def getCentroid(attribute_variants, comparator, sort_values=None):
"""
Takes in a list of attribute values for a field,
evaluates the centroid using the comparator,
& returns the centroid (i.e. the 'best' value for the field)
"""
n = len(attribute_variants)
distance_matrix = numpy.zeros([n, n])
# populate distance matrix by looping through elements of matrix triangle
for i in range(0, n):
for j in range(0, i):
distance = comparator(attribute_variants[i], attribute_variants[j])
distance_matrix[i, j] = distance_matrix[j, i] = distance
average_distance = distance_matrix.mean(0)
# there can be ties for minimum, average distance string
min_dist_indices = numpy.where(
average_distance == average_distance.min())[0]
if len(min_dist_indices) > 1:
centroid = breakCentroidTie(attribute_variants,
min_dist_indices,
sort_values=sort_values)
else:
centroid_index = min_dist_indices[0]
centroid = attribute_variants[centroid_index]
return centroid
def getCentroid(attribute_variants, comparator):
"""
Takes in a list of attribute values for a field,
evaluates the centroid using the comparator,
& returns the centroid (i.e. the 'best' value for the field)
"""
n = len(attribute_variants)
distance_matrix = numpy.zeros([n,n])
# populate distance matrix by looping through elements of matrix triangle
for i in range (0,n):
for j in range (0, i):
distance = comparator(attribute_variants[i], attribute_variants[j])
distance_matrix[i,j] = distance_matrix[j,i] = distance
average_distance = distance_matrix.mean(0)
# there can be ties for minimum, average distance string
min_dist_indices = numpy.where(average_distance==average_distance.min())[0]
if len(min_dist_indices) > 1:
centroid = breakCentroidTie(attribute_variants, min_dist_indices)
else :
centroid_index = min_dist_indices[0]
centroid = attribute_variants[centroid_index]
return centroid