Skip to content

Commit

Permalink
FIX: lint error and refactor k_vars to dim
Browse files Browse the repository at this point in the history
  • Loading branch information
@tupui
tupui committed Jun 21, 2020
1 parent 5a3181f commit 9fb8360
Show file tree
Hide file tree
Showing 2 changed files with 22 additions and 29 deletions.
1 change: 0 additions & 1 deletion scipy/stats/__init__.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -377,7 +377,6 @@
qmc.NormalQMC
qmc.MultivariateNormalQMC
qmc.discrepancy
qmc.discrepancy_star_L2
qmc.scale
Expand Down
50 changes: 22 additions & 28 deletions scipy/stats/tests/test_qmc.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -104,7 +104,7 @@ def test_van_der_corput(self):

def test_halton(self):
# without bounds
engine = qmc.Halton(k_vars=2)
engine = qmc.Halton(dim=2)
sample = engine.random(n_samples=5)

out = np.array([[1 / 2, 1 / 3], [1 / 4, 2 / 3], [3 / 4, 1 / 9], [1 / 8, 4 / 9], [5 / 8, 7 / 9]])
Expand All @@ -120,7 +120,7 @@ def test_halton(self):
assert_almost_equal(sample, out, decimal=1)

# continuing
engine = qmc.Halton(k_vars=2)
engine = qmc.Halton(dim=2)
engine.fast_forward(2)
sample = engine.random(n_samples=3)
sample = qmc.scale(sample, bounds=corners)
Expand All @@ -135,14 +135,14 @@ def test_lhs(self):

corners = np.array([[0, 2], [10, 5]])

lhs = qmc.LatinHypercube(k_vars=2)
lhs = qmc.LatinHypercube(dim=2)
sample = lhs.random(n_samples=5)
sample = qmc.scale(sample, bounds=corners)
out = np.array([[5.7, 3.2], [5.5, 3.9], [5.2, 3.6],
[5.1, 3.3], [5.8, 4.1]])
assert_almost_equal(sample, out, decimal=1)

lhs = qmc.LatinHypercube(k_vars=2, centered=True)
lhs = qmc.LatinHypercube(dim=2, centered=True)
sample = lhs.random(n_samples=5)
out = np.array([[0.1, 0.5], [0.3, 0.1], [0.7, 0.1],
[0.1, 0.1], [0.3, 0.7]])
Expand All @@ -153,7 +153,7 @@ def test_orthogonal_lhs(self):

corners = np.array([[0, 2], [10, 5]])

olhs = qmc.OrthogonalLatinHypercube(k_vars=2)
olhs = qmc.OrthogonalLatinHypercube(dim=2)
sample = olhs.random(n_samples=5)
sample = qmc.scale(sample, bounds=corners)
out = np.array([[3.933, 2.670], [7.794, 4.031], [4.520, 2.129],
Expand All @@ -175,23 +175,23 @@ def test_orthogonal_lhs(self):
def test_optimal_design(self):
seed = 123456

olhs = qmc.OrthogonalLatinHypercube(k_vars=2, seed=seed)
olhs = qmc.OrthogonalLatinHypercube(dim=2, seed=seed)
sample_ref = olhs.random(n_samples=20)
disc_ref = qmc.discrepancy(sample_ref)

optimal_1 = qmc.OptimalDesign(k_vars=2, start_design=sample_ref, seed=seed)
optimal_1 = qmc.OptimalDesign(dim=2, start_design=sample_ref, seed=seed)
sample_1 = optimal_1.random(n_samples=20)
disc_1 = qmc.discrepancy(sample_1)

assert disc_1 < disc_ref

optimal_ = qmc.OptimalDesign(k_vars=2, seed=seed)
optimal_ = qmc.OptimalDesign(dim=2, seed=seed)
sample_ = optimal_.random(n_samples=20)
disc_ = qmc.discrepancy(sample_)

assert disc_ < disc_ref

optimal_2 = qmc.OptimalDesign(k_vars=2, start_design=sample_ref, niter=2, seed=seed)
optimal_2 = qmc.OptimalDesign(dim=2, start_design=sample_ref, niter=2, seed=seed)
sample_2 = optimal_2.random(n_samples=20)
disc_2 = qmc.discrepancy(sample_2)
assert disc_2 < disc_1
Expand All @@ -200,13 +200,13 @@ def test_optimal_design(self):
sample_1 = optimal_1.random(n_samples=20)
assert_allclose(sample_1, sample_2)

optimal_3 = qmc.OptimalDesign(k_vars=2, start_design=sample_ref, force=True)
optimal_3 = qmc.OptimalDesign(dim=2, start_design=sample_ref, force=True)
sample_3 = optimal_3.random(n_samples=20)
disc_3 = qmc.discrepancy(sample_3)
assert disc_3 < disc_ref

