def __init__(self, vertices=None, kdt=None, falloff=None, negate=False, power=2):
self.falloff = falloff
self.negate = negate
if kdt is not None:
self.kdt = kdt
elif vertices is not None:
self.kdt = SvKdTree.new(SvKdTree.best_available_implementation(), vertices, power=power)
else:
raise Exception("Either kdt or vertices must be provided")
self.__description__ = "KDT Attractor"
def __init__(self, vertices=None, kdt=None, falloff=None, power=2):
self.falloff = falloff
if kdt is not None:
self.kdt = kdt
elif vertices is not None:
self.kdt = SvKdTree.new(SvKdTree.best_available_implementation(),
vertices,
power=power)
else:
raise Exception("Either kdt or vertices must be provided")
def to_point(self, centers, falloff):
if self.metric == 'DISTANCE':
metric_single = 'EUCLIDEAN'
else:
metric_single = self.metric
n = len(centers)
if n == 1:
sfield = SvScalarFieldPointDistance(centers[0],
falloff=falloff,
metric=metric_single,
power=self.get_power())
vfield = SvVectorFieldPointDistance(centers[0],
falloff=falloff,
metric=metric_single,
power=self.get_power())
elif self.merge_mode == 'AVG':
sfields = [
SvScalarFieldPointDistance(center,
falloff=falloff,
metric=metric_single,
power=self.get_power())
for center in centers
]
sfield = SvMergedScalarField('AVG', sfields)
vfields = [
SvVectorFieldPointDistance(center,
falloff=falloff,
metric=metric_single,
power=self.get_power())
for center in centers
]
vfield = SvAverageVectorField(vfields)
elif self.merge_mode == 'MIN':
kdt = SvKdTree.new(SvKdTree.best_available_implementation(),
centers,
power=self.get_power())
vfield = SvKdtVectorField(kdt=kdt, falloff=falloff)
sfield = SvKdtScalarField(kdt=kdt, falloff=falloff)
else: # SEP
sfield = [
SvScalarFieldPointDistance(center,
falloff=falloff,
metric=metric_single,
power=self.get_power())
for center in centers
]
vfield = [
SvVectorFieldPointDistance(center,
falloff=falloff,
metric=metric_single,
power=self.get_power())
for center in centers
]
return vfield, sfield
def __init__(self, sites, metric='DISTANCE', power=None):
self.sites = np.asarray(sites)
self.metric = metric
if metric == 'DISTANCE':
self.power = 2
elif metric == 'MANHATTAN':
self.power = 1
elif metric == 'CHEBYSHEV':
self.power = np.inf
elif metric == 'CUSTOM':
self.power = power
else:
raise Exception("Unsupported metric")
self.implementation = SvKdTree.best_available_implementation()
self.kdtree = SvKdTree.new(self.implementation,
sites,
power=self.power)
def _check_min_distance(v_new, vs_old, min_r):
if not vs_old:
return True
kdt = SvKdTree.new(SvKdTree.BLENDER, vs_old)
nearest, idx, dist = kdt.query(v_new)
if dist is None:
return True
ok = (dist >= min_r)
#if not ok:
# print(f"V {v_new} => {nearest}, {dist} >= {min_r}")
return ok