mirror of
https://gh.wpcy.net/https://github.com/discourse/discourse.git
synced 2026-06-19 03:23:50 +08:00
628520e541
Each KPI requested its report with a `prev_period` facet, which runs a second scan over the previous period just to produce the comparison value. This PR now requests the report once over the combined [prev_start, end] window and split the daily series by date: current is [start, end], previous is the equal-length window before it, so we make one scan instead of two.
56 lines
1.7 KiB
Ruby
Vendored
56 lines
1.7 KiB
Ruby
Vendored
# frozen_string_literal: true
|
|
|
|
module Reports::DauByMau
|
|
extend ActiveSupport::Concern
|
|
|
|
class_methods do
|
|
def report_dau_by_mau(report)
|
|
report.labels = [
|
|
{ type: :date, property: :x, title: I18n.t("reports.default.labels.day") },
|
|
{ type: :percent, property: :y, title: I18n.t("reports.default.labels.percent") },
|
|
]
|
|
|
|
report.average = true
|
|
report.percent = true
|
|
|
|
data_start = report.facets.include?(:prev_period) ? report.prev_start_date : report.start_date
|
|
data_points = UserVisit.count_by_active_users(data_start, report.end_date)
|
|
|
|
report.data = []
|
|
|
|
compute_dau_by_mau =
|
|
Proc.new do |data_point|
|
|
if data_point["mau"] == 0
|
|
0
|
|
else
|
|
((data_point["dau"].to_f / data_point["mau"].to_f) * 100).ceil(2)
|
|
end
|
|
end
|
|
|
|
dau_avg =
|
|
Proc.new do |points|
|
|
if !points.empty?
|
|
sum = points.sum { |data_point| compute_dau_by_mau.call(data_point) }
|
|
(sum.to_f / points.count.to_f).ceil(2)
|
|
end
|
|
end
|
|
|
|
if report.facets.include?(:prev_period)
|
|
current_start = report.start_date.to_date
|
|
data_points, prev_data_points =
|
|
data_points.partition { |data_point| data_point["date"].to_date >= current_start }
|
|
report.prev_period = dau_avg.call(prev_data_points)
|
|
end
|
|
|
|
data_points.each do |data_point|
|
|
report.data << { x: data_point["date"], y: compute_dau_by_mau.call(data_point) }
|
|
end
|
|
|
|
if report.facets.include?(:prev30Days)
|
|
prev30_days_data =
|
|
UserVisit.count_by_active_users(report.start_date - 30.days, report.start_date)
|
|
report.prev30Days = dau_avg.call(prev30_days_data)
|
|
end
|
|
end
|
|
end
|
|
end
|
