smbinterp/interp/baker/__init__.py

from collections import namedtuple
from functools import wraps
import itertools

import numpy as np

import interp

AGGRESSIVE_ERROR_SOLVE = True
RAISE_PATHOLOGICAL_EXCEPTION = False

__version__ = interp.__version__

Answer = namedtuple("Answer", ['qlin', 'final', 'error', 'abc'])


def get_phis(X, R):
    """
      The get_phis function is used to get barycentric coordonites for a
      point on a triangle or tetrahedron (Equation (*\ref{eq:qlinarea}*))

      in 2D:

      X - the destination point (2D)
          X = [0,0]
      R - the three points that make up the 2-D triangular simplex
          R = [[-1, -1], [0, 2], [1, -1]]

      this will return [0.333, 0.333, 0.333]


      in 3D:

      X - the destination point (3D)
          X = [0,0,0]
      R - the four points that make up the 3-D simplex (tetrahedron)
          R = [
             [ 0.0000,  0.0000,  1.0000],
             [ 0.9428,  0.0000, -0.3333],
             [-0.4714,  0.8165, -0.3333],
             [-0.4714, -0.8165, -0.3333],
           ]

      this will return [0.25, 0.25, 0.25, 0.25]
    """

    # equations (*\ref{eq:lin3d}*) and (*\ref{eq:lin2d}*)
    if len(X) == 2:
        A = np.array([
                        [1, 1, 1],
                        [R[0][0], R[1][0], R[2][0]],
                        [R[0][1], R[1][1], R[2][1]],
                      ])
        b = np.array([1, X[0], X[1]])
    elif len(X) == 3:
        A = np.array([
                    [1, 1, 1, 1],
                    [R[0][0], R[1][0], R[2][0], R[3][0]],
                    [R[0][1], R[1][1], R[2][1], R[3][1]],
                    [R[0][2], R[1][2], R[2][2], R[3][2]],
                    ])
        b = np.array([1, X[0], X[1], X[2]])
    else:
        raise Exception("inapropriate demension on X")
    phi = np.linalg.solve(A, b)
    return phi


def qlinear(X, R, q):
    """
        this calculates the linear portion of q from R to X

        This is equation (*\ref{eq:qlinbasis}*)

        X = destination point
        R = a inter.grid object; must have R.points and R.q
    """

    phis = get_phis(X, R)
    qlin = np.sum([q_i * phi_i for q_i, phi_i in zip(q, phis)])

    return phis, qlin


def get_error(phi, R, R_q, S, S_q, order=2):
    """
        Calculate the error approximation terms, returning the unknowns
        a,b, and c in equation (*\ref{eq:quadratic2d}*).
    """
    B = []   # equation ((*\ref{eq:B2d}*)
    w = []   # equation ((*\ref{eq:w}*)

    cur_pattern = pattern(len(phi), order)

    for (s, cur_q) in zip(S, S_q):
        cur_phi, cur_qlin = qlinear(s, R, R_q)
        l = []
        for i in cur_pattern:
            cur_sum = cur_phi[i[0]]
            for j in i[1:]:
                cur_sum *= cur_phi[j]
            l.append(cur_sum)

        B.append(l)
        w.append(cur_q - cur_qlin)

    B = np.array(B)
    w = np.array(w)

    A = np.dot(B.T, B)
    b = np.dot(B.T, w)

    try:
        abc = np.linalg.solve(A, b)
    except np.linalg.LinAlgError:
        if not AGGRESSIVE_ERROR_SOLVE:
            return None, None
        abc = np.dot(np.linalg.pinv(A), b)

    error_term = 0.0
    for (a, i) in zip(abc, cur_pattern):
        cur_sum = a
        for j in i:
            cur_sum *= phi[j]
        error_term += cur_sum

    return error_term, abc


def interpolate(X, R, R_q, S=None, S_q=None, order=2):
    """
        This is the main function to call to get an interpolation to X from the
        input meshes

        X -- the destination point

        R = Simplex
        R_q = q values at R
        S = extra points
        S_q = q values at S

        order - order of interpolation - 1
    """

    qlin = None
    error_term = None
    final = None
    abc = {}

    # calculate values only for the simplex triangle
    phi, qlin = qlinear(X, R, R_q)

    if order in xrange(2, 11) and S is not None:
        error_term, abc = get_error(phi, R, R_q, S, S_q, order)

        # if a pathological vertex configuration was encountered and
        # AGGRESSIVE_ERROR_SOLVE is False, get_error will return (None, None)
        # indicating that only linear interpolation should be performed
        if (error_term is None) and (abc is None):
            if RAISE_PATHOLOGICAL_EXCEPTION:
                raise np.linalg.LinAlgError("Pathological Vertex Config")
        else:
            final = qlin + error_term
    elif order not in xrange(2, 11):
        raise Exception('unsupported order "%d" for baker method' % order)

    return Answer(qlin=qlin, error=error_term, final=final, abc=abc)


def memoize(f):
    """
        for more information on what I'm doing here, please read:
        http://en.wikipedia.org/wiki/Memoize
    """
    cache = {}

    @wraps(f)
    def memf(simplex_size, nu):
        x = (simplex_size, nu)
        if x not in cache:
            cache[x] = f(simplex_size, nu)
        return cache[x]
    return memf


@memoize
def pattern(simplex_size, nu):
    """
        This function returns the pattern requisite to compose the error
        approximation function, and the matrix B.
    """

    r = []
    for i in itertools.product(xrange(simplex_size), repeat=nu):
        if len(set(i)) != 1:
            r.append(tuple(sorted(i)))
    unique_r = list(set(r))
    return unique_r