def make_tables(self):
# overall promoted link traffic
impressions = traffic.AdImpressionsByCodename.historical_totals("day")
clicks = traffic.ClickthroughsByCodename.historical_totals("day")
data = traffic.zip_timeseries(impressions, clicks)
columns = [
dict(color=COLORS.UPVOTE_ORANGE,
title=_("total impressions by day"),
shortname=_("impressions")),
dict(color=COLORS.DOWNVOTE_BLUE,
title=_("total clicks by day"),
shortname=_("clicks")),
]
self.totals = TimeSeriesChart("traffic-ad-totals",
_("ad totals"),
"day",
columns,
data,
self.traffic_last_modified,
classes=["traffic-table"])
# get summary of top ads
advert_summary = traffic.AdImpressionsByCodename.top_last_month()
things = AdvertTrafficSummary.get_things(ad for ad, data
in advert_summary)
self.advert_summary = []
for id, data in advert_summary:
name = AdvertTrafficSummary.get_ad_name(id, things=things)
url = AdvertTrafficSummary.get_ad_url(id, things=things)
self.advert_summary.append(((name, url), data))
def traffic_totals():
from r2.models import traffic
impressions = traffic.AdImpressionsByCodename.historical_totals("day")
clicks = traffic.ClickthroughsByCodename.historical_totals("day")
traffic_data = traffic.zip_timeseries(impressions, clicks)
return [(d.date(), v) for d, v in traffic_data]
def make_tables(self):
# overall promoted link traffic
impressions = traffic.AdImpressionsByCodename.historical_totals("day")
clicks = traffic.ClickthroughsByCodename.historical_totals("day")
data = traffic.zip_timeseries(impressions, clicks)
columns = [
dict(color=COLORS.UPVOTE_ORANGE,
title=_("total impressions by day"),
shortname=_("impressions")),
dict(color=COLORS.DOWNVOTE_BLUE,
title=_("total clicks by day"),
shortname=_("clicks")),
]
self.totals = TimeSeriesChart("traffic-ad-totals",
_("ad totals"),
"day",
columns,
data,
self.traffic_last_modified,
classes=["traffic-table"])
# get summary of top ads
advert_summary = traffic.AdImpressionsByCodename.top_last_month()
things = AdvertTrafficSummary.get_things(
ad for ad, data in advert_summary)
self.advert_summary = []
for id, data in advert_summary:
name = AdvertTrafficSummary.get_ad_name(id, things=things)
url = AdvertTrafficSummary.get_ad_url(id, things=things)
self.advert_summary.append(((name, url), data))
def get_promo_traffic(thing, start, end):
"""Get traffic for a Promoted Link or PromoCampaign"""
use_adserver_reporting = _use_adserver_reporting(thing)
spent_fn = None
if isinstance(thing, Link):
if use_adserver_reporting:
imp_fn = traffic.AdserverImpressionsByCodename.promotion_history
click_fn = traffic.AdserverClickthroughsByCodename.promotion_history
spent_fn = traffic.AdserverSpentPenniesByCodename.promotion_history
else:
imp_fn = traffic.AdImpressionsByCodename.promotion_history
click_fn = traffic.ClickthroughsByCodename.promotion_history
elif isinstance(thing, PromoCampaign):
if use_adserver_reporting:
imp_fn = traffic.AdserverTargetedImpressionsByCodename.promotion_history
click_fn = traffic.AdserverTargetedClickthroughsByCodename.promotion_history
spent_fn = traffic.AdserverTargetedSpentPenniesByCodename.promotion_history
else:
imp_fn = traffic.TargetedImpressionsByCodename.promotion_history
click_fn = traffic.TargetedClickthroughsByCodename.promotion_history
start = start.replace(tzinfo=None)
end = end.replace(tzinfo=None)
