def test_percent_build_with_masks(self):
for n_portfolio in self.n_portfolios:
x = np.random.randn(self.n_samples, n_portfolio)
choices = np.random.choice(self.n_samples,
self.n_mask,
replace=False)
masks = np.full(self.n_samples, True, dtype=bool)
masks[choices] = False
calc_weights = percent_build(x, self.p_included, masks=masks)
expected_weights = np.zeros((len(x), n_portfolio))
filtered_index = np.arange(len(x))[masks]
filtered_x = x[masks]
big_boolen = np.full(x.shape, False, dtype=bool)
n_included = int(self.p_included * len(x))
chosen = (-filtered_x).argsort(axis=0).argsort(axis=0) < n_included
big_boolen[filtered_index] = chosen
for j in range(x.shape[1]):
expected_weights[big_boolen[:, j], j] = 1.
np.testing.assert_array_almost_equal(calc_weights,
expected_weights)
def benchmark_build_percent_with_group(n_samples: int, n_loops: int, p_included: float, n_groups: int) -> None:
print("-" * 60)
print("Starting portfolio construction by percent with group-by values benchmarking")
print("Parameters(n_samples: {0}, p_included: {1}, n_loops: {2}, n_groups: {3})".format(n_samples, p_included, n_loops, n_groups))
n_portfolio = 10
x = np.random.randn(n_samples, n_portfolio)
groups = np.random.randint(n_groups, size=n_samples)
start = dt.datetime.now()
for _ in range(n_loops):
calc_weights = percent_build(x, p_included, groups=groups)
impl_model_time = dt.datetime.now() - start
print('{0:20s}: {1}'.format('Implemented model', impl_model_time))
start = dt.datetime.now()
for _ in range(n_loops):
grouped_ordering = pd.DataFrame(-x).groupby(groups).rank()
grouped_count = pd.DataFrame(-x).groupby(groups).transform(lambda x: x.count())
exp_weights = np.zeros((len(x), n_portfolio))
n_included = (grouped_count * p_included).astype(int)
masks = (grouped_ordering <= n_included).values
for j in range(n_portfolio):
exp_weights[masks[:, j], j] = 1.
benchmark_model_time = dt.datetime.now() - start
np.testing.assert_array_almost_equal(calc_weights, exp_weights)
print('{0:20s}: {1}'.format('Benchmark model', benchmark_model_time))
def benchmark_build_percent(n_samples: int, n_loops: int, p_included: float) -> None:
print("-" * 60)
print("Starting portfolio construction by percent benchmarking")
print("Parameters(n_samples: {0}, p_included: {1}, n_loops: {2})".format(n_samples, p_included, n_loops))
n_portfolio = 10
x = np.random.randn(n_samples, n_portfolio)
start = dt.datetime.now()
for _ in range(n_loops):
calc_weights = percent_build(x, p_included)
impl_model_time = dt.datetime.now() - start
print('{0:20s}: {1}'.format('Implemented model', impl_model_time))
start = dt.datetime.now()
for _ in range(n_loops):
exp_weights = np.zeros((len(x), n_portfolio))
n_incuded = int(p_included * len(x))
choosed_index = (-x).argsort(axis=0).argsort(axis=0) < n_incuded
for j in range(n_portfolio):
exp_weights[choosed_index[:, j], j] = 1.
benchmark_model_time = dt.datetime.now() - start
np.testing.assert_array_almost_equal(calc_weights, exp_weights)
print('{0:20s}: {1}'.format('Benchmark model', benchmark_model_time))
def test_percent_build(self):
n_include = int(self.n_samples * self.p_included)
for n_portfolio in self.n_portfolios:
x = np.random.randn(self.n_samples, n_portfolio)
calc_weights = percent_build(x, self.p_included)
expected_weights = np.zeros((len(x), n_portfolio))
masks = (-x).argsort(axis=0).argsort(axis=0) < n_include
for j in range(x.shape[1]):
expected_weights[masks[:, j], j] = 1.
np.testing.assert_array_almost_equal(calc_weights, expected_weights)
def test_percent_build_with_group(self):
for n_portfolio in self.n_portfolios:
x = np.random.randn(self.n_samples, n_portfolio)
groups = np.random.randint(self.n_groups, size=self.n_samples)
calc_weights = percent_build(x, self.p_included, groups)
grouped_ordering = pd.DataFrame(-x).groupby(groups).rank()
grouped_count = pd.DataFrame(-x).groupby(groups).transform(lambda x: x.count())
expected_weights = np.zeros((len(x), n_portfolio))
n_include = (grouped_count * self.p_included).astype(int)
masks = (grouped_ordering <= n_include).values
for j in range(x.shape[1]):
expected_weights[masks[:, j], j] = 1.
np.testing.assert_array_almost_equal(calc_weights, expected_weights)
def build_portfolio(er: np.ndarray,
builder: Optional[str] = 'long_short',
**kwargs) -> np.ndarray:
builder = builder.lower()
if builder == 'ls' or builder == 'long_short':
return long_short_build(er, **kwargs).flatten()
elif builder == 'rank':
return rank_build(er, **kwargs).flatten()
elif builder == 'percent':
return percent_build(er, **kwargs).flatten()
elif builder == 'linear_prog' or builder == 'linear':
status, _, weight = linear_build(er, **kwargs)
if status != 'optimal':
raise ValueError(
'linear programming optimizer in status: {0}'.format(status))
else:
return weight