EntityDataRequestHandler::getIncomingRedirects() - Code Metrics - wikimedia/mediawiki-extensions-Wikibase - Measure and Improve Code Quality continuously with Scrutinizer

EntityDataRequestHandler::getIncomingRedirects() A
last analyzed 2020-11-13 17:34 UTC

↳ Parent: EntityDataRequestHandler

Complexity

Conditions	2
Paths	2

Size

Total Lines

Duplication

Lines	0
Ratio	0 %

Importance

Changes

Metric	Value
dl	0
loc	17
rs	9.7
c	0
b	0
f	0
cc	2
nc	2
nop	1

<?php

namespace Wikibase\Repo\LinkedData;

use HtmlCacheUpdater;
use HttpError;
use OutputPage;
use Psr\Log\LoggerInterface;
use WebRequest;
use WebResponse;
use Wikibase\DataModel\Entity\EntityId;
use Wikibase\DataModel\Entity\EntityIdParser;
use Wikibase\DataModel\Entity\EntityIdParsingException;
use Wikibase\DataModel\Entity\EntityRedirect;
use Wikibase\DataModel\Services\Lookup\EntityRedirectLookup;
use Wikibase\DataModel\Services\Lookup\EntityRedirectLookupException;
use Wikibase\Lib\Store\BadRevisionException;
use Wikibase\Lib\Store\EntityRevision;
use Wikibase\Lib\Store\EntityRevisionLookup;
use Wikibase\Lib\Store\EntityTitleLookup;
use Wikibase\Lib\Store\RedirectRevision;
use Wikibase\Lib\Store\RevisionedUnresolvedRedirectException;
use Wikibase\Lib\Store\StorageException;
use Wikimedia\Http\HttpAcceptNegotiator;
use Wikimedia\Http\HttpAcceptParser;

/**
 * Request handler implementing a linked data interface for Wikibase entities.
 *
 * @license GPL-2.0-or-later
 * @author Daniel Kinzler
 * @author Thomas Pellissier Tanon
 * @author Anja Jentzsch < [email protected] >
 */
class EntityDataRequestHandler {

	/**
	 * Allowed smallest and biggest number of seconds for the "max-age=..." and "s-maxage=..." cache
	 * control parameters.
	 *
	 * @todo Hard maximum could be configurable somehow.
	 */
	const MINIMUM_MAX_AGE = 0;
	const MAXIMUM_MAX_AGE = 2678400; // 31 days

	/**
	 * @var EntityDataSerializationService
	 */
	private $serializationService;

	/**
	 * @var EntityDataUriManager
	 */
	private $uriManager;

	/**
	 * @var EntityIdParser
	 */
	private $entityIdParser;

	/**
	 * @var EntityRevisionLookup
	 */
	private $entityRevisionLookup;

	/**
	 * @var EntityRedirectLookup
	 */
	private $entityRedirectLookup;

	/**
	 * @var EntityTitleLookup
	 */
	private $entityTitleLookup;

	/**
	 * @var EntityDataFormatProvider
	 */
	private $entityDataFormatProvider;

	/**
	 * @var HtmlCacheUpdater
	 */
	private $htmlCacheUpdater;

	/**
	 * @var LoggerInterface
	 */
	private $logger;

	/**
	 * @var string
	 */
	private $defaultFormat;

	/**
	 * @var int Number of seconds to cache entity data.
	 */
	private $maxAge;

	/**
	 * @var bool
	 */
	private $useCdn;

	/**
	 * @var string|null
	 */
	private $frameOptionsHeader;

	/**
	 * @var string[]
	 */
	private $entityTypesWithoutRdfOutput;

	/**
	 * @param EntityDataUriManager $uriManager
	 * @param HtmlCacheUpdater $htmlCacheUpdater
	 * @param EntityTitleLookup $entityTitleLookup
	 * @param EntityIdParser $entityIdParser
	 * @param EntityRevisionLookup $entityRevisionLookup
	 * @param EntityRedirectLookup $entityRedirectLookup
	 * @param EntityDataSerializationService $serializationService
	 * @param EntityDataFormatProvider $entityDataFormatProvider
	 * @param LoggerInterface $logger
	 * @param string[] $entityTypesWithoutRdfOutput
	 * @param string $defaultFormat The format as a file extension or MIME type.
	 * @param int $maxAge number of seconds to cache entity data
	 * @param bool $useCdn do we have web caches configured?
	 * @param string|null $frameOptionsHeader for X-Frame-Options
	 */
	public function __construct(
		EntityDataUriManager $uriManager,
		HtmlCacheUpdater $htmlCacheUpdater,
		EntityTitleLookup $entityTitleLookup,
		EntityIdParser $entityIdParser,
		EntityRevisionLookup $entityRevisionLookup,
		EntityRedirectLookup $entityRedirectLookup,
		EntityDataSerializationService $serializationService,
		EntityDataFormatProvider $entityDataFormatProvider,
		LoggerInterface $logger,
		array $entityTypesWithoutRdfOutput,
		$defaultFormat,
		$maxAge,
		$useCdn,
		$frameOptionsHeader
	) {
		$this->uriManager = $uriManager;
		$this->htmlCacheUpdater = $htmlCacheUpdater;
		$this->entityTitleLookup = $entityTitleLookup;
		$this->entityIdParser = $entityIdParser;
		$this->entityRevisionLookup = $entityRevisionLookup;
		$this->entityRedirectLookup = $entityRedirectLookup;
		$this->serializationService = $serializationService;
		$this->entityDataFormatProvider = $entityDataFormatProvider;
		$this->logger = $logger;
		$this->entityTypesWithoutRdfOutput = $entityTypesWithoutRdfOutput;
		$this->defaultFormat = $defaultFormat;
		$this->maxAge = $maxAge;
		$this->useCdn = $useCdn;
		$this->frameOptionsHeader = $frameOptionsHeader;
	}

	/**
	 * Checks whether the request is complete, i.e. whether it contains all information needed
	 * to reply with entity data.
	 *
	 * This does not check whether the request is valid and will actually produce a successful
	 * response.
	 *
	 * @param string|null $doc Document name, e.g. Q5 or Q5.json or Q5:33.xml
	 * @param WebRequest $request
	 *
	 * @return bool
	 */
	public function canHandleRequest( $doc, WebRequest $request ) {
		if ( $doc === '' || $doc === null ) {
			if ( $request->getText( 'id', '' ) === '' ) {
				return false;
			}
		}

		return true;
	}

	/**
	 * Main method for handling requests.
	 *
	 * @param string $doc Document name, e.g. Q5 or Q5.json or Q5:33.xml
	 * @param WebRequest $request The request parameters. Known parameters are:
	 *        - id: the entity ID
	 *        - format: the format
	 *        - oldid|revision: the revision ID
	 *        - action=purge: to purge cached data from (web) caches
	 * @param OutputPage $output
	 *
	 * @note Instead of an output page, a WebResponse could be sufficient, but
	 *        redirect logic is currently implemented in OutputPage.
	 *
	 * @throws HttpError
	 * @suppress SecurityCheck-DoubleEscaped
	 */
	public function handleRequest( $doc, WebRequest $request, OutputPage $output ) {
		// No matter what: The response is always public
		$output->getRequest()->response()->header( 'Access-Control-Allow-Origin: *' );

		$revision = 0;

		list( $id, $format ) = $this->uriManager->parseDocName( $doc );

		// get entity id and format from request parameter
		$format = $request->getText( 'format', $format );
		$id = $request->getText( 'id', $id );
		$revision = $request->getInt( 'revision', $revision );
		$redirectMode = $request->getText( 'redirect' );
		//TODO: malformed revision IDs should trigger a code 400

		// If there is no ID, fail
		if ( $id === null || $id === '' ) {
			//TODO: different error message?
			throw new HttpError( 400, $output->msg( 'wikibase-entitydata-bad-id', $id ) );
		}

		try {
			$entityId = $this->entityIdParser->parse( $id );
		} catch ( EntityIdParsingException $ex ) {
			throw new HttpError( 400, $output->msg( 'wikibase-entitydata-bad-id', $id ) );
		}

		if ( $this->entityDataFormatProvider->isRdfFormat( $format ) &&
			in_array( $entityId->getEntityType(), $this->entityTypesWithoutRdfOutput )
		) {
			throw new HttpError( 406, $output->msg( 'wikibase-entitydata-rdf-not-available', $entityId->getEntityType() ) );
		}

		//XXX: allow for logged in users only?
		if ( $request->getText( 'action' ) === 'purge' ) {
			$this->purgeWebCache( $entityId, $revision );
			//XXX: Now what? Proceed to show the data?
		}

		if ( $format === null || $format === '' ) {
			// if no format is given, apply content negotiation and return.
			$this->httpContentNegotiation( $request, $output, $entityId, $revision );
			return;
		}

		//NOTE: will trigger a 415 if the format is not supported
		$format = $this->getCanonicalFormat( $format );

		if ( $doc !== null && $doc !== '' ) {
			// if subpage syntax is used, always enforce the canonical form
			$canonicalDoc = $this->uriManager->getDocName( $entityId, $format );

			if ( $doc !== $canonicalDoc ) {
				$url = $this->uriManager->getDocUrl( $entityId, $format, $revision );
				$output->redirect( $url, 301 );
				return;
			}
		}

		// if the format is HTML, redirect to the entity's wiki page
		if ( $format === 'html' ) {
			$url = $this->uriManager->getDocUrl( $entityId, 'html', $revision );
			$output->redirect( $url, 303 );
			return;
		}

		// if redirection was force, redirect
		if ( $redirectMode === 'force' ) {
			$url = $this->uriManager->getDocUrl( $entityId, $format, $revision );
			$output->redirect( $url, 303 );
			return;
		}

		$this->showData( $request, $output, $format, $entityId, $revision );
	}

	/**
	 * Returns the canonical format name for the given format.
	 *
	 * @param string $format
	 *
	 * @return string
	 * @throws HttpError code 415 if the format is not supported.
	 */
	public function getCanonicalFormat( $format ) {
		$format = strtolower( $format );

		// we always support html, it's handled by the entity's wiki page.
		if ( $format === 'html' || $format === 'htm' || $format === 'text/html' ) {
			return 'html';
		}

		// normalize format name (note that HTML may not be known to the service)
		$canonicalFormat = $this->entityDataFormatProvider->getFormatName( $format );

		if ( $canonicalFormat === null ) {
			$msg = wfMessage( 'wikibase-entitydata-unsupported-format', $format );
			// @phan-suppress-next-line SecurityCheck-DoubleEscaped
			throw new HttpError( 415, $msg );
		}

		return $canonicalFormat;
	}

	/**
	 * Purges the entity data identified by the doc parameter from any HTTP caches.
	 * Does nothing if $wgUseCdn is not set.
	 *
	 * @param EntityId $id The entity ID for which to purge all data.
	 * @param int $revision The revision ID (0 for current/unspecified)
	 */
	public function purgeWebCache( EntityId $id, int $revision ) {
		$urls = $this->uriManager->getPotentiallyCachedUrls( $id, $revision );
		if ( $urls !== [] ) {
			$this->htmlCacheUpdater->purgeUrls( $urls );
		}
	}

	/**
	 * Applies HTTP content negotiation.
	 * If the negotiation is successfull, this method will set the appropriate redirect
	 * in the OutputPage object and return. Otherwise, an HttpError is thrown.
	 *
	 * @param WebRequest $request
	 * @param OutputPage $output
	 * @param EntityId $id The ID of the entity to show
	 * @param int      $revision The desired revision
	 *
	 * @throws HttpError
	 */
	public function httpContentNegotiation( WebRequest $request, OutputPage $output, EntityId $id, $revision = 0 ) {
		$headers = $request->getAllHeaders();
		if ( isset( $headers['ACCEPT'] ) ) {
			$parser = new HttpAcceptParser();
			$accept = $parser->parseWeights( $headers['ACCEPT'] );
		} else {
			// anything goes
			$accept = [
				'*' => 0.1 // just to make extra sure
			];

			$defaultFormat = $this->entityDataFormatProvider->getFormatName( $this->defaultFormat );
			$defaultMime = $this->entityDataFormatProvider->getMimeType( $defaultFormat );

			// prefer the default
			if ( $defaultMime != null ) {
				$accept[$defaultMime] = 1;
			}
		}

		$mimeTypes = $this->entityDataFormatProvider->getSupportedMimeTypes();
		$mimeTypes[] = 'text/html'; // HTML is handled by the normal page URL

		$negotiator = new HttpAcceptNegotiator( $mimeTypes );
		$format = $negotiator->getBestSupportedKey( $accept, null );

		if ( $format === null ) {
			$mimeTypes = implode( ', ', $this->entityDataFormatProvider->getSupportedMimeTypes() );
			$msg = $output->msg( 'wikibase-entitydata-not-acceptable', $mimeTypes );
			// @phan-suppress-next-line SecurityCheck-DoubleEscaped
			throw new HttpError( 406, $msg );
		}

		$format = $this->getCanonicalFormat( $format );

		$url = $this->uriManager->getDocUrl( $id, $format, $revision );
		$output->redirect( $url, 303 );
	}

	/**
	 * Loads the requested Entity. Redirects are resolved if no specific revision
	 * is requested or they are explicitly allowed by $allowRedirects.
	 *
	 * @param EntityId $id
	 * @param int $revision The revision ID (use 0 for the current revision).
	 * @param bool $allowRedirects Can we fetch redirects when revision is set?
	 *
	 * @return array list( EntityRevision, RedirectRevision|null )
	 * @throws HttpError
	 * @suppress SecurityCheck-DoubleEscaped
	 */
	private function getEntityRevision( EntityId $id, $revision, $allowRedirects = false ) {
		$prefixedId = $id->getSerialization();
		$redirectRevision = null;

		try {
			$entityRevision = $this->entityRevisionLookup->getEntityRevision( $id, $revision );

			if ( $entityRevision === null ) {
				$this->logger->debug(
					'{method}: entity not found: {prefixedId}',
					[
						'method' => __METHOD__,
						'prefixedId' => $prefixedId,
					]
				);

				$msg = wfMessage( 'wikibase-entitydata-not-found', $prefixedId );
				throw new HttpError( 404, $msg );
			}
		} catch ( RevisionedUnresolvedRedirectException $ex ) {
			$redirectRevision = new RedirectRevision(
				new EntityRedirect( $id, $ex->getRedirectTargetId() ),
				$ex->getRevisionId(), $ex->getRevisionTimestamp()
			);

			if ( $revision === 0 || $allowRedirects ) {
				// If no specific revision is requested or redirects are explicitly allowed, resolve the redirect.
				list( $entityRevision, ) = $this->getEntityRevision( $ex->getRedirectTargetId(), 0 );
			} else {
				// The requested revision is a redirect
				$this->logger->debug(
					'{method}: revision {revision} of {prefixedId} is a redirect: {exMsg}',
					[
						'method' => __METHOD__,
						'revision' => $revision,
						'prefixedId' => $prefixedId,
						'exMsg' => strval( $ex ),
					]
				);

				$msg = wfMessage( 'wikibase-entitydata-bad-revision', $prefixedId, $revision );
				throw new HttpError( 400, $msg );
			}
		} catch ( BadRevisionException $ex ) {
			$this->logger->debug(
				'{method}: could not load revision {revision} or {prefixedId}: {exMsg}',
				[
					'method' => __METHOD__,
					'revision' => $revision,
					'prefixedId' => $prefixedId,
					'exMsg' => strval( $ex ),
				]
			);

			$msg = wfMessage( 'wikibase-entitydata-bad-revision', $prefixedId, $revision );
			throw new HttpError( 404, $msg );
		} catch ( StorageException $ex ) {
			$this->logger->debug(
				'{method}: failed to load {prefixedId}: {exMsg} (revision {revision})',
				[
					'method' => __METHOD__,
					'prefixedId' => $prefixedId,
					'exMsg' => strval( $ex ),
					'revision' => $revision,
				]
			);

