Maintained package (scipy) implementation

[alexandreCameron]

For those looking to upgrade the function to a maintained package:

def get_sequence_using_sobol_seq(n_samples, d):
    from sobol_seq import i4_sobol_generate
    return i4_sobol_generate(d, n_samples).flatten()

is equivalent to

def get_sequence_using_scipy(n_samples, d):
    from scipy.stats.qmc import Sobol
    support_generator = Sobol(d, scramble=False, seed=None).reset()
    # The first value is remove because it's 0
    support = support_generator.random(n=n_samples+1).flatten()[1:]
    return  support

[keithbriggs]

Correction:

from scipy.stats.qmc import Sobol

def get_sequence_using_scipy(n_samples, d):
    support_generator = Sobol(d, scramble=False, seed=None).reset()
    # The first value is removed because it's 0
    return support_generator.random(n=n_samples+1).flatten()[1:]