Numba Utils

compute_bin(x: float, bin_edges: ndarray) → int[source]

Compute determine bin for a given observation, x

Parameters:

  • x (float) – Observation.

  • bin_edges (np.ndarray) – Available bins.

Returns:

bin – Index of bin to add observation to.

Return type:

int

dirichlet(x: ndarray) → ndarray[source]

Dirichlet distribution sampler

Parameters:

x (np.ndarray) – Parameters of dirichlet distribution to sample from.

Returns:

dist – Random sample from dirichlet distribution with parameters x.

Return type:

np.ndarray

get_delta_r(r: ndarray) → float[source]

Calculates increment of a sorted array with evenly distributed floats

Parameters:

r (np.ndarray) – Array of equally spaced consecutive numbers.

Returns:

delta_r – Distance between numbers in r.

Return type:

float

histogram(a: ndarray, weights: ndarray, r: ndarray) → ndarray[source]

Calculate histogram for observations, a , with weights over the domain, r, as precursor for KDE of distance distribution.

Parameters:

  • a (np.ndarray) – Array of relevant pairwise distances between NO midpoint libraries.

  • weights (np.ndarray) – Weights array corresponding to distances in a .

  • r (np.ndarray) – Domain of distance distribution.

Returns:

hist – Weighted histogram of pairwise distances.

Return type:

np.ndarray

jaccard(p: ndarray, q: ndarray) → float[source]

Calculate the jaccard index between the two provided vectors (distance distributions)

Parameters:

  • p (np.ndarray) – Sets to calculate the jaccard index between.

  • q (np.ndarray) – Sets to calculate the jaccard index between.

Returns:

jac – Jaccard index between q and p .

Return type:

float

kl_divergence(p: ndarray, q: ndarray) → float[source]

Compute the Kullback–Leibler divergence (KLD) of p and q .

Parameters:

  • p (np.ndarray) – The distributions to calculate the KLD between.

  • q (np.ndarray) – The distributions to calculate the KLD between.

Returns:

kld – The Kullback–Leibler divergence between p and q .

Return type:

float

normdist(delta_r: float, mu: float = 0.0, sigma: float = 1.0) → Tuple[ndarray, ndarray][source]

Calculate normal distribution for convolution with histogram of distances between two spin label ensembles

Parameters:

  • delta_r (float) – Space between points in distance domain.

  • mu (float) – Mean for normal distribution.

  • sigma (float) – Standard deviation of normal distribution.

Returns:

  • x (np.ndarray) – Domain of the distribution.

  • y (np.ndarray) – PDF values of the distribution.

pairwise_dist(x: ndarray, y: ndarray) → ndarray[source]

Calculate the pairwise (Euclidean) distance between the coordinate sets.

Parameters:

  • x (np.ndarray) – Coordinate sets to calculate the distance between

  • y (np.ndarray) – Coordinate sets to calculate the distance between

Returns:

distances – Pairwise distance matrix between coordinates of x and y.

Return type:

np.ndarray

batch_ic2cart(IC_idx_Array: ndarray, ICArray: ndarray)[source]

Vectorization of ic2cart along the ICArray argument to calculate cartesian coordinates for multiple conformations.

Parameters:

  • IC_idx_Array (np.ndarray) –

    Array of indexes corresponding to the atoms defining the internal coordinate.

    IC_idx_Array[i, 0] # index of the atom that atom i is bonded to. IC_idx_Array[i, 1] # index of the atom that atom i creates an angle with. IC_idx_Array[i, 2] # index of the atom that atom i creates a dihedral angle with.

    A value of -1 indicates that there is no precursor atom to define the bond, angle or dihedral and that atom i is one of the coordinate system defining atoms.

  • ICArray –

    Internal coordinate values for n conformers

    IC_idx_Array[n, i, 0] # bond distance of the atom that atom i and one of its precursor atoms. IC_idx_Array[n, i, 1] # bond angle between atom i two of its precursor atoms. IC_idx_Array[n, i, 2] # index of the atom that atom i creates a dihedral angle with three of it’s precursor atoms.

Returns:

coords – Array of cartesian coordinates corresponding to ICAtom list atoms

Return type:

np.ndarray

np_all_axis1(x: ndarray) → ndarray[source]

Numba compatible version of np.all(x, axis=1).

Parameters:

x ((M, N) np.ndarray) – Boolean input array.

Returns:

out – Boolean array indicating if all values in the second axis of x are true.

Return type:

(M,) np.ndarray

get_sasa(atom_coords: ndarray, atom_radii: ndarray, environment_coords: ndarray | None = None, environment_radii: ndarray | None = None, probe_radius: float = 1.4, npoints: int = 1024, by_atom: bool = False) → ndarray[source]
Parameters:

  • atom_coords ((M, N, 3) np.ndarray) – M frames of the 3D cartesian coordinates of the N atoms to calculate the SASA of.

  • atom_radii ((N,) np.ndarray) – vdW radii + solvent radii of the atoms in atom_coords

  • environment_coords ((L, 3) np.ndarray) – 3D cartesian coordinates of the atoms nearby the atoms for which the SASA is desired.

  • environment_radii ((L,) np.ndarray) – vdW radii + solvent radii of the atoms in env_coords

  • probe_radius (flaot) – Radius of the solvent probe.

  • npoints (int) – Number of points to use in sphere used for SASA calculation

  • by_atom (bool) – Return SASA of each atom rather than the set of atoms

Returns:

sasa – Solvent accessible surface areas of the provided coords in the provided environment.

Return type:

np.ndarray:

fibonacci_points(n: int) → ndarray[source]

Get n evenly spaced points on the unit sphere using the fibonacci method.

Parameters:

n (int) – Number of points to return.

Returns:

coords – 3D coordinates of the points on the unit sphere.

Return type:

np.ndarray