lvl0/fedi-feed-router
1
0
Fork 0
Code Issues 27 Pull requests Projects Releases 13 Packages Wiki Activity Actions

115 - Switch Belga provider to website discovery via JSON API

Browse source
This commit is contained in:
myrmidex 2026-08-01 16:05:45 +02:00
parent 2a4d4ad84d
commit 299ec3151a
7 changed files with 118 additions and 12 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
app/Services/Parsers/BelgaHomepageParser.php
View file

@ -16,8 +16,8 @@ public static function extractArticleUrls(string $json): array
return collect($decoded['data'])
->pluck('id')
->filter(fn($id) => is_int($id) || (is_string($id) && ctype_digit($id)))
->map(fn($id) => sprintf(self::ARTICLE_URL_TEMPLATE, $id))
->filter(fn ($id) => is_int($id) || (is_string($id) && ctype_digit($id)))
->map(fn ($id) => sprintf(self::ARTICLE_URL_TEMPLATE, $id))
->unique()
->values()
->toArray();

7
app/Services/Parsers/BelgaHomepageParserAdapter.php
View file

@ -6,6 +6,11 @@
class BelgaHomepageParserAdapter implements HomepageParserInterface
{
/**
* Must stay in sync with config('feed.providers.belga.languages.en.url').
* Nothing calls getHomepageUrl() at runtime — the fetcher uses $feed->url,
* seeded from that config value — so drift here would go unnoticed.
*/
private const API_URL = 'https://capi.belga.press/belgapress/api/public/pressreleases?offset=1&count=6&search=&start=&end=&newsroomId=70&language=EN';
public function __construct(
@ -24,7 +29,7 @@ public function extractArticleUrls(string $html): array
public function getHomepageUrl(): string
{
return str_replace('language=EN', 'language=' . strtoupper($this->language), self::API_URL);
return str_replace('language=EN', 'language='.strtoupper($this->language), self::API_URL);
}
public function getSourceName(): string

6
config/feed.php
View file

@ -2,6 +2,7 @@
use App\Services\Parsers\BelgaArticlePageParser;
use App\Services\Parsers\BelgaArticleParser;
use App\Services\Parsers\BelgaHomepageParserAdapter;
use App\Services\Parsers\GuardianArticlePageParser;
use App\Services\Parsers\GuardianArticleParser;
use App\Services\Parsers\VrtArticlePageParser;
@ -41,12 +42,13 @@
'code' => 'belga',
'name' => 'Belga News Agency',
'description' => 'Belgian national news agency',
'type' => 'rss',
'type' => 'website',
'is_active' => true,
'languages' => [
'en' => ['url' => 'https://www.belganewsagency.eu/feed'],
'en' => ['url' => 'https://capi.belga.press/belgapress/api/public/pressreleases?offset=1&count=6&search=&start=&end=&newsroomId=70&language=EN'],
],
'parsers' => [
'homepage' => BelgaHomepageParserAdapter::class,
'article' => BelgaArticleParser::class,
'article_page' => BelgaArticlePageParser::class,
],

10
tests/Feature/Http/Controllers/Api/V1/FeedsControllerTest.php
View file

@ -92,22 +92,24 @@ public function test_store_creates_belga_feed_successfully(): void
$response = $this->postJson('/api/v1/feeds', $feedData);
$expectedUrl = config('feed.providers.belga.languages.en.url');
$response->assertStatus(201)
->assertJson([
'success' => true,
'message' => 'Feed created successfully!',
'data' => [
'name' => 'Belga Test Feed',
'url' => 'https://www.belganewsagency.eu/feed',
'type' => 'rss',
'url' => $expectedUrl,
'type' => 'website',
'is_active' => true,
],
]);
$this->assertDatabaseHas('feeds', [
'name' => 'Belga Test Feed',
'url' => 'https://www.belganewsagency.eu/feed',
'type' => 'rss',
'url' => $expectedUrl,
'type' => 'website',
]);
}

