Function discrete_sequence

networkx/utils/random_sequence.py:142–170

Return sample sequence of length n from a given discrete distribution or discrete cumulative distribution. One of the following must be specified. distribution = histogram of values, will be normalized cdistribution = normalized discrete cumulative distribution

(n, distribution=None, cdistribution=None, seed=None)

@py_random_state(3)
def discrete_sequence(n, distribution=None, cdistribution=None, seed=None):
    """
    Return sample sequence of length n from a given discrete distribution
    or discrete cumulative distribution.

    One of the following must be specified.

    distribution = histogram of values, will be normalized

    cdistribution = normalized discrete cumulative distribution

    """
    import bisect

    if cdistribution is not None:
        cdf = cdistribution
    elif distribution is not None:
        cdf = cumulative_distribution(distribution)
    else:
        raise nx.NetworkXError(
            "discrete_sequence: distribution or cdistribution missing"
        )

    # get a uniform random number
    inputseq = [seed.random() for i in range(n)]

    # choose from CDF
    seq = [bisect.bisect_left(cdf, s) - 1 for s in inputseq]
    return seq


@py_random_state(2)