Add functionality to delete old versions of documents

This commit introduces a feature that automatically deletes old versions of StoredObjects in the Chill application. A cron job, "RemoveOldVersionCronJob", has been implemented to delete versions older than 90 days. A message handler, "RemoveOldVersionMessageHandler", has been added to handle deletion requests. Furthermore, unit tests for the new functionality have been provided.
This commit is contained in:
Julien Fastré 2024-07-15 15:54:26 +02:00
parent 67d24cb951
commit c38f7c1179
Signed by: julienfastre
GPG Key ID: BDE2190974723FCB
10 changed files with 476 additions and 0 deletions

View File

@ -0,0 +1,92 @@
<?php
declare(strict_types=1);
/*
* Chill is a software for social workers
*
* For the full copyright and license information, please view
* the LICENSE file that was distributed with this source code.
*/
namespace Chill\DocStoreBundle\Repository;
use Chill\DocStoreBundle\Entity\StoredObjectVersion;
use Doctrine\DBAL\Connection;
use Doctrine\DBAL\Types\Types;
use Doctrine\ORM\EntityManagerInterface;
use Doctrine\ORM\EntityRepository;
use Doctrine\Persistence\ObjectRepository;
/**
* @implements ObjectRepository<StoredObjectVersion>
*/
class StoredObjectVersionRepository implements ObjectRepository
{
private EntityRepository $repository;
private Connection $connection;
public function __construct(EntityManagerInterface $entityManager)
{
$this->repository = $entityManager->getRepository(StoredObjectVersion::class);
$this->connection = $entityManager->getConnection();
}
public function find($id): ?StoredObjectVersion
{
return $this->repository->find($id);
}
public function findAll(): array
{
return $this->repository->findAll();
}
public function findBy(array $criteria, ?array $orderBy = null, ?int $limit = null, ?int $offset = null): array
{
return $this->repository->findBy($criteria, $orderBy, $limit, $offset);
}
public function findOneBy(array $criteria): ?StoredObjectVersion
{
return $this->repository->findOneBy($criteria);
}
/**
* Finds the IDs of versions older than a given date and that are not the last version.
*
* Those version are good candidates for a deletion.
*
* @param \DateTimeImmutable $beforeDate the date to compare versions against
*
* @return iterable returns an iterable with the IDs of the versions
*/
public function findIdsByVersionsOlderThanDateAndNotLastVersion(\DateTimeImmutable $beforeDate): iterable
{
$results = $this->connection->executeQuery(
self::QUERY_FIND_IDS_BY_VERSIONS_OLDER_THAN_DATE_AND_NOT_LAST_VERSION,
[$beforeDate],
[Types::DATETIME_IMMUTABLE]
);
foreach ($results->iterateAssociative() as $row) {
yield $row['sov_id'];
}
}
private const QUERY_FIND_IDS_BY_VERSIONS_OLDER_THAN_DATE_AND_NOT_LAST_VERSION = <<<'SQL'
SELECT
sov.id AS sov_id
FROM chill_doc.stored_object_version sov
WHERE
sov.createdat < ?::timestamp
AND
sov.version < (SELECT MAX(sub_sov.version) FROM chill_doc.stored_object_version sub_sov WHERE sub_sov.stored_object_id = sov.stored_object_id)
SQL;
public function getClassName(): string
{
return StoredObjectVersion::class;
}
}

