Skip to content

Commit 4996970

Browse files
kyteinskykyteinsky
committed
feat: Migrate LinkReferenceProvider to OCP
This would be useful to eleminate the need for using this OC class when an app dev wants to implement a custom reference provider for the web client but wants to fall back on opengraph for mobile clients. Signed-off-by: Anupam Kumar <kyteinsky@gmail.com>
1 parent 3fb1674 commit 4996970

File tree

2 files changed

+219
-155
lines changed

2 files changed

+219
-155
lines changed

‎lib/private/Collaboration/Reference/LinkReferenceProvider.php‎

Lines changed: 4 additions & 155 deletions
Original file line numberDiff line numberDiff line change
@@ -5,6 +5,7 @@
55
* @copyright Copyright (c) 2022 Julius Härtl <jus@bitgrid.net>
66
*
77
* @author Julius Härtl <jus@bitgrid.net>
8+
* @author Anupam Kumar <kyteinsky@gmail.com>
89
*
910
* @license GNU AGPL version 3 or any later version
1011
*
@@ -24,160 +25,8 @@
2425

2526
namespace OC\Collaboration\Reference;
2627

27-
use Fusonic\OpenGraph\Consumer;
28-
use GuzzleHttp\Exception\GuzzleException;
29-
use GuzzleHttp\Psr7\LimitStream;
30-
use GuzzleHttp\Psr7\Utils;
31-
use OC\Security\RateLimiting\Exception\RateLimitExceededException;
32-
use OC\Security\RateLimiting\Limiter;
33-
use OC\SystemConfig;
34-
use OCP\Collaboration\Reference\IReference;
35-
use OCP\Collaboration\Reference\IReferenceProvider;
36-
use OCP\Collaboration\Reference\Reference;
37-
use OCP\Files\AppData\IAppDataFactory;
38-
use OCP\Files\NotFoundException;
39-
use OCP\Http\Client\IClientService;
40-
use OCP\IRequest;
41-
use OCP\IURLGenerator;
42-
use OCP\IUserSession;
43-
use Psr\Log\LoggerInterface;
28+
use OCP\Collaboration\Reference\LinkReferenceProvider as OCPLinkReferenceProvider;
4429

