accessibleSolrConnections = [];
            if (isset($solrConnections[0])) {
                $accessibleSolrConnections[0] = $solrConnections[0];
            }

            $translationOverlays = $this->pagesRepository->findTranslationOverlaysByPageId((int)$page['uid']);

            foreach ($translationOverlays as $overlay) {
                $languageId = $overlay['sys_language_uid'];
                if (array_key_exists($languageId, $solrConnections)) {
                    $accessibleSolrConnections[$languageId] = $solrConnections[$languageId];
                }
            }

            $solrConnections = $accessibleSolrConnections;
        }

        return $solrConnections;
    }

    /**
     * Finds the FE user groups used on a page including all groups of content
     * elements and groups of records of extensions that have correctly been
     * pushed through ContentObjectRenderer during rendering.
     *
     * @param Item $item Index queue item representing the current page to get the user groups from
     * @param int $language The sys_language_uid language ID
     * @return array Array of user group IDs
     */
    protected function getAccessGroupsFromContent(Item $item, $language = 0)
    {
        static $accessGroupsCache;

        $accessGroupsCacheEntryId = $item->getRecordUid() . '|' . $language;
        if (!isset($accessGroupsCache[$accessGroupsCacheEntryId])) {
            $request = $this->buildBasePageIndexerRequest();
            $request->setIndexQueueItem($item);
            $request->addAction('findUserGroups');

            $indexRequestUrl = $this->getDataUrl($item, $language);
            $response = $request->send($indexRequestUrl);

            $groups = $response->getActionResult('findUserGroups');
            if (is_array($groups)) {

                $accessGroupsCache[$accessGroupsCacheEntryId] = $groups;
            }

            if ($this->loggingEnabled) {
                $this->logger->log(
                    SolrLogManager::INFO,
                    'Page Access Groups',
                    [
                        'item' => (array)$item,
                        'language' => $language,
                        'index request url' => $indexRequestUrl,
                        'request' => (array)$request,
                        'response' => (array)$response,
                        'groups' => $groups
                    ]
                );
            }
        }

        return $accessGroupsCache[$accessGroupsCacheEntryId];
    }

    // Utility methods

    /**
     * Builds a base page indexer request with configured headers and other
     * parameters.
     *
     * @return PageIndexerRequest Base page indexer request
     */
    protected function buildBasePageIndexerRequest()
    {
        $request = GeneralUtility::makeInstance(PageIndexerRequest::class);
        $request->setParameter('loggingEnabled', $this->loggingEnabled);

        if (!empty($this->options['authorization.'])) {
            $request->setAuthorizationCredentials(
                $this->options['authorization.']['username'],
                $this->options['authorization.']['password']
            );
        }

        if (!empty($this->options['frontendDataHelper.']['headers.'])) {
            foreach ($this->options['frontendDataHelper.']['headers.'] as $headerValue) {
                $request->addHeader($headerValue);
            }
        }

        if (!empty($this->options['frontendDataHelper.']['requestTimeout'])) {
            $request->setTimeout((float)$this->options['frontendDataHelper.']['requestTimeout']);
        }

        return $request;
    }

    /**
     * Determines a page ID's URL.
     *
     * Tries to find a domain record to use to build an URL for a given page ID
     * and then actually build and return the page URL.
     *
     * @param Item $item Item to index
     * @param int $language The language id
     * @return string URL to send the index request to
     * @throws \RuntimeException
     */
    protected function getDataUrl(Item $item, $language = 0)
    {
        $scheme = 'http';
        $host = $item->getSite()->getDomain();
        $path = '/';
        $pageId = $item->getRecordUid();

        // deprecated
        if (!empty($this->options['scheme'])) {
            $this->logger->log(
                SolrLogManager::INFO,
                'Using deprecated option "scheme" to set the scheme (http / https) for the page indexer frontend helper. Use plugin.tx_solr.index.queue.pages.indexer.frontendDataHelper.scheme instead'
            );
            $scheme = $this->options['scheme'];
        }