# adserver is daily aggregate.
if use_adserver_reporting:
start = start.replace(hour=0)
end = end.replace(hour=0)
imps = imp_fn(thing._fullname, start, end)
clicks = click_fn(thing._fullname, start, end)
spent_pennies = spent_fn and spent_fn(thing._fullname, start, end)
if imps and not clicks:
clicks = [(imps[0][0], (0, ))]
if clicks and not imps:
imps = [(clicks[0][0], (0, ))]
if imps and not spent_pennies:
spent_pennies = [(imps[0][0], (0, ))]
elif clicks and not spent_pennies:
spent_pennies = [(clicks[0][0], (0, ))]
args = [
imps,
clicks,
]
if use_adserver_reporting:
spent = [(date, (values[0] / 100., ))
for date, values in spent_pennies]
args.append(spent)
history = traffic.zip_timeseries(*args, order="ascending")
return history
def get_data_for_interval(self, interval, columns):
columns[1]["title"] = _("impressions by %s" % interval)
columns[1]["shortname"] = _("impressions")
columns += [
dict(shortname=_("unique clicks")),
dict(color=COLORS.MISCELLANEOUS,
title=_("clicks by %s" % interval),
shortname=_("total clicks")),
]
imps = traffic.AdImpressionsByCodename.history(interval, self.code)
clicks = traffic.ClickthroughsByCodename.history(interval, self.code)
return traffic.zip_timeseries(imps, clicks)
def make_tables(self):
start, end = promote.get_total_run(self.thing)
if not start or not end:
self.history = []
return
cutoff = end - datetime.timedelta(days=31)
start = max(start, cutoff)
fullname = self.thing._fullname
imps = traffic.AdImpressionsByCodename.promotion_history(fullname,
start, end)
clicks = traffic.ClickthroughsByCodename.promotion_history(fullname,
start, end)
# promotion might have no clicks, zip_timeseries needs valid columns
if imps and not clicks:
clicks = [(imps[0][0], (0, 0))]
history = traffic.zip_timeseries(imps, clicks, order="ascending")
computed_history = []
self.total_impressions, self.total_clicks = 0, 0
for date, data in history:
u_imps, imps, u_clicks, clicks = data
u_ctr = _clickthrough_rate(u_imps, u_clicks)
ctr = _clickthrough_rate(imps, clicks)
self.total_impressions += imps
self.total_clicks += clicks
computed_history.append((date, data + (u_ctr, ctr)))
self.history = computed_history
if self.total_impressions > 0:
self.total_ctr = _clickthrough_rate(self.total_impressions,
self.total_clicks)
# XXX: _is_promo_preliminary correctly expects tz-aware datetimes
# because it's also used with datetimes from promo code. this hack
# relies on the fact that we're storing UTC w/o timezone info.
# TODO: remove this when traffic is correctly using timezones.
end_aware = end.replace(tzinfo=g.tz)
self.is_preliminary = _is_promo_preliminary(end_aware)
# we should only graph a sane number of data points (not everything)
self.max_points = traffic.points_for_interval("hour")
return computed_history
def get_data_for_interval(self, interval, columns):
pageviews = traffic.PageviewsBySubreddit.history(interval, c.site.name)
if interval == "day":
columns.append(dict(color=COLORS.MISCELLANEOUS,
title=_("subscriptions by day"),
shortname=_("subscriptions")))
sr_name = c.site.name
subscriptions = traffic.SubscriptionsBySubreddit.history(interval,
sr_name)
return traffic.zip_timeseries(pageviews, subscriptions)
else:
return pageviews
def get_tables(self):
start, end = promote.get_total_run(self.thing)
if not start or not end:
self.history = []
return
cutoff = end - datetime.timedelta(days=31)
start = max(start, cutoff)
fullname = self.thing._fullname
imps = traffic.AdImpressionsByCodename.promotion_history(
fullname, start, end)
clicks = traffic.ClickthroughsByCodename.promotion_history(
fullname, start, end)
# promotion might have no clicks, zip_timeseries needs valid columns
if imps and not clicks:
clicks = [(imps[0][0], (0, 0))]
history = traffic.zip_timeseries(imps, clicks, order="ascending")
computed_history = []
self.total_impressions, self.total_clicks = 0, 0
for date, data in history:
u_imps, imps, u_clicks, clicks = data
u_ctr = self.calculate_clickthrough_rate(u_imps, u_clicks)
ctr = self.calculate_clickthrough_rate(imps, clicks)
self.total_impressions += imps
self.total_clicks += clicks
computed_history.append((date, data + (u_ctr, ctr)))
self.history = computed_history
if self.total_impressions > 0:
self.total_ctr = (
(float(self.total_clicks) / self.total_impressions) * 100.)
