def __init__(self,
data_mgr: DataManager = DataManager(),
calendars: Union[str, Tuple[str, ...]] = 'Weekend',
tz: timezone = timezone('America/New_York'),
valuation_method: ValuationMethod = ValuationMethod(ValuationFixingType.PRICE),
action_impl_map=None):
if action_impl_map is None:
action_impl_map = {Action: SubmitOrderActionImpl}
self.action_impl_map = action_impl_map
self.calendars = calendars
self.tz = tz
self.data_handler = DataHandler(data_mgr, tz=tz)
self.valuation_method = valuation_method
self.execution_engine = None
def test_backtest_predefined_timezone_aware():
tz = 'Europe/London'
start_dt = '2021-01-01T08:00'
end_dt = '2021-12-31T17:00'
states = pd.bdate_range(start_dt, end_dt,
freq='1H', tz=tz).to_series().between_time(
'08:00', '17:00').index.tolist()
trigger_dates = pd.bdate_range(
start_dt, end_dt, freq='1H',
tz=tz).to_series().at_time('17:00').index.tolist()
data = np.random.randn(len(states))
s_rt = pd.Series(index=states, data=data)
s_eod = s_rt.at_time('17:00')
s_eod.index = s_eod.index.date
generic_bond_future = IRBondFuture(currency='EUR')
add_trade_action = AddTradeAction(generic_bond_future)
simple_date_trigger_requirement = DateTriggerRequirements(
dates=trigger_dates)
simple_date_trigger = DateTrigger(
trigger_requirements=simple_date_trigger_requirement,
actions=[add_trade_action])
data_manager = DataManager()
data_manager.add_data_source(pd.Series(index=states, data=data),
DataFrequency.REAL_TIME, generic_bond_future,
ValuationFixingType.PRICE)
data_manager.add_data_source(s_eod, DataFrequency.DAILY,
generic_bond_future,
ValuationFixingType.PRICE)
# instantiate a new strategy
strategy = Strategy(None, triggers=simple_date_trigger)
engine = PredefinedAssetEngine(data_mgr=data_manager,
tz=timezone(tz),
calendars='Weekend')
backtest = engine.run_backtest(strategy=strategy,
start=states[0],
end=states[-1],
states=states)
assert len(backtest.trade_ledger()) == 364
def test_backtest_predefined():
# Test simple MOC order
trigger = ExampleTestTrigger()
strategy = Strategy(initial_portfolio=None, triggers=[trigger])
start = dt.date(2021, 1, 4)
mid = dt.date(2021, 1, 5)
end = dt.date(2021, 1, 6)
# these mocks are needed as the date functions need a GSSession
gs_quant.backtests.predefined_asset_engine.is_business_day = mock.Mock(
return_value=True)
gs_quant.backtests.predefined_asset_engine.business_day_offset = mock.Mock(
return_value=mid)
data_mgr = DataManager()
underlying = Security(ric='TestRic')
close_prices = pd.Series(dtype=float)
close_prices[start] = 1
close_prices[mid] = 1.5
close_prices[end] = 2
data_mgr.add_data_source(close_prices, DataFrequency.DAILY, underlying,
ValuationFixingType.PRICE)
engine = PredefinedAssetEngine(data_mgr=data_mgr)
backtest = engine.run_backtest(strategy, start=start, end=end)
perf = backtest.performance
holdings = backtest.historical_holdings
cash_asset = backtest.cash_asset
# 100 on the initial date
assert perf[start] == 100
# 100 on the next day as we traded MOC
assert perf[mid] == 100
assert holdings[mid][cash_asset] == 100 - 1.5
assert holdings[mid][underlying] == 1
# 100.5 = 98.5 (cash) + 2 ( test asset)
assert holdings[end][cash_asset] == 100 - 1.5
assert holdings[end][underlying] == 1
assert perf[end] == 100.5
# Test TWAP orders with no ON positions
twap_entry_mid = 16
twap_exit_mid = 25
twap_entry_end = 30
twap_exit_end = 40
data_twap = {
dt.datetime.combine(mid, dt.time(10, 30)): twap_entry_mid,
dt.datetime.combine(mid, dt.time(14, 30)): twap_exit_mid,
dt.datetime.combine(end, dt.time(10, 30)): twap_entry_end,
dt.datetime.combine(end, dt.time(14, 30)): twap_exit_end
}
data_mgr.add_data_source(pd.Series(data_twap), DataFrequency.REAL_TIME,
underlying, ValuationFixingType.PRICE)
trigger = FuturesExample()
strategy = Strategy(initial_portfolio=None, triggers=[trigger])
engine = PredefinedAssetEngine(data_mgr=data_mgr,
tz=timezone('Europe/London'))
backtest = engine.run_backtest(strategy, start=start, end=end)
perf = backtest.performance
holdings = backtest.historical_holdings
weights = backtest.historical_weights
cash_asset = backtest.cash_asset
# start: 100
assert perf[start] == 100
# mid: 100 + (twap_exit - twap_entry)
assert perf[mid] == 100 - twap_entry_mid + twap_exit_mid
assert holdings[mid][cash_asset] == perf[mid]
assert underlying not in holdings[mid]
assert weights[mid][cash_asset] == 1
assert underlying not in weights[mid]
# 100.5 = 98.5 (cash) + 2 ( test asset)
assert perf[
end] == 100 - twap_entry_mid + twap_exit_mid - twap_entry_end + twap_exit_end
assert holdings[end][cash_asset] == perf[end]
assert underlying not in holdings[end]
assert weights[end][cash_asset] == 1
assert underlying not in weights[end]