Function test_solve

test/test_solvers.py:105–148

(nx)

def test_solve(nx):
    n_samples_s = 10
    n_samples_t = 7
    n_features = 2
    rng = np.random.RandomState(0)

    x = rng.randn(n_samples_s, n_features)
    y = rng.randn(n_samples_t, n_features)
    a = ot.utils.unif(n_samples_s)
    b = ot.utils.unif(n_samples_t)

    M = ot.dist(x, y)
    reg = 1e-1

    # solve unif weights
    sol0 = ot.solve(M, reg=reg)

    print(sol0)

    # solve signe weights
    sol = ot.solve(M, a, b, reg=reg)

    # check some attributes
    sol.potentials
    sol.sparse_plan
    sol.marginals
    sol.status

    # print("dual = {}".format(sol.potentials))
    # assert_allclose_sol(sol0, sol)

    # solve in backend
    ab, bb, Mb = nx.from_numpy(a, b, M)
    solb = ot.solve(M, a, b)

    assert_allclose_sol(sol, solb)

    # test not implemented unbalanced and check raise
    with pytest.raises(NotImplementedError):
        sol0 = ot.solve(M, unbalanced=1, unbalanced_type="cryptic divergence")

    # test not implemented reg_type and check raise
    with pytest.raises(NotImplementedError):
        sol0 = ot.solve(M, reg=1, reg_type="cryptic divergence")


@pytest.mark.skipif(not torch, reason="torch no installed")