def test_repr(self):
m = BSLookbackOption()
assert repr(m) == "BSLookbackOption()"
liability = LookbackOption(BrownianStock(), strike=1.1)
m = BSLookbackOption(liability)
assert repr(m) == "BSLookbackOption(strike=1.1)"
def test_implied_volatility(self):
x = torch.tensor(
[[0.0, 0.0, 0.1, 0.01], [0.0, 0.0, 0.1, 0.02], [0.0, 0.0, 0.1, 0.03]]
)
m = BSLookbackOption()
iv = m.implied_volatility(x[:, 0], x[:, 1], x[:, 2], x[:, 3])
result = BSLookbackOption().price(x[:, 0], x[:, 1], x[:, 2], iv)
expect = x[:, -1]
assert torch.allclose(result, expect, atol=1e-4)
def test_features(self):
m = BSLookbackOption()
assert m.features() == [
"log_moneyness",
"max_log_moneyness",
"expiry_time",
"volatility",
]
_ = [get_feature(f) for f in m.features()]
def test_example(self):
from pfhedge import Hedger
from pfhedge.instruments import BrownianStock
from pfhedge.instruments import LookbackOption
deriv = LookbackOption(BrownianStock(), strike=1.03)
model = BSLookbackOption(deriv)
hedger = Hedger(model, model.features())
price = hedger.price(deriv)
assert torch.allclose(price, torch.tensor(0.0177), atol=1e-4)
def test_shape(self):
torch.distributions.Distribution.set_default_validate_args(False)
m = BSLookbackOption()
self.assert_shape_delta(m)
self.assert_shape_gamma(m)
self.assert_shape_price(m)
self.assert_shape_forward(m)
def test_put_notimplemented(self):
with pytest.raises(ValueError):
# not yet supported
BSLookbackOption(call=False)
def test_gamma(self):
m = BSLookbackOption()
s = np.log(1.00 / 1.03)
result = m.gamma(s, s, 1.0, 0.2).item()
assert np.isclose(result, 4.466, atol=0.8)
def test_delta(self):
m = BSLookbackOption()
s = np.log(1.00 / 1.03)
result = m.delta(s, s, 1.0, 0.2).item()
assert np.isclose(result, 1.037, atol=1e-2)
def test_price(self):
m = BSLookbackOption(strike=1.03)
s = np.log(1.00 / 1.03)
result = m.price(s, s, 1.0, 0.2).item()
assert np.isclose(result, 0.14, atol=1e-2)
def test_forward(self):
m = BSLookbackOption()
s = np.log(1.00 / 1.03)
x = torch.tensor([s, s, 1.0, 0.2]).reshape(1, -1)
result = m(x).item()
assert np.isclose(result, 1.037, atol=1e-2)