			$msg = wfMessage( 'wikibase-entitydata-storage-error', $prefixedId, $revision );
			throw new HttpError( 500, $msg );
		}

		return [ $entityRevision, $redirectRevision ];
	}

	/**
	 * Loads incoming redirects referring to the given entity ID.
	 *
	 * @param EntityId $id
	 *
	 * @return EntityId[]
	 * @throws HttpError
	 */
	private function getIncomingRedirects( EntityId $id ) {
		try {
			return $this->entityRedirectLookup->getRedirectIds( $id );
		} catch ( EntityRedirectLookupException $ex ) {
			$prefixedId = $id->getSerialization();
			$this->logger->debug(
				'{method}: failed to load incoming redirects of {prefixedId}: {exMsg}',
				[
					'method' => __METHOD__,
					'prefixedId' => $prefixedId,
					'exMsg' => strval( $ex ),
				]
			);

			return [];
		}
	}

	/**
	 * Output entity data.
	 *
	 * @param WebRequest $request
	 * @param OutputPage $output
	 * @param string $format The name (mime type of file extension) of the format to use
	 * @param EntityId $id The entity ID
	 * @param int $revision The revision ID (use 0 for the current revision).
	 *
	 * @throws HttpError
	 */
	public function showData( WebRequest $request, OutputPage $output, $format, EntityId $id, $revision ) {
		$flavor = $request->getRawVal( 'flavor' );

		/** @var EntityRevision $entityRevision */
		/** @var RedirectRevision $followedRedirectRevision */
		// If flavor is "dump", we allow fetching redirects by revision, since we won't
		// be dumping the content of the target revision.
		list( $entityRevision, $followedRedirectRevision ) = $this->getEntityRevision( $id, $revision, $flavor === 'dump' );

		// handle If-Modified-Since
		$imsHeader = $request->getHeader( 'IF-MODIFIED-SINCE' );
		if ( $imsHeader !== false ) {
			$ims = wfTimestamp( TS_MW, $imsHeader );

			if ( $entityRevision->getTimestamp() <= $ims ) {
				$response = $output->getRequest()->response();
				$response->header( 'Status: 304', true, 304 );
				$output->setArticleBodyOnly( true );
				return;
			}
		}

		if ( $flavor === 'dump' || $revision > 0 ) {
			// In dump mode and when fetching a specific revision, don't include incoming redirects.
			$incomingRedirects = [];
		} else {
			// Get the incoming redirects of the entity (if we followed a redirect, use the target id).
			$incomingRedirects = $this->getIncomingRedirects( $entityRevision->getEntity()->getId() );
		}

		list( $data, $contentType ) = $this->serializationService->getSerializedData(
			$format,
			$entityRevision,
			$followedRedirectRevision,
			$incomingRedirects,
			$flavor
		);

		$output->disable();
		$this->outputData(
			$request,
			$id,
			$revision,
			$output->getRequest()->response(),
			$data,
			$contentType,
			$entityRevision->getTimestamp()
		);
	}

	/**
	 * Output the entity data and set the appropriate HTTP response headers.
	 *
	 * @param WebRequest  $request
	 * @param EntityId    $requestId       the original entity ID of the request
	 * @param int         $requestRevision the original revision ID of the request (0 for latest)
	 * @param WebResponse $response
	 * @param string      $data        the data to output
	 * @param string      $contentType the data's mime type
	 * @param string      $lastModified
	 */
	public function outputData(
		WebRequest $request,
		EntityId $requestId,
		int $requestRevision,
		WebResponse $response,
		string $data,
		string $contentType,
		string $lastModified
	) {
		// NOTE: similar code as in RawAction::onView, keep in sync.

		$maxAge = $request->getInt( 'maxage', $this->maxAge );
		$sMaxAge = $request->getInt( 'smaxage', $this->maxAge );

		$maxAge  = max( self::MINIMUM_MAX_AGE, min( self::MAXIMUM_MAX_AGE, $maxAge ) );
		$sMaxAge = max( self::MINIMUM_MAX_AGE, min( self::MAXIMUM_MAX_AGE, $sMaxAge ) );

		$response->header( 'Content-Type: ' . $contentType . '; charset=UTF-8' );

		if ( $lastModified ) {
			$response->header( 'Last-Modified: ' . wfTimestamp( TS_RFC2822, $lastModified ) );
		}

		//Set X-Frame-Options API results (bug T41180)
		if ( $this->frameOptionsHeader !== null && $this->frameOptionsHeader !== '' ) {
			$response->header( "X-Frame-Options: $this->frameOptionsHeader" );
		}

		$cacheableUrls = $this->uriManager->getCacheableUrls( $requestId, $requestRevision );
		if ( in_array( $request->getFullRequestURL(), $cacheableUrls ) ) {
			$response->header( 'Cache-Control: public, s-maxage=' . $sMaxAge . ', max-age=' . $maxAge );
		} else {
			$response->header( 'Cache-Control: private, no-cache, s-maxage=0' );
		}

		ob_clean(); // remove anything that might already be in the output buffer.

		print $data;

