Crawler-Parser: og/product:price-Meta (toom 799 €, Hornbach) + Hornbach.ch als erste Ausland-Live-Quelle (cc)

livecrawler/app.py

@@ -36,6 +36,22 @@ def _num(s):
     g=m.groups()
     return float(g[0]) + (float(g[1])/100 if len(g)>1 and g[1] else 0)
 
+def clean_price(v):
+    v=(v or "").strip()
+    if not v: return None
+    if "," in v: v=v.replace(".","").replace(",",".")
+    m=re.search(r"\d+(?:\.\d+)?", v)
+    return float(m.group()) if m else None
+
+def price_from_meta(h):
+    for pat in ("product:price:amount","og:price:amount"):
+        m=re.search(r'<meta[^>]+(?:property|name)=["\']'+pat+r'["\'][^>]+content=["\']([^"\']+)', h, re.I) or \
+          re.search(r'<meta[^>]+content=["\']([^"\']+)["\'][^>]+(?:property|name)=["\']'+pat+r'["\']', h, re.I)
+        if m:
+            p=clean_price(m.group(1))
+            if p: return p
+    return None
+
 def parse(htmltext):
     # 1) JSON-LD Offer
     price=None; avail=None
@@ -56,6 +72,8 @@ def parse(htmltext):
     low=htmltext.lower()
     in_store = ("vorrätig" in low or "verfügbar" in low) and not ("nicht vorrätig" in low or "nicht verfügbar" in low)
     online = ("in den warenkorb" in low) or ("lieferung" in low and "nicht möglich" not in low) or (avail and "instock" in avail.lower())
+    if price is None:
+        price=price_from_meta(htmltext)
     if price is None:
         mp=_num(htmltext[:200000])
         price=mp
@@ -75,7 +93,7 @@ async def fetch_one(client, src):
         if r.status_code!=200:
             return {**src,"price":None,"availability":"unknown","status":r.status_code}
         price, avail = parse(r.text)
-        return {"chain":src["chain"],"kind":src["kind"],"productId":src["productId"],"name":src["name"],
+        return {"chain":src["chain"],"kind":src["kind"],"cc":src.get("cc","DE"),"productId":src["productId"],"name":src["name"],
                 "url":src["url"],"price":price,"availability":avail}
     except Exception as e:
         log.warning("fetch fail %s: %s", src["url"], e)