45-
class LinkReferenceProvider implements IReferenceProvider {
46-
public const MAX_PREVIEW_SIZE = 1024 * 1024;
47-
48-
public const ALLOWED_CONTENT_TYPES = [
49-
'image/png',
50-
'image/jpg',
51-
'image/jpeg',
52-
'image/gif',
53-
'image/svg+xml',
54-
'image/webp'
55-
];
56-
57-
public function __construct(
58-
private IClientService $clientService,
59-
private LoggerInterface $logger,
60-
private SystemConfig $systemConfig,
61-
private IAppDataFactory $appDataFactory,
62-
private IURLGenerator $urlGenerator,
63-
private Limiter $limiter,
64-
private IUserSession $userSession,
65-
private IRequest $request,
66-
) {
67-
}
68-
69-
public function matchReference(string $referenceText): bool {
70-
if ($this->systemConfig->getValue('reference_opengraph', true) !== true) {
71-
return false;
72-
}
73-
74-
return (bool)preg_match(IURLGenerator::URL_REGEX, $referenceText);
75-
}
76-
77-
public function resolveReference(string $referenceText): ?IReference {
78-
if ($this->matchReference($referenceText)) {
79-
$reference = new Reference($referenceText);
80-
$this->fetchReference($reference);
81-
return $reference;
82-
}
83-
84-
return null;
85-
}
86-
87-
private function fetchReference(Reference $reference): void {
88-
try {
89-
$user = $this->userSession->getUser();
90-
if ($user) {
91-
$this->limiter->registerUserRequest('opengraph', 10, 120, $user);
92-
} else {
93-
$this->limiter->registerAnonRequest('opengraph', 10, 120, $this->request->getRemoteAddress());
94-
}
95-
} catch (RateLimitExceededException $e) {
96-
return;
97-
}
98-
99-
$client = $this->clientService->newClient();
100-
try {
101-
$headResponse = $client->head($reference->getId(), [ 'timeout' => 10 ]);
102-
} catch (\Exception $e) {
103-
$this->logger->debug('Failed to perform HEAD request to get target metadata', ['exception' => $e]);
104-
return;
105-
}
106-
$linkContentLength = $headResponse->getHeader('Content-Length');
107-
if (is_numeric($linkContentLength) && (int) $linkContentLength > 5 * 1024 * 1024) {
108-
$this->logger->debug('Skip resolving links pointing to content length > 5 MB');
109-
return;
110-
}
111-
$linkContentType = $headResponse->getHeader('Content-Type');
112-
$expectedContentType = 'text/html';
113-
$suffixedExpectedContentType = $expectedContentType . ';';
114-
$startsWithSuffixed = str_starts_with($linkContentType, $suffixedExpectedContentType);
115-
// check the header begins with the expected content type
116-
if ($linkContentType !== $expectedContentType && !$startsWithSuffixed) {
117-
$this->logger->debug('Skip resolving links pointing to content type that is not "text/html"');
118-
return;
119-
}
120-
try {
121-
$response = $client->get($reference->getId(), [ 'timeout' => 10 ]);
122-
} catch (\Exception $e) {
123-
$this->logger->debug('Failed to fetch link for obtaining open graph data', ['exception' => $e]);
124-
return;
125-
}
126-
127-
$responseBody = (string)$response->getBody();
128-
129-
// OpenGraph handling
130-
$consumer = new Consumer();
131-
$consumer->useFallbackMode = true;
132-
$object = $consumer->loadHtml($responseBody);
133-
134-
$reference->setUrl($reference->getId());
135-
136-
if ($object->title) {
137-
$reference->setTitle($object->title);
138-
}
139-
140-
if ($object->description) {
141-
$reference->setDescription($object->description);
142-
}
143-
144-
if ($object->images) {
145-
try {
146-
$host = parse_url($object->images[0]->url, PHP_URL_HOST);
147-
if ($host === false || $host === null) {
148-
$this->logger->warning('Could not detect host of open graph image URI for ' . $reference->getId());
149-
} else {
150-
$appData = $this->appDataFactory->get('core');
151-
try {
152-
$folder = $appData->getFolder('opengraph');
153-
} catch (NotFoundException $e) {
154-
$folder = $appData->newFolder('opengraph');
155-
}
156-
$response = $client->get($object->images[0]->url, ['timeout' => 10]);
157-
$contentType = $response->getHeader('Content-Type');
158-
$contentLength = $response->getHeader('Content-Length');
159-
160-
if (in_array($contentType, self::ALLOWED_CONTENT_TYPES, true) && $contentLength < self::MAX_PREVIEW_SIZE) {
161-
$stream = Utils::streamFor($response->getBody());
162-
$bodyStream = new LimitStream($stream, self::MAX_PREVIEW_SIZE, 0);
163-
$reference->setImageContentType($contentType);
164-
$folder->newFile(md5($reference->getId()), $bodyStream->getContents());
165-
$reference->setImageUrl($this->urlGenerator->linkToRouteAbsolute('core.Reference.preview', ['referenceId' => md5($reference->getId())]));
166-
}
167-
}
168-
} catch (GuzzleException $e) {
169-
$this->logger->info('Failed to fetch and store the open graph image for ' . $reference->getId(), ['exception' => $e]);
170-
} catch (\Throwable $e) {
171-
$this->logger->error('Failed to fetch and store the open graph image for ' . $reference->getId(), ['exception' => $e]);
172-
}
173-
}
174-
}
175-
176-
public function getCachePrefix(string $referenceId): string {
177-
return $referenceId;
178-
}
179-
180-
public function getCacheKey(string $referenceId): ?string {
181-
return null;
182-
}
30+
/** @deprecated 29.0.0 Use OCP\Collaboration\Reference\LinkReferenceProvider instead */
31+
class LinkReferenceProvider extends OCPLinkReferenceProvider {
18332
}