		// exit normally here, keeping all levels of output buffering.
	}

}


1			<?php
2
3			namespace Wikibase\Repo\LinkedData;
4
5			use HtmlCacheUpdater;
6			use HttpError;
7			use OutputPage;
8			use Psr\Log\LoggerInterface;
9			use WebRequest;
10			use WebResponse;
11			use Wikibase\DataModel\Entity\EntityId;
12			use Wikibase\DataModel\Entity\EntityIdParser;
13			use Wikibase\DataModel\Entity\EntityIdParsingException;
14			use Wikibase\DataModel\Entity\EntityRedirect;
15			use Wikibase\DataModel\Services\Lookup\EntityRedirectLookup;
16			use Wikibase\DataModel\Services\Lookup\EntityRedirectLookupException;
17			use Wikibase\Lib\Store\BadRevisionException;
18			use Wikibase\Lib\Store\EntityRevision;
19			use Wikibase\Lib\Store\EntityRevisionLookup;
20			use Wikibase\Lib\Store\EntityTitleLookup;
21			use Wikibase\Lib\Store\RedirectRevision;
22			use Wikibase\Lib\Store\RevisionedUnresolvedRedirectException;
23			use Wikibase\Lib\Store\StorageException;
24			use Wikimedia\Http\HttpAcceptNegotiator;
25			use Wikimedia\Http\HttpAcceptParser;
26
27			/**
28			* Request handler implementing a linked data interface for Wikibase entities.
29			*
30			* @license GPL-2.0-or-later
31			* @author Daniel Kinzler
32			* @author Thomas Pellissier Tanon
33			* @author Anja Jentzsch < [email protected] >
34			*/
35			class EntityDataRequestHandler {
36
37			/**
38			* Allowed smallest and biggest number of seconds for the "max-age=..." and "s-maxage=..." cache
39			* control parameters.
40			*
41			* @todo Hard maximum could be configurable somehow.
42			*/
43			const MINIMUM_MAX_AGE = 0;
44			const MAXIMUM_MAX_AGE = 2678400; // 31 days
45
46			/**
47			* @var EntityDataSerializationService
48			*/
49			private $serializationService;
50
51			/**
52			* @var EntityDataUriManager
53			*/
54			private $uriManager;
55
56			/**
57			* @var EntityIdParser
58			*/
59			private $entityIdParser;
60
61			/**
62			* @var EntityRevisionLookup
63			*/
64			private $entityRevisionLookup;
65
66			/**
67			* @var EntityRedirectLookup
68			*/
69			private $entityRedirectLookup;
70
71			/**
72			* @var EntityTitleLookup
73			*/
74			private $entityTitleLookup;
75
76			/**
77			* @var EntityDataFormatProvider
78			*/
79			private $entityDataFormatProvider;
80
81			/**
82			* @var HtmlCacheUpdater
83			*/
84			private $htmlCacheUpdater;
85
86			/**
87			* @var LoggerInterface
88			*/
89			private $logger;
90
91			/**
92			* @var string
93			*/
94			private $defaultFormat;
95
96			/**
97			* @var int Number of seconds to cache entity data.
98			*/
99			private $maxAge;
100
101			/**
102			* @var bool
103			*/
104			private $useCdn;
105
106			/**
107			* @var string\|null
108			*/
109			private $frameOptionsHeader;
110
111			/**
112			* @var string[]
113			*/
114			private $entityTypesWithoutRdfOutput;
115
116			/**
117			* @param EntityDataUriManager $uriManager
118			* @param HtmlCacheUpdater $htmlCacheUpdater
119			* @param EntityTitleLookup $entityTitleLookup
120			* @param EntityIdParser $entityIdParser
121			* @param EntityRevisionLookup $entityRevisionLookup
122			* @param EntityRedirectLookup $entityRedirectLookup
123			* @param EntityDataSerializationService $serializationService
124			* @param EntityDataFormatProvider $entityDataFormatProvider
125			* @param LoggerInterface $logger
126			* @param string[] $entityTypesWithoutRdfOutput
127			* @param string $defaultFormat The format as a file extension or MIME type.
128			* @param int $maxAge number of seconds to cache entity data
129			* @param bool $useCdn do we have web caches configured?
130			* @param string\|null $frameOptionsHeader for X-Frame-Options
131			*/
132			public function __construct(
133			EntityDataUriManager $uriManager,
134			HtmlCacheUpdater $htmlCacheUpdater,
135			EntityTitleLookup $entityTitleLookup,
136			EntityIdParser $entityIdParser,
137			EntityRevisionLookup $entityRevisionLookup,
138			EntityRedirectLookup $entityRedirectLookup,
139			EntityDataSerializationService $serializationService,
140			EntityDataFormatProvider $entityDataFormatProvider,
141			LoggerInterface $logger,
142			array $entityTypesWithoutRdfOutput,
143			$defaultFormat,
144			$maxAge,
145			$useCdn,
146			$frameOptionsHeader
147			) {
148			$this->uriManager = $uriManager;
149			$this->htmlCacheUpdater = $htmlCacheUpdater;
150			$this->entityTitleLookup = $entityTitleLookup;
151			$this->entityIdParser = $entityIdParser;
152			$this->entityRevisionLookup = $entityRevisionLookup;
153			$this->entityRedirectLookup = $entityRedirectLookup;
154			$this->serializationService = $serializationService;
155			$this->entityDataFormatProvider = $entityDataFormatProvider;
156			$this->logger = $logger;
157			$this->entityTypesWithoutRdfOutput = $entityTypesWithoutRdfOutput;
158			$this->defaultFormat = $defaultFormat;
159			$this->maxAge = $maxAge;
160			$this->useCdn = $useCdn;
161			$this->frameOptionsHeader = $frameOptionsHeader;
162			}
163
164			/**
165			* Checks whether the request is complete, i.e. whether it contains all information needed
166			* to reply with entity data.
167			*
168			* This does not check whether the request is valid and will actually produce a successful
169			* response.
