def update_news():
recent_news = get_news()
authors = [news['author'] for news in recent_news]
titles = s.query(News.title).filter(News.author.in_(authors)).subquery()
existing_news = s.query(News).filter(News.title.in_(titles)).all()
for news in recent_news:
if not existing_news or news not in existing_news:
fill(news)
redirect("/news")
def update_news():
recent_news = get_news()
authors = [news['author'] for news in recent_news]
titles = s.query(News.title).filter(News.author.in_(authors)).subquery()
existing_news = s.query(News).filter(News.title.in_(titles)).all()
titles_bd = [i.title for i in existing_news]
authors_bd = [i.author for i in existing_news]
for news in recent_news:
if not existing_news or (news['title'] not in titles_bd
and news["author"] not in authors_bd):
fill(news)
redirect("/news")
def get_categories(model, x_data, row_data):
pred = model.predict(x_data)
proba = model.predict_proba(x_data)
cnt = [0, 0]
for i, p in enumerate(pred):
indices = np.where(p)
clear = True
for e in proba[i]:
if 0.050 < e < 0.800:
clear = False
break
cats = [0, 0, 0, 0, 0]
for idx in indices[0]:
cats[int(idx)] = 1
fill(row_data[i], cats[0], cats[1], cats[2], cats[3], cats[4], int(clear))
cnt[int(clear)] += 1
print("Reviews successfully classified with {} clear and {} confusing reviews\n".format(cnt[1], cnt[0]))
def collect():
url = request.forms.get("url")
for r in get_reviews(url):
fill(r)
redirect("/reviews")