def test_chain(seed=42):
np.random.seed(seed)
nsteps = np.random.randint(200, 400)
ndim = np.random.randint(2, 200)
nwalkers = 2 * ndim
s = samples(ndim, nwalkers)
s.extend(nsteps)
for i in range(nsteps):
x = np.random.rand(nwalkers, ndim)
s.save(x)
assert np.shape(s.chain) == (nwalkers, nsteps, ndim)
def test_flatten(seed=42):
np.random.seed(seed)
nsteps = np.random.randint(200, 400)
ndim = np.random.randint(2, 200)
nwalkers = 2 * ndim
s = samples(ndim, nwalkers)
s.extend(nsteps)
for i in range(nsteps):
x = np.random.rand(nwalkers, ndim)
s.save(x)
burn = np.random.randint(2, 100)
thin = np.random.randint(1, 10)
assert np.shape(s.flatten(burn,
thin)) == ((nsteps - burn) * nwalkers / thin,
ndim)
def test_flatten(seed=42):
np.random.seed(seed)
nsteps = np.random.randint(200, 400)
ndim = np.random.randint(2, 200)
nwalkers = 2 * ndim
s = samples(ndim, nwalkers)
s.extend(nsteps, None)
for i in range(nsteps):
x = np.random.rand(nwalkers, ndim)
z = np.random.rand(nwalkers)
s.save(x, z, None)
assert np.shape(s.flatten()) == (nsteps * nwalkers, ndim)
burn = np.random.randint(2, 100)
thin = np.random.randint(1, 10)
assert np.shape(s.flatten(burn, thin)) == (np.ceil(
(nsteps - burn) / thin) * nwalkers, ndim)
assert np.shape(s.flatten_logprob(burn, thin)) == (np.ceil(
(nsteps - burn) / thin) * nwalkers)