feat(search): Treffer ohne Recipe-JSON-LD rausfiltern

Wir fetchen die Trefferseite sowieso schon fürs Thumbnail — prüfen
jetzt in der gleichen HTML-Parse-Runde, ob überhaupt ein
schema.org/Recipe JSON-LD vorhanden ist. Fehlt es, wird der Treffer
aus der Liste entfernt, weil der Importer auf dieser Seite später
sowieso mit „Diese Seite enthält kein Rezept" scheitern würde.

- Migration 007: thumbnail_cache.has_recipe (NULL=unbekannt, 0=nein, 1=ja).
- Fetch-Fehler hinterlassen NULL → Treffer bleibt konservativ sichtbar.
- Neue export `hasRecipeJsonLd(html)` in json-ld-recipe.ts.
- Alle Cache-Reads/Writes nehmen den neuen Wert mit.

Tests: +2 für Filter/Failover, bestehende Thumbnail-Tests mit
Recipe-JSON-LD-Stub ergänzt, damit sie nicht selber rausgefiltert
werden.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

src/lib/server/db/migrations/007_thumbnail_cache_has_recipe.sql

@@ -0,0 +1,11 @@
+-- Erweitert thumbnail_cache um ein has_recipe-Flag. Beim Thumbnail-
+-- Enrichment checken wir, ob die Seite überhaupt ein schema.org/Recipe
+-- JSON-LD enthält — sonst kann der Importer das Rezept später sowieso
+-- nicht extrahieren, und der User sieht nur die „Diese Seite enthält
+-- kein Rezept"-Fehlermeldung.
+--
+-- NULL = unbekannt (vor dieser Migration gecached oder Fetch schlug fehl,
+--                   dann behalten wir den Treffer konservativ);
+-- 0    = gesicherter Nicht-Treffer (ausblenden);
+-- 1    = Rezept vorhanden.
+ALTER TABLE thumbnail_cache ADD COLUMN has_recipe INTEGER;

src/lib/server/parsers/json-ld-recipe.ts

@@ -106,6 +106,14 @@ function findRecipeNode(html: string): JsonLdNode | null {
   return null;
 }
 
+export function hasRecipeJsonLd(html: string): boolean {
+  try {
+    return findRecipeNode(html) !== null;
+  } catch {
+    return false;
+  }
+}
+
 export function extractRecipeFromHtml(html: string): Recipe | null {
   const node = findRecipeNode(html);
   if (!node) return null;

src/lib/server/search/searxng.ts

@@ -2,6 +2,7 @@ import type Database from 'better-sqlite3';
 import { parseHTML } from 'linkedom';
 import { listDomains, normalizeDomain } from '../domains/repository';
 import { fetchText } from '../http';
+import { hasRecipeJsonLd } from '../parsers/json-ld-recipe';
 
 export type WebHit = {
   url: string;
@@ -182,70 +183,91 @@ function extractPageImage(html: string, baseUrl: string): string | null {
 const THUMB_TTL_DAYS = Number(process.env.KOCHWAS_THUMB_TTL_DAYS ?? 30);
 const THUMB_TTL_MS = THUMB_TTL_DAYS * 24 * 60 * 60 * 1000;
 
-function readCachedThumbnail(
+type PageMeta = {
+  image: string | null;
+  hasRecipe: 0 | 1 | null;
+};
+
+function readCachedPageMeta(
   db: Database.Database,
   url: string
-): { image: string | null } | null {
+): PageMeta | null {
   const row = db
-    .prepare<[string, string], { image: string | null }>(
+    .prepare<
-      "SELECT image FROM thumbnail_cache WHERE url = ? AND expires_at > ?"
+      [string, string],
+      { image: string | null; has_recipe: 0 | 1 | null }
+    >(
+      'SELECT image, has_recipe FROM thumbnail_cache WHERE url = ? AND expires_at > ?'
     )
     .get(url, new Date().toISOString());
-  return row ?? null;
+  if (!row) return null;
+  return { image: row.image, hasRecipe: row.has_recipe };
 }
 
-function writeCachedThumbnail(
+function writeCachedPageMeta(
   db: Database.Database,
   url: string,
-  image: string | null
+  meta: PageMeta
 ): void {
   const expiresAt = new Date(Date.now() + THUMB_TTL_MS).toISOString();
   db.prepare(
-    'INSERT OR REPLACE INTO thumbnail_cache (url, image, expires_at) VALUES (?, ?, ?)'
+    'INSERT OR REPLACE INTO thumbnail_cache (url, image, expires_at, has_recipe) VALUES (?, ?, ?, ?)'
-  ).run(url, image, expiresAt);
+  ).run(url, meta.image, expiresAt, meta.hasRecipe);
 }
 
-async function enrichThumbnail(
+async function enrichPageMeta(
   db: Database.Database,
   url: string
-): Promise<string | null> {
+): Promise<PageMeta> {
-  const cached = readCachedThumbnail(db, url);
+  const cached = readCachedPageMeta(db, url);
-  if (cached) return cached.image;
+  if (cached) return cached;
-  let image: string | null = null;
+  let meta: PageMeta = { image: null, hasRecipe: null };
   try {
     const html = await fetchText(url, { timeoutMs: 4_000, maxBytes: 512 * 1024 });
-    image = extractPageImage(html, url);
+    meta = {
+      image: extractPageImage(html, url),
+      hasRecipe: hasRecipeJsonLd(html) ? 1 : 0
+    };
   } catch {
-    image = null;
+    // Fetch failed — leave hasRecipe null (unknown) so we don't permanently
+    // hide a temporary-network-error URL.
   }
-  writeCachedThumbnail(db, url, image);
+  writeCachedPageMeta(db, url, meta);
-  return image;
+  return meta;
 }
 
-async function enrichAllThumbnails(
+async function enrichAndFilterHits(
   db: Database.Database,
   hits: WebHit[]
-): Promise<void> {
+): Promise<WebHit[]> {
-  // Always fetch the page image even when SearXNG gave us a thumbnail —
+  // Always fetch the page even when SearXNG gave us a thumbnail — we need
-  // the search engine's thumbnail is typically 150-200px, while og:image
+  // the HTML anyway for the high-res og:image AND to confirm a Recipe
-  // / JSON-LD image on the page is the full-resolution recipe photo.
+  // JSON-LD actually exists. The thumbnail_cache table (default 30-day TTL)
-  // The thumbnail_cache table (default 30-day TTL) makes repeat searches instant.
+  // makes repeat searches instant.
-  if (hits.length === 0) return;
+  if (hits.length === 0) return hits;
   // Lazy cleanup of expired entries — O(log n) index scan, cheap.
   db.prepare('DELETE FROM thumbnail_cache WHERE expires_at <= ?').run(
     new Date().toISOString()
   );
+  const metas = new Map<string, PageMeta>();
   const queue = [...hits];
   const LIMIT = 6;
   const workers = Array.from({ length: Math.min(LIMIT, queue.length) }, async () => {
     while (queue.length > 0) {
       const h = queue.shift();
       if (!h) break;
-      const image = await enrichThumbnail(db, h.url);
+      metas.set(h.url, await enrichPageMeta(db, h.url));
-      if (image) h.thumbnail = image;
     }
   });
   await Promise.all(workers);
+  // Drop confirmed-non-recipe pages (hasRecipe === 0). Keep unknown (null)
+  // and confirmed recipes (1).
+  return hits
+    .filter((h) => metas.get(h.url)?.hasRecipe !== 0)
+    .map((h) => {
+      const image = metas.get(h.url)?.image;
+      return image ? { ...h, thumbnail: image } : h;
+    });
 }
 
 export async function searchWeb(
@@ -310,7 +332,7 @@ export async function searchWeb(
     if (hits.length >= limit) break;
   }
   if (opts.enrichThumbnails !== false) {
-    await enrichAllThumbnails(db, hits);
+    return await enrichAndFilterHits(db, hits);
   }
   return hits;
 }

tests/integration/searxng.test.ts

@@ -100,11 +100,62 @@ describe('searchWeb', () => {
     expect(receivedPageno).toBe(null);
   });
 
