def recommendations():
s = session()
labeled = s.query(News).filter(
News.label != None).all() #новости, которые разметили
news_title = [news.title for news in labeled]
news_label = [news.label for news in labeled]
bayes = NaiveBayesClassifier(
) #пропускаем данные через наивный классификатор
bayes.fit(news_title, news_label)
notlabeled = s.query(News).filter(
News.label == None).all() #новости, которые не разметили
new_news_title = [news.title for news in notlabeled]
new_news_label = bayes.predict(new_news_title)
good_news, maybe_news, never_news = [], [], []
#Добавляем метки нерзамечанным новостям
for id, label in enumerate(new_news_label):
if label == 'good':
good_news.append(notlabeled[id])
elif label == 'maybe':
maybe_news.append(notlabeled[id])
elif label == 'never':
never_news.append(notlabeled[id])
classified_news = [good_news, maybe_news, never_news]
return template('news_ranked',
good_news=good_news,
maybe_news=maybe_news,
never_news=never_news)
def add_label():
s = session()
label = request.GET['label']
newid = request.GET['id']
headline = s.query(News).filter(News.id == int(newid)).all()[0]
headline.label = label
s.commit()
redirect("/news")
def add_label():
label = request.query.label
id = request.query.id
s = session()
changing_news = s.query(News).get(id)
changing_news.label = label
s.commit()
redirect("/news")
def add_label():
# 1. Получить значения параметров label и id из GET-запроса
id = request.query.get('id')
label = request.query.get('label')
s = session()
entry = s.query(News).get(id)
# 3. Изменить значение метки записи на значение label
entry.label = label
# 4. Сохранить результат в БД
s.commit()
redirect("/news")
def update_news():
news_s = get_news('https://news.ycombinator.com/newest', n_pages=5)
s = session()
news_bd = s.query(News).all()
news_bd = [(n.title, n.author) for n in news_bd]
for new in news_s:
if (new['title'], new['author']) not in news_bd:
news = News(title=new['title'],
author=new['author'],
url=new['url'],
comments=new['comments'],
points=new['points'])
s.add(news)
s.commit()
redirect('/news')
def update_news():
new_news_list = get_news("https://news.ycombinator.com/", n_pages=5)
s = session()
old_news = s.query(News).all()
for news in new_news_list:
for o_news in old_news:
if (news['title'] == o_news.title
and news['author'] == o_news.author):
break
else:
new_news = News(title=news['title'],
author=news['author'],
url=news['url'],
comments=news['comments'],
points=news['points'])
s.add(new_news)
s.commit()
redirect("/news")
def update_news():
news_list = get_news("https://news.ycombinator.com/newest", n_pages=1)
cnt = 0
for elem in news_list:
s = session()
if len(
s.query(News).filter(News.title == elem['title']).filter(
News.author == elem['author']).all()) > 0:
print('Not writing')
continue
cnt += 1
print('Writing data element:', cnt)
news = News(title=elem['title'],
author=elem['author'],
url=elem['url'],
comments=elem['comments'],
points=elem['points'])
s.add(news)
s.commit()
redirect("/news")
def recommendations():
# 1. Получить список неразмеченных новостей из БД
# 2. Получить прогнозы для каждой новости
# 3. Вывести ранжированную таблицу с новостями
s = session()
none_news = []
rows = s.query(News).filter(News.label == None).all()
learn_news = s.query(News).filter(News.label != None).all()
X_train, y_train = [], []
for news in learn_news:
X_train.append(news.title)
y_train.append(news.label)
X_train = [clean(x).lower() for x in X_train]
model = NaiveBayesClassifier(alpha=1)
model.fit(X_train, y_train)
for news in rows:
none_news.append(news.title)
predict_labels = model.predict(none_news)
for news, label in zip(rows, predict_labels):
news.label = label
classified_news = sorted(rows, key=lambda news: news.label)
return template('./news_recommendations.html', rows=classified_news)
def classify_news():
# 1. Получить список неразмеченных новостей из БД
# 2. Получить прогнозы для каждой новости
# 3. Вывести ранжированную таблицу с новостями
s = session()
good, maybe, never, X_train, y_train = [], [], [], [], []
headlines_classified = s.query(News).filter(News.label != None)
headlines_nonclassified = s.query(News).filter(News.label == None)
for topic in headlines_classified:
X_train.append(topic.title)
y_train.append(topic.label)
if topic.label == 'good':
good.append(topic)
elif topic.label == 'maybe':
maybe.append(topic)
else:
never.append(topic)
model = NaiveBayesClassifier()
model.fit(X_train, y_train)
for topic in headlines_nonclassified:
topic.label = model.predict(topic.title)
if topic.label == 'good':
good.append(topic)
elif topic.label == 'maybe':
maybe.append(topic)
else:
never.append(topic)
rows = good + maybe + never
return template('news_template.tpl', rows=rows)
def news_list():
TEMPLATE_PATH.insert(0, '')
s = session()
rows = s.query(News).filter(News.label == None).all()
return template('news_template', rows=rows)
def news_list():
s = session()
rows = s.query(News).filter(News.label == None).all()
return template('news_template.tpl', rows=rows)