4659511182

Build Type

github

Committed by GitHub

Commit Message

Merge 3320693ee into 52de2ab68

Pull Request Pull Request #108: Sphinx build

Run Details

1589 of 3151 relevant lines covered (50.43%)

4.46 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

28.57

/UncertainSCI/sampling.py

import numpy as np
from numpy import random

from UncertainSCI.indexing import MultiIndexSet

def mixture_tensor_discrete_sampling(M, grids, weights, iset):
    """
    Constructs M random samples from a mixture of tensorial discrete measures.
    The measure is a uniform mixture of measures, each supported on the tensor
    product of grids (a list of arrays).

    The input weights define a collection of tensorial measures on the discrete
    support.

    The MultiIndexSet iset defines which measures are used in the mixture.

    weights is a list of arrays, where weights[j][:,i] corresponds to the
    i'th distribution on dimension j.
    """

    assert len(grids) == len(weights) == iset.dim

    N = [grids[j].size for j in range(iset.dim)]

    x = np.zeros([M, iset.dim])
    indices = iset.choice(M)

    max_indices = iset.maximum_unvivariate_degree()

    # Loop over possible values for the indices for each dimension.
    # This should be faster than looping over samples.

    for d in range(iset.dim):
        for k in range(max_indices[d]):
            mask = indices[:,d] == k
            x[mask,d] = random.choice(grids[d], probabilities=weights[d][:,k])

    return x

1	import numpy as np	8✔
2	from numpy import random	8✔
3
4	from UncertainSCI.indexing import MultiIndexSet	8✔
5
6	def mixture_tensor_discrete_sampling(M, grids, weights, iset):	8✔
7	"""
8	Constructs M random samples from a mixture of tensorial discrete measures.
9	The measure is a uniform mixture of measures, each supported on the tensor
10	product of grids (a list of arrays).
11
12	The input weights define a collection of tensorial measures on the discrete
13	support.
14
15	The MultiIndexSet iset defines which measures are used in the mixture.
16
17	weights is a list of arrays, where weights[j][:,i] corresponds to the
18	i'th distribution on dimension j.
19	"""
20
21	assert len(grids) == len(weights) == iset.dim	×
22
23	N = [grids[j].size for j in range(iset.dim)]	×
24
25	x = np.zeros([M, iset.dim])	×
26	indices = iset.choice(M)	×
27
28	max_indices = iset.maximum_unvivariate_degree()	×
29
30	# Loop over possible values for the indices for each dimension.
31	# This should be faster than looping over samples.
32
33	for d in range(iset.dim):	×
34	for k in range(max_indices[d]):	×
35	mask = indices[:,d] == k	×
36	x[mask,d] = random.choice(grids[d], probabilities=weights[d][:,k])	×
37
38	return x	×

SCIInstitute / UncertainSCI / 4659511182

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous