def test_cauchy_add_sub(self):
loc1 = random.randint(1, 11)
sc1 = random.expovariate(1 / 10)
loc2 = random.randint(1, 11)
sc2 = random.expovariate(1 / 10)
a = random.expovariate(1 / 10)
A = rvpy.Cauchy(loc1, sc1)
B = rvpy.Cauchy(loc2, sc2)
Cplus = A + B
self.assertIsInstance(Cplus, rvpy.Cauchy)
self.assertEqual(Cplus.loc, loc1 + loc2)
self.assertEqual(Cplus.scale, sc1 + sc2)
Cminus = A - B
self.assertIsInstance(Cminus, rvpy.Cauchy)
self.assertEqual(Cminus.loc, loc1 - loc2)
self.assertEqual(Cminus.scale, sc1 + sc2)
Aplusa = A + a
self.assertIsInstance(Aplusa, rvpy.Cauchy)
self.assertEqual(Aplusa.loc, loc1 + a)
self.assertEqual(Aplusa.scale, sc1)
Aminusa = A - a
self.assertIsInstance(Aminusa, rvpy.Cauchy)
self.assertEqual(Aminusa.loc, loc1 - a)
self.assertEqual(Aminusa.scale, sc1)
def test_cauchy_conversion(self):
Z1 = rvpy.Cauchy().to_standard()
self.assertIsInstance(Z1, rvpy.StandardCauchy)
Z2 = self.X.to_nonstandard()
self.assertIsInstance(Z2, rvpy.Cauchy)
self.assertNotIsInstance(Z2, rvpy.StandardCauchy)
Z3 = self.X.to_t()
self.assertIsInstance(Z3, rvpy.T)
self.assertEqual(Z3.df, 1)
def test_cauchy_mul_div(self):
c = random.randint(-10, 10)
c = c if c != 0 else 1
Ymul = c * self.Y
self.assertIsInstance(Ymul, rvpy.Cauchy)
self.assertEqual(Ymul.loc, c * self.Y.loc)
self.assertEqual(Ymul.scale, abs(c) * self.Y.scale)
Ydiv = self.Y / c
self.assertIsInstance(Ydiv, rvpy.Cauchy)
self.assertAlmostEqual(Ydiv.loc, self.Y.loc / c)
self.assertAlmostEqual(Ydiv.scale, self.Y.scale / abs(c))
d = random.randint(1, 10)
Zinv = 1 / rvpy.Cauchy(0, d)
self.assertIsInstance(Zinv, rvpy.Cauchy)
self.assertEqual(Zinv.loc, 0)
self.assertAlmostEqual(Zinv.scale, 1 / d)
def test_cauchy_errors(self):
with self.assertRaises(AssertionError):
rvpy.Cauchy(0, -1)
with self.assertRaises(AssertionError):
rvpy.Cauchy(3, 5).to_standard()
def setUp(self):
self.X = rvpy.StandardCauchy()
self.Y = rvpy.Cauchy(random.randint(1, 11), random.expovariate(1 / 10))