Hermes Agent
199 lines
8.0 KiB
Python
199 lines
8.0 KiB
Python
#!/usr/bin/env python3
|
|
"""Spiele-Offensive Top-30 mit korrekten Bildern aus den Listing-Seiten."""
|
|
import re, sys, urllib.request
|
|
from datetime import datetime
|
|
from html import unescape
|
|
|
|
UA = "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36"
|
|
NOW = datetime.now().strftime("%d.%m.%Y %H:%M")
|
|
PID = "505"
|
|
OUT = "/home/hermes/workspace/spiele_offensive_top30.html"
|
|
|
|
def fetch(url, timeout=15):
|
|
req = urllib.request.Request(url, headers={"User-Agent": UA})
|
|
with urllib.request.urlopen(req, timeout=timeout) as r:
|
|
raw = r.read()
|
|
try: return raw.decode("utf-8")
|
|
except: return raw.decode("iso-8859-1", errors="replace")
|
|
|
|
def fix(s):
|
|
for a,b in [("ü","ü"),("ö","ö"),("ä","ä"),("ß","ß"),("Ãœ","Ü"),("Ä","Ä"),("Ö","Ö"),("é","é")]:
|
|
s = s.replace(a,b)
|
|
return unescape(s.replace("&","&").replace(""",'"')).strip()
|
|
|
|
# ── Build aid→image_filename map from Sonderangebote listing pages ──
|
|
# The HTML pattern: <a onclick="registerItemSelection(dataLayer, AID, …)" …><center><img src="https://mediaservice…/size=100x100/FILENAME" …></center></a>
|
|
print("Scraping Sonderangebote...", file=sys.stderr)
|
|
aid_img = {}
|
|
|
|
for page in range(10):
|
|
try:
|
|
h = fetch(f"https://www.spiele-offensive.de/Kat/Sonderangebote.html?sse={page}", 15)
|
|
except Exception as e:
|
|
print(f" p{page}: {e}", file=sys.stderr); continue
|
|
|
|
# Clean regex: capture aid from onclick + filename from img src
|
|
for m in re.finditer(
|
|
r'registerItemSelection\(dataLayer,\s*(\d+)\s*,.*?'
|
|
r'<img src="https://mediaservice\.happyshops\.com/ANY/Article/\d+/size=100x100/([^"]+)"',
|
|
h, re.DOTALL
|
|
):
|
|
aid, fn = m.group(1), m.group(2).strip()
|
|
fn = fn.split("?")[0]
|
|
if aid not in aid_img:
|
|
aid_img[aid] = fn
|
|
|
|
print(f" Map: {len(aid_img)} entries", file=sys.stderr)
|
|
|
|
# ── Also scrape individual product pages for any missing ──
|
|
def get_img_from_product_page(aid):
|
|
"""Fallback: fetch product detail page and extract main image filename."""
|
|
try:
|
|
h = fetch(f"https://www.spiele-offensive.de/index.php?cmd=artikel_anzeigen&aid={aid}", 10)
|
|
# Main product image: class='abildg' src='.../width=450/FILENAME'
|
|
m = re.search(r"class='abildg'[^>]*src='[^']*/([^/\"'\s?]+)'", h)
|
|
if m:
|
|
return m.group(1).strip()
|
|
# Alternative: any width=450 image
|
|
m = re.search(r"width=450/([^\"'\s?]+)", h)
|
|
if m:
|
|
return m.group(1).strip()
|
|
# Last resort: first mediaservice image that's not 100x100
|
|
for m in re.finditer(r'src="https://mediaservice[^"]*/([^/"\s?]+)"', h):
|
|
fn = m.group(1)
|
|
if 'size=100x100' not in fn and len(fn) > 3:
|
|
return fn.strip()
|
|
except:
|
|
pass
|
|
return None
|
|
|
|
# ── Extract deals from existing deals HTML ──
|
|
print("Reading existing deals...", file=sys.stderr)
|
|
with open("/home/hermes/workspace/spiele_offensive_deals.html") as f:
|
|
html = f.read()
|
|
|
|
deals = []
|
|
for m in re.finditer(
|
|
r'<div class="card-name">([^<]+)</div>.*?'
