feat(scraping): implémenter le health check de tous les scrapers

- Commande CheckAllScrapersHealth + handler avec ports dédiés
- Value Object ContentSourceHealthCheckData
- Resource API Platform et State Processor
- Adapters InMemory et tests unitaires + fonctionnels
This commit is contained in:
ext.jeremy.guillot@maxicoffee.domains
2026-03-16 00:08:57 +01:00
parent 795cbeccc3
commit 01474c264b
11 changed files with 453 additions and 0 deletions

View File

@@ -0,0 +1,7 @@
<?php
namespace App\Domain\Scraping\Application\Command;
readonly class CheckAllScrapersHealth
{
}

View File

@@ -0,0 +1,64 @@
<?php
namespace App\Domain\Scraping\Application\CommandHandler;
use App\Domain\Scraping\Application\Command\CheckAllScrapersHealth;
use App\Domain\Scraping\Application\Command\TestScraperConfiguration;
use App\Domain\Scraping\Domain\Contract\Repository\ContentSourceForHealthCheckInterface;
use App\Domain\Scraping\Domain\Contract\Repository\ContentSourceHealthRepositoryInterface;
use Psr\Log\LoggerInterface;
readonly class CheckAllScrapersHealthHandler
{
public function __construct(
private ContentSourceForHealthCheckInterface $contentSourceForHealthCheckRepo,
private ContentSourceHealthRepositoryInterface $contentSourceHealthRepo,
private TestScraperConfigurationHandler $testScraperConfigurationHandler,
private LoggerInterface $logger,
) {
}
public function handle(CheckAllScrapersHealth $command): void
{
$sources = $this->contentSourceForHealthCheckRepo->getAll();
foreach ($sources as $source) {
if ($source->testSlug === null || $source->testChapterNumber === null) {
$this->logger->warning('ContentSource {id} has no test config, skipping health check.', ['id' => $source->id]);
continue;
}
try {
$this->contentSourceHealthRepo->markAsTesting($source->id);
$testUrl = str_replace(
['{slug}', '{chapterNumber}'],
[$source->testSlug, $source->testChapterNumber],
$source->chapterUrlFormat
);
$testCommand = new TestScraperConfiguration(
baseUrl: $source->baseUrl,
chapterUrlFormat: $source->chapterUrlFormat,
scrapingType: $source->scrapingType,
testUrl: $testUrl,
mangaSlug: $source->testSlug,
chapterNumber: $source->testChapterNumber,
imageSelector: $source->imageSelector,
nextPageSelector: $source->nextPageSelector,
chapterSelector: $source->chapterSelector,
);
$response = $this->testScraperConfigurationHandler->handle($testCommand);
if ($response->success) {
$this->contentSourceHealthRepo->markAsHealthy($source->id, new \DateTimeImmutable());
} else {
$firstError = $response->errors[0]['message'] ?? 'Erreur inconnue';
$this->contentSourceHealthRepo->markAsUnhealthy($source->id, new \DateTimeImmutable(), $firstError);
}
} catch (\Exception $e) {
$this->contentSourceHealthRepo->markAsUnhealthy($source->id, new \DateTimeImmutable(), $e->getMessage());
}
}
}
}

View File

@@ -0,0 +1,11 @@
<?php
namespace App\Domain\Scraping\Domain\Contract\Repository;
use App\Domain\Scraping\Domain\Model\ValueObject\ContentSourceHealthCheckData;
interface ContentSourceForHealthCheckInterface
{
/** @return ContentSourceHealthCheckData[] */
public function getAll(): array;
}

View File

@@ -0,0 +1,12 @@
<?php
namespace App\Domain\Scraping\Domain\Contract\Repository;
interface ContentSourceHealthRepositoryInterface
{
public function markAsTesting(int $sourceId): void;
public function markAsHealthy(int $sourceId, \DateTimeImmutable $testedAt): void;
public function markAsUnhealthy(int $sourceId, \DateTimeImmutable $testedAt, string $error): void;
}

View File

@@ -0,0 +1,19 @@
<?php
namespace App\Domain\Scraping\Domain\Model\ValueObject;
readonly class ContentSourceHealthCheckData
{
public function __construct(
public int $id,
public string $baseUrl,
public string $chapterUrlFormat,
public string $scrapingType,
public ?string $imageSelector,
public ?string $nextPageSelector,
public ?string $chapterSelector,
public ?string $testSlug,
public ?float $testChapterNumber,
) {
}
}

View File

@@ -0,0 +1,23 @@
<?php
namespace App\Domain\Scraping\Infrastructure\ApiPlatform\Resource;
use ApiPlatform\Metadata\ApiResource;
use ApiPlatform\Metadata\Post;
use App\Domain\Scraping\Infrastructure\ApiPlatform\State\Processor\CheckAllScrapersHealthStateProcessor;
#[ApiResource(
shortName: 'Scraping',
operations: [
new Post(
uriTemplate: '/scraping/check-all-health',
processor: CheckAllScrapersHealthStateProcessor::class,
output: false,
status: 202,
description: 'Déclenche le test de santé de tous les scrapers configurés avec testSlug',
),
]
)]
class CheckAllScrapersHealthResource
{
}

View File

@@ -0,0 +1,23 @@
<?php
namespace App\Domain\Scraping\Infrastructure\ApiPlatform\State\Processor;
use ApiPlatform\Metadata\Operation;
use ApiPlatform\State\ProcessorInterface;
use App\Domain\Scraping\Application\Command\CheckAllScrapersHealth;
use App\Domain\Scraping\Application\CommandHandler\CheckAllScrapersHealthHandler;
readonly class CheckAllScrapersHealthStateProcessor implements ProcessorInterface
{
public function __construct(
private CheckAllScrapersHealthHandler $handler,
) {
}
public function process(mixed $data, Operation $operation, array $uriVariables = [], array $context = []): null
{
$this->handler->handle(new CheckAllScrapersHealth());
return null;
}
}