fix: Rezeptverlauf-Bilder lokal speichern und Windows-Subprocess-Fix

- Bilder beim Extrahieren lokal cachen (data/images/{id}.ext) und über GET /api/history/{id}/image ausliefern; Löschen bei Verlaufseintrag-Entfernung
- history.db in ./data/ Volume verschoben, damit sie Container-Rebuilds überlebt
- yt-dlp Subprocess-Aufrufe auf Windows gefixt: run_in_executor statt asyncio.create_subprocess_exec (SelectorEventLoop-Kompatibilität)
- ./data:/app/data Docker-Volume für persistente Datenhaltung hinzugefügt

app/api/extract.py

@@ -1,5 +1,6 @@
 """GET /api/extract/stream – SSE endpoint."""
 from __future__ import annotations
+import asyncio
 import json
 import logging
 from typing import AsyncGenerator
@@ -13,12 +14,22 @@ from app.extractors.video import extract_video
 from app.models import Recipe, StatusEvent, ErrorEvent, ExtractResponse
 from app.services.nutrition import enrich_nutrition
 from app.services.tandoor_service import recipe_to_tandoor
-from app.services.history_service import save_recipe as save_to_history
+from app.services.history_service import save_recipe as save_to_history, update_image_url
+from app.services import image_service
 
 log = logging.getLogger(__name__)
 router = APIRouter()
 
 
+async def _cache_image_background(history_id: int, remote_url: str) -> None:
+    local_path = await image_service.download_and_cache(history_id, remote_url)
+    if local_path:
+        try:
+            update_image_url(history_id, local_path)
+        except Exception as e:
+            log.warning("Failed to update image_url in DB: %s", e)
+
+
 def _sse(event: str, data: str) -> str:
     return f"event: {event}\ndata: {data}\n\n"
 
@@ -51,6 +62,8 @@ async def _stream(url: str) -> AsyncGenerator[str, None]:
             history_id = save_to_history(base.model_dump())
         except Exception as db_err:
             log.warning("History save failed: %s", db_err)
+        if history_id and recipe.image_url:
+            asyncio.create_task(_cache_image_background(history_id, recipe.image_url))
         response = ExtractResponse(recipe=recipe, tandoor_json=tandoor_json, history_id=history_id)
         yield _sse("result", response.model_dump_json())