def add_label():
s = session()
new_id = request.query['id']
label = request.query['label']
new = s.query(News).get(new_id)
new.label = label
s.commit()
redirect("/news")
def update_news():
news_lst = get_news('https://news.ycombinator.com', 10)
s = session()
for i in range(len(news_lst)):
if len(
s.query(News).filter(
News.title == news_lst[i]['title'],
News.author == news_lst[i]['author']).all()) == 0:
new_news = News(title=news_lst[i]['title'],
author=news_lst[i]['author'],
points=news_lst[i]['points'],
comments=news_lst[i]['comments'],
url=news_lst[i]['url'])
s.add(new_news)
s.commit()
redirect("/news")
def recommendations():
s = session()
rows_unlabelled = s.query(News).filter(News.label == None).all()
X = [clean(row.title).lower() for row in rows_unlabelled]
predictions = model.predict(X)
rows_good = [
rows_unlabelled[i] for i in range(len(rows_unlabelled))
if predictions[i] == 'good'
]
rows_maybe = [
rows_unlabelled[i] for i in range(len(rows_unlabelled))
if predictions[i] == 'maybe'
]
rows_never = [
rows_unlabelled[i] for i in range(len(rows_unlabelled))
if predictions[i] == 'never'
]
return template('recommendations_template',
rows_good=rows_good,
rows_maybe=rows_maybe,
rows_never=rows_never)
def news_list():
s = session()
rows = s.query(News).filter(News.label == None).all()
return template('news_template', rows=rows)
]
rows_maybe = [
rows_unlabelled[i] for i in range(len(rows_unlabelled))
if predictions[i] == 'maybe'
]
rows_never = [
rows_unlabelled[i] for i in range(len(rows_unlabelled))
if predictions[i] == 'never'
]
return template('recommendations_template',
rows_good=rows_good,
rows_maybe=rows_maybe,
rows_never=rows_never)
def clean(s):
translator = str.maketrans("", "", string.punctuation)
return s.translate(translator)
if __name__ == "__main__":
s = session()
rows = s.query(News).filter(News.label != None).all()
X_train = [clean(row.title).lower() for row in rows]
y_train = [row.label for row in rows]
model = NaiveBayesClassifier(alpha=0.05)
model.fit(X_train, y_train)
run(host="localhost", port=9998)