def test_spam_trends_stats(self, mock_analytics_upageviews):
"""Test that the correct stats show up on the spam trends dashboard."""
# Period length
days_in_week = 7
days_in_month = 28
days_in_quarter = 91
# Dates
today = datetime.datetime.today()
yesterday = today - datetime.timedelta(days=1)
three_days_ago = today - datetime.timedelta(days=3)
weekly_start_date = today - datetime.timedelta(days=days_in_week)
ten_days_ago = today - datetime.timedelta(days=10)
monthly_start_date = today - datetime.timedelta(days=days_in_month)
thirtyfive_days_ago = today - datetime.timedelta(days=35)
quarterly_start_date = today - datetime.timedelta(days=days_in_quarter)
hundred_days_ago = today - datetime.timedelta(days=100)
# Revisions made by self.testuser: 3 made today, 3 made 3 days ago,
# 3 made 10 days ago, 3 made 35 days ago, 3 made 100 days ago
revs = self.create_revisions(num=15, creator=self.testuser, document=self.document)
for i in range(0, 3):
revs[i].created = today
for i in range(3, 6):
revs[i].created = three_days_ago
for i in range(6, 9):
revs[i].created = ten_days_ago
for i in range(9, 12):
revs[i].created = thirtyfive_days_ago
for i in range(12, 15):
revs[i].created = hundred_days_ago
for rev in revs:
rev.save()
# Published spam by self.testuser
spam_rev_today = revs[2]
spam_rev_3_days_ago = revs[5]
spam_rev_10_days_ago = revs[8]
spam_rev_35_days_ago = revs[11]
spam_rev_100_days_ago = revs[14]
spam_revs = [spam_rev_today, spam_rev_3_days_ago, spam_rev_10_days_ago,
spam_rev_35_days_ago, spam_rev_100_days_ago]
# Summary of spam submissions
spam_weekly = [spam_rev_3_days_ago]
spam_monthly = [spam_rev_3_days_ago, spam_rev_10_days_ago]
spam_quarterly = [spam_rev_3_days_ago, spam_rev_10_days_ago,
spam_rev_35_days_ago]
# All of the spam_revs were published and then marked as spam
for rev in spam_revs:
rev.save()
rev.akismet_submissions.add(RevisionAkismetSubmission(
sender=self.admin, type="spam")
)
# Summary of self.testuser's ham submissions
ham_weekly = revs[3:5]
ham_monthly = revs[3:5] + revs[6:8]
ham_quarterly = revs[3:5] + revs[6:8] + revs[9:11]
# There were 2 correctly blocked spam attempts 3 days ago (within past week)
true_blocked_spam_num = 2
for i in range(0, true_blocked_spam_num):
document_spam_rev_3_days_ago = DocumentSpamAttempt(
user=self.testuser,
title='A spam revision',
slug='spam-revision-slug',
document=self.document,
review=DocumentSpamAttempt.SPAM
)
document_spam_rev_3_days_ago.save()
document_spam_rev_3_days_ago.created = three_days_ago
document_spam_rev_3_days_ago.save()
# There was 1 incorrectly blocked spam attempt 3 days ago
false_blocked_spam_num = 1
for i in range(0, false_blocked_spam_num):
document_ham_rev_3_days_ago = DocumentSpamAttempt(
user=self.testuser,
title='Not a spam revision',
slug='ham-revision-slug',
document=self.document,
review=DocumentSpamAttempt.HAM
)
document_ham_rev_3_days_ago.save()
document_ham_rev_3_days_ago.created = three_days_ago
document_ham_rev_3_days_ago.save()
page_views = {}
# The spam from 3 days ago was seen 3 times, from 10 days ago see 10 times,
# and from 35 days ago seen 35 times
page_views[spam_rev_3_days_ago.id] = 3
page_views[spam_rev_10_days_ago.id] = 10
page_views[spam_rev_35_days_ago.id] = 35
# The mock Google Analytics return values for page views
mock_analytics_upageviews.return_value = page_views
self.client.login(username='******', password='******')