‎lib/public/Collaboration/Reference/LinkReferenceProvider.php‎

Lines changed: 215 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,215 @@
1+
<?php
2+
3+
declare(strict_types=1);
4+
/**
5+
* @copyright Copyright (c) 2022 Julius Härtl <jus@bitgrid.net>
6+
*
7+
* @author Julius Härtl <jus@bitgrid.net>
8+
* @author Anupam Kumar <kyteinsky@gmail.com>
9+
*
10+
* @license GNU AGPL version 3 or any later version
11+
*
12+
* This program is free software: you can redistribute it and/or modify
13+
* it under the terms of the GNU Affero General Public License as
14+
* published by the Free Software Foundation, either version 3 of the
15+
* License, or (at your option) any later version.
16+
*
17+
* This program is distributed in the hope that it will be useful,
18+
* but WITHOUT ANY WARRANTY; without even the implied warranty of
19+
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
20+
* GNU Affero General Public License for more details.
21+
*
22+
* You should have received a copy of the GNU Affero General Public License
23+
* along with this program. If not, see <http://www.gnu.org/licenses/>.
24+
*/
25+
26+
namespace OCP\Collaboration\Reference;
27+
28+
use Fusonic\OpenGraph\Consumer;
29+
use GuzzleHttp\Exception\GuzzleException;
30+
use GuzzleHttp\Psr7\LimitStream;
31+
use GuzzleHttp\Psr7\Utils;
32+
use OC\Security\RateLimiting\Exception\RateLimitExceededException;
33+
use OC\Security\RateLimiting\Limiter;
34+
use OC\SystemConfig;
35+
use OCP\Files\AppData\IAppDataFactory;
36+
use OCP\Files\NotFoundException;
37+
use OCP\Http\Client\IClientService;
38+
use OCP\IRequest;
39+
use OCP\IURLGenerator;
40+
use OCP\IUserSession;
41+
use Psr\Log\LoggerInterface;
42+
43+
/**
44+
* @since 29.0.0
45+
*/
46+
class LinkReferenceProvider implements IReferenceProvider {
47+
48+
/* for image size and webpage header */
49+
private const MAX_CONTENT_LENGTH = 5 * 1024 * 1024;
50+
51+
private const ALLOWED_CONTENT_TYPES = [
52+
'image/png',
53+
'image/jpg',
54+
'image/jpeg',
55+
'image/gif',
56+
'image/svg+xml',
57+
'image/webp'
58+
];
59+
60+
/**
61+
* @since 29.0.0
62+
*/
63+
public function __construct(
64+
private IClientService $clientService,
65+
private LoggerInterface $logger,
66+
private SystemConfig $systemConfig,
67+
private IAppDataFactory $appDataFactory,
68+
private IURLGenerator $urlGenerator,
69+
private Limiter $limiter,
70+
private IUserSession $userSession,
71+
private IRequest $request,
72+
) {
73+
}
74+
75+
/**
76+
* @inheritDoc
77+
* @since 29.0.0
78+
*/
79+
public function matchReference(string $referenceText): bool {
80+
if ($this->systemConfig->getValue('reference_opengraph', true) !== true) {
81+
return false;
82+
}
83+
84+
return (bool)preg_match(IURLGenerator::URL_REGEX, $referenceText);
85+
}
86+
87+
/**
88+
* @inheritDoc
89+
* @since 29.0.0
90+
*/
91+
public function resolveReference(string $referenceText): ?IReference {
92+
if ($this->matchReference($referenceText)) {
93+
$reference = new Reference($referenceText);
94+
$this->fetchReference($reference);
95+
return $reference;
96+
}
97+
98+
return null;
99+
}
100+
101+
/**
102+
* Populates the reference with OpenGraph data
103+
*
104+
* @param Reference $reference
105+
* @since 29.0.0
106+
*/
107+
private function fetchReference(Reference $reference): void {
108+
try {
109+
$user = $this->userSession->getUser();
110+
if ($user) {
111+
$this->limiter->registerUserRequest('opengraph', 10, 120, $user);
112+
} else {
113+
$this->limiter->registerAnonRequest('opengraph', 10, 120, $this->request->getRemoteAddress());
114+
}
115+
} catch (RateLimitExceededException $e) {