5
tests/Unit/Actions/CreateFeedActionTest.php
View file

@ -42,8 +42,9 @@ public function test_creates_belga_feed_with_correct_url(): void
$feed = $this->action->execute('Belga News', 'belga', $language->id);
$this->assertEquals('https://www.belganewsagency.eu/feed', $feed->url);
$this->assertEquals('rss', $feed->type);
$this->assertStringStartsWith('https://capi.belga.press/belgapress/api/public/pressreleases?', $feed->url);
$this->assertStringContainsString('newsroomId=70', $feed->url);
$this->assertEquals('website', $feed->type);
$this->assertEquals('belga', $feed->provider);
$this->assertNull($feed->description);
}

96
tests/Unit/Services/ArticleFetcherBelgaTest.php Normal file
View file

@ -0,0 +1,96 @@
<?php
namespace Tests\Unit\Services;
use App\Models\Feed;
use App\Models\Language;
use Illuminate\Foundation\Testing\RefreshDatabase;
use Illuminate\Support\Facades\Http;
use Mockery;
use Tests\TestCase;
use Tests\Traits\CreatesArticleFetcher;
/**
* Belga discovery runs through the website path against a JSON API rather than
* a rendered homepage. This lives in its own file because ArticleFetcherTest
* registers a catch-all Http::fake in setUp() that a per-test fake cannot
* override.
*/
class ArticleFetcherBelgaTest extends TestCase
{
use CreatesArticleFetcher, RefreshDatabase;
private function apiResponse(): string
{
return file_get_contents(__DIR__.'/../../Fixtures/belga-pressreleases.json');
}
private function belgaFeed(): Feed
{
$language = Language::factory()->create(['short_code' => 'en']);
return Feed::factory()->create([
'type' => 'website',
'provider' => 'belga',
'language_id' => $language->id,
'url' => config('feed.providers.belga.languages.en.url'),
]);
}
public function test_creates_articles_from_belga_api_response(): void
{
Http::fake(['*' => Http::response($this->apiResponse(), 200)]);
$result = $this->createArticleFetcher()->getArticlesFromFeed($this->belgaFeed());
$this->assertCount(6, $result);
$this->assertDatabaseHas('articles', [
'url' => 'https://www.belganewsagency.eu/press-releases/35285/',
]);
$this->assertDatabaseHas('articles', [
'url' => 'https://www.belganewsagency.eu/press-releases/35269/',
]);
}
public function test_associates_created_articles_with_the_feed(): void
{
Http::fake(['*' => Http::response($this->apiResponse(), 200)]);
$feed = $this->belgaFeed();
$this->createArticleFetcher()->getArticlesFromFeed($feed);
$this->assertDatabaseHas('articles', [
'url' => 'https://www.belganewsagency.eu/press-releases/35285/',
'feed_id' => $feed->id,
]);
}
public function test_returns_empty_collection_when_api_returns_no_articles(): void
{
// A well-formed envelope with nothing in it is legitimate (quiet day),
// and must be a no-op rather than an error.
Http::fake(['*' => Http::response('{"data":[],"_meta":{"total":0}}', 200)]);
$result = $this->createArticleFetcher()->getArticlesFromFeed($this->belgaFeed());
$this->assertEmpty($result);
$this->assertDatabaseCount('articles', 0);
}
public function test_returns_empty_collection_when_api_returns_an_error_page(): void
{
// The failure mode that caused #115: a 404 HTML body reaching the parser.
Http::fake(['*' => Http::response('<html><body>404 Not Found</body></html>', 200)]);
$result = $this->createArticleFetcher()->getArticlesFromFeed($this->belgaFeed());
$this->assertEmpty($result);
}
protected function tearDown(): void
{
Mockery::close();
parent::tearDown();
}
}

2
tests/Unit/Services/Parsers/BelgaHomepageParserTest.php
View file

@ -10,7 +10,7 @@ class BelgaHomepageParserTest extends TestCase
{
private function fixture(): string
{
return file_get_contents(__DIR__ . '/../../../Fixtures/belga-pressreleases.json');
return file_get_contents(__DIR__.'/../../../Fixtures/belga-pressreleases.json');
}
public function test_extracts_article_urls_from_api_json(): void