# The first response will say that the report is being processed
response = self.client.get(reverse('dashboards.spam', locale='en-US'))
eq_(200, response.status_code)
response2 = self.client.get(reverse('dashboards.spam', locale='en-US'))
page = pq(response2.content)
row_daily = page.find('.spam-trends-table tbody tr')[0].text_content().replace(' ', '').strip('\n').split('\n')
row_weekly = page.find('.spam-trends-table tbody tr')[1].text_content().replace(' ', '').strip('\n').split('\n')
row_monthly = page.find('.spam-trends-table tbody tr')[2].text_content().replace(' ', '').strip('\n').split('\n')
row_quarterly = page.find('.spam-trends-table tbody tr')[3].text_content().replace(' ', '').strip('\n').split('\n')
# These are the columns in the spam dashboard spam trends table
period = 0
start_date = 1
spam_viewers_change_percent = 2
spam_viewers = 3
daily_average_viewers = 4
published_spam = 5
blocked_spam = 6
blocked_ham = 7
true_positive_rate = 8
true_negative_rate = 9
# The periods are identified as 'Daily', 'Weekly', 'Monthly', 'Quarterly'
eq_(row_daily[period], 'Daily')
eq_(row_weekly[period], 'Weekly')
eq_(row_monthly[period], 'Monthly')
eq_(row_quarterly[period], 'Quarterly')
# The start dates for each period are correct
eq_(row_daily[start_date], yesterday.strftime('%Y-%m-%d'))
eq_(row_weekly[start_date], weekly_start_date.strftime('%Y-%m-%d'))
eq_(row_monthly[start_date], monthly_start_date.strftime('%Y-%m-%d'))
eq_(row_quarterly[start_date], quarterly_start_date.strftime('%Y-%m-%d'))
# The page views during the week, month, quarter
spam_views_week = page_views[spam_rev_3_days_ago.id]
spam_views_month = spam_views_week + page_views[spam_rev_10_days_ago.id]
spam_views_month_exclude_week = page_views[spam_rev_10_days_ago.id]
spam_views_quarter = spam_views_month + page_views[spam_rev_35_days_ago.id]
spam_views_quarter_exclude_month = page_views[spam_rev_35_days_ago.id]
# The percentage change in spam viewers
weekly_spam_change_percent = '{:.1%}'.format(
float(spam_views_week - spam_views_month_exclude_week) / spam_views_month_exclude_week
)
monthly_spam_change_percent = '{:.1%}'.format(
float(spam_views_month - spam_views_quarter_exclude_month) / spam_views_quarter_exclude_month
)
eq_(row_daily[spam_viewers_change_percent], '0.0%')
eq_(row_weekly[spam_viewers_change_percent], weekly_spam_change_percent)
eq_(row_monthly[spam_viewers_change_percent], monthly_spam_change_percent)
eq_(row_quarterly[spam_viewers_change_percent], '0.0%')
# The spam viewers
eq_(int(row_daily[spam_viewers]), 0)
eq_(int(row_weekly[spam_viewers]), spam_views_week)
eq_(int(row_monthly[spam_viewers]), spam_views_month)
eq_(int(row_quarterly[spam_viewers]), spam_views_quarter)
# The daily average of spam viewers
eq_(float(row_daily[daily_average_viewers]), 0.0)
eq_(row_weekly[daily_average_viewers],
'{:.1f}'.format(float(spam_views_week) / days_in_week))
eq_(row_monthly[daily_average_viewers],
'{:.1f}'.format(float(spam_views_month) / days_in_month))
eq_(row_quarterly[daily_average_viewers],
'{:.1f}'.format(float(spam_views_quarter) / days_in_quarter))
# The published spam: 1 this week, 2 this month, 3 this quarter
eq_(int(row_daily[published_spam]), len([]))
eq_(int(row_weekly[published_spam]), len(spam_weekly))
eq_(int(row_monthly[published_spam]), len(spam_monthly))
eq_(int(row_quarterly[published_spam]), len(spam_quarterly))
# The blocked spam: there were 2 correctly blocked spam attempts 3 days ago
eq_(int(row_daily[blocked_spam]), 0)
