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/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 }>(
-      "SELECT image FROM thumbnail_cache WHERE url = ? AND expires_at > ?"
+    .prepare<
+      [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 (?, ?, ?)'
-  ).run(url, image, expiresAt);
+    'INSERT OR REPLACE INTO thumbnail_cache (url, image, expires_at, has_recipe) VALUES (?, ?, ?, ?)'
+  ).run(url, meta.image, expiresAt, meta.hasRecipe);
 }
 
-async function enrichThumbnail(
+async function enrichPageMeta(
   db: Database.Database,
   url: string
-): Promise<string | null> {
-  const cached = readCachedThumbnail(db, url);
-  if (cached) return cached.image;
-  let image: string | null = null;
+): Promise<PageMeta> {
+  const cached = readCachedPageMeta(db, url);
+  if (cached) return cached;
+  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);
-  return image;
+  writeCachedPageMeta(db, url, meta);
+  return meta;
 }
 
-async function enrichAllThumbnails(
+async function enrichAndFilterHits(
   db: Database.Database,
   hits: WebHit[]
-): Promise<void> {
-  // Always fetch the page image even when SearXNG gave us a thumbnail —
-  // the search engine's thumbnail is typically 150-200px, while og:image
-  // / JSON-LD image on the page is the full-resolution recipe photo.
-  // The thumbnail_cache table (default 30-day TTL) makes repeat searches instant.
-  if (hits.length === 0) return;
+): Promise<WebHit[]> {
+  // Always fetch the page even when SearXNG gave us a thumbnail — we need
+  // the HTML anyway for the high-res og:image AND to confirm a Recipe
+  // JSON-LD actually exists. The thumbnail_cache table (default 30-day TTL)
+  // makes repeat searches instant.
+  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);
-      if (image) h.thumbnail = image;
+      metas.set(h.url, await enrichPageMeta(db, h.url));
     }
   });
   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;
 }