View File

@ -0,0 +1,60 @@
<?php
declare(strict_types=1);
/*
* Chill is a software for social workers
*
* For the full copyright and license information, please view
* the LICENSE file that was distributed with this source code.
*/
namespace Chill\DocStoreBundle\Service\StoredObjectCleaner;
use Chill\DocStoreBundle\Repository\StoredObjectVersionRepository;
use Chill\MainBundle\Cron\CronJobInterface;
use Chill\MainBundle\Entity\CronJobExecution;
use Symfony\Component\Clock\ClockInterface;
use Symfony\Component\Messenger\MessageBusInterface;
final readonly class RemoveOldVersionCronJob implements CronJobInterface
{
public const KEY = 'remove-old-stored-object-version';
private const LAST_DELETED_KEY = 'last-deleted-stored-object-version-id';
public const KEEP_INTERVAL = 'P90D';
public function __construct(
private ClockInterface $clock,
private MessageBusInterface $messageBus,
private StoredObjectVersionRepository $storedObjectVersionRepository,
) {}
public function canRun(?CronJobExecution $cronJobExecution): bool
{
if (null === $cronJobExecution) {
return true;
}
return $this->clock->now() >= $cronJobExecution->getLastEnd()->add(new \DateInterval('P1D'));
}
public function getKey(): string
{
return self::KEY;
}
public function run(array $lastExecutionData): ?array
{
$deleteBeforeDate = $this->clock->now()->sub(new \DateInterval(self::KEEP_INTERVAL));
$maxDeleted = $lastExecutionData[self::LAST_DELETED_KEY] ?? 0;
foreach ($this->storedObjectVersionRepository->findIdsByVersionsOlderThanDateAndNotLastVersion($deleteBeforeDate) as $id) {
$this->messageBus->dispatch(new RemoveOldVersionMessage($id));
$maxDeleted = max($maxDeleted, $id);
}
return [self::LAST_DELETED_KEY => $maxDeleted];
}
}

View File

@ -0,0 +1,19 @@
<?php
declare(strict_types=1);
/*
* Chill is a software for social workers
*
* For the full copyright and license information, please view
* the LICENSE file that was distributed with this source code.
*/
namespace Chill\DocStoreBundle\Service\StoredObjectCleaner;
final readonly class RemoveOldVersionMessage
{
public function __construct(
public int $storedObjectVersionId
) {}
}

View File

@ -0,0 +1,54 @@
<?php
declare(strict_types=1);
/*
* Chill is a software for social workers
*
* For the full copyright and license information, please view
* the LICENSE file that was distributed with this source code.
*/
namespace Chill\DocStoreBundle\Service\StoredObjectCleaner;
use Chill\DocStoreBundle\Exception\StoredObjectManagerException;
use Chill\DocStoreBundle\Repository\StoredObjectVersionRepository;
use Chill\DocStoreBundle\Service\StoredObjectManagerInterface;
use Doctrine\ORM\EntityManagerInterface;
use Psr\Log\LoggerInterface;
use Symfony\Component\Messenger\Handler\MessageHandlerInterface;
final readonly class RemoveOldVersionMessageHandler implements MessageHandlerInterface
{
private const LOG_PREFIX = '[RemoveOldVersionMessageHandler] ';
public function __construct(
private StoredObjectVersionRepository $storedObjectVersionRepository,
private LoggerInterface $logger,
private EntityManagerInterface $entityManager,
private StoredObjectManagerInterface $storedObjectManager,
) {}
/**
* @throws StoredObjectManagerException
*/
public function __invoke(RemoveOldVersionMessage $message): void
{
$this->logger->info(self::LOG_PREFIX.'Received one message', ['storedObjectVersionId' => $message->storedObjectVersionId]);
$storedObjectVersion = $this->storedObjectVersionRepository->find($message->storedObjectVersionId);
if (null === $storedObjectVersion) {
$this->logger->error(self::LOG_PREFIX.'StoredObjectVersion not found in database', ['storedObjectVersionId' => $message->storedObjectVersionId]);
throw new \RuntimeException('StoredObjectVersion not found with id '.$message->storedObjectVersionId);
}
$this->storedObjectManager->delete($storedObjectVersion);
$this->entityManager->remove($storedObjectVersion);
$this->entityManager->flush();
// clear the entity manager for future usage
$this->entityManager->clear();
}
}

View File

@ -217,6 +217,23 @@ final class StoredObjectManager implements StoredObjectManagerInterface
return $version; return $version;
} }
/**
* @throws StoredObjectManagerException
*/
public function delete(StoredObjectVersion $storedObjectVersion): void
{
$signedUrl = $this->tempUrlGenerator->generate('DELETE', $storedObjectVersion->getFilename());
try {
$response = $this->client->request('DELETE', $signedUrl->url);
if (Response::HTTP_NO_CONTENT !== $response->getStatusCode()) {
throw StoredObjectManagerException::invalidStatusCode($response->getStatusCode());
}
} catch (TransportExceptionInterface $exception) {
throw StoredObjectManagerException::errorDuringHttpRequest($exception);
}
}
public function clearCache(): void public function clearCache(): void
{ {
$this->inMemory = []; $this->inMemory = [];