eq_(int(row_weekly[blocked_spam]), true_blocked_spam_num)
eq_(int(row_monthly[blocked_spam]), true_blocked_spam_num)
eq_(int(row_quarterly[blocked_spam]), true_blocked_spam_num)
# The blocked ham: there was 1 incorrectly blocked spam attempt 3 days ago
eq_(int(row_daily[blocked_ham]), 0)
eq_(int(row_weekly[blocked_ham]), false_blocked_spam_num)
eq_(int(row_monthly[blocked_ham]), false_blocked_spam_num)
eq_(int(row_quarterly[blocked_ham]), false_blocked_spam_num)
# The true positive rate == blocked_spam / total spam
tpr_weekly = '{:.1%}'.format(
true_blocked_spam_num / float(true_blocked_spam_num + len(spam_weekly))
)
tpr_monthly = '{:.1%}'.format(
true_blocked_spam_num / float(true_blocked_spam_num + len(spam_monthly))
)
tpr_quarterly = '{:.1%}'.format(
true_blocked_spam_num / float(true_blocked_spam_num + len(spam_quarterly))
)
eq_(row_daily[true_positive_rate], '100.0%')
eq_(row_weekly[true_positive_rate], tpr_weekly)
eq_(row_monthly[true_positive_rate], tpr_monthly)
eq_(row_quarterly[true_positive_rate], tpr_quarterly)
# The true negative rate == published ham / total ham
tnr_weekly = '{:.1%}'.format(
len(ham_weekly) / float(false_blocked_spam_num + len(ham_weekly))
)
tnr_monthly = '{:.1%}'.format(
len(ham_monthly) / float(false_blocked_spam_num + len(ham_monthly))
)
tnr_quarterly = '{:.1%}'.format(
len(ham_quarterly) / float(false_blocked_spam_num + len(ham_quarterly))
)
eq_(row_daily[true_negative_rate], '100.0%')
eq_(row_weekly[true_negative_rate], tnr_weekly)
eq_(row_monthly[true_negative_rate], tnr_monthly)
eq_(row_quarterly[true_negative_rate], tnr_quarterly)
def test_spam_trends_stats(self, mock_analytics_upageviews):
"""Test that the correct stats show up on the spam trends dashboard."""
# Period length
days_in_week = 7
days_in_month = 28
days_in_quarter = 91
# Dates
today = datetime.datetime.today()
yesterday = today - datetime.timedelta(days=1)
three_days_ago = today - datetime.timedelta(days=3)
weekly_start_date = today - datetime.timedelta(days=days_in_week)
ten_days_ago = today - datetime.timedelta(days=10)
monthly_start_date = today - datetime.timedelta(days=days_in_month)
thirtyfive_days_ago = today - datetime.timedelta(days=35)
quarterly_start_date = today - datetime.timedelta(days=days_in_quarter)
hundred_days_ago = today - datetime.timedelta(days=100)
# Revisions made by self.testuser: 3 made today, 3 made 3 days ago,
# 3 made 10 days ago, 3 made 35 days ago, 3 made 100 days ago
revs = self.create_revisions(num=15,
creator=self.testuser,
document=self.document)
for i in range(0, 3):
revs[i].created = today
for i in range(3, 6):
revs[i].created = three_days_ago
for i in range(6, 9):
revs[i].created = ten_days_ago
for i in range(9, 12):
revs[i].created = thirtyfive_days_ago
for i in range(12, 15):
revs[i].created = hundred_days_ago
for rev in revs:
rev.save()
# Published spam by self.testuser
spam_rev_today = revs[2]
spam_rev_3_days_ago = revs[5]
spam_rev_10_days_ago = revs[8]
spam_rev_35_days_ago = revs[11]
spam_rev_100_days_ago = revs[14]
spam_revs = [
spam_rev_today,
spam_rev_3_days_ago,
spam_rev_10_days_ago,
spam_rev_35_days_ago,
spam_rev_100_days_ago,
]
# Summary of spam submissions
spam_weekly = [spam_rev_3_days_ago]
spam_monthly = [spam_rev_3_days_ago, spam_rev_10_days_ago]
spam_quarterly = [
spam_rev_3_days_ago,
spam_rev_10_days_ago,
spam_rev_35_days_ago,
]
# All of the spam_revs were published and then marked as spam
for rev in spam_revs:
rev.save()
rev.akismet_submissions.create(sender=self.admin, type="spam")
# Summary of self.testuser's ham submissions
ham_weekly = revs[3:5]
ham_monthly = revs[3:5] + revs[6:8]
ham_quarterly = revs[3:5] + revs[6:8] + revs[9:11]
# There were 2 correctly blocked spam attempts 3 days ago (within past week)
true_blocked_spam_num = 2
for i in range(0, true_blocked_spam_num):
document_spam_rev_3_days_ago = DocumentSpamAttempt(
user=self.testuser,
title="A spam revision",
slug="spam-revision-slug",
document=self.document,
review=DocumentSpamAttempt.SPAM,
)
document_spam_rev_3_days_ago.save()
document_spam_rev_3_days_ago.created = three_days_ago
document_spam_rev_3_days_ago.save()
# There was 1 incorrectly blocked spam attempt 3 days ago
false_blocked_spam_num = 1
for i in range(0, false_blocked_spam_num):
document_ham_rev_3_days_ago = DocumentSpamAttempt(
user=self.testuser,
title="Not a spam revision",
slug="ham-revision-slug",
document=self.document,
review=DocumentSpamAttempt.HAM,
)
document_ham_rev_3_days_ago.save()
document_ham_rev_3_days_ago.created = three_days_ago
document_ham_rev_3_days_ago.save()
page_views = {}
# The spam from 3 days ago was seen 3 times, from 10 days ago see 10 times,
# and from 35 days ago seen 35 times
page_views[spam_rev_3_days_ago.id] = 3
page_views[spam_rev_10_days_ago.id] = 10
page_views[spam_rev_35_days_ago.id] = 35
# The mock Google Analytics return values for page views
mock_analytics_upageviews.return_value = page_views
self.client.login(username="******", password="******")
# The first response will say that the report is being processed
response = self.client.get(reverse("dashboards.spam"),
HTTP_HOST=settings.WIKI_HOST)
assert 200 == response.status_code
response2 = self.client.get(reverse("dashboards.spam"),
HTTP_HOST=settings.WIKI_HOST)
page = pq(response2.content)
row_daily = (
page.find(".spam-trends-table tbody tr")[0].text_content().replace(
" ", "").strip("\n").split("\n"))
row_weekly = (
page.find(".spam-trends-table tbody tr")[1].text_content().replace(
" ", "").strip("\n").split("\n"))
row_monthly = (
page.find(".spam-trends-table tbody tr")[2].text_content().replace(
" ", "").strip("\n").split("\n"))
row_quarterly = (
page.find(".spam-trends-table tbody tr")[3].text_content().replace(
" ", "").strip("\n").split("\n"))
# These are the columns in the spam dashboard spam trends table
period = 0
start_date = 1
spam_viewers_change_percent = 2
spam_viewers = 3
daily_average_viewers = 4
published_spam = 5
blocked_spam = 6
blocked_ham = 7
true_positive_rate = 8
true_negative_rate = 9
# The periods are identified as 'Daily', 'Weekly', 'Monthly', 'Quarterly'
assert "Daily" == row_daily[period]
assert "Weekly" == row_weekly[period]
assert "Monthly" == row_monthly[period]
assert "Quarterly" == row_quarterly[period]
# The start dates for each period are correct
assert yesterday.strftime("%Y-%m-%d") == row_daily[start_date]
assert weekly_start_date.strftime("%Y-%m-%d") == row_weekly[start_date]
assert monthly_start_date.strftime(
"%Y-%m-%d") == row_monthly[start_date]
assert quarterly_start_date.strftime(
"%Y-%m-%d") == row_quarterly[start_date]
# The page views during the week, month, quarter
spam_views_week = page_views[spam_rev_3_days_ago.id]
spam_views_month = spam_views_week + page_views[
spam_rev_10_days_ago.id]
spam_views_month_exclude_week = page_views[spam_rev_10_days_ago.id]
