def test_portfolio(monkeypatch):
monkeypatch.setattr(config, "TURNOVER_PERIOD", 7)
monkeypatch.setattr(config, "TURNOVER_CUT_OFF", 0.6)
port = portfolio.Portfolio(**PARAMS)
assert "Дата - 2018-03-19" in str(port)
assert port.date == pd.Timestamp("2018-03-19")
assert port.index.tolist() == ["GAZP", "TTLK", "VSMO", CASH, PORTFOLIO]
assert np.allclose(port.shares, [6820, 1_230_000, 145, 1000, 1])
assert np.allclose(port.lot_size, [10, 10000, 1, 1, 1])
assert np.allclose(port.lots, [682, 123, 145, 1000, 1])
assert np.allclose(port.price, [139.91, 0.1525, 17630, 1, 3_699_111])
assert np.allclose(port.value,
[954_186, 187_575, 2_556_350, 1000, 3_699_111])
assert np.allclose(
port.weight,
[
0.257_950_085_844_95,
0.050_708_129_601_95,
0.691_071_449_329_312,
0.000_270_335_223_788,
1,
],
)
assert np.allclose(port.turnover_factor,
[1, 0, 0.693_130_358_498_49, 1, 1])
def test_std_gradient():
pos = dict(PRTK=100, RTKM=200, SIBN=300)
port = portfolio.Portfolio("2018-12-17", 1000, pos)
metrics3 = metrics.Metrics(port, months=3)
metrics12 = metrics.Metrics(port, months=12)
assert metrics12.std_gradient == pytest.approx(metrics12.std[PORTFOLIO])
assert metrics3.std_gradient == pytest.approx(metrics12.std[PORTFOLIO] / 2)
assert metrics3.std_gradient == pytest.approx(metrics3.std[PORTFOLIO] / 2)
def test_find_better_model_fake_base(monkeypatch, capsys):
monkeypatch.setattr(cv, "MAX_SEARCHES", 10)
monkeypatch.setattr(cv, "print_result", lambda x, y, z: 1
if x == "Базовая модель" else 0)
pos = dict(LSNGP=10, KZOS=20, GMKN=30)
port = portfolio.Portfolio(pd.Timestamp("2018-12-19"), 100, pos)
cv.find_better_model(port)
captured = capsys.readouterr()
assert "ЛУЧШАЯ МОДЕЛЬ - Базовая модель" in captured.out
def test_update_data_bad_investor_name():
port = portfolio.Portfolio(name=["test"],
date=DATE_NEW,
cash=CASH,
positions=POSITIONS)
with pytest.raises(POptimizerError) as error:
pdf.update_data("test", port.date, port.value[PORTFOLIO],
dict(WLMike1=1000), 1234)
assert str(error.value) == "Неверное имя инвестора - WLMike1"
def test_find_better_model(monkeypatch, capsys):
monkeypatch.setattr(cv, "MAX_SEARCHES", 10)
monkeypatch.setattr(cv, "MAX_DEPTH", 7)
pos = dict(LSNGP=10, KZOS=20, GMKN=30)
port = portfolio.Portfolio(pd.Timestamp("2018-12-19"), 100, pos)
cv.find_better_model(port, PARAMS)
captured = capsys.readouterr()
assert "Базовая модель" in captured.out
assert "Найденная модель" in captured.out
assert "ЛУЧШАЯ МОДЕЛЬ - Найденная модель" in captured.out
def test_forecast_func(monkeypatch):
monkeypatch.setattr(config, "ML_PARAMS", ML_PARAMS)
pos = dict(SIBN=300, PRTK=100, RTKM=200)
port = portfolio.Portfolio("2018-12-17", 1000, pos)
result = metrics.Metrics(port)
# noinspection PyProtectedMember
forecast = result._forecast_func()
assert isinstance(forecast, Forecast)
assert forecast.date == pd.Timestamp("2018-12-17")
assert forecast.tickers == ("PRTK", "RTKM", "SIBN")
assert forecast.shrinkage == pytest.approx(0.6972112123349591)
def test_update_data_bad_date():
port = portfolio.Portfolio(date=DATE_OLD, cash=CASH, positions=POSITIONS)
with pytest.raises(POptimizerError) as error:
pdf.update_data(
"test",
port.date,
port.value[PORTFOLIO],
dict(WLMike=1000, Igor=-2000),
1234,
)
assert str(error.value) == "В этом месяце данные уже вносились в отчет"
def test_report_new_month():
port = portfolio.Portfolio(date=DATE_NEW, cash=CASH, positions=POSITIONS)
pdf.report("test", port, dict(WLMike=10000, Igor=-5000), 4321)
df = pdf.read_data("test")
date = port.date
assert df.index[-1] == date
assert df.loc[date, "WLMike"] == pytest.approx(10000)
assert df.loc[date, "Igor"] == pytest.approx(-5000)
assert df.loc[date, "Value_WLMike"] == pytest.approx(387_550.829_708_377)
assert df.loc[date, "Value_Igor"] == pytest.approx(4972.170_292)
assert df.loc[date, "Value"] == pytest.approx(392_523)
assert df.loc[date, "Dividends"] == pytest.approx(4321)
def test_update_data():
port = portfolio.Portfolio(date=DATE_NEW, cash=CASH, positions=POSITIONS)
pdf.update_data("test", port.date, port.value[PORTFOLIO],
dict(WLMike=1000, Igor=-2000), 1234)
df = pdf.read_data("test")
assert df.shape == (241, 6)
assert df.index[-1] == pd.Timestamp("2018-05-07")
assert df.loc["2018-05-07", "WLMike"] == pytest.approx(1000)
assert df.loc["2018-05-07", "Igor"] == pytest.approx(-2000)
assert df.loc["2018-05-07",
"Value_WLMike"] == pytest.approx(384_396.431_074_62)
assert df.loc["2018-05-07",
"Value_Igor"] == pytest.approx(8126.568_925_380_33)
assert df.loc["2018-05-07", "Value"] == pytest.approx(392_523)
assert df.loc["2018-05-07", "Dividends"] == pytest.approx(1234)
def make_metrics():
"""Подготовка тестовых данных."""
