def test_pamr_agg_error(self):
"""
Tests ValueError if aggressiveness is less than 0.
"""
# Initialize PAMR.
pamr4 = PAMR(optimization_method=2, epsilon=0.5, agg=-5)
with self.assertRaises(ValueError):
# Running allocate will raise ValueError.
pamr4.allocate(self.data)
def test_pamr_epsilon_error(self):
"""
Tests ValueError if epsilon is less than 0.
"""
# Initialize PAMR.
pamr3 = PAMR(optimization_method=2, epsilon=-1, agg=10)
with self.assertRaises(ValueError):
# Running allocate will raise ValueError.
pamr3.allocate(self.data)
def test_pamr_method_error(self):
"""
Tests ValueError if optimization method is not 0, 1, or 2.
"""
# Initialize PAMR.
pamr5 = PAMR(optimization_method=5, epsilon=0.5, agg=10)
with self.assertRaises(ValueError):
# Running allocate will raise ValueError.
pamr5.allocate(self.data)
def test_pamr2_solution(self):
"""
Test the calculation of passive aggressive mean reversion with the PAMR-2 optimization method
"""
# Initialize PAMR-2.
pamr2 = PAMR(optimization_method=2, epsilon=0.5, agg=10)
# Allocates asset prices to PAMR.
pamr2.allocate(self.data, resample_by='M')
# Create np.array of all_weights.
all_weights = np.array(pamr2.all_weights)
# Check if all weights sum to 1.
for i in range(all_weights.shape[0]):
weights = all_weights[i]
assert (weights >= 0).all()
assert len(weights) == self.data.shape[1]
np.testing.assert_almost_equal(np.sum(weights), 1)