        // check whether we should use ssl / https
        if (!empty($this->options['frontendDataHelper.']['scheme'])) {
            $scheme = $this->options['frontendDataHelper.']['scheme'];
        }

        // overwriting the host
        if (!empty($this->options['frontendDataHelper.']['host'])) {
            $host = $this->options['frontendDataHelper.']['host'];
        }

        // setting a path if TYPO3 is installed in a sub directory
        if (!empty($this->options['frontendDataHelper.']['path'])) {
            $path = $this->options['frontendDataHelper.']['path'];
        }

        $mountPointParameter = $this->getMountPageDataUrlParameter($item);
        $dataUrl = $scheme . '://' . $host . $path . 'index.php?id=' . $pageId;
        $dataUrl .= ($mountPointParameter !== '') ? '&MP=' . $mountPointParameter : '';
        $dataUrl .= '&L=' . $language;

        if (!GeneralUtility::isValidUrl($dataUrl)) {
            $this->logger->log(
                SolrLogManager::ERROR,
                'Could not create a valid URL to get frontend data while trying to index a page.',
                [
                    'item' => (array)$item,
                    'constructed URL' => $dataUrl,
                    'scheme' => $scheme,
                    'host' => $host,
                    'path' => $path,
                    'page ID' => $pageId,
                    'indexer options' => $this->options
                ]
            );

            throw new \RuntimeException(
                'Could not create a valid URL to get frontend data while trying to index a page. Created URL: ' . $dataUrl,
                1311080805
            );
        }

        if ($GLOBALS['TYPO3_CONF_VARS']['EXTCONF']['solr']['IndexQueuePageIndexer']['dataUrlModifier']) {
            $dataUrlModifier = GeneralUtility::makeInstance($GLOBALS['TYPO3_CONF_VARS']['EXTCONF']['solr']['IndexQueuePageIndexer']['dataUrlModifier']);

            if ($dataUrlModifier instanceof PageIndexerDataUrlModifier) {
                $dataUrl = $dataUrlModifier->modifyDataUrl($dataUrl, [
                    'item' => $item,
                    'scheme' => $scheme,
                    'host' => $host,
                    'path' => $path,
                    'pageId' => $pageId,
                    'language' => $language
                ]);
            } else {
                throw new \RuntimeException(
                    $GLOBALS['TYPO3_CONF_VARS']['EXTCONF']['solr']['IndexQueuePageIndexer']['dataUrlModifier']
                    . ' is not an implementation of ApacheSolrForTypo3\Solr\IndexQueue\PageIndexerDataUrlModifier',
                    1290523345
                );
            }
        }

        return $dataUrl;
    }

    /**
     * Generates the MP URL parameter needed to access mount pages. If the item
     * is identified as being a mounted page, the &MP parameter is generated.
     *
     * @param Item $item Item to get an &MP URL parameter for
     * @return string &MP URL parameter if $item is a mounted page
     */
    protected function getMountPageDataUrlParameter(Item $item)
    {
        $mountPageUrlParameter = '';

        if ($item->hasIndexingProperty('isMountedPage')) {
            $mountPageUrlParameter =
                $item->getIndexingProperty('mountPageSource')
                . '-'
                . $item->getIndexingProperty('mountPageDestination');
        }

        return $mountPageUrlParameter;
    }

    #
    # Frontend User Groups Access
    #

    /**
     * Creates a single Solr Document for a page in a specific language and for
     * a specific frontend user group.
     *
     * @param Item $item The index queue item representing the page.
     * @param int $language The language to use.
     * @param int $userGroup The frontend user group to use.
     * @return PageIndexerResponse Page indexer response
     * @throws \RuntimeException if indexing an item failed
     */
    protected function indexPage(Item $item, $language = 0, $userGroup = 0)
    {
        $accessRootline = $this->getAccessRootline($item, $language,
            $userGroup);