116+
return;
117+
}
118+
119+
$client = $this->clientService->newClient();
120+
try {
121+
$headResponse = $client->head($reference->getId(), [ 'timeout' => 10 ]);
122+
} catch (\Exception $e) {
123+
$this->logger->debug('Failed to perform HEAD request to get target metadata', ['exception' => $e]);
124+
return;
125+
}
126+
127+
$linkContentLength = $headResponse->getHeader('Content-Length');
128+
if (is_numeric($linkContentLength) && (int) $linkContentLength > self::MAX_CONTENT_LENGTH) {
129+
$this->logger->debug('Skip resolving links pointing to content length > 5 MiB');
130+
return;
131+
}
132+
133+
$linkContentType = $headResponse->getHeader('Content-Type');
134+
$expectedContentTypeRegex = '/^text\/html;?/i';
135+
136+
// check the header begins with the expected content type
137+
if (!preg_match($expectedContentTypeRegex, $linkContentType)) {
138+
$this->logger->debug('Skip resolving links pointing to content type that is not "text/html"');
139+
return;
140+
}
141+
142+
try {
143+
$response = $client->get($reference->getId(), [ 'timeout' => 10 ]);
144+
} catch (\Exception $e) {
145+
$this->logger->debug('Failed to fetch link for obtaining open graph data', ['exception' => $e]);
146+
return;
147+
}
148+
149+
$responseBody = (string)$response->getBody();
150+
151+
// OpenGraph handling
152+
$consumer = new Consumer();
153+
$consumer->useFallbackMode = true;
154+
$object = $consumer->loadHtml($responseBody);
155+
156+
$reference->setUrl($reference->getId());
157+
158+
if ($object->title) {
159+
$reference->setTitle($object->title);
160+
}
161+
162+
if ($object->description) {
163+
$reference->setDescription($object->description);
164+
}
165+
166+
if ($object->images) {
167+
try {
168+
$host = parse_url($object->images[0]->url, PHP_URL_HOST);
169+
if ($host === false || $host === null) {
170+
$this->logger->warning('Could not detect host of open graph image URI for ' . $reference->getId());
171+
return;
172+
}
173+
174+
$appData = $this->appDataFactory->get('core');
175+
try {
176+
$folder = $appData->getFolder('opengraph');
177+
} catch (NotFoundException $e) {
178+
$folder = $appData->newFolder('opengraph');
179+
}
180+
181+
$response = $client->get($object->images[0]->url, ['timeout' => 10]);
182+
$contentType = $response->getHeader('Content-Type');
183+
$contentLength = $response->getHeader('Content-Length');
184+
185+
if (in_array($contentType, self::ALLOWED_CONTENT_TYPES, true) && $contentLength < self::MAX_CONTENT_LENGTH) {
186+
$stream = Utils::streamFor($response->getBody());
187+
$bodyStream = new LimitStream($stream, self::MAX_CONTENT_LENGTH, 0);
188+
$reference->setImageContentType($contentType);
189+
$folder->newFile(md5($reference->getId()), $bodyStream->getContents());
190+
$reference->setImageUrl($this->urlGenerator->linkToRouteAbsolute('core.Reference.preview', ['referenceId' => md5($reference->getId())]));
191+
}
192+
} catch (GuzzleException $e) {
193+
$this->logger->info('Failed to fetch and store the open graph image for ' . $reference->getId(), ['exception' => $e]);
194+
} catch (\Throwable $e) {
195+
$this->logger->error('Failed to fetch and store the open graph image for ' . $reference->getId(), ['exception' => $e]);
196+
}
197+
}
198+
}
199+
200+
/**
201+
* @inheritDoc
202+
* @since 29.0.0
203+
*/
204+
public function getCachePrefix(string $referenceId): string {
205+
return $referenceId;
206+
}
207+
208+
/**
209+
* @inheritDoc
210+
* @since 29.0.0
211+
*/
212+
public function getCacheKey(string $referenceId): ?string {
213+
return null;
214+
}
215+
}

0 commit comments

Comments
 (0)