170			*
171			* @param string\|null $doc Document name, e.g. Q5 or Q5.json or Q5:33.xml
172			* @param WebRequest $request
173			*
174			* @return bool
175			*/
176			public function canHandleRequest( $doc, WebRequest $request ) {
177			if ( $doc === '' \|\| $doc === null ) {
178			if ( $request->getText( 'id', '' ) === '' ) {
179			return false;
180			}
181			}
182
183			return true;
184			}
185
186			/**
187			* Main method for handling requests.
188			*
189			* @param string $doc Document name, e.g. Q5 or Q5.json or Q5:33.xml
190			* @param WebRequest $request The request parameters. Known parameters are:
191			* - id: the entity ID
192			* - format: the format
193			* - oldid\|revision: the revision ID
194			* - action=purge: to purge cached data from (web) caches
195			* @param OutputPage $output
196			*
197			* @note Instead of an output page, a WebResponse could be sufficient, but
198			* redirect logic is currently implemented in OutputPage.
199			*
200			* @throws HttpError
201			* @suppress SecurityCheck-DoubleEscaped
202			*/
203			public function handleRequest( $doc, WebRequest $request, OutputPage $output ) {
204			// No matter what: The response is always public
205			$output->getRequest()->response()->header( 'Access-Control-Allow-Origin: *' );
206
207			$revision = 0;
208
209			list( $id, $format ) = $this->uriManager->parseDocName( $doc );
210
211			// get entity id and format from request parameter
212			$format = $request->getText( 'format', $format );
213			$id = $request->getText( 'id', $id );
214			$revision = $request->getInt( 'revision', $revision );
215			$redirectMode = $request->getText( 'redirect' );
216			//TODO: malformed revision IDs should trigger a code 400
217
218			// If there is no ID, fail
219			if ( $id === null \|\| $id === '' ) {
220			//TODO: different error message?
221			throw new HttpError( 400, $output->msg( 'wikibase-entitydata-bad-id', $id ) );
222			}
223
224			try {
225			$entityId = $this->entityIdParser->parse( $id );
226			} catch ( EntityIdParsingException $ex ) {
227			throw new HttpError( 400, $output->msg( 'wikibase-entitydata-bad-id', $id ) );
228			}
229
230			if ( $this->entityDataFormatProvider->isRdfFormat( $format ) &&
231			in_array( $entityId->getEntityType(), $this->entityTypesWithoutRdfOutput )
232			) {
233			throw new HttpError( 406, $output->msg( 'wikibase-entitydata-rdf-not-available', $entityId->getEntityType() ) );
234			}
235
236			//XXX: allow for logged in users only?
237			if ( $request->getText( 'action' ) === 'purge' ) {
238			$this->purgeWebCache( $entityId, $revision );
239			//XXX: Now what? Proceed to show the data?
240			}
241
242			if ( $format === null \|\| $format === '' ) {
243			// if no format is given, apply content negotiation and return.
244			$this->httpContentNegotiation( $request, $output, $entityId, $revision );
245			return;
246			}
247
248			//NOTE: will trigger a 415 if the format is not supported
249			$format = $this->getCanonicalFormat( $format );
250
251			if ( $doc !== null && $doc !== '' ) {
252			// if subpage syntax is used, always enforce the canonical form
253			$canonicalDoc = $this->uriManager->getDocName( $entityId, $format );
254
255			if ( $doc !== $canonicalDoc ) {
256			$url = $this->uriManager->getDocUrl( $entityId, $format, $revision );
257			$output->redirect( $url, 301 );
258			return;
259			}
260			}
261
262			// if the format is HTML, redirect to the entity's wiki page
263			if ( $format === 'html' ) {
264			$url = $this->uriManager->getDocUrl( $entityId, 'html', $revision );
265			$output->redirect( $url, 303 );
266			return;
267			}
268
269			// if redirection was force, redirect
270			if ( $redirectMode === 'force' ) {
271			$url = $this->uriManager->getDocUrl( $entityId, $format, $revision );
272			$output->redirect( $url, 303 );
273			return;
274			}
275
276			$this->showData( $request, $output, $format, $entityId, $revision );
277			}
278
279			/**
280			* Returns the canonical format name for the given format.
281			*
282			* @param string $format
283			*
284			* @return string
285			* @throws HttpError code 415 if the format is not supported.
286			*/
287			public function getCanonicalFormat( $format ) {
288			$format = strtolower( $format );
289
290			// we always support html, it's handled by the entity's wiki page.
291			if ( $format === 'html' \|\| $format === 'htm' \|\| $format === 'text/html' ) {
292			return 'html';
293			}
294
295			// normalize format name (note that HTML may not be known to the service)
296			$canonicalFormat = $this->entityDataFormatProvider->getFormatName( $format );
297
298			if ( $canonicalFormat === null ) {
299			$msg = wfMessage( 'wikibase-entitydata-unsupported-format', $format );
300			// @phan-suppress-next-line SecurityCheck-DoubleEscaped
301			throw new HttpError( 415, $msg );
302			}
303
304			return $canonicalFormat;
305			}
306
307			/**
308			* Purges the entity data identified by the doc parameter from any HTTP caches.
309			* Does nothing if $wgUseCdn is not set.
310			*
311			* @param EntityId $id The entity ID for which to purge all data.
312			* @param int $revision The revision ID (0 for current/unspecified)
313			*/
