.gitignore

@@ -24,3 +24,4 @@ yarn-error.log
 /coverage-report*
 /coverage.xml
 /.php-cs-fixer.dist.php
+/.php-cs-fixer.cache

app/Services/Parsers/BelgaArticlePageParser.php

@@ -110,12 +110,12 @@ public static function extractThumbnail(string $html): ?string
     {
         // Try OpenGraph image first
         if (preg_match('/<meta property="og:image" content="([^"]+)"/i', $html, $matches)) {
-            return $matches[1];
+            return html_entity_decode($matches[1], ENT_QUOTES, 'UTF-8');
         }
 
         // Try first image in article content
         if (preg_match('/<img[^>]+src="([^"]+)"/i', $html, $matches)) {
-            return $matches[1];
+            return html_entity_decode($matches[1], ENT_QUOTES, 'UTF-8');
         }
 
         return null;

app/Services/Parsers/BelgaHomepageParser.php

@@ -0,0 +1,28 @@
+<?php
+
+namespace App\Services\Parsers;
+
+class BelgaHomepageParser
+{
+    private const ARTICLE_URL_TEMPLATE = 'https://www.belganewsagency.eu/press-releases/%s/';
+
+    /**
+     * @return array<int, string>
+     */
+    public static function extractArticleUrls(string $json): array
+    {
+        $decoded = json_decode($json, true);
+
+        if (! is_array($decoded) || ! isset($decoded['data']) || ! is_array($decoded['data'])) {
+            return [];
+        }
+
+        return collect($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))
+            ->unique()
+            ->values()
+            ->toArray();
+    }
+}

app/Services/Parsers/BelgaHomepageParserAdapter.php

@@ -0,0 +1,39 @@
+<?php
+
+namespace App\Services\Parsers;
+
+use App\Contracts\HomepageParserInterface;
+
+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(
+        private string $language = 'en',
+    ) {}
+
+    public function canParse(string $url): bool
+    {
+        return str_contains($url, 'belganewsagency.eu') || str_contains($url, 'capi.belga.press');
+    }
+
+    public function extractArticleUrls(string $html): array
+    {
+        return BelgaHomepageParser::extractArticleUrls($html);
+    }
+
+    public function getHomepageUrl(): string
+    {
+        return str_replace('language=EN', 'language='.strtoupper($this->language), self::API_URL);
+    }
+
+    public function getSourceName(): string
+    {
+        return 'Belga News Agency';
+    }
+}

app/Services/Parsers/GuardianArticlePageParser.php

@@ -68,12 +68,12 @@ public static function extractThumbnail(string $html): ?string
     {
         // Try OpenGraph image first
         if (preg_match('/<meta property="og:image" content="([^"]+)"/i', $html, $matches)) {
-            return $matches[1];
+            return html_entity_decode($matches[1], ENT_QUOTES, 'UTF-8');
         }
 
         // Try first image in content
         if (preg_match('/<img[^>]+src="([^"]+)"/i', $html, $matches)) {
-            return $matches[1];
+            return html_entity_decode($matches[1], ENT_QUOTES, 'UTF-8');
         }
 
         return null;

app/Services/Parsers/VrtArticlePageParser.php

@@ -67,11 +67,11 @@ public static function extractFullArticle(string $html): ?string
     public static function extractThumbnail(string $html): ?string
     {
         if (preg_match('/<meta property="og:image" content="([^"]+)"/i', $html, $matches)) {
-            return $matches[1];
+            return html_entity_decode($matches[1], ENT_QUOTES, 'UTF-8');
         }
 
         if (preg_match('/<img[^>]+src="([^"]+)"/i', $html, $matches)) {
-            return $matches[1];
+            return html_entity_decode($matches[1], ENT_QUOTES, 'UTF-8');
         }
 
         return null;

config/app.php

@@ -13,7 +13,7 @@
     |
     */
 
-    'name' => env('APP_NAME', 'Laravel'),
+    'name' => 'Fedi Feed Router',
 
     /*
     |--------------------------------------------------------------------------

config/feed.php

@@ -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,
             ],

pint.json

@@ -0,0 +1,3 @@
+{
+    "preset": "laravel"
+}

resources/views/layouts/app.blade.php

@@ -5,7 +5,7 @@
         <meta name="viewport" content="width=device-width, initial-scale=1">
         <meta name="csrf-token" content="{{ csrf_token() }}">
 
