115 - Read Belga API url from config and sync the feed row via migration

app/Services/Parsers/BelgaHomepageParserAdapter.php

@@ -7,15 +7,13 @@
 class BelgaHomepageParserAdapter implements HomepageParserInterface
 {
     /**
-     * Must stay in sync with config('feed.providers.belga.languages.en.url').
+     * Belga is English-only for now (#124). The unused $language argument keeps
-     * Nothing calls getHomepageUrl() at runtime — the fetcher uses $feed->url,
+     * the constructor compatible with HomepageParserFactory, which passes one
-     * seeded from that config value — so drift here would go unnoticed.
+     * positionally to every homepage parser.
+     *
+     * @phpstan-ignore constructor.unusedParameter
      */
-    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(string $language = 'en') {}
-
-    public function __construct(
-        private string $language = 'en',
-    ) {}
 
     public function canParse(string $url): bool
     {
@@ -29,7 +27,9 @@ public function extractArticleUrls(string $html): array
 
     public function getHomepageUrl(): string
     {
-        return str_replace('language=EN', 'language='.strtoupper($this->language), self::API_URL);
+        $url = config('feed.providers.belga.languages.en.url');
+
+        return is_string($url) ? $url : '';
     }
 
     public function getSourceName(): string

config/feed.php

@@ -45,7 +45,7 @@
             'type' => 'website',
             'is_active' => true,
             'languages' => [
-                'en' => ['url' => 'https://capi.belga.press/belgapress/api/public/pressreleases?offset=1&count=6&search=&start=&end=&newsroomId=70&language=EN'],
+                'en' => ['url' => 'https://capi.belga.press/belgapress/api/public/pressreleases?offset=1&count=50&search=&start=&end=&newsroomId=70&language=EN'],
             ],
             'parsers' => [
                 'homepage' => BelgaHomepageParserAdapter::class,

database/migrations/2024_01_01_000012_sync_belga_feed_to_website_discovery.php

@@ -0,0 +1,65 @@
+<?php
+
+use Illuminate\Database\Migrations\Migration;
+use Illuminate\Support\Facades\DB;
+
+/**
+ * Belga retired its public RSS feed, so the provider switched to website
+ * discovery against their public JSON API (see #115).
+ *
+ * System feeds are maintained by the platform, not the end user, so their
+ * stored url/type are updated through migrations rather than the UI.
+ *
+ * Keyed on `provider`, never on the previous url: environments are at
+ * different points in this feed's history (some on the dead RSS url, some
+ * already on an earlier API url), and a url-keyed update would silently
+ * match nothing on most of them.
+ *
+ * `feeds.url` is unique while `feeds.provider` is not, so a blanket update
+ * across several belga rows would collide. CreateFeedAction looks feeds up by
+ * url, so a config url change can leave a second belga row behind: adopt the
+ * one already on the target url and deactivate the superseded ones.
+ *
+ * Superseded rows are deactivated rather than deleted — routes.feed_id
+ * cascades on delete, so removing a feed would take its routing rules with it.
+ */
+return new class extends Migration
+{
+    public function up(): void
+    {
+        $url = config('feed.providers.belga.languages.en.url');
+
+        if (! is_string($url) || $url === '') {
+            return;
+        }
+
+        $keepId = DB::table('feeds')->where('provider', 'belga')->where('url', $url)->min('id')
+            ?? DB::table('feeds')->where('provider', 'belga')->min('id');
+
+        if ($keepId === null) {
+            return;
+        }
+
+        DB::table('feeds')
+            ->where('provider', 'belga')
+            ->where('id', '!=', $keepId)
+            ->update([
+                'is_active' => false,
+                'updated_at' => now(),
+            ]);
+
+        DB::table('feeds')
+            ->where('id', $keepId)
+            ->update([
+                'type' => 'website',
+                'url' => $url,
+                'updated_at' => now(),
+            ]);
+    }
+
+    public function down(): void
+    {
+        // No rollback: the previous RSS endpoint 404s, so restoring it would
+        // only reinstate a feed that silently yields no articles.
+    }
+};

tests/Unit/Services/Parsers/BelgaHomepageParserTest.php

@@ -4,7 +4,7 @@
 
 use App\Services\Parsers\BelgaHomepageParser;
 use App\Services\Parsers\BelgaHomepageParserAdapter;
-use PHPUnit\Framework\TestCase;
+use Tests\TestCase;
 
 class BelgaHomepageParserTest extends TestCase
 {
@@ -98,10 +98,14 @@ public function test_adapter_delegates_and_reports_source(): void
         $this->assertSame('Belga News Agency', $adapter->getSourceName());
     }
 
-    public function test_adapter_uppercases_language_in_api_url(): void
+    public function test_adapter_url_comes_from_config(): void
     {
-        $this->assertStringContainsString('language=EN', (new BelgaHomepageParserAdapter('en'))->getHomepageUrl());
+        config(['feed.providers.belga.languages.en.url' => 'https://example.test/pressreleases?count=7']);
-        $this->assertStringContainsString('language=NL', (new BelgaHomepageParserAdapter('nl'))->getHomepageUrl());
+
+        $this->assertSame(
+            'https://example.test/pressreleases?count=7',
+            (new BelgaHomepageParserAdapter('en'))->getHomepageUrl()
+        );
     }
 
     public function test_adapter_api_url_pins_load_bearing_query_params(): void
@@ -111,6 +115,7 @@ public function test_adapter_api_url_pins_load_bearing_query_params(): void
         $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);
+        $this->assertStringContainsString('count=50', $url);
+        $this->assertStringContainsString('language=EN', $url);
     }
 }