def test_accountant(self):
from diffprivlib.accountant import BudgetAccountant
acc = BudgetAccountant(1.5, 0)
a = np.random.random((1000, 5)).round()
count_nonzero(a, epsilon=1, accountant=acc)
self.assertEqual((1.0, 0), acc.total())
with acc:
with self.assertRaises(BudgetError):
count_nonzero(a, epsilon=1)
def test_nan(self):
a = np.random.random((5, 5)).round()
a[2, 2] = np.nan
res = count_nonzero(a)
self.assertFalse(np.isnan(res))
def test_keepdims(self):
a = np.random.random((5, 5)).round()
self.assertEqual(count_nonzero(a, axis=0).ndim, 1)
self.assertEqual(count_nonzero(a, axis=0, keepdims=True).ndim, 2)
self.assertEqual(count_nonzero(a, axis=0, keepdims=True).shape, (1, 5))
def test_axis(self):
a = np.random.random((1000, 5)).round()
res_dp = count_nonzero(a, epsilon=1, axis=0)
self.assertEqual(res_dp.shape, (5, ))
def test_strings(self):
a = ["", "", "Diffprivlib", "Diffprivlib", "Python", "Numpy"]
res = np.count_nonzero(a)
res_dp = count_nonzero(a, epsilon=float("inf"))
self.assertEqual(res, res_dp)
def test_array_like(self):
self.assertIsNotNone(count_nonzero([-1, 0, 1]))
self.assertIsNotNone(count_nonzero((-1, 0, 1)))
def test_large_epsilon(self):
a = np.random.random(1000).round()
res = float(np.count_nonzero(a))
res_dp = count_nonzero(a, epsilon=5)
self.assertAlmostEqual(res, res_dp, delta=0.01 * res)
def test_inf_epsilon(self):
a = np.random.random(1000).round()
res = float(np.count_nonzero(a))
res_dp = count_nonzero(a, epsilon=float("inf"))
self.assertAlmostEqual(res, res_dp)
def test_no_epsilon(self):
a = np.array([1, 2, 3])
self.assertIsNotNone(count_nonzero(a))
def test_no_params(self):
a = np.array([1, 2, 3])
res = count_nonzero(a)
self.assertIsNotNone(res)