-        <title>{{ config('app.name', 'FFR') }}</title>
+        <title>{{ config('app.name') }}</title>
 
         <!-- Fonts -->
         <link rel="preconnect" href="https://fonts.bunny.net">

resources/views/layouts/guest.blade.php

@@ -5,7 +5,7 @@
         <meta name="viewport" content="width=device-width, initial-scale=1">
         <meta name="csrf-token" content="{{ csrf_token() }}">
 
-        <title>{{ config('app.name', 'FFR') }}</title>
+        <title>{{ config('app.name') }}</title>
 
         <!-- Fonts -->
         <link rel="preconnect" href="https://fonts.bunny.net">

resources/views/layouts/onboarding.blade.php

@@ -5,7 +5,7 @@
         <meta name="viewport" content="width=device-width, initial-scale=1">
         <meta name="csrf-token" content="{{ csrf_token() }}">
 
-        <title>{{ config('app.name', 'FFR') }} - Setup</title>
+        <title>{{ config('app.name') }} - Setup</title>
 
         <!-- Fonts -->
         <link rel="preconnect" href="https://fonts.bunny.net">

tests/Feature/Http/Controllers/Api/V1/FeedsControllerTest.php

@@ -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',
         ]);
     }
 

tests/Unit/Actions/CreateFeedActionTest.php

@@ -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);
     }

tests/Unit/Services/ArticleFetcherBelgaTest.php

@@ -0,0 +1,100 @@
+<?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
+    {
+        $contents = file_get_contents(__DIR__.'/../../Fixtures/belga-pressreleases.json');
+
+        $this->assertNotFalse($contents, 'Belga fixture could not be read.');
+
+        return $contents;
+    }
+
+    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();
+    }
+}

tests/Unit/Services/Parsers/BelgaArticlePageParserTest.php

@@ -242,6 +242,36 @@ public function test_extract_thumbnail_prefers_og_image(): void
         $this->assertEquals('https://example.com/og-image.jpg', $thumbnail);
     }
 
+    public function test_extract_thumbnail_decodes_html_entities_in_og_image(): void
+    {
+        // Real Belga CDN URLs carry query params, so the & in the raw HTML
+        // must be decoded or the URL Lemmy receives is malformed. See #117.
+        $html = '<html><head><meta property="og:image" content="https://cdn.example.com/img:12345:full?v=6a6c828f&amp;m=cdodnjha"/></head></html>';
+
+        $thumbnail = BelgaArticlePageParser::extractThumbnail($html);
+
+        $this->assertEquals('https://cdn.example.com/img:12345:full?v=6a6c828f&m=cdodnjha', $thumbnail);
+        $this->assertStringNotContainsString('&amp;', (string) $thumbnail);
+    }
+
+    public function test_extract_thumbnail_decodes_html_entities_in_img_tag(): void
+    {
+        $html = '<html><body><img src="https://cdn.example.com/pic.png?a=1&amp;b=2" alt="test"/></body></html>';
+
+        $thumbnail = BelgaArticlePageParser::extractThumbnail($html);
+
+        $this->assertEquals('https://cdn.example.com/pic.png?a=1&b=2', $thumbnail);
+    }
+
+    public function test_extract_data_returns_decoded_thumbnail(): void
+    {
+        $html = '<html><head><meta property="og:image" content="https://cdn.example.com/x.jpg?v=1&amp;m=2"/></head></html>';
+
+        $data = BelgaArticlePageParser::extractData($html);
+
+        $this->assertEquals('https://cdn.example.com/x.jpg?v=1&m=2', $data['thumbnail']);
+    }
+
     public function test_extract_thumbnail_returns_null_when_not_found(): void
     {
         $html = '<html><body><div>No images here</div></body></html>';

