def _rhs(x, s, eps, powers, diff, basis):
'''
Returns the differentiated RBF and polynomial terms evaluated at x
'''
x = x[None, :]
# number of nodes in the stencil and the number of dimensions
Ns, Ndim = s.shape
# number of monomial terms
Np = powers.shape[0]
d = np.empty(Ns + Np, dtype=float)
d[:Ns] = basis(x, s, eps=eps, diff=diff)[0, :]
d[Ns:] = rbf.poly.mvmonos(x, powers, diff=diff)[0, :]
return d
def __init__(self,
y,
d,
sigma=None,
eps=1.0,
basis=rbf.basis.phs3,
order=1,
extrapolate=True):
y = np.asarray(y)
assert_shape(y, (None, None), 'y')
d = np.asarray(d)
assert_shape(d, (y.shape[0], ), 'd')
q, dim = y.shape
if sigma is None:
# if sigma is not specified then it is zeros
sigma = np.zeros(q)
elif np.isscalar(sigma):
# if a float is specified then use it as the uncertainties for
# all observations
sigma = np.repeat(sigma, q)
else:
sigma = np.asarray(sigma)
assert_shape(sigma, (y.shape[0], ), 'sigma')
# form block consisting of the RBF and uncertainties on the
# diagonal
K = basis(y, y, eps=eps)
Cd = scipy.sparse.diags(sigma**2)
# form the block consisting of the monomials
powers = rbf.poly.powers(order, dim)
P = rbf.poly.mvmonos(y, powers)
# create zeros vector for the right-hand-side
z = np.zeros((powers.shape[0], ))
# solve for the RBF and mononomial coefficients
basis_coeff, poly_coeff = PartitionedSolver(K + Cd, P).solve(d, z)
self._y = y
self._basis = basis
self._order = order
self._eps = eps
self._basis_coeff = basis_coeff
self._poly_coeff = poly_coeff
self._powers = powers
self.extrapolate = extrapolate
def _lhs(s, eps, powers, basis):
'''
Returns the transposed RBF alternant matrix with added polynomial
terms and constraints
'''
# number of nodes in the stencil and the number of dimensions
Ns, Ndim = s.shape
# number of monomial terms
Np = powers.shape[0]
# deriviative orders
diff = np.zeros(Ndim, dtype=int)
A = np.zeros((Ns + Np, Ns + Np), dtype=float)
A[:Ns, :Ns] = basis(s, s, eps=eps, diff=diff).T
Ap = rbf.poly.mvmonos(s, powers, diff=diff)
A[Ns:, :Ns] = Ap.T
A[:Ns, Ns:] = Ap
return A
def _interpolation_matrix(xitp,x,diff,eps,basis,order): ''' returns the matrix that maps the coefficients to the function values at the interpolation points ''' # number of interpolation points and spatial dimensions I,D = xitp.shape # number of observation points N = x.shape[0] # powers for the additional polynomials powers = rbf.poly.powers(order,D) # number of polynomial terms P = powers.shape[0] # allocate array A = np.zeros((I,N+P)) A[:,:N] = basis(xitp,x,eps=eps,diff=diff) A[:,N:] = rbf.poly.mvmonos(xitp,powers,diff=diff) return A
def _interpolation_matrix(xitp,x,diff,eps,basis,order): ''' returns the matrix that maps the coefficients to the function values at the interpolation points ''' # number of interpolation points and spatial dimensions I,D = xitp.shape # number of observation points N = x.shape[0] # powers for the additional polynomials powers = rbf.poly.powers(order,D) # number of polynomial terms P = powers.shape[0] # allocate array A = np.zeros((I,N+P)) A[:,:N] = basis(xitp,x,eps=eps,diff=diff) A[:,N:] = rbf.poly.mvmonos(xitp,powers,diff=diff) return A
def _coefficient_matrix(x, eps, basis, order):
'''
returns the matrix used to compute the radial basis function
coefficients
'''
# number of observation points and spatial dimensions
N, D = x.shape
# powers for the additional polynomials
powers = rbf.poly.monomial_powers(order, D)
# number of polynomial terms
P = powers.shape[0]
# allocate array
A = np.zeros((N + P, N + P))
A[:N, :N] = basis(x, x, eps=eps)
Ap = rbf.poly.mvmonos(x, powers)
A[N:, :N] = Ap.T
A[:N, N:] = Ap
return A
def _coefficient_matrix(x,eps,basis,order): ''' returns the matrix used to compute the radial basis function coefficients ''' # number of observation points and spatial dimensions N,D = x.shape # powers for the additional polynomials powers = rbf.poly.powers(order,D) # number of polynomial terms P = powers.shape[0] # allocate array A = np.zeros((N+P,N+P)) A[:N,:N] = basis(x,x,eps=eps) Ap = rbf.poly.mvmonos(x,powers) A[N:,:N] = Ap.T A[:N,N:] = Ap return A
def weights(x,
s,
diffs,
coeffs=None,
basis=rbf.basis.phs3,
order=None,
eps=1.0):
'''
Returns the weights which map a functions values at `s` to an
approximation of that functions derivative at `x`. The weights are
computed using the RBF-FD method described in [1]. In this function
`x` is a single point in D-dimensional space. Use `weight_matrix` to
compute the weights for multiple point.
Parameters
----------
x : (D,) array
Target point. The weights will approximate the derivative at this
point.
s : (N, D) array
Stencil points. The derivative will be approximated with a
weighted sum of the function values at this point.
diffs : (D,) int array or (K, D) int array
Derivative orders for each spatial dimension. For example `[2, 0]`
indicates that the weights should approximate the second
derivative with respect to the first spatial dimension in
two-dimensional space. diffs can also be a (K, D) array, where
each (D,) sub-array is a term in a differential operator. For
example the two-dimensional Laplacian can be represented as
`[[2, 0], [0, 2]]`.
coeffs : (K,) array, optional
Coefficients for each term in the differential operator specified
with `diffs`. Defaults to an array of ones. If `diffs` was
specified as a (D,) array then `coeffs` should be a length 1
array.
basis : rbf.basis.RBF, optional
Type of RBF. Select from those available in `rbf.basis` or create
your own.
order : int, optional
Order of the added polynomial. This defaults to the highest
derivative order. For example, if `diffs` is `[[2, 0], [0, 1]]`, then
order is set to 2.
eps : float or (N,) array, optional
Shape parameter for each RBF, which have centers `s`. This only
makes a difference when using RBFs that are not scale invariant.
All the predefined RBFs except for the odd order polyharmonic
splines are not scale invariant.
Returns
-------
out : (N,) array
RBF-FD weights
Examples
--------
Calculate the weights for a one-dimensional second order derivative.
>>> x = np.array([1.0])
>>> s = np.array([[0.0], [1.0], [2.0]])
>>> diff = (2,)
>>> weights(x, s, diff)
array([ 1., -2., 1.])
Calculate the weights for estimating an x derivative from three
points in a two-dimensional plane
>>> x = np.array([0.25, 0.25])
>>> s = np.array([[0.0, 0.0],
[1.0, 0.0],
[0.0, 1.0]])
>>> diff = (1, 0)
>>> weights(x, s, diff)
array([ -1., 1., 0.])
Notes
-----
This function may become unstable with high order polynomials (i.e.,
`order` is high). This can be somewhat remedied by shifting the
coordinate system so that x is zero
References
----------
[1] Fornberg, B. and N. Flyer. A Primer on Radial Basis
Functions with Applications to the Geosciences. SIAM, 2015.
'''
x = np.asarray(x, dtype=float)
assert_shape(x, (None, ), 'x')
s = np.asarray(s, dtype=float)
assert_shape(s, (None, x.shape[0]), 's')
diffs = np.asarray(diffs, dtype=int)
diffs = _reshape_diffs(diffs)
# stencil size and number of dimensions
size, dim = s.shape
if coeffs is None:
coeffs = np.ones(diffs.shape[0], dtype=float)
else:
coeffs = np.asarray(coeffs, dtype=float)
assert_shape(coeffs, (diffs.shape[0], ), 'coeffs')
max_order = _max_poly_order(size, dim)
if order is None:
order = _default_poly_order(diffs)
order = min(order, max_order)
if order > max_order:
raise ValueError('Polynomial order is too high for the stencil size')
# get the powers for the added monomials
powers = rbf.poly.powers(order, dim)
# evaluate the RBF and monomials at each point in the stencil. This
# becomes the left-hand-side
A = basis(s, s, eps=eps)
P = rbf.poly.mvmonos(s, powers)
