Source code for SALib.sample.finite_diff

from typing import Dict

import numpy as np

from . import common_args
from . import sobol_sequence
from ..util import scale_samples, read_param_file




def sample(
    problem: Dict,
    N: int,
    delta: float = 0.01,
    seed: int = None,
    skip_values: int = 1024,
) -> np.ndarray:
    """
    Generate matrix of samples for Derivative-based Global Sensitivity Measure (DGSM).

    Start from a QMC (Sobol') sequence and finite difference with delta % steps

    Parameters
    ----------
    problem : dict
        SALib problem specification
    N : int
        Number of samples
    delta : float
        Finite difference step size (percent)
    seed : int or None
        Random seed value
    skip_values : int
        How many values of the Sobol sequence to skip

    Returns
    -------
    np.array : DGSM sequence

    References
    ----------
    1. Sobol', I.M., Kucherenko, S., 2009.
       Derivative based global sensitivity measures and their link with
       global sensitivity indices.
       Mathematics and Computers in Simulation 79, 3009-3017.
       https://doi.org/10.1016/j.matcom.2009.01.023

    2. Sobol', I.M., Kucherenko, S., 2010.
       Derivative based global sensitivity measures.
       Procedia - Social and Behavioral Sciences 2, 7745-7746.
       https://doi.org/10.1016/j.sbspro.2010.05.208
    """
    if seed:
        np.random.seed(seed)

    D = problem["num_vars"]
    bounds = problem["bounds"]

    # Create base sequence - could be any type of sampling
    base_sequence = sobol_sequence.sample(N + skip_values, D)

    # scale before finite differencing
    base_sequence = scale_samples(base_sequence, problem)

    dgsm_sequence = np.empty([N * (D + 1), D])

    index = 0
    for i in range(skip_values, N + skip_values):
        # Copy the initial point
        dgsm_sequence[index, :] = base_sequence[i, :]

        index += 1
        for j in range(D):
            temp = np.zeros(D)
            bnd_j = bounds[j]
            temp[j] = base_sequence[i, j] * delta
            dgsm_sequence[index, :] = base_sequence[i, :] + temp
            dgsm_sequence[index, j] = min(dgsm_sequence[index, j], bnd_j[1])
            dgsm_sequence[index, j] = max(dgsm_sequence[index, j], bnd_j[0])
            index += 1

    return dgsm_sequence





def cli_parse(parser):
    """Add method specific options to CLI parser.

    Parameters
    ----------
    parser : argparse object

    Returns
    -------
    Updated argparse object
    """
    parser.add_argument(
        "-d",
        "--delta",
        type=float,
        required=False,
        default=0.01,
        help="Finite difference step size (percent)",
    )
    return parser





def cli_action(args):
    """Run sampling method

    Parameters
    ----------
    args : argparse namespace
    """
    problem = read_param_file(args.paramfile)
    param_values = sample(problem, args.samples, args.delta, seed=args.seed)
    np.savetxt(
        args.output,
        param_values,
        delimiter=args.delimiter,
        fmt="%." + str(args.precision) + "e",
    )



if __name__ == "__main__":
    common_args.run_cli(cli_parse, cli_action)