tests/Unit/Services/Parsers/BelgaHomepageParserTest.php

@@ -0,0 +1,116 @@
+<?php
+
+namespace Tests\Unit\Services\Parsers;
+
+use App\Services\Parsers\BelgaHomepageParser;
+use App\Services\Parsers\BelgaHomepageParserAdapter;
+use PHPUnit\Framework\TestCase;
+
+class BelgaHomepageParserTest extends TestCase
+{
+    private function fixture(): string
+    {
+        $contents = file_get_contents(__DIR__.'/../../../Fixtures/belga-pressreleases.json');
+
+        $this->assertNotFalse($contents, 'Belga fixture could not be read.');
+
+        return $contents;
+    }
+
+    /**
+     * @param  array<int, array<string, mixed>>  $data
+     */
+    private function jsonWith(array $data): string
+    {
+        return (string) json_encode(['data' => $data]);
+    }
+
+    public function test_extracts_article_urls_from_api_json(): void
+    {
+        $urls = BelgaHomepageParser::extractArticleUrls($this->fixture());
+
+        $this->assertCount(6, $urls);
+        $this->assertSame('https://www.belganewsagency.eu/press-releases/35285/', $urls[0]);
+        $this->assertContains('https://www.belganewsagency.eu/press-releases/35269/', $urls);
+    }
+
+    public function test_deduplicates_urls(): void
+    {
+        $json = $this->jsonWith([
+            ['id' => 100],
+            ['id' => 100],
+            ['id' => '100'],
+            ['id' => 101],
+        ]);
+
+        $urls = BelgaHomepageParser::extractArticleUrls($json);
+
+        $this->assertSame([
+            'https://www.belganewsagency.eu/press-releases/100/',
+            'https://www.belganewsagency.eu/press-releases/101/',
+        ], $urls);
+    }
+
+    public function test_returns_empty_array_for_malformed_json(): void
+    {
+        $this->assertSame([], BelgaHomepageParser::extractArticleUrls('<html>404 not found</html>'));
+        $this->assertSame([], BelgaHomepageParser::extractArticleUrls('{"data": ['));
+        $this->assertSame([], BelgaHomepageParser::extractArticleUrls(''));
+    }
+
+    public function test_returns_empty_array_when_data_key_missing_or_empty(): void
+    {
+        $this->assertSame([], BelgaHomepageParser::extractArticleUrls('{}'));
+        $this->assertSame([], BelgaHomepageParser::extractArticleUrls('{"data": []}'));
+        $this->assertSame([], BelgaHomepageParser::extractArticleUrls('{"data": "not-an-array"}'));
+    }
+
+    public function test_skips_entries_without_usable_id(): void
+    {
+        $json = $this->jsonWith([
+            ['id' => 200],
+            ['id' => null],
+            ['publishDate' => '2026-08-01T06:24:19'],
+            ['id' => 'not-numeric'],
+            ['id' => 201],
+        ]);
+
+        $urls = BelgaHomepageParser::extractArticleUrls($json);
+
+        $this->assertSame([
+            'https://www.belganewsagency.eu/press-releases/200/',
+            'https://www.belganewsagency.eu/press-releases/201/',
+        ], $urls);
+    }
+
+    public function test_adapter_delegates_and_reports_source(): void
+    {
+        $adapter = new BelgaHomepageParserAdapter('en');
+
+        $this->assertSame(
+            BelgaHomepageParser::extractArticleUrls($this->fixture()),
+            $adapter->extractArticleUrls($this->fixture())
+        );
+
+        $this->assertTrue($adapter->canParse('https://www.belganewsagency.eu/press-releases/35285/'));
+        $this->assertTrue($adapter->canParse('https://capi.belga.press/belgapress/api/public/pressreleases'));
+        $this->assertFalse($adapter->canParse('https://www.vrt.be/vrtnws/en/'));
+        $this->assertSame('Belga News Agency', $adapter->getSourceName());
+    }
+
+    public function test_adapter_uppercases_language_in_api_url(): void
+    {
+        $this->assertStringContainsString('language=EN', (new BelgaHomepageParserAdapter('en'))->getHomepageUrl());
+        $this->assertStringContainsString('language=NL', (new BelgaHomepageParserAdapter('nl'))->getHomepageUrl());
+    }
+
+    public function test_adapter_api_url_pins_load_bearing_query_params(): void
+    {
+        $url = (new BelgaHomepageParserAdapter('en'))->getHomepageUrl();
+
+        $this->assertStringStartsWith('https://capi.belga.press/belgapress/api/public/pressreleases?', $url);
+        $this->assertStringContainsString('newsroomId=70', $url);
+        $this->assertStringContainsString('offset=1', $url);
+        $this->assertStringContainsString('count=6', $url);
+    }
+}

