prop/test/microwaveprop/backtest_test.exs

defmodule Microwaveprop.BacktestTest do
  use Microwaveprop.DataCase, async: true

  alias Microwaveprop.Backtest
  alias Microwaveprop.Radio.Contact

  defp create_contact(attrs \\ %{}) do
    default = %{
      station1: "W5XD",
      station2: "K5TR",
      qso_timestamp: ~U[2026-03-28 18:00:00Z],
      mode: "CW",
      band: Decimal.new("10000"),
      grid1: "EM12",
      grid2: "EM00",
      pos1: %{"lat" => 32.9, "lon" => -97.0},
      pos2: %{"lat" => 30.3, "lon" => -97.7},
      distance_km: Decimal.new("295")
    }

    {:ok, contact} =
      %Contact{}
      |> Contact.changeset(Map.merge(default, attrs))
      |> Repo.insert()

    contact
  end

  describe "evaluate/2" do
    test "returns counts and a qso distribution for a trivial feature" do
      create_contact()

      feature = fn _lat, _lon, _time -> 1.0 end
      report = Backtest.evaluate(feature, sample_size: 10, baseline_size: 10)

      assert report.qso_count == 1
      assert report.baseline_count == 10
      assert %Backtest.Distribution{} = report.qso_distribution
      assert %Backtest.Distribution{} = report.baseline_distribution
      assert report.qso_distribution.count == 1
      assert report.qso_distribution.mean == 1.0
    end

    test "skips contacts with nil pos1" do
      create_contact(%{pos1: nil, pos2: nil})

      feature = fn _lat, _lon, _time -> 1.0 end
      report = Backtest.evaluate(feature, sample_size: 10, baseline_size: 0)

      assert report.qso_count == 0
    end

    test "excludes nil feature values from the distribution" do
      create_contact(%{station1: "A"})
      create_contact(%{station1: "B"})

      feature = fn _lat, _lon, _time -> nil end
      report = Backtest.evaluate(feature, sample_size: 10, baseline_size: 0)

      assert report.qso_count == 2
      assert report.qso_distribution.count == 0
    end
  end

  describe "lift_by_distance/2" do
    test "bins QSOs by distance_km and summarizes the feature per bin" do
      create_contact(%{station1: "A", distance_km: Decimal.new("50"), pos1: %{"lat" => 1.0, "lon" => 0.0}})
      create_contact(%{station1: "B", distance_km: Decimal.new("300"), pos1: %{"lat" => 5.0, "lon" => 0.0}})
      create_contact(%{station1: "C", distance_km: Decimal.new("800"), pos1: %{"lat" => 9.0, "lon" => 0.0}})

      f = fn lat, _lon, _time -> lat end
      bins = Backtest.lift_by_distance(f, sample_size: 10)

      assert bins["0-100"].count == 1
      assert bins["0-100"].mean == 1.0
      assert bins["100-250"].count == 0
      assert bins["250-500"].count == 1
      assert bins["250-500"].mean == 5.0
      assert bins["500-1000"].count == 1
      assert bins["500-1000"].mean == 9.0
      assert bins["1000+"].count == 0
    end

    test "drops nil feature values from the bin stats" do
      create_contact(%{station1: "A", distance_km: Decimal.new("50"), pos1: %{"lat" => 1.0, "lon" => 0.0}})
      create_contact(%{station1: "B", distance_km: Decimal.new("60"), pos1: %{"lat" => 2.0, "lon" => 0.0}})

      f = fn lat, _lon, _time -> if lat == 1.0, do: nil, else: 2.0 end
      bins = Backtest.lift_by_distance(f, sample_size: 10)

      assert bins["0-100"].count == 1
      assert bins["0-100"].mean == 2.0
    end
  end

  describe "random_baseline/2" do
    test "generates N samples drawn from the contact locations" do
      create_contact(%{
        station1: "A",
        qso_timestamp: ~U[2026-03-01 00:00:00Z],
        pos1: %{"lat" => 1.0, "lon" => 10.0}
      })

      create_contact(%{
        station1: "B",
        qso_timestamp: ~U[2026-06-15 12:00:00Z],
        pos1: %{"lat" => 2.0, "lon" => 20.0}
      })

      samples = Backtest.random_baseline(5, sample_size: 10)
      assert length(samples) == 5

      Enum.each(samples, fn {lat, lon, time} ->
        assert lat in [1.0, 2.0]
        assert lon in [10.0, 20.0]
        assert %DateTime{} = time
      end)
    end

    test "keeps timestamps within ±30 days of a source QSO" do
      create_contact(%{
        station1: "A",
        qso_timestamp: ~U[2026-03-01 00:00:00Z],
        pos1: %{"lat" => 1.0, "lon" => 10.0}
      })

      samples = Backtest.random_baseline(200, sample_size: 10)
      source = ~U[2026-03-01 00:00:00Z]

      Enum.each(samples, fn {_, _, time} ->
        delta_seconds = abs(DateTime.diff(time, source))
        assert delta_seconds <= 30 * 24 * 3600
      end)
    end

    test "returns an empty list when there are no contacts with a position" do
      assert Backtest.random_baseline(10) == []
    end
  end

  describe "lift_by_band/2" do
    test "groups feature values by band and summarizes per band" do
      create_contact(%{
        station1: "A",
        band: Decimal.new("10000"),
        pos1: %{"lat" => 1.0, "lon" => 0.0}
      })

      create_contact(%{
        station1: "B",
        band: Decimal.new("24000"),
        pos1: %{"lat" => 7.0, "lon" => 0.0}
      })

      create_contact(%{
        station1: "C",
        band: Decimal.new("24000"),
        pos1: %{"lat" => 9.0, "lon" => 0.0}
      })

      f = fn lat, _lon, _time -> lat end
      bands = Backtest.lift_by_band(f, sample_size: 10)

      assert bands[Decimal.new("10000")].count == 1
      assert bands[Decimal.new("10000")].mean == 1.0
      assert bands[Decimal.new("24000")].count == 2
      assert bands[Decimal.new("24000")].mean == 8.0
    end
  end
end