def todays_sentiment():
"""Get a mean of sentiment for all of today's articles."""
t = time.time()
sentiment = db.session.query(db.func.avg(Article.sent)).filter(Article.date==date.today()).all()
timing(t, "article.todays_sentiment")
if sentiment[0][0]:
return round(sentiment[0][0]*100, 2)
return 0
def rawwordcloud(text, scale, width=150, height=80, POS=POS):
t = time.time()
tokenized = nltk.word_tokenize(text)
tokenized = nltk.pos_tag(tokenized)
text = []
for token in tokenized:
if token[1] in POS:
text.append(token[0])
text = ' '.join(text)
t = timing(t, "generating texts")
# Make wordcloud
wordcloud = WordCloud(
width=width, height=height, background_color='white',
stopwords=stopwords, scale=scale)\
.generate(text).to_array()
t = timing(t, "generating wordcloud")
# Make image
img = Image.fromarray(wordcloud.astype('uint8'))
timing(t, "generating image")
return img
def status():
t = time.time()
agencies = Agency.query.order_by(Agency.name).all()
ids = [a.id for a in agencies]
articles = Article.query.filter(Article.date == date.today(),
Article.agency_id.in_(ids)).all()
t = timing(t, "getting articles")
for article in articles:
if not hasattr(article.agency, 'todays_articles_speed'):
article.agency.todays_articles_speed = []
article.agency.todays_articles_speed.append(article)
t = timing(t, "hanging articles")
for agency in agencies:
try:
agency.todays_sentiment_speed = speedround(
[a.sentiment for a in agency.todays_articles_speed])
agency.todays_neutrality_speed = speedround(
[a.neutrality for a in agency.todays_articles_speed])
except:
agency.todays_sentiment_speed = 0.0
agency.todays_neutrality_speed = 0.0
agency.todays_articles_speed = []
t = timing(t, "Calculating todays stats")
return render_template('status.html', agencies=agencies, title='Status')
def todays_count(self):
"""Count of today's articles for the agency"""
t = time.time()
count = self.articles.filter(Article.date==date.today()).count()
timing(t, "agency.todays_count")
return count
def todays_articles(self):
t = time.time()
articles = self.articles.filter(Article.date==date.today())\
.order_by(Article.sent.desc())
timing(t, "agency.todays_articles")
return articles
def index():
"""Generates an index of today's articles, sorted by sentiment, colored by
sentiment.
"""
t = time.time()
sort = request.args.get('sort', default='Name')
sorts = {
'Name':
Agency.query.order_by(Agency.name),
'Articles': (Agency.query.join(Article).filter(
Article.date == date.today()).group_by(Agency.id).order_by(
db.func.count(Article.id).desc())),
'Cumulative Sentiment':
Agency.query.order_by(Agency.cum_sent.desc()),
'Cumulative Neutrality':
Agency.query.order_by(Agency.cum_neut.desc()),
}
page = request.args.get('page', default=1)
try:
page = int(page)
except:
abort(404)
PER_PAGE = 3
pages = list(range(1, math.ceil(Agency.query.count() / PER_PAGE) + 1))
t = timing(t, "getting sorts and pages")
try:
with open('/status.json', 'rt') as fin:
jobs = json.load(fin)
except Exception as e:
app.logger.info(e)
jobs = {'running': []}
t = timing(t, "getting jobs")
try:
agencies = sorts[sort].paginate(page, PER_PAGE).items
except KeyError:
abort(404)
t = timing(t, "getting agencies")
ids = [a.id for a in agencies]
articles = Article.query.filter(Article.date == date.today(),
Article.agency_id.in_(ids)).all()
t = timing(t, "getting articles")
for article in articles:
if not hasattr(article.agency, 'todays_articles_speed'):
article.agency.todays_articles_speed = []
article.agency.todays_articles_speed.append(article)
t = timing(t, "hanging articles")
for agency in agencies:
try:
agency.todays_sentiment_speed = speedround(
[a.sentiment for a in agency.todays_articles_speed])
agency.todays_neutrality_speed = speedround(
[a.neutrality for a in agency.todays_articles_speed])
except:
agency.todays_sentiment_speed = 0.0
agency.todays_neutrality_speed = 0.0
agency.todays_articles_speed = []
t = timing(t, "calculating stats")
html = render_template(
'index.html',
count=Article.query.filter(Article.date == date.today()).count(),
dbcount=Article.query.count(),
agencies=agencies,
sorts=sorts,
sort=sort,
pages=pages,
page=page,
overall=Article.todays_sentiment(),
running=jobs['running'],
active='index')
timing(t, "rendering template")
return html