|
|
r'<div class="card-badge"[^>]*>([^<]+)</div>.*?'
|
|
r'<span class="card-old">([^<]*)</span>.*?'
|
|
r'<span class="card-new">([^<]+)</span>.*?'
|
|
r'<a href="[^"]*aid=(\d+)',
|
|
html, re.DOTALL
|
|
):
|
|
name = m.group(1).strip()
|
|
badge = m.group(2).strip()
|
|
old = m.group(3).strip().replace(" €", "")
|
|
new = m.group(4).strip().replace(" €", "")
|
|
aid = m.group(5).strip()
|
|
rabatt = int(re.search(r'(\d+)', badge).group(1)) if re.search(r'(\d+)', badge) else 0
|
|
deals.append({"name": name, "rabatt": rabatt, "old_price": old, "new_price": new, "aid": aid})
|
|
|
|
# Filter & sort
|
|
SKIP = ['puzzle', 'jigsaw', 'wundertüte', 'rätselpuzzle', 'tragetasche']
|
|
deals = [d for d in deals if not any(w in d["name"].lower() for w in SKIP)]
|
|
seen = set()
|
|
uniq = []
|
|
for d in deals:
|
|
if d["aid"] not in seen:
|
|
seen.add(d["aid"])
|
|
uniq.append(d)
|
|
uniq.sort(key=lambda d: -d["rabatt"])
|
|
top30 = uniq[:30]
|
|
|
|
# ── Ensure all products have images ──
|
|
for d in top30:
|
|
aid = d["aid"]
|
|
if aid not in aid_img:
|
|
fn = get_img_from_product_page(aid)
|
|
if fn:
|
|
aid_img[aid] = fn
|
|
print(f" Fallback {aid}: {fn}", file=sys.stderr)
|
|
else:
|
|
print(f" NO IMAGE: {aid} - {d['name'][:50]}", file=sys.stderr)
|
|
|
|
# ── Generate HTML ──
|
|
def img_url(aid, fn):
|
|
padded = f"{int(aid):012d}"
|
|
return f"https://mediaservice.happyshops.com/ANY/Article/{padded}/width=200/{fn}"
|
|
|
|
cards = ""
|
|
for d in top30:
|
|
name = d["name"]
|
|
aid = d["aid"]
|
|
old = d["old_price"]
|
|
new = d["new_price"]
|
|
rabatt = d["rabatt"]
|
|
|
|
try:
|
|
saved = float(old.replace(",",".")) - float(new.replace(",","."))
