# Blog (EN) und alte DE-Meldungen paaren; Nummer nach Blog-Uhrzeit import sys, json, re sys.path.insert(0, sys.argv[1]); from blog2wiki import blog2wiki blog = json.load(open(sys.argv[2])); old = json.load(open(sys.argv[3])) manual = { # EN-Titel-Anfang -> DE-Titel-Anfang, wo der Wortabgleich nicht reicht 'Emperor Sickness Worsens': 'Krankheit des Kaisers', 'Patreus Attacks Durius': 'Patreus greift Durius', } def words(s): return set(re.findall(r'[a-zäöüß0-9]{4,}', s.lower())) pairs, used = [], set() from collections import defaultdict cnt = defaultdict(int) for b in blog: y, m, d = b['published'][:10].split('-') cnt[(m, d)] += 1 pid = f'2014:{m}:{d}_{cnt[(m, d)]:02d}' cands = [o for o in old if o['mm'] == m and o['dd'] == d and o['id'] not in used] en = blog2wiki(b['html']) # Abgleich ueber Eigennamen/Zahlen im Text def score(o): return len(words(en) & words(o['body'])) + 3 * len(words(b['title']) & words(o['title'])) fix = next((v for k, v in manual.items() if b['title'].startswith(k)), None) o = next(o for o in cands if o['title'].startswith(fix)) if fix else max(cands, key=score) used.add(o['id']) pairs.append({'id': pid, 'title_en': b['title'], 'body_en': en, 'url': b['url'], 'labels': b['labels'], 'title': o['title'], 'body': o['body'], 'old_img': o['old_img'], 'old_id': o['id']}) json.dump(pairs, open(sys.argv[4], 'w'), ensure_ascii=False, indent=1) for p in pairs: print(p['id'], p['old_id'][-5:], '|', p['title_en'][:45], '|', p['title'][:45])