Hermes Agent
86 lines
2.7 KiB
Python
86 lines
2.7 KiB
Python
#!/usr/bin/env python3
|
|
"""BSN Article QC Checker — auto-fix umlauts, flag forbidden words & missing affiliate links.
|
|
QUICK CHECK: python3 qc_check.py <article.html>
|
|
FULL WORKFLOW: Use korrekturschleife skill (Writer → Fact-Checker → Korrektur)
|
|
"""
|
|
import sys, re
|
|
|
|
UMLAUT_FIXES = {
|
|
'uber ': 'über ', ' uber ': ' über ', 'fur ': 'für ', ' fur ': ' für ',
|
|
'Fur ': 'Für ', 'fuhrt ': 'führt ', 'Mazene': 'Mäzene', 'duster': 'düster',
|
|
'kunstlich': 'künstlich', 'Anderung': 'Änderung', 'Verfugbarkeit': 'Verfügbarkeit',
|
|
'berucksichtigt': 'berücksichtigt', 'Berucksichtigt': 'Berücksichtigt',
|
|
'eingeschrankt': 'eingeschränkt', 'Außergewohnliches': 'Außergewöhnliches',
|
|
}
|
|
|
|
FORBIDDEN = ['XML API', 'API v2', 'xmlapi2', 'Betriebsgeheimnis',
|
|
'FlareSolverr', 'Scraping', 'Proxy', 'Cronjob']
|
|
|
|
LOGIC_FIXES = {
|
|
'Nicht-Basisspielen': 'Basisspielen',
|
|
}
|
|
|
|
def main():
|
|
if len(sys.argv) < 2:
|
|
print("Usage: qc_check.py <article.html>")
|
|
sys.exit(1)
|
|
|
|
path = sys.argv[1]
|
|
with open(path) as f:
|
|
html = f.read()
|
|
|
|
fixes = []
|
|
errors = []
|
|
|
|
# 1. Auto-fix umlauts
|
|
for wrong, correct in UMLAUT_FIXES.items():
|
|
if wrong in html:
|
|
count = html.count(wrong)
|
|
html = html.replace(wrong, correct)
|
|
fixes.append(f' UMLAUT: {wrong} → {correct} ({count}x)')
|
|
|
|
# 2. Auto-fix logic errors
|
|
for wrong, correct in LOGIC_FIXES.items():
|
|
if wrong in html:
|
|
html = html.replace(wrong, correct)
|
|
fixes.append(f' LOGIK: {wrong} → {correct}')
|
|
|
|
# 3. Flag forbidden words (manual fix needed)
|
|
for word in FORBIDDEN:
|
|
if word.lower() in html.lower():
|
|
errors.append(f' VERBOTEN: "{word}" im Artikel!')
|
|
|
|
# 4. Check deal/buy links (NOT Quellen) for affiliate tracking
|
|
# Deal links: inside .buy-link, infobox, or table td.buy
|
|
deal_sections = re.findall(r'(?:class="buy-link"|class="infobox"|Zum Deal).*?</a>', html, re.DOTALL)
|
|
for section in deal_sections:
|
|
links = re.findall(r'href="([^"]*milan-spiele\.de[^"]*)"', section)
|
|
for link in links:
|
|
if 'awinmid' not in link:
|
|
errors.append(f' AFFILIATE: Milan-Link ohne AWIN: {link[:80]}')
|
|
|
|
# Report
|
|
if fixes:
|
|
print(f'QC: {path}')
|
|
for f in fixes:
|
|
print(f)
|
|
print(f' → {len(fixes)} Fixes angewendet.')
|
|
|
|
if errors:
|
|
for e in errors:
|
|
print(e)
|
|
print(f'\n FEHLER: {len(errors)} manuelle Korrekturen notig!')
|
|
sys.exit(1)
|
|
|
|
if not fixes and not errors:
|
|
print(f'QC: {path} — OK')
|
|
|
|
if fixes:
|
|
with open(path, 'w') as f:
|
|
f.write(html)
|
|
|
|
sys.exit(0)
|
|
|
|
if __name__ == '__main__':
|
|
main()
|