View File

@ -51,6 +51,11 @@ interface StoredObjectManagerInterface
*/ */
public function write(StoredObject $document, string $clearContent, ?string $contentType = null): StoredObjectVersion; public function write(StoredObject $document, string $clearContent, ?string $contentType = null): StoredObjectVersion;
/**
* @throws StoredObjectManagerException
*/
public function delete(StoredObjectVersion $storedObjectVersion): void;
/** /**
* return or compute the etag for the document. * return or compute the etag for the document.
* *

View File

@ -0,0 +1,43 @@
<?php
declare(strict_types=1);
/*
* Chill is a software for social workers
*
* For the full copyright and license information, please view
* the LICENSE file that was distributed with this source code.
*/
namespace Chill\DocStoreBundle\Tests\Repository;
use Chill\DocStoreBundle\Repository\StoredObjectVersionRepository;
use Doctrine\ORM\EntityManagerInterface;
use Symfony\Bundle\FrameworkBundle\Test\KernelTestCase;
/**
* @internal
*
* @coversNothing
*/
class StoredObjectVersionRepositoryTest extends KernelTestCase
{
private EntityManagerInterface $entityManager;
protected function setUp(): void
{
self::bootKernel();
$this->entityManager = self::getContainer()->get(EntityManagerInterface::class);
}
public function testFindIdsByVersionsOlderThanDateAndNotLastVersion(): void
{
$repository = new StoredObjectVersionRepository($this->entityManager);
// get old version, to get a chance to get one
$actual = $repository->findIdsByVersionsOlderThanDateAndNotLastVersion(new \DateTimeImmutable('1970-01-01'));
self::assertIsIterable($actual);
self::assertContainsOnly('int', $actual);
}
}

View File

@ -0,0 +1,104 @@
<?php
declare(strict_types=1);
/*
* Chill is a software for social workers
*
* For the full copyright and license information, please view
* the LICENSE file that was distributed with this source code.
*/
namespace Tests\Service\StoredObjectCleaner;
use Chill\DocStoreBundle\Repository\StoredObjectVersionRepository;
use Chill\DocStoreBundle\Service\StoredObjectCleaner\RemoveOldVersionCronJob;
use Chill\DocStoreBundle\Service\StoredObjectCleaner\RemoveOldVersionMessage;
use Chill\MainBundle\Entity\CronJobExecution;
use Symfony\Bundle\FrameworkBundle\Test\KernelTestCase;
use Symfony\Component\Clock\MockClock;
use Symfony\Component\Messenger\Envelope;
use Symfony\Component\Messenger\MessageBusInterface;
/**
* @internal
*
* @coversNothing
*/
class RemoveOldVersionCronJobTest extends KernelTestCase
{
/**
* @dataProvider buildTestCanRunData
*/
public function testCanRun(?CronJobExecution $cronJobExecution, bool $expected): void
{
$repository = $this->createMock(StoredObjectVersionRepository::class);
$clock = new MockClock(new \DateTimeImmutable('2024-01-01 00:00:00', new \DateTimeZone('+00:00')));
$cronJob = new RemoveOldVersionCronJob($clock, $this->buildMessageBus(), $repository);
self::assertEquals($expected, $cronJob->canRun($cronJobExecution));
}
public function testRun(): void
{
// we create a clock in the future. This led us a chance to having stored object to delete
$clock = new MockClock(new \DateTimeImmutable('2024-01-01 00:00:00', new \DateTimeZone('+00:00')));
$repository = $this->createMock(StoredObjectVersionRepository::class);
$repository->expects($this->once())
->method('findIdsByVersionsOlderThanDateAndNotLastVersion')
->with(new \DateTime('2023-10-03 00:00:00', new \DateTimeZone('+00:00')))
->willReturnCallback(function ($arg) {
yield 1;
yield 3;
yield 2;
})
;
$cronJob = new RemoveOldVersionCronJob($clock, $this->buildMessageBus(true), $repository);
$results = $cronJob->run([]);
self::assertArrayHasKey('last-deleted-stored-object-version-id', $results);
self::assertIsInt($results['last-deleted-stored-object-version-id']);
}
public static function buildTestCanRunData(): iterable
{
yield [
(new CronJobExecution('last-deleted-stored-object-version-id'))->setLastEnd(new \DateTimeImmutable('2023-12-31 00:00:00', new \DateTimeZone('+00:00'))),
true,
];
yield [
(new CronJobExecution('last-deleted-stored-object-version-id'))->setLastEnd(new \DateTimeImmutable('2023-12-30 23:59:59', new \DateTimeZone('+00:00'))),
true,
];
yield [
(new CronJobExecution('last-deleted-stored-object-version-id'))->setLastEnd(new \DateTimeImmutable('2023-12-31 00:00:01', new \DateTimeZone('+00:00'))),
false,
];
yield [
null,
true,
];
}
private function buildMessageBus(bool $expectDistpatchAtLeastOnce = false): MessageBusInterface
{
$messageBus = $this->createMock(MessageBusInterface::class);
$methodDispatch = match ($expectDistpatchAtLeastOnce) {
true => $messageBus->expects($this->atLeastOnce())->method('dispatch')->with($this->isInstanceOf(RemoveOldVersionMessage::class)),
false => $messageBus->method('dispatch'),
};
$methodDispatch->willReturnCallback(function (RemoveOldVersionMessage $message) {
return new Envelope($message);
});
return $messageBus;
}
}

