def test_1d():
B = gen_BrownianMotion()
Bs = aliased_function("B", B)
t = sympy.Symbol('t')
def_t = sympy.DeferredVector('t')
def_expr = 3*sympy.exp(Bs(def_t)) + 4
expected = 3*np.exp(B.y)+4
ee_lam = lambdify(def_t, def_expr)
yield assert_almost_equal(ee_lam(B.x), expected)
# use vectorize to do the deferred stuff
expr = 3*sympy.exp(Bs(t)) + 4
ee_vec = vectorize(expr)
yield assert_almost_equal(ee_vec(B.x), expected)
# with any arbitrary symbol
b = sympy.Symbol('b')
expr = 3*sympy.exp(Bs(b)) + 4
ee_vec = vectorize(expr, b)
yield assert_almost_equal(ee_vec(B.x), expected)
def test_vectorize():
theta = sympy.Symbol('theta')
num_func = lambdify(theta, sympy.cos(theta))
yield assert_equal(num_func(0), 1)
# we don't need to do anything for a naturally numpy'ed function
yield assert_array_almost_equal(num_func([0, np.pi]), [1, -1])
# but we do for single valued functions
func = aliased_function('f', lambda x: x**2)
num_func = lambdify(theta, func(theta))
yield assert_equal(num_func(2), 4)
# ** on a list raises a type error
yield assert_raises(TypeError, num_func, [2, 3])
# so vectorize
num_func = vectorize(func(theta), theta)