# the results are preliminary until 1 day after the promotion ends
now = datetime.datetime.utcnow()
self.is_preliminary = end + datetime.timedelta(days=1) > now
# we should only graph a sane number of data points (not everything)
self.max_points = traffic.points_for_interval("hour")
return computed_history
def get_promo_traffic(thing, start, end):
"""Get traffic for a Promoted Link or PromoCampaign"""
if isinstance(thing, Link):
imp_fn = traffic.AdImpressionsByCodename.promotion_history
click_fn = traffic.ClickthroughsByCodename.promotion_history
elif isinstance(thing, PromoCampaign):
imp_fn = traffic.TargetedImpressionsByCodename.promotion_history
click_fn = traffic.TargetedClickthroughsByCodename.promotion_history
imps = imp_fn(thing._fullname, start.replace(tzinfo=None),
end.replace(tzinfo=None))
clicks = click_fn(thing._fullname, start.replace(tzinfo=None),
end.replace(tzinfo=None))
if imps and not clicks:
clicks = [(imps[0][0], (0, ))]
history = traffic.zip_timeseries(imps, clicks, order="ascending")
return history
def traffic_totals():
from r2.models import traffic
impressions = traffic.AdImpressionsByCodename.historical_totals("day")
clicks = traffic.ClickthroughsByCodename.historical_totals("day")
traffic_data = traffic.zip_timeseries(impressions, clicks)
return [(d.date(), v) for d, v in traffic_data]
def make_tables(self):
now = datetime.datetime.utcnow().replace(minute=0, second=0,
microsecond=0)
promo_start, promo_end = promote.get_total_run(self.thing)
promo_end = min(now, promo_end)
if not promo_start or not promo_end:
self.history = []
return
if self.period:
start = self.after
end = self.before
if not start and not end:
end = promo_end
start = end - self.period
elif not end:
end = start + self.period
elif not start:
start = end - self.period
if start > promo_start:
p = request.get.copy()
p.update({'after':None, 'before':start.strftime('%Y%m%d%H')})
self.prev = '%s?%s' % (request.path, urllib.urlencode(p))
else:
start = promo_start
if end < promo_end:
p = request.get.copy()
p.update({'after':end.strftime('%Y%m%d%H'), 'before':None})
self.next = '%s?%s' % (request.path, urllib.urlencode(p))
else:
end = promo_end
else:
start, end = promo_start, promo_end
fullname = self.thing._fullname
imps = traffic.AdImpressionsByCodename.promotion_history(fullname,
start, end)
clicks = traffic.ClickthroughsByCodename.promotion_history(fullname,
start, end)
# promotion might have no clicks, zip_timeseries needs valid columns
if imps and not clicks:
clicks = [(imps[0][0], (0, 0))]
history = traffic.zip_timeseries(imps, clicks, order="ascending")
computed_history = []
self.total_impressions, self.total_clicks = 0, 0
for date, data in history:
u_imps, imps, u_clicks, clicks = data
u_ctr = _clickthrough_rate(u_imps, u_clicks)
ctr = _clickthrough_rate(imps, clicks)
self.total_impressions += imps
self.total_clicks += clicks
computed_history.append((date, data + (u_ctr, ctr)))
self.history = computed_history
if self.total_impressions > 0:
self.total_ctr = _clickthrough_rate(self.total_impressions,
self.total_clicks)
# XXX: _is_promo_preliminary correctly expects tz-aware datetimes
# because it's also used with datetimes from promo code. this hack
# relies on the fact that we're storing UTC w/o timezone info.
# TODO: remove this when traffic is correctly using timezones.
end_aware = end.replace(tzinfo=g.tz)
self.is_preliminary = _is_promo_preliminary(end_aware)
# we should only graph a sane number of data points (not everything)
self.max_points = traffic.points_for_interval("hour")
return computed_history