View File

@ -0,0 +1,51 @@
<?php
declare(strict_types=1);
/*
* Chill is a software for social workers
*
* For the full copyright and license information, please view
* the LICENSE file that was distributed with this source code.
*/
namespace Tests\Service\StoredObjectCleaner;
use Chill\DocStoreBundle\Entity\StoredObject;
use Chill\DocStoreBundle\Repository\StoredObjectVersionRepository;
use Chill\DocStoreBundle\Service\StoredObjectCleaner\RemoveOldVersionMessage;
use Chill\DocStoreBundle\Service\StoredObjectCleaner\RemoveOldVersionMessageHandler;
use Chill\DocStoreBundle\Service\StoredObjectManagerInterface;
use Doctrine\ORM\EntityManagerInterface;
use PHPUnit\Framework\TestCase;
use Psr\Log\NullLogger;
/**
* @internal
*
* @coversNothing
*/
class RemoveOldVersionMessageHandlerTest extends TestCase
{
public function testInvoke(): void
{
$object = new StoredObject();
$version = $object->registerVersion();
$storedObjectVersionRepository = $this->createMock(StoredObjectVersionRepository::class);
$storedObjectVersionRepository->expects($this->once())->method('find')
->with($this->identicalTo(1))
->willReturn($version);
$entityManager = $this->createMock(EntityManagerInterface::class);
$entityManager->expects($this->once())->method('remove')->with($this->identicalTo($version));
$entityManager->expects($this->once())->method('flush');
$entityManager->expects($this->once())->method('clear');
$storedObjectManager = $this->createMock(StoredObjectManagerInterface::class);
$storedObjectManager->expects($this->once())->method('delete')->with($this->identicalTo($version));
$handler = new RemoveOldVersionMessageHandler($storedObjectVersionRepository, new NullLogger(), $entityManager, $storedObjectManager);
$handler(new RemoveOldVersionMessage(1));
}
}

View File

@ -198,6 +198,37 @@ final class StoredObjectManagerTest extends TestCase
self::assertSame($storedObject->getCurrentVersion(), $newVersion); self::assertSame($storedObject->getCurrentVersion(), $newVersion);
} }
public function testDelete(): void
{
$storedObject = new StoredObject();
$version = $storedObject->registerVersion(filename: 'object_name');
$httpClient = new MockHttpClient(function ($method, $url, $options) {
self::assertEquals('DELETE', $method);
self::assertEquals('https://example.com/object_name', $url);
return new MockResponse('', [
'http_code' => 204,
]);
});
$tempUrlGenerator = $this->createMock(TempUrlGeneratorInterface::class);
$tempUrlGenerator
->expects($this->once())
->method('generate')
->with($this->identicalTo('DELETE'), $this->identicalTo('object_name'))
->willReturnCallback(function (string $method, string $objectName) {
return new SignedUrl(
$method,
'https://example.com/'.$objectName,
new \DateTimeImmutable('1 hours')
);
});
$storedObjectManager = new StoredObjectManager($httpClient, $tempUrlGenerator);
$storedObjectManager->delete($version);
}
public function testWriteWithDeleteAt() public function testWriteWithDeleteAt()
{ {
$storedObject = new StoredObject(); $storedObject = new StoredObject();