# Evaluate the RBF and monomials for each term in the differential
# operator. This becomes the right-hand-side.
a = coeffs[0] * basis(x[None, :], s, eps=eps, diff=diffs[0])
p = coeffs[0] * rbf.poly.mvmonos(x[None, :], powers, diff=diffs[0])
for c, d in zip(coeffs[1:], diffs[1:]):
a += c * basis(x[None, :], s, eps=eps, diff=d)
p += c * rbf.poly.mvmonos(x[None, :], powers, diff=d)
# squeeze `a` and `p` into 1d arrays. `a` is ran through as_array
# because it may be sparse.
a = rbf.linalg.as_array(a)[0]
p = p[0]
# attempt to compute the RBF-FD weights
try:
w = PartitionedSolver(A, P).solve(a, p)[0]
return w
except np.linalg.LinAlgError:
raise np.linalg.LinAlgError(
'An error was raised while computing the RBF-FD weights at '
'point %s with the RBF %s and the polynomial order %s. This '
'may be due to a stencil with duplicate or collinear points. '
'The stencil contains the following points:\n%s' %
(x, basis, order, s))
forcing = sympy.lambdify((x, y), forcing_sym, 'numpy') # define a circular domain vert, smp = rbf.domain.circle() nodes, smpid = menodes(N, vert, smp) # smpid describes which boundary simplex, if any, the nodes are # attached to. If it is -1, then the node is in the interior boundary, = (smpid >= 0).nonzero() interior, = (smpid == -1).nonzero() # create the left-hand-side matrix which is the Laplacian of the basis # function for interior nodes and the undifferentiated basis functions # for the boundary nodes A = np.zeros((N, N)) A[interior] = basis(nodes[interior], nodes, diff=[2, 0]) A[interior] += basis(nodes[interior], nodes, diff=[0, 2]) A[boundary, :] = basis(nodes[boundary], nodes) # create the right-hand-side vector, consisting of the forcing term # for the interior nodes and zeros for the boundary nodes d = np.zeros(N) d[interior] = forcing(nodes[interior, 0], nodes[interior, 1]) d[boundary] = true_soln(nodes[boundary, 0], nodes[boundary, 1]) # find the RBF coefficients that solve the PDE coeff = np.linalg.solve(A, d) # create a collection of interpolation points to evaluate the # solution. It is easiest to just call menodes again itp, dummy = menodes(10000, vert, smp, itr=0)
forcing = sympy.lambdify((x,y),forcing_sym,'numpy') # define a circular domain vert,smp = rbf.domain.circle() nodes,smpid = menodes(N,vert,smp) # smpid describes which boundary simplex, if any, the nodes are # attached to. If it is -1, then the node is in the interior boundary, = (smpid>=0).nonzero() interior, = (smpid==-1).nonzero() # create the left-hand-side matrix which is the Laplacian of the basis # function for interior nodes and the undifferentiated basis functions # for the boundary nodes A = np.zeros((N,N)) A[interior] = basis(nodes[interior],nodes,diff=[2,0]) A[interior] += basis(nodes[interior],nodes,diff=[0,2]) A[boundary,:] = basis(nodes[boundary],nodes) # create the right-hand-side vector, consisting of the forcing term # for the interior nodes and zeros for the boundary nodes d = np.zeros(N) d[interior] = forcing(nodes[interior,0],nodes[interior,1]) d[boundary] = true_soln(nodes[boundary,0],nodes[boundary,1]) # find the RBF coefficients that solve the PDE coeff = np.linalg.solve(A,d) # create a collection of interpolation points to evaluate the # solution. It is easiest to just call menodes again itp,dummy = menodes(10000,vert,smp,itr=0)