pele.mindist.exact_match

class pele.mindist.exact_match.StandardClusterAlignment

Source

class StandardClusterAlignment:

class to iterate over standard alignments for atomic clusters

Quickly determines alignments of clusters which are possible exact matches. It uses atoms which are far away from the center to determine possible rotations. The algorithm does the following:

1. Get 2 reference atoms from structure 1 which are farthest away from center and are not linear

2. Determine candidates from structure 2 which are in same shell as reference atoms from structure 1 (+- accuracy)

3. loop over all candidate combinations to determine orientation and check for match. Skip directly if angle of candidates does not match angle of reference atoms in structure 1.

Parameters:

coords1np.array

first coordinates

coords2np.array

second coordinates

accuracyfloat

accuracy of shell for atom candidates in standard alignment

can_invertboolean

is an inversion possible?

Examples

>> for rot, invert in StandardClusterAlignment(X1, X2): >> print “possible rotation:”,rot,”inversion:”,invert

class pele.mindist.exact_match.ClusterTransoformation

Source

class ClusterTransoformation:

an object that defines a transformation on a cluster

class pele.mindist.exact_match.ExactMatchCluster

Source

class ExactMatchCluster:

Deterministic check if 2 clusters are a perfect match

Determines quickly if 2 clusters are a perfect match. It uses check_standard_alignment_cluster to get possible orientations.

Parameters:

accuracy: float, optional

maximum deviation of atoms to still consider cluster as a match

check_inversion: boolean, optional

check for inversion symmetry, default is True

permlist: iteratable, optional

list of allowed permutations. Default is None which means all particles can be permuted

align_com: boolean, optional

Flag if com should be removed before comparison

Examples

>>> x1 = np.random.random(3*natoms)
>>> x2 = x1 + 1e-4*np.random.random(x1.shape)
>>> matches = ExactClusterMatch(accuracy=1e-3)
>>> if match(x1, x2):
>>>     print "the two structures are identical

ExactMatchCluster.standard_alignments()

Source

def standard_alignments(self, coords1, coords2):

return an iterator over the standard alignments

ExactMatchCluster.__call__()

Source

def __call__(self, coords1, coords2, check_inversion=True):

return True if the structures are an exact mach, False otherwise

ExactMatchCluster.find_transformation()

Source

def find_transformation(self, coords1, coords2, check_inversion=True):

Return None if the two structures are different, else return the transformations that brings them into alignment

Returns:

None:

if the two structures are different

namedtuple:

with fields: rotation, permutation, inversion

ExactMatchCluster.check_match()

Source

def check_match(self, x1, x2, rot, invert):

Give a rotation and inversion, make a more detailed comparison if the 2 structures match

Parameters:

rot: np.array, 3x3

guessed rotation based on reference atoms

invert: boolean

True do match for inverted coordinates

returns: boolean

True or False for match