positions = {"BSPB": 4890, "FESH": 1300, "KZOS": 5080}
port = portfolio.Portfolio(["test"], "2020-05-14", 84449, positions)
mean = pd.Series([0.09, 0.06, 0.07], index=list(positions))
cov = np.array([
[0.04, 0.005, 0.01],
[0.005, 0.0625, 0.00625],
[0.01, 0.00625, 0.0625],
], )
fake_forecast = SimpleNamespace()
fake_forecast.mean = mean
fake_forecast.cov = cov
fake_forecast.max_std = 0.11
# noinspection PyTypeChecker
yield metrics.MetricsSingle(port, fake_forecast)
def make_resample():
"""Данне для тестов."""
positions = {"BSPB": 4890, "FESH": 1300}
port = portfolio.Portfolio("test", "2020-05-14", 84449, positions)
mean1 = pd.Series([0.09, 0.06], index=list(positions))
cov1 = np.array([[0.04, 0.005], [0.005, 0.0625]])
mean2 = pd.Series([0.05, 0.09], index=list(positions))
cov12 = np.array([[0.0225, 0.0042], [0.0042, 0.0196]])
def fake_get_forecasts(*_):
yield from (
SimpleNamespace(
mean=mean1,
cov=cov1,
history_days=1,
cor=0.4,
shrinkage=0.3,
risk_aversion=1,
error_tolerance=0,
),
SimpleNamespace(
mean=mean2,
cov=cov12,
history_days=2,
cor=0.5,
shrinkage=0.2,
risk_aversion=1,
error_tolerance=0,
),
)
saved_get_forecast = metrics.evolve.get_forecasts
metrics.evolve.get_forecasts = fake_get_forecasts
yield metrics.MetricsResample(port)
metrics.evolve.get_forecasts = saved_get_forecast
def test_portfolio_wrong_date():
PARAMS["date"] = "2018-12-09"
with pytest.raises(POptimizerError) as error:
portfolio.Portfolio(**PARAMS)
assert "Для даты 2018-12-09 отсутствуют исторические котировки" == str(
error.value)
def test_portfolio_wrong_value():
with pytest.raises(POptimizerError) as error:
PARAMS["value"] = 123
portfolio.Portfolio(**PARAMS)
assert "Введенная стоимость портфеля 123" in str(error.value)
def make_portfolio():
return portfolio.Portfolio(**PARAMS)
MTSS=7490,
MVID=3260,
NMTP=0,
PHOR=0,
PMSBP=17290,
RSTIP=87000,
RTKM=0,
RTKMP=182_600,
SNGSP=23500,
TTLK=0,
UPRO=1_267_000,
VSMO=73,
)
CASH = 1_548_264
DATE = "2018-04-19"
TEST_PORTFOLIO = portfolio.Portfolio(date=DATE, cash=CASH, positions=POSITIONS)
def test_drop_small_positions():
df = pdf_lower.drop_small_positions(TEST_PORTFOLIO)
index = df.index
assert len(df) == pdf_lower.MAX_TABLE_ROWS + 2
assert index[-1] == PORTFOLIO
assert df[PORTFOLIO] == pytest.approx(38525478)
assert index[-2] == OTHER
assert df[OTHER] == pytest.approx(5822654)
assert index[0] == "RTKMP"
assert df.iloc[0] == pytest.approx(11_167_816)
assert index[-3] == "MVID"
assert df.iloc[-3] == pytest.approx(1_310_520)