# no optimization
optimal_4 = qmc.OptimalDesign(k_vars=2, start_design=sample_ref,
optimal_4 = qmc.OptimalDesign(dim=2, start_design=sample_ref,
optimization=False, niter=100, seed=seed)
sample_4 = optimal_4.random(n_samples=20)
disc_4 = qmc.discrepancy(sample_4)
Expand Down Expand Up @@ -254,47 +254,47 @@ def test_other_engine(self):
class TestNormalQMC:
def test_NormalQMC(self):
# d = 1
engine = qmc.NormalQMC(k_vars=1)
engine = qmc.NormalQMC(dim=1)
samples = engine.random()
assert_equal(samples.shape, (1, 1))
samples = engine.random(n_samples=5)
assert_equal(samples.shape, (5, 1))
# d = 2
engine = qmc.NormalQMC(k_vars=2)
engine = qmc.NormalQMC(dim=2)
samples = engine.random()
assert_equal(samples.shape, (1, 2))
samples = engine.random(n_samples=5)
assert_equal(samples.shape, (5, 2))

def test_NormalQMCInvTransform(self):
# d = 1
engine = qmc.NormalQMC(k_vars=1, inv_transform=True)
engine = qmc.NormalQMC(dim=1, inv_transform=True)
samples = engine.random()
assert_equal(samples.shape, (1, 1))
samples = engine.random(n_samples=5)
assert_equal(samples.shape, (5, 1))
# d = 2
engine = qmc.NormalQMC(k_vars=2, inv_transform=True)
engine = qmc.NormalQMC(dim=2, inv_transform=True)
samples = engine.random()
assert_equal(samples.shape, (1, 2))
samples = engine.random(n_samples=5)
assert_equal(samples.shape, (5, 2))

def test_other_engine(self):
engine = qmc.NormalQMC(k_vars=2, engine=qmc.Sobol(k_vars=2, scramble=True), inv_transform=True)
engine = qmc.NormalQMC(dim=2, engine=qmc.Sobol(dim=2, scramble=True), inv_transform=True)
samples = engine.random()
assert_equal(samples.shape, (1, 2))

def test_NormalQMCSeeded(self):
# test even dimension
engine = qmc.NormalQMC(k_vars=2, seed=12345)
engine = qmc.NormalQMC(dim=2, seed=12345)
samples = engine.random(n_samples=2)
samples_expected = np.array(
[[-0.63099602, -1.32950772], [0.29625805, 1.86425618]]
)
assert_array_almost_equal(samples, samples_expected)
# test odd dimension
engine = qmc.NormalQMC(k_vars=3, seed=12345)
engine = qmc.NormalQMC(dim=3, seed=12345)
samples = engine.random(n_samples=2)
samples_expected = np.array(
[
Expand All @@ -306,14 +306,14 @@ def test_NormalQMCSeeded(self):

def test_NormalQMCSeededInvTransform(self):
# test even dimension
engine = qmc.NormalQMC(k_vars=2, seed=12345, inv_transform=True)
engine = qmc.NormalQMC(dim=2, seed=12345, inv_transform=True)
samples = engine.random(n_samples=2)
samples_expected = np.array(
[[-0.41622922, 0.46622792], [-0.96063897, -0.75568963]]
)
assert_array_almost_equal(samples, samples_expected)
# test odd dimension
engine = qmc.NormalQMC(k_vars=3, seed=12345, inv_transform=True)
engine = qmc.NormalQMC(dim=3, seed=12345, inv_transform=True)
samples = engine.random(n_samples=2)
samples_expected = np.array(
[
Expand All @@ -324,7 +324,7 @@ def test_NormalQMCSeededInvTransform(self):
assert_array_almost_equal(samples, samples_expected)

def test_NormalQMCShapiro(self):
engine = qmc.NormalQMC(k_vars=2, seed=12345)
engine = qmc.NormalQMC(dim=2, seed=12345)
samples = engine.random(n_samples=250)
assert_(all(np.abs(samples.mean(axis=0)) < 1e-2))
assert_(all(np.abs(samples.std(axis=0) - 1) < 1e-2))
Expand All @@ -337,7 +337,7 @@ def test_NormalQMCShapiro(self):
assert_(np.abs(cov[0, 1]) < 1e-2)

def test_NormalQMCShapiroInvTransform(self):
engine = qmc.NormalQMC(k_vars=2, seed=12345, inv_transform=True)
engine = qmc.NormalQMC(dim=2, seed=12345, inv_transform=True)
samples = engine.random(n_samples=250)
assert_(all(np.abs(samples.mean(axis=0)) < 1e-2))
assert_(all(np.abs(samples.std(axis=0) - 1) < 1e-2))
Expand Down Expand Up @@ -370,12 +370,6 @@ def test_MultivariateNormalQMCSymmetric(self):
ValueError, qmc.MultivariateNormalQMC, [0, 0], [[1, 0], [2, 1]]
)

def test_MultivariateNormalQMCCov(self):
# non square covariance matrix
assert_raises(
ValueError, qmc.MultivariateNormalQMC, [0, 0], [[1], [2, 1]]
)

def test_MultivariateNormalQMCDim(self):
# incompatible dimension of mean/cov
assert_raises(
Expand Down

0 comments on commit 9fb8360

Please sign in to comment.