|
|
saved_s = f"({saved:.2f} € gespart)".replace(".", ",")
|
|
except:
|
|
saved_s = ""
|
|
|
|
old_h = f'<span class="cold">UVP {old} €</span>'
|
|
bc = "#b12704" if rabatt >= 80 else ("#e63946" if rabatt >= 60 else "#cc0c39")
|
|
|
|
if aid in aid_img:
|
|
isrc = img_url(aid, aid_img[aid])
|
|
igh = f'<img src="{isrc}" alt="" class="dimg" loading="lazy" onerror="this.style.display=\'none\'">'
|
|
else:
|
|
igh = '<div class="dimg" style="display:flex;align-items:center;justify-content:center;color:#ccc;font-size:2rem;background:#eee">?</div>'
|
|
|
|
cards += f"""<div class="d">
|
|
{igh}
|
|
<div class="dinfo">
|
|
<span class="dname">{name}</span>
|
|
<span class="dprice">{new} €</span>{old_h}
|
|
<span class="dbadge" style="background:{bc}">-{rabatt}%</span>
|
|
<span class="dsave">{saved_s}</span>
|
|
<a href="https://www.spiele-offensive.de/index.php?cmd=artikel_anzeigen&aid={aid}&pid={PID}" target="_blank" class="dlink">Zu Spiele-Offensive →</a>
|
|
</div></div>
|
|
"""
|
|
|
|
out_html = f"""<!DOCTYPE html>
|
|
<html lang="de">
|
|
<head>
|
|
<meta charset="UTF-8">
|
|
<meta name="viewport" content="width=device-width,initial-scale=1.0">
|
|
<title>Top 30 Brettspiel-Angebote — Spiele-Offensive.de</title>
|
|
<style>
|
|
body{{margin:0;padding:0;background:#f9f7f4;font-family:-apple-system,BlinkMacSystemFont,'Segoe UI',Roboto,sans-serif;color:#2c2416;line-height:1.6}}
|
|
article{{max-width:720px;margin:0 auto;padding:2rem 1.5rem;background:#fff;box-shadow:0 2px 12px rgba(0,0,0,.06)}}
|
|
h1{{font-size:1.5rem;color:#1a1a1a}}.meta{{color:#999;font-size:.85rem;margin-bottom:1.5rem}}
|
|
a{{color:#007185;text-decoration:none}}a:hover{{text-decoration:underline}}
|
|
.d{{display:flex;gap:1rem;padding:1rem 0;border-bottom:1px solid #eee;align-items:flex-start}}
|
|
.dimg{{width:100px;height:100px;object-fit:contain;flex-shrink:0;border-radius:4px;background:#f5f5f5}}
|
|
.dinfo{{flex:1;min-width:0}}
|
|
.dname{{font-weight:600;color:#1a1a1a;font-size:.95rem;display:block;line-height:1.3;margin-bottom:3px}}
|
|
.dprice{{color:#b12704;font-size:1.2rem;font-weight:bold}}
|
|
.cold{{text-decoration:line-through;color:#999;margin-left:.5rem;font-size:.9rem}}
|
|
.dbadge{{color:#fff;padding:2px 6px;border-radius:3px;margin-left:.5rem;font-size:.85rem;font-weight:600}}
|
|
.dsave{{color:#b12704;margin-left:.3rem;font-size:.9rem}}
|
|
.dlink{{display:inline-block;margin-top:4px;color:#e63946;font-size:.9rem;font-weight:500}}
|
|
.dlink:hover{{text-decoration:underline}}
|
|
.footer{{margin-top:2rem;padding-top:1rem;border-top:1px solid #eee;font-size:.85rem;color:#888}}
|
|
</style>
|
|
</head>
|
|
<body>
|
|
<article>
|
|
<h1>Top 30 Brettspiel-Angebote auf Spiele-Offensive.de</h1>
|
|
<p class="meta">Stand: {NOW} · Sortiert nach Rabatt · Preise inkl. MwSt. · Affiliate-Links (PID {PID}) · Bilder: HappyShops CDN</p>
|
|
|
|
{cards}
|
|
|
|
<div class="footer">
|
|
<p>Alle Preise und Verfügbarkeiten zum Zeitpunkt der Erhebung. Zwischenverkauf und Preisänderungen vorbehalten.</p>
|
|
<p>Dieser Beitrag enthält Affiliate-Links. Bei einem Kauf über diese Links erhalten wir eine kleine Provision — für dich entstehen keine Mehrkosten.</p>
|
|
</div>
|
|
</article></body></html>"""
|
|
|
|
with open(OUT, "w") as f:
|
|
f.write(out_html)
|
|
|
|
import subprocess
|
|
c = subprocess.run(["grep", "-c", "mediaservice", OUT], capture_output=True, text=True)
|
|
imgs = c.stdout.strip()
|
|
print(f"Done: {len(top30)} deals, {imgs} with images, {len(out_html)} bytes", file=sys.stderr)
|
|
print(f"FILE:{OUT}")
|