+  it('drops hits whose page lacks a Recipe JSON-LD', async () => {
+    const pageServer = createServer((req, res) => {
+      res.writeHead(200, { 'content-type': 'text/html; charset=utf-8' });
+      if (req.url === '/with-recipe') {
+        res.end(`<html><head>
+          <script type="application/ld+json">${JSON.stringify({
+            '@type': 'Recipe',
+            name: 'Pie',
+            image: 'https://cdn.example/pie.jpg'
+          })}</script>
+        </head></html>`);
+      } else {
+        // forum page: no Recipe JSON-LD
+        res.end('<html><head><title>Forum</title></head><body>Diskussion</body></html>');
+      }
+    });
+    await new Promise<void>((r) => pageServer.listen(0, '127.0.0.1', r));
+    const addr = pageServer.address() as AddressInfo;
+    try {
+      const db = openInMemoryForTest();
+      addDomain(db, '127.0.0.1');
+      respondWith([
+        { url: `http://127.0.0.1:${addr.port}/with-recipe`, title: 'Recipe', content: '' },
+        { url: `http://127.0.0.1:${addr.port}/forum-thread`, title: 'Forum', content: '' }
+      ]);
+      const hits = await searchWeb(db, 'x', { searxngUrl: baseUrl });
+      expect(hits.length).toBe(1);
+      expect(hits[0].url.endsWith('/with-recipe')).toBe(true);
+    } finally {
+      await new Promise<void>((r) => pageServer.close(() => r()));
+    }
+  });
+
+  it('keeps hit when page fetch fails (unknown recipe status)', async () => {
+    const db = openInMemoryForTest();
+    addDomain(db, '127.0.0.1');
+    // URL points to a port nobody listens on → fetch fails
+    respondWith([
+      { url: 'http://127.0.0.1:1/unreachable', title: 'Unreachable', content: '' }
+    ]);
+    const hits = await searchWeb(db, 'x', { searxngUrl: baseUrl });
+    expect(hits.length).toBe(1);
+  });
+
+  // Minimal Recipe-JSON-LD stub so enrichAndFilterHits doesn't drop test hits
+  // as non-recipe pages. Used in tests that focus on thumbnail extraction.
+  const RECIPE_LD = `<script type="application/ld+json">${JSON.stringify({
+    '@type': 'Recipe',
+    name: 'stub'
+  })}</script>`;
+
   it('enriches missing thumbnails from og:image', async () => {
     const pageServer = createServer((_req, res) => {
       res.writeHead(200, { 'content-type': 'text/html; charset=utf-8' });
       res.end(
-        '<html><head><meta property="og:image" content="https://cdn.example/foo.jpg" /></head><body></body></html>'
+        `<html><head><meta property="og:image" content="https://cdn.example/foo.jpg" />${RECIPE_LD}</head><body></body></html>`
       );
     });
     await new Promise<void>((r) => pageServer.listen(0, '127.0.0.1', r));
@@ -151,7 +202,7 @@ describe('searchWeb', () => {
     const pageServer = createServer((_req, res) => {
       res.writeHead(200, { 'content-type': 'text/html; charset=utf-8' });
       res.end(
-        '<html><body><article><img src="/uploads/dish.jpg" alt=""></article></body></html>'
+        `<html><head>${RECIPE_LD}</head><body><article><img src="/uploads/dish.jpg" alt=""></article></body></html>`
       );
     });
     await new Promise<void>((r) => pageServer.listen(0, '127.0.0.1', r));
@@ -172,7 +223,7 @@ describe('searchWeb', () => {
     const pageServer = createServer((_req, res) => {
       res.writeHead(200, { 'content-type': 'text/html; charset=utf-8' });
       res.end(
-        '<html><head><meta property="og:image" content="https://cdn.example/hq.jpg" /></head></html>'
+        `<html><head><meta property="og:image" content="https://cdn.example/hq.jpg" />${RECIPE_LD}</head></html>`
       );
     });
     await new Promise<void>((r) => pageServer.listen(0, '127.0.0.1', r));
@@ -194,7 +245,7 @@ describe('searchWeb', () => {
   it('keeps SearXNG thumbnail when page has no image', async () => {
     const pageServer = createServer((_req, res) => {
       res.writeHead(200, { 'content-type': 'text/html; charset=utf-8' });
-      res.end('<html><head></head><body>no images here</body></html>');
+      res.end(`<html><head>${RECIPE_LD}</head><body>no images here</body></html>`);
     });
     await new Promise<void>((r) => pageServer.listen(0, '127.0.0.1', r));
     const addr = pageServer.address() as AddressInfo;
@@ -217,7 +268,7 @@ describe('searchWeb', () => {
     const pageServer = createServer((_req, res) => {
       pageHits += 1;
       res.writeHead(200, { 'content-type': 'text/html; charset=utf-8' });
-      res.end('<html><head><meta property="og:image" content="https://cdn.example/c.jpg"></head></html>');
+      res.end(`<html><head><meta property="og:image" content="https://cdn.example/c.jpg">${RECIPE_LD}</head></html>`);
     });
     await new Promise<void>((r) => pageServer.listen(0, '127.0.0.1', r));
     const addr = pageServer.address() as AddressInfo;