test_stochastic_collocation_sampler.py

import chaospy as cp
import easyvvuq as uq
import pytest
import numpy as np


@pytest.fixture
def sc_sampler():
    vary = {
        "Pe": cp.Uniform(100.0, 200.0),
        "f": cp.Uniform(0.95, 1.05)
    }
    sampler = uq.sampling.SCSampler(vary=vary, polynomial_order=[2, 5], quadrature_rule="G")
    return sampler


def test_sampler(sc_sampler):
    result = [{'Pe': 111.27016653792582, 'f': 0.9533765242898424},
              {'Pe': 111.27016653792582, 'f': 0.9669395306766867},
              {'Pe': 111.27016653792582, 'f': 0.98806904069584},
              {'Pe': 111.27016653792582, 'f': 1.01193095930416},
              {'Pe': 111.27016653792582, 'f': 1.033060469323313},
              {'Pe': 111.27016653792582, 'f': 1.046623475710158},
              {'Pe': 150.0, 'f': 0.9533765242898424},
              {'Pe': 150.0, 'f': 0.9669395306766867},
              {'Pe': 150.0, 'f': 0.98806904069584},
              {'Pe': 150.0, 'f': 1.01193095930416},
              {'Pe': 150.0, 'f': 1.033060469323313},
              {'Pe': 150.0, 'f': 1.046623475710158},
              {'Pe': 188.72983346207417, 'f': 0.9533765242898424},
              {'Pe': 188.72983346207417, 'f': 0.9669395306766867},
              {'Pe': 188.72983346207417, 'f': 0.98806904069584},
              {'Pe': 188.72983346207417, 'f': 1.01193095930416},
              {'Pe': 188.72983346207417, 'f': 1.033060469323313},
              {'Pe': 188.72983346207417, 'f': 1.046623475710158}]
    for sample, reference in zip(sc_sampler, result):
        np.testing.assert_array_almost_equal(
            np.array(list(sample.values())), np.array(list(reference.values())))


def test_generate_grid(sc_sampler):
    grid = sc_sampler.generate_grid(np.array([[2, 5]]))
    np.testing.assert_array_almost_equal(grid, np.array(
        [(111.27016653792582, 0.9533765242898424),
         (111.27016653792582, 0.9669395306766867),
         (111.27016653792582, 0.98806904069584),
         (111.27016653792582, 1.01193095930416),
         (111.27016653792582, 1.033060469323313),
         (111.27016653792582, 1.046623475710158),
         (150.0, 0.9533765242898424),
         (150.0, 0.9669395306766867),
         (150.0, 0.98806904069584),
         (150.0, 1.01193095930416),
         (150.0, 1.033060469323313),
         (150.0, 1.046623475710158),
         (188.72983346207417, 0.9533765242898424),
         (188.72983346207417, 0.9669395306766867),
         (188.72983346207417, 0.98806904069584),
         (188.72983346207417, 1.01193095930416),
         (188.72983346207417, 1.033060469323313),
         (188.72983346207417, 1.046623475710158)]))


# Removed test, already tested in test_hierarchical_sparse_grid_sc.py
# def test_cmpute_sparse_multi_idx(sc_sampler):
#     assert((sc_sampler.compute_sparse_multi_idx(5, 2) ==
#             np.array([[1, 1], [1, 2], [1, 3], [1, 4], [1, 5],
#                       [2, 1], [2, 2], [2, 3], [2, 4],
#                       [3, 1], [3, 2], [3, 3],
#                       [4, 1], [4, 2],
#                       [5, 1]])).all())