        $request = $this->buildBasePageIndexerRequest();
        $request->setIndexQueueItem($item);
        $request->addAction('indexPage');
        $request->setParameter('accessRootline', (string)$accessRootline);

        $indexRequestUrl = $this->getDataUrl($item, $language);
        $response = $request->send($indexRequestUrl);
        $indexActionResult = $response->getActionResult('indexPage');

        if ($this->loggingEnabled) {
            $logSeverity = SolrLogManager::INFO;
            $logStatus = 'Info';
            if ($indexActionResult['pageIndexed']) {
                $logSeverity = SolrLogManager::NOTICE;
                $logStatus = 'Success';
            }

            $this->logger->log(
                $logSeverity,
                'Page Indexer: ' . $logStatus,
                [
                    'item' => (array)$item,
                    'language' => $language,
                    'user group' => $userGroup,
                    'index request url' => $indexRequestUrl,
                    'request' => (array)$request,
                    'request headers' => $request->getHeaders(),
                    'response' => (array)$response
                ]
            );
        }

        if (!$indexActionResult['pageIndexed']) {
            $message = 'Failed indexing page Index Queue item: ' . $item->getIndexQueueUid() . ' url: ' . $indexRequestUrl;

            throw new \RuntimeException($message, 1331837081);
        }

        return $response;
    }

    /**
     * Generates a page document's "Access Rootline".
     *
     * The Access Rootline collects frontend user group access restrictions set
     * for pages up in a page's rootline extended to sub-pages.
     *
     * The format is like this:
     * pageId1:group1,group2|groupId2:group3|c:group1,group4,groupN
     *
     * The single elements of the access rootline are separated by a pipe
     * character. All but the last elements represent pages, the last element
     * defines the access restrictions applied to the page's content elements
     * and records shown on the page.
     * Each page element is composed by the page ID of the page setting frontend
     * user access restrictions, a colon, and a comma separated list of frontend
     * user group IDs restricting access to the page.
     * The content access element does not have a page ID, instead it replaces
     * the ID by a lower case C.
     *
     * @param Item $item Index queue item representing the current page
     * @param int $language The sys_language_uid language ID
     * @param int $contentAccessGroup The user group to use for the content access rootline element. Optional, will be determined automatically if not set.
     * @return string An Access Rootline.
     */
    protected function getAccessRootline(
        Item $item,
        $language = 0,
        $contentAccessGroup = null
    ) {
        static $accessRootlineCache;

        $mountPointParameter = $this->getMountPageDataUrlParameter($item);

        $accessRootlineCacheEntryId = $item->getRecordUid() . '|' . $language;
        if ($mountPointParameter !== '') {
            $accessRootlineCacheEntryId .= '|' . $mountPointParameter;
        }
        if (!is_null($contentAccessGroup)) {
            $accessRootlineCacheEntryId .= '|' . $contentAccessGroup;
        }

        if (!isset($accessRootlineCache[$accessRootlineCacheEntryId])) {
            $accessRootline = Rootline::getAccessRootlineByPageId(
                $item->getRecordUid(),
                $mountPointParameter
            );

            // current page's content access groups
            $contentAccessGroups = [$contentAccessGroup];
            if (is_null($contentAccessGroup)) {
                $contentAccessGroups = $this->getAccessGroupsFromContent($item, $language);
            }
            $element = GeneralUtility::makeInstance(RootlineElement::class, 'c:' . implode(',', $contentAccessGroups));
            $element = GeneralUtility::makeInstance(RootlineElement::class, /** @scrutinizer ignore-type */ 'c:' . implode(',', $contentAccessGroups));
            $accessRootline->push($element);

            $accessRootlineCache[$accessRootlineCacheEntryId] = $accessRootline;
        }

        return $accessRootlineCache[$accessRootlineCacheEntryId];
    }
}

TYPO3-Solr / ext-solr

Push — master ( 02f8b8...776b21 )

PageIndexer::getSolrConnectionsByItem() B

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like