314			public function purgeWebCache( EntityId $id, int $revision ) {
315			$urls = $this->uriManager->getPotentiallyCachedUrls( $id, $revision );
316			if ( $urls !== [] ) {
317			$this->htmlCacheUpdater->purgeUrls( $urls );
318			}
319			}
320
321			/**
322			* Applies HTTP content negotiation.
323			* If the negotiation is successfull, this method will set the appropriate redirect
324			* in the OutputPage object and return. Otherwise, an HttpError is thrown.
325			*
326			* @param WebRequest $request
327			* @param OutputPage $output
328			* @param EntityId $id The ID of the entity to show
329			* @param int $revision The desired revision
330			*
331			* @throws HttpError
332			*/
333			public function httpContentNegotiation( WebRequest $request, OutputPage $output, EntityId $id, $revision = 0 ) {
334			$headers = $request->getAllHeaders();
335			if ( isset( $headers['ACCEPT'] ) ) {
336			$parser = new HttpAcceptParser();
337			$accept = $parser->parseWeights( $headers['ACCEPT'] );
338			} else {
339			// anything goes
340			$accept = [
341			'*' => 0.1 // just to make extra sure
342			];
343
344			$defaultFormat = $this->entityDataFormatProvider->getFormatName( $this->defaultFormat );
345			$defaultMime = $this->entityDataFormatProvider->getMimeType( $defaultFormat );
346
347			// prefer the default
348			if ( $defaultMime != null ) {
349			$accept[$defaultMime] = 1;
350			}
351			}
352
353			$mimeTypes = $this->entityDataFormatProvider->getSupportedMimeTypes();
354			$mimeTypes[] = 'text/html'; // HTML is handled by the normal page URL
355
356			$negotiator = new HttpAcceptNegotiator( $mimeTypes );
357			$format = $negotiator->getBestSupportedKey( $accept, null );
358
359			if ( $format === null ) {
360			$mimeTypes = implode( ', ', $this->entityDataFormatProvider->getSupportedMimeTypes() );
361			$msg = $output->msg( 'wikibase-entitydata-not-acceptable', $mimeTypes );
362			// @phan-suppress-next-line SecurityCheck-DoubleEscaped
363			throw new HttpError( 406, $msg );
364			}
365
366			$format = $this->getCanonicalFormat( $format );
367
368			$url = $this->uriManager->getDocUrl( $id, $format, $revision );
369			$output->redirect( $url, 303 );
370			}
371
372			/**
373			* Loads the requested Entity. Redirects are resolved if no specific revision
374			* is requested or they are explicitly allowed by $allowRedirects.
375			*
376			* @param EntityId $id
377			* @param int $revision The revision ID (use 0 for the current revision).
378			* @param bool $allowRedirects Can we fetch redirects when revision is set?
379			*
380			* @return array list( EntityRevision, RedirectRevision\|null )
381			* @throws HttpError
382			* @suppress SecurityCheck-DoubleEscaped
383			*/
384			private function getEntityRevision( EntityId $id, $revision, $allowRedirects = false ) {
385			$prefixedId = $id->getSerialization();
386			$redirectRevision = null;
387
388			try {
389			$entityRevision = $this->entityRevisionLookup->getEntityRevision( $id, $revision );
390
391			if ( $entityRevision === null ) {
392			$this->logger->debug(
393			'{method}: entity not found: {prefixedId}',
394			[
395			'method' => __METHOD__,
396			'prefixedId' => $prefixedId,
397			]
398			);
399
400			$msg = wfMessage( 'wikibase-entitydata-not-found', $prefixedId );
401			throw new HttpError( 404, $msg );
402			}
403			} catch ( RevisionedUnresolvedRedirectException $ex ) {
404			$redirectRevision = new RedirectRevision(
405			new EntityRedirect( $id, $ex->getRedirectTargetId() ),
406			$ex->getRevisionId(), $ex->getRevisionTimestamp()
407			);
408
409			if ( $revision === 0 \|\| $allowRedirects ) {
410			// If no specific revision is requested or redirects are explicitly allowed, resolve the redirect.
411			list( $entityRevision, ) = $this->getEntityRevision( $ex->getRedirectTargetId(), 0 );
412			} else {
413			// The requested revision is a redirect
414			$this->logger->debug(
415			'{method}: revision {revision} of {prefixedId} is a redirect: {exMsg}',
416			[
417			'method' => __METHOD__,
418			'revision' => $revision,
419			'prefixedId' => $prefixedId,
420			'exMsg' => strval( $ex ),
421			]
422			);
423
424			$msg = wfMessage( 'wikibase-entitydata-bad-revision', $prefixedId, $revision );
425			throw new HttpError( 400, $msg );
426			}
427			} catch ( BadRevisionException $ex ) {
428			$this->logger->debug(
429			'{method}: could not load revision {revision} or {prefixedId}: {exMsg}',
430			[
431			'method' => __METHOD__,
432			'revision' => $revision,
433			'prefixedId' => $prefixedId,
434			'exMsg' => strval( $ex ),
435			]
436			);
437
438			$msg = wfMessage( 'wikibase-entitydata-bad-revision', $prefixedId, $revision );
439			throw new HttpError( 404, $msg );
440			} catch ( StorageException $ex ) {
441			$this->logger->debug(
442			'{method}: failed to load {prefixedId}: {exMsg} (revision {revision})',
443			[
444			'method' => __METHOD__,
445			'prefixedId' => $prefixedId,
446			'exMsg' => strval( $ex ),
447			'revision' => $revision,
448			]
449			);
450
451			$msg = wfMessage( 'wikibase-entitydata-storage-error', $prefixedId, $revision );
452			throw new HttpError( 500, $msg );
453			}
454
455			return [ $entityRevision, $redirectRevision ];
456			}
457
458			/**
