| pairDist |
index |
where the sum is over two sets of atoms with positions qi and
qj, and associated weights wi and wj. where rc is the probWidth member value, and
NP = 105 ⁄ (32πr7c) The value N = 1 ⁄ (Δr∑iwi∑jwj), where
Δr
is the grid spacing of rVals, unless normalize=False, in which case
N = 1 ⁄ Δr.
Class to generate a pairwise distance distribution and the associated
gradients with respect to atomic coordinates.
Constructor:
PairDist(rVals,
atomSel1,
atomSel2,
weights1,
weights2,
probWidth=0.5,
probType="quartic",
calcGradient=False)
rVals - an array containing a uniform grid of distance values
atomSel1, atomSel2 - atomSel.AtomSel objects containing the atoms
whose positions are sed to generate the distance
distribution.
weights1, weights2 - arrays specifying weights for atoms in atomSel1 and
and atomSel2, repsectively.
probWidth - the width of the probability distribution contribution
from a single pair of atoms.
probType - a string describing the type of probability
distribution. Currently, only "quartic" is accepted.
calcGradient - specifies whether the gradient is calculated.
Members:
Accessors:
weights1()
weights2() - return/set weights. If normalize==True, setWeights1/2 will
normalize input weights.
probWidth - probWidth, as specified in the constructor.
probType - as specified in the constructor.
calcGradient - a specified in the constructor.
calcWeightDerivs - calculate derivative of Pr with respect to weights1
and weights2, If True, the values are stored in the
members d_dw1 and d_dw2, respectively. These are two-
dimensional arrays with the first dimension denoting
the atom specified in the respective atomSel, and
the second specifying bin number.
normalize - by default, the probability distribution is normalized by
the product of the sums of the two weights arrays. Set
normalize to False to disable this normalization.
Methods:
updateValues() - force a recalculation of the P(r) (and gradient, if the
current value of calcGradient is True).
Pr() - return an array containing the pairwise distance distribution the
distance at the center of each bin corresponding to the associated
value of rVals.
Computed as
/ N_P (r^2 - r_c^2)^2 , for r<r_c
Pr_{i,j}(r) = |
\ 0 , for r>=r_c
getGradient(binNum) - return a list of (index,Vec3) pairs specifying all of
the gradient contributions to binNum of the distance
distribution.
removeAtoms1(sel) - remove the atoms from the PairDist - specified by an
removeAtoms2(sel) atomSel.AtomSel argument, along the the associated
weight(s).
# This file was automatically generated by SWIG (http://www.swig.org).
# Version 4.0.2
#
# Do not make changes to this file unless you know what you are doing--modify
# the SWIG interface file instead.
| Classes | ||||||||||||||||||||||||||||||||||||||||||
|
| ||||||||||||||||||||||||||||||||||||||||||
| Functions | ||
| ||
| Data | ||
| __CDSVector_hh__ = 1 | ||