def extract_topics_from_url(db):
articles = db.articles.find()
topic_re = re.compile('\/\d+\/(?P<topic>[a-z]+)\/(?P<subtopic>[a-z]+)?\/')
topics = TopicsSet()
for article in articles:
match = topic_re.search(article['url'])
if match:
t = Topic(match.group('topic'))
if t not in topics:
topics.add(t)
else:
t = topics.getelement(t)
subtopic = match.group('subtopic')
if subtopic:
t.add_topic(subtopic)
return topics