459			* Loads incoming redirects referring to the given entity ID.
460			*
461			* @param EntityId $id
462			*
463			* @return EntityId[]
464			* @throws HttpError
465			*/
466			private function getIncomingRedirects( EntityId $id ) {
467			try {
468			return $this->entityRedirectLookup->getRedirectIds( $id );
469			} catch ( EntityRedirectLookupException $ex ) {
470			$prefixedId = $id->getSerialization();
471			$this->logger->debug(
472			'{method}: failed to load incoming redirects of {prefixedId}: {exMsg}',
473			[
474			'method' => __METHOD__,
475			'prefixedId' => $prefixedId,
476			'exMsg' => strval( $ex ),
477			]
478			);
479
480			return [];
481			}
482			}
483
484			/**
485			* Output entity data.
486			*
487			* @param WebRequest $request
488			* @param OutputPage $output
489			* @param string $format The name (mime type of file extension) of the format to use
490			* @param EntityId $id The entity ID
491			* @param int $revision The revision ID (use 0 for the current revision).
492			*
493			* @throws HttpError
494			*/
495			public function showData( WebRequest $request, OutputPage $output, $format, EntityId $id, $revision ) {
496			$flavor = $request->getRawVal( 'flavor' );
497
498			/** @var EntityRevision $entityRevision */
499			/** @var RedirectRevision $followedRedirectRevision */
500			// If flavor is "dump", we allow fetching redirects by revision, since we won't
501			// be dumping the content of the target revision.
502			list( $entityRevision, $followedRedirectRevision ) = $this->getEntityRevision( $id, $revision, $flavor === 'dump' );
503
504			// handle If-Modified-Since
505			$imsHeader = $request->getHeader( 'IF-MODIFIED-SINCE' );
506			if ( $imsHeader !== false ) {
507			$ims = wfTimestamp( TS_MW, $imsHeader );
508
509			if ( $entityRevision->getTimestamp() <= $ims ) {
510			$response = $output->getRequest()->response();
511			$response->header( 'Status: 304', true, 304 );
512			$output->setArticleBodyOnly( true );
513			return;
514			}
515			}
516
517			if ( $flavor === 'dump' \|\| $revision > 0 ) {
518			// In dump mode and when fetching a specific revision, don't include incoming redirects.
519			$incomingRedirects = [];
520			} else {
521			// Get the incoming redirects of the entity (if we followed a redirect, use the target id).
522			$incomingRedirects = $this->getIncomingRedirects( $entityRevision->getEntity()->getId() );
523			}
524
525			list( $data, $contentType ) = $this->serializationService->getSerializedData(
526			$format,
527			$entityRevision,
528			$followedRedirectRevision,
529			$incomingRedirects,
530			$flavor
531			);
532
533			$output->disable();
534			$this->outputData(
535			$request,
536			$id,
537			$revision,
538			$output->getRequest()->response(),
539			$data,
540			$contentType,
541			$entityRevision->getTimestamp()
542			);
543			}
544
545			/**
546			* Output the entity data and set the appropriate HTTP response headers.
547			*
548			* @param WebRequest $request
549			* @param EntityId $requestId the original entity ID of the request
550			* @param int $requestRevision the original revision ID of the request (0 for latest)
551			* @param WebResponse $response
552			* @param string $data the data to output
553			* @param string $contentType the data's mime type
554			* @param string $lastModified
555			*/
556			public function outputData(
557			WebRequest $request,
558			EntityId $requestId,
559			int $requestRevision,
560			WebResponse $response,
561			string $data,
562			string $contentType,
563			string $lastModified
564			) {
565			// NOTE: similar code as in RawAction::onView, keep in sync.
566
567			$maxAge = $request->getInt( 'maxage', $this->maxAge );
568			$sMaxAge = $request->getInt( 'smaxage', $this->maxAge );
569
570			$maxAge = max( self::MINIMUM_MAX_AGE, min( self::MAXIMUM_MAX_AGE, $maxAge ) );
571			$sMaxAge = max( self::MINIMUM_MAX_AGE, min( self::MAXIMUM_MAX_AGE, $sMaxAge ) );
572
573			$response->header( 'Content-Type: ' . $contentType . '; charset=UTF-8' );
574
575			if ( $lastModified ) {
576			$response->header( 'Last-Modified: ' . wfTimestamp( TS_RFC2822, $lastModified ) );
577			}
578
579			//Set X-Frame-Options API results (bug T41180)
580			if ( $this->frameOptionsHeader !== null && $this->frameOptionsHeader !== '' ) {
581			$response->header( "X-Frame-Options: $this->frameOptionsHeader" );
582			}
583
584			$cacheableUrls = $this->uriManager->getCacheableUrls( $requestId, $requestRevision );
585			if ( in_array( $request->getFullRequestURL(), $cacheableUrls ) ) {
586			$response->header( 'Cache-Control: public, s-maxage=' . $sMaxAge . ', max-age=' . $maxAge );
587			} else {
588			$response->header( 'Cache-Control: private, no-cache, s-maxage=0' );
589			}
590
591			ob_clean(); // remove anything that might already be in the output buffer.
592
593			print $data;
594
595			// exit normally here, keeping all levels of output buffering.
596			}
597
598			}
599

wikimedia / mediawiki-extensions-Wikibase

EntityDataRequestHandler::getIncomingRedirects() A last analyzed 2020-11-13 17:34 UTC

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like

EntityDataRequestHandler::getIncomingRedirects() A
last analyzed 2020-11-13 17:34 UTC