tests/Unit/Services/Parsers/GuardianArticlePageParserTest.php

@@ -282,4 +282,24 @@ public function test_extract_full_article_with_realistic_guardian_html(): void
         $this->assertStringContainsString("\n\n", $fullArticle);
         $this->assertStringNotContainsString('<strong>', $fullArticle);
     }
+
+    public function test_extract_thumbnail_decodes_html_entities(): void
+    {
+        // Entity-encoded ampersands make the URL unusable downstream. See #117.
+        $html = '<html><head><meta property="og:image" content="https://i.guim.co.uk/img/media/abc.jpg?width=1200&amp;quality=85"/></head></html>';
+
+        $thumbnail = GuardianArticlePageParser::extractThumbnail($html);
+
+        $this->assertEquals('https://i.guim.co.uk/img/media/abc.jpg?width=1200&quality=85', $thumbnail);
+        $this->assertStringNotContainsString('&amp;', (string) $thumbnail);
+    }
+
+    public function test_extract_thumbnail_decodes_html_entities_in_img_tag(): void
+    {
+        $html = '<html><body><img src="https://i.guim.co.uk/img/pic.png?a=1&amp;b=2" alt="test"/></body></html>';
+
+        $thumbnail = GuardianArticlePageParser::extractThumbnail($html);
+
+        $this->assertEquals('https://i.guim.co.uk/img/pic.png?a=1&b=2', $thumbnail);
+    }
 }

tests/Unit/Services/Parsers/VrtArticlePageParserTest.php

@@ -0,0 +1,75 @@
+<?php
+
+namespace Tests\Unit\Services\Parsers;
+
+use App\Services\Parsers\VrtArticlePageParser;
+use Tests\TestCase;
+
+class VrtArticlePageParserTest extends TestCase
+{
+    public function test_extract_thumbnail_from_og_image(): void
+    {
+        $html = '<html><head><meta property="og:image" content="https://images.vrt.be/image.jpg"/></head></html>';
+
+        $this->assertEquals(
+            'https://images.vrt.be/image.jpg',
+            VrtArticlePageParser::extractThumbnail($html)
+        );
+    }
+
+    public function test_extract_thumbnail_falls_back_to_img_tag(): void
+    {
+        $html = '<html><body><img src="https://images.vrt.be/body.png" alt="test"/></body></html>';
+
+        $this->assertEquals(
+            'https://images.vrt.be/body.png',
+            VrtArticlePageParser::extractThumbnail($html)
+        );
+    }
+
+    public function test_extract_thumbnail_prefers_og_image(): void
+    {
+        $html = '<html><head><meta property="og:image" content="https://images.vrt.be/og.jpg"/></head>'
+            .'<body><img src="https://images.vrt.be/body.png" alt="test"/></body></html>';
+
+        $this->assertEquals(
+            'https://images.vrt.be/og.jpg',
+            VrtArticlePageParser::extractThumbnail($html)
+        );
+    }
+
+    public function test_extract_thumbnail_returns_null_when_not_found(): void
+    {
+        $this->assertNull(VrtArticlePageParser::extractThumbnail('<html><body>No images</body></html>'));
+    }
+
+    public function test_extract_thumbnail_decodes_html_entities(): void
+    {
+        // Entity-encoded ampersands make the URL unusable downstream. See #117.
+        $html = '<html><head><meta property="og:image" content="https://images.vrt.be/orig.jpg?crop=1&amp;w=1200"/></head></html>';
+
+        $thumbnail = VrtArticlePageParser::extractThumbnail($html);
+
+        $this->assertEquals('https://images.vrt.be/orig.jpg?crop=1&w=1200', $thumbnail);
+        $this->assertStringNotContainsString('&amp;', (string) $thumbnail);
+    }
+
+    public function test_extract_thumbnail_decodes_html_entities_in_img_tag(): void
+    {
+        $html = '<html><body><img src="https://images.vrt.be/pic.png?a=1&amp;b=2" alt="test"/></body></html>';
+
+        $this->assertEquals(
+            'https://images.vrt.be/pic.png?a=1&b=2',
+            VrtArticlePageParser::extractThumbnail($html)
+        );
+    }
+
+    public function test_extract_data_returns_decoded_thumbnail(): void
+    {
+        $html = '<html><head><meta property="og:image" content="https://images.vrt.be/x.jpg?v=1&amp;m=2"/></head></html>';
+
+        $data = VrtArticlePageParser::extractData($html);
+
+        $this->assertEquals('https://images.vrt.be/x.jpg?v=1&m=2', $data['thumbnail']);
+    }
+}