spam_views_quarter = spam_views_month + page_views[
spam_rev_35_days_ago.id]
spam_views_quarter_exclude_month = page_views[spam_rev_35_days_ago.id]
# The percentage change in spam viewers
weekly_spam_change_percent = "{:.1%}".format(
float(spam_views_week - spam_views_month_exclude_week) /
spam_views_month_exclude_week)
monthly_spam_change_percent = "{:.1%}".format(
float(spam_views_month - spam_views_quarter_exclude_month) /
spam_views_quarter_exclude_month)
assert "0.0%" == row_daily[spam_viewers_change_percent]
assert weekly_spam_change_percent == row_weekly[
spam_viewers_change_percent]
assert monthly_spam_change_percent == row_monthly[
spam_viewers_change_percent]
assert "0.0%" == row_quarterly[spam_viewers_change_percent]
# The spam viewers
assert 0 == int(row_daily[spam_viewers])
assert spam_views_week == int(row_weekly[spam_viewers])
assert spam_views_month == int(row_monthly[spam_viewers])
assert spam_views_quarter == int(row_quarterly[spam_viewers])
# The daily average of spam viewers
assert float(row_daily[daily_average_viewers]) == 0.0
assert row_weekly[daily_average_viewers] == "{:.1f}".format(
float(spam_views_week) / days_in_week)
assert row_monthly[daily_average_viewers] == "{:.1f}".format(
float(spam_views_month) / days_in_month)
assert row_quarterly[daily_average_viewers] == "{:.1f}".format(
float(spam_views_quarter) / days_in_quarter)
# The published spam: 1 this week, 2 this month, 3 this quarter
assert not int(row_daily[published_spam])
assert len(spam_weekly) == int(row_weekly[published_spam])
assert len(spam_monthly) == int(row_monthly[published_spam])
assert len(spam_quarterly) == int(row_quarterly[published_spam])
# The blocked spam: there were 2 correctly blocked spam attempts 3 days ago
assert 0 == int(row_daily[blocked_spam])
assert true_blocked_spam_num == int(row_weekly[blocked_spam])
assert true_blocked_spam_num == int(row_monthly[blocked_spam])
assert true_blocked_spam_num == int(row_quarterly[blocked_spam])
# The blocked ham: there was 1 incorrectly blocked spam attempt 3 days ago
assert 0 == int(row_daily[blocked_ham])
assert false_blocked_spam_num == int(row_weekly[blocked_ham])
assert false_blocked_spam_num == int(row_monthly[blocked_ham])
assert false_blocked_spam_num == int(row_quarterly[blocked_ham])
# The true positive rate == blocked_spam / total spam
tpr_weekly = "{:.1%}".format(
true_blocked_spam_num /
float(true_blocked_spam_num + len(spam_weekly)))
tpr_monthly = "{:.1%}".format(
true_blocked_spam_num /
float(true_blocked_spam_num + len(spam_monthly)))
tpr_quarterly = "{:.1%}".format(
true_blocked_spam_num /
float(true_blocked_spam_num + len(spam_quarterly)))
assert "100.0%" == row_daily[true_positive_rate]
assert tpr_weekly == row_weekly[true_positive_rate]
assert tpr_monthly == row_monthly[true_positive_rate]
assert tpr_quarterly == row_quarterly[true_positive_rate]
# The true negative rate == published ham / total ham
tnr_weekly = "{:.1%}".format(
len(ham_weekly) / float(false_blocked_spam_num + len(ham_weekly)))
tnr_monthly = "{:.1%}".format(
len(ham_monthly) /
float(false_blocked_spam_num + len(ham_monthly)))
tnr_quarterly = "{:.1%}".format(
len(ham_quarterly) /
float(false_blocked_spam_num + len(ham_quarterly)))
assert "100.0%" == row_daily[true_negative_rate]
assert tnr_weekly == row_weekly[true_negative_rate]
assert tnr_monthly == row_monthly[true_negative_rate]
assert tnr_quarterly == row_quarterly[true_negative_rate]
