def __init__(self, a, b):
assert isinstance(a, (list, tuple))
assert isinstance(b, (list, tuple))
assert len(a) == len(b)
CompositeDistribution.__init__(
self,
[DrawFrom(random.beta, a=a_p, b=b_p) for (a_p, b_p) in zip(a, b)])
def generate(self, box, n, sampling=None):
if len(box) > 0:
if is_io_process():
if sampling is None:
sampling = UniformDistribution()
elif isinstance(sampling, tuple):
assert len(sampling) == len(box)
sampling = CompositeDistribution(sampling)
self._list = sampling.sample(box, n)
self._list = is_io_process.mpi_comm.bcast(self._list, root=0)
else:
for i in range(n):
self._list.append(tuple())
def generate(self, box, n, sampling=None):
if len(box) > 0:
if sampling is None:
sampling = UniformDistribution()
elif isinstance(sampling, tuple):
assert len(sampling) == len(box)
sampling = CompositeDistribution(sampling)
def run_sampling():
return sampling.sample(box, n)
self._list = parallel_generate(run_sampling, self.mpi_comm)
else:
for i in range(n):
self._list.append(tuple())