Introduce `active_visitors` to `imported_pages` and start populating it with `activeUsers` from GA4 imports (#4027)

* Import `activeUsers` into `imported_pages.active_visitors` for GA4

* Add test for active visitors

* Simplify assertion in active visitors test

* Improve assertion for active visitors further

lib/plausible/google/ga4/report_request.ex

@@ -60,7 +60,13 @@ defmodule Plausible.Google.GA4.ReportRequest do
         dataset: "imported_pages",
         dimensions: ["date", "hostName", "pagePath"],
         # NOTE: no exits as GA4 DATA API does not provide that metric
-        metrics: ["totalUsers", "screenPageViews", "sessions", "userEngagementDuration"]
+        metrics: [
+          "totalUsers",
+          "activeUsers",
+          "screenPageViews",
+          "sessions",
+          "userEngagementDuration"
+        ]
       },
       %__MODULE__{
         dataset: "imported_entry_pages",

lib/plausible/imported/google_analytics4.ex

@@ -124,6 +124,7 @@ defmodule Plausible.Imported.GoogleAnalytics4 do
       hostname: row.dimensions |> Map.fetch!("hostName") |> String.replace_prefix("www.", ""),
       page: row.dimensions |> Map.fetch!("pagePath") |> URI.parse() |> Map.get(:path),
       visitors: row.metrics |> Map.fetch!("totalUsers") |> parse_number(),
+      active_visitors: row.metrics |> Map.fetch!("activeUsers") |> parse_number(),
       visits: row.metrics |> Map.fetch!("sessions") |> parse_number(),
       pageviews: row.metrics |> Map.fetch!("screenPageViews") |> parse_number(),
       # NOTE: no exits metric in GA4 API currently

lib/plausible/imported/page.ex

@@ -11,6 +11,7 @@ defmodule Plausible.Imported.Page do
     field :page, :string
     field :visits, Ch, type: "UInt64"
     field :visitors, Ch, type: "UInt64"
+    field :active_visitors, Ch, type: "UInt64"
     field :pageviews, Ch, type: "UInt64"
     field :exits, Ch, type: "UInt64"
     field :time_on_page, Ch, type: "UInt64"

test/plausible/imported/google_analytics4_test.exs

@@ -4,6 +4,7 @@ defmodule Plausible.Imported.GoogleAnalytics4Test do
   import Mox
   import Ecto.Query, only: [from: 2]
 
+  alias Plausible.ClickhouseRepo
   alias Plausible.Repo
   alias Plausible.Imported.GoogleAnalytics4
 
@@ -131,9 +132,38 @@ defmodule Plausible.Imported.GoogleAnalytics4Test do
       assert_browsers(conn, breakdown_params)
       assert_os(conn, breakdown_params)
       assert_os_versions(conn, breakdown_params)
+      assert_active_visitors(site_import)
     end
   end
 
+  defp assert_active_visitors(site_import) do
+    result =
+      ClickhouseRepo.query!(
+        "SELECT date, sum(visitors) AS all_visitors, sum(active_visitors) AS all_active_visitors " <>
+          "FROM imported_pages WHERE site_id = #{site_import.site_id} AND import_id = #{site_import.id} GROUP BY date"
+      )
+      |> Map.fetch!(:rows)
+      |> Enum.map(fn [date, all_visitors, all_active_visitors] ->
+        %{date: date, visitors: all_visitors, active_visitors: all_active_visitors}
+      end)
+
+    assert length(result) == 31
+
+    Enum.each(result, fn row ->
+      assert row.visitors > 100 and row.active_visitors > 100
+      assert row.active_visitors <= row.visitors
+    end)
+
+    ClickhouseRepo.query!(
+      "SELECT time_on_page FROM imported_pages WHERE active_visitors = 0 AND " <>
+        "site_id = #{site_import.site_id} AND import_id = #{site_import.id}"
+    )
+    |> Map.fetch!(:rows)
+    |> Enum.each(fn [time_on_page] ->
+      assert time_on_page == 0
+    end)
+  end
+
   defp assert_timeseries(conn, params) do
     params =
       Map.put(