import numpy as np
try:
from scipy.spatial.distance import cdist
scipy_present = True
except ImportError:
scipy_present = False
from .units import Units
__all__ = ['rotation_matrix', 'axis_rotation_matrix', 'distance_array', 'angle','dihedral','cell_shape']
[docs]def rotation_matrix(vec1, vec2):
"""
Calculate the rotation matrix rotating *vec1* to *vec2*. Vectors can be any containers with 3 numerical values. They don't need to be normalized. Returns 3x3 numpy array.
"""
vec1, vec2 = np.array(vec1), np.array(vec2)
a = vec1/np.linalg.norm(vec1)
b = vec2/np.linalg.norm(vec2)
# avoid division by zero in case of antiparallel vectors
if abs(1+np.dot(a,b)) < 1E-8:
return -np.eye(3)
v1,v2,v3 = np.cross(a,b)
M = np.array([[0, -v3, v2], [v3, 0, -v1], [-v2, v1, 0]])
return (np.identity(3) + M + np.dot(M,M)/(1+np.dot(a,b)))
[docs]def axis_rotation_matrix(vector, angle, unit='radian'):
"""
Calculate the rotation matrix rotating along the *vector* by *angle* expressed in *unit*.
*vector* can be any container with 3 numerical values. They don't need to be normalized. A positive angle denotes counterclockwise rotation, when looking along *vector*. Returns 3x3 numpy array.
"""
vector /= np.linalg.norm(vector)
v0, v1, v2 = vector
W = np.array([[0, -v2, v1],
[v2, 0, -v0],
[-v1, v0, 0]])
angle = Units.convert(angle, unit, 'radian')
a1 = np.sin(angle)
a2 = 1.0 - np.cos(angle)
return np.identity(3) + a1 * W + a2 * W@W
[docs]def distance_array(array1, array2):
"""Calculates distance between each pair of points in *array1* and *array2*. Returns 2D ``numpy`` array.
Uses fast ``cdist`` function if ``scipy`` is present, otherwise falls back to slightly slower ``numpy`` loop. Arguments should be 2-dimensional ``numpy`` arrays with the same second dimension. If *array1* is A x N and *array2* is B x N, the returned array is A x B.
"""
return cdist(array1, array2) if scipy_present else np.array([np.linalg.norm(i - array2, axis=1) for i in array1])
[docs]def angle(vec1, vec2, result_unit='radian'):
"""Calculate an angle between vectors *vec1* and *vec2*.
*vec1* and *vec2* should be iterable containers of length 3 (for example: tuple, list, numpy array). Values stored in them are expressed in Angstrom. Returned value is expressed in *result_unit*.
This method requires all atomic coordinates to be numerical values, :exc:`~exceptions.TypeError` is raised otherwise.
"""
vec1 = np.array([*vec1], dtype=float)
vec2 = np.array([*vec2], dtype=float)
num = np.dot(vec1, vec2)
den = np.sqrt(((vec1)**2).sum()) * np.sqrt(((vec2)**2).sum())
return Units.convert(np.arccos(num/den), 'radian', result_unit)
[docs]def dihedral(p1, p2, p3, p4, unit='radian'):
"""Calculate the value of diherdal angle formed by points *p1*, *p2*, *p3* and *p4* in a 3D space. Arguments can be any containers with 3 numerical values, also instances of |Atom|. Returned value is always non-negative, measures the angle clockwise (looking along *p2-p3* vector) and is expressed in *unit*."""
p1 = np.array([*p1], dtype=float)
p2 = np.array([*p2], dtype=float)
p3 = np.array([*p3], dtype=float)
p4 = np.array([*p4], dtype=float)
b0 = p1 - p2
b1 = p3 - p2
b2 = p4 - p3
b1 /= np.linalg.norm(b1)
v = b0 - np.dot(b0, b1)*b1
w = b2 - np.dot(b2, b1)*b1
x = np.dot(v, w)
y = np.dot(np.cross(b1, v), w)
ret = np.arctan2(y, x)
ret = 2*np.pi+ret if ret < 0 else ret
return Units.convert(ret, 'radian', unit)
[docs]def cell_shape (lattice) :
"""
Converts lattice vectors to lengths and angles
Sets internal cell size data, based on set of cell vectors.
*cellvectors* is list containing three cell vectors (a 3x3 matrix)
"""
lattice = np.asarray(lattice)
a,b,c = np.sqrt((lattice**2).sum(axis=1))
if a == 0. and b == 0. and c == 0. :
return
alpha,beta,gamma = (90.,90.,90.)
if c != 0 :
alpha = angle (lattice[1],lattice[2])
beta = angle (lattice[0],lattice[2])
if b != 0 :
gamma = angle (lattice[0],lattice[1])
return [a,b,c,alpha,beta,gamma]