Only extract html (#26)

* Update astro

* Fix recent article statement, again

* Only extract html

package.json

@@ -4,13 +4,13 @@
 	"dependencies": {
 		"@ai-sdk/openai": "^0.0.48",
 		"@astrojs/check": "^0.9.4",
-		"@astrojs/node": "^9.0.0",
-		"@astrojs/tailwind": "^5.1.3",
+		"@astrojs/node": "9.0.1",
+		"@astrojs/tailwind": "5.1.4",
 		"@iconify-json/lucide": "^1.2.10",
 		"@mozilla/readability": "^0.5.0",
 		"@tailwindcss/typography": "^0.5.13",
 		"ai": "^3.3.9",
-		"astro": "^5.0.1",
+		"astro": "5.1.6",
 		"astro-icon": "^1.1.1",
 		"dompurify": "^3.1.6",
 		"jsdom": "^24.1.1",

src/services/cache.ts

@@ -68,7 +68,10 @@ class CacheService {
 		this.#getRecentArticlesStatement = db.query<
 			SerializedReadablePage,
 			{ limit: number }
-		>("SELECT * FROM articles ORDER BY createdAt, published DESC LIMIT :limit");
+		>(`SELECT * FROM articles
+			ORDER BY
+			CASE WHEN createdAt IS NULL THEN published ELSE createdAt END DESC
+			LIMIT :limit`);
 
 		console.log(
 			"CacheService initialized [article count: %d]",

src/services/clipper.ts

@@ -15,6 +15,10 @@ async function fetchPage(url: URL): Promise<JSDOM> {
 	if (!response.ok) {
 		throw new Error(`Failed to fetch ${url.toString()}`);
 	}
+	const contentType = response.headers.get("content-type");
+	if (!contentType || !contentType.includes("text/html")) {
+		throw new Error(`URL "${url.toString()}" is not an HTML page`);
+	}
 	const page = new JSDOM(await response.text(), { url: url.toString() });
 	// force lazy-loaded images to load
 	const LAZY_DATA_ATTRS = [