def __init__(self):
def tensor_points(P):
PP0, PP1 = np.array(np.meshgrid(P, P))
return np.array((PP0.ravel(), PP1.ravel())).T
def tensor_weights(W):
return np.dot(W[:, np.newaxis], W[np.newaxis, :]).ravel()
self._quadrature_points = [tensor_points(GaussQuadratures.quadrature(npoints=p + 1)[0])
for p in xrange(GaussQuadratures.maxpoints())]
self._quadrature_weights = [tensor_weights(GaussQuadratures.quadrature(npoints=p + 1)[1])
for p in xrange(GaussQuadratures.maxpoints())]
self._quadrature_npoints = np.arange(1, GaussQuadratures.maxpoints() + 1) ** 2
self._quadrature_orders = GaussQuadratures.orders
self._quadrature_order_map = GaussQuadratures.order_map
def quadrature(self, order=None, npoints=None, quadrature_type='default'):
if quadrature_type == 'default' or quadrature_type == 'gauss':
P, W = GaussQuadratures.quadrature(order, npoints)
return P[:, np.newaxis], W
else:
raise NotImplementedError(
'quadrature_type must be "default" or "gauss"')
def __init__(self, flux, flux_derivative, gausspoints=5, intervals=1):
self.__auto_init(locals())
self.build_parameter_type(flux, flux_derivative)
points, weights = GaussQuadratures.quadrature(npoints=self.gausspoints)
points = points / intervals
points = ((np.arange(self.intervals, dtype=np.float)[:, np.newaxis] *
(1 / intervals)) + points[np.newaxis, :]).ravel()
weights = np.tile(weights, intervals) * (1 / intervals)
self.points = points
self.weights = weights
def __init__(self):
def tensor_points(P):
PP0, PP1 = np.array(np.meshgrid(P, P))
return np.array((PP0.ravel(), PP1.ravel())).T
def tensor_weights(W):
return np.dot(W[:, np.newaxis], W[np.newaxis, :]).ravel()
self._quadrature_points = [
tensor_points(GaussQuadratures.quadrature(npoints=p + 1)[0])
for p in range(GaussQuadratures.maxpoints())
]
self._quadrature_weights = [
tensor_weights(GaussQuadratures.quadrature(npoints=p + 1)[1])
for p in range(GaussQuadratures.maxpoints())
]
self._quadrature_npoints = np.arange(1,
GaussQuadratures.maxpoints() +
1)**2
self._quadrature_orders = GaussQuadratures.orders
self._quadrature_order_map = GaussQuadratures.order_map
def __init__(self, flux, flux_derivative, gausspoints=5, intervals=1):
self.flux = flux
self.flux_derivative = flux_derivative
self.gausspoints = gausspoints
self.intervals = intervals
self.build_parameter_type(inherits=(flux, flux_derivative))
points, weights = GaussQuadratures.quadrature(npoints=self.gausspoints)
points = points / intervals
points = ((np.arange(self.intervals, dtype=np.float)[:, np.newaxis] * (1 / intervals))
+ points[np.newaxis, :]).ravel()
weights = np.tile(weights, intervals) * (1 / intervals)
self.points = points
self.weights = weights
def test_points(self):
for order in GaussQuadratures.orders:
P, _ = GaussQuadratures.quadrature(order)
assert float_cmp_all(P, np.sort(P))
assert 0.0 < P[0]
assert P[-1] < 1.0
def test_weights(self):
for order in GaussQuadratures.orders:
_, W = GaussQuadratures.quadrature(order)
assert float_cmp(sum(W), 1)
def quadrature(self, order=None, npoints=None, quadrature_type='default'):
if quadrature_type == 'default' or quadrature_type == 'gauss':
P, W = GaussQuadratures.quadrature(order, npoints)
return P[:, np.newaxis], W
else:
raise NotImplementedError('quadrature_type must be "default" or "gauss"')
def test_points(self):
for order in GaussQuadratures.orders:
P, _ = GaussQuadratures.quadrature(order)
assert float_cmp_all(P, np.sort(P))
assert 0.0 < P[0]
assert P[-1] < 1.0
def test_weights(self):
for order in GaussQuadratures.orders:
_, W = GaussQuadratures.quadrature(order)
assert float_cmp(sum(W), 1)
def test_points(self):
for order in GaussQuadratures.orders:
P, _ = GaussQuadratures.quadrature(order)
np.testing.assert_array_equal(P, np.sort(P))
self.assertLess(0.0, P[0])
self.assertLess(P[-1], 1.0)
def test_weights(self):
for order in GaussQuadratures.orders:
_, W = GaussQuadratures.quadrature(order)
self.assertAlmostEqual(sum(W), 1)