def test_bernoulli_kullback_leibler():
assert bernoulli_kullback_leibler(0, 1) == np.infty
q = np.random.random()
assert bernoulli_kullback_leibler(0, q) > 0
assert bernoulli_kullback_leibler(q, q) == 0
x = np.random.uniform(0, 1, 10)
x.sort()
for i in range(np.size(x) - 1):
assert bernoulli_kullback_leibler(
x[0], x[i]) < bernoulli_kullback_leibler(x[0], x[i + 1])
def test_kl_upper_bound():
assert kl_upper_bound(0.5 * 1, 1, 10, c=1,
eps=1e-3) == pytest.approx(0.997, abs=1e-3)
assert kl_upper_bound(0.5 * 10, 10, 20, c=1,
eps=1e-3) == pytest.approx(0.835, abs=1e-3)
assert kl_upper_bound(0.5 * 20, 20, 40, c=1,
eps=1e-3) == pytest.approx(0.777, abs=1e-3)
rands = np.random.randint(1, 500, 2)
rands.sort()
mu, count, time = np.random.random(), rands[0], rands[1]
ucb = kl_upper_bound(mu * count, count, time, c=1, eps=1e-3)
assert not np.isnan(ucb)
d_max = 1 * np.log(time) / count
assert bernoulli_kullback_leibler(mu, ucb) == pytest.approx(d_max,
abs=1e-2)
def test_d_bernoulli_kullback_leibler_dq():
x = np.random.uniform(0, 1, 2)
p, q = x
eps = 1e-6
assert d_bernoulli_kullback_leibler_dq(p, q) == \
pytest.approx((bernoulli_kullback_leibler(p, q+eps) - bernoulli_kullback_leibler(p, q-eps)) / (2*eps), 1e-3)