Add a helper utility for generating gaussian data for `ParticleGroup`s

[ken-lauer]

Initial attempt here:

from typing import Optional
import numpy as np


def gaussian_data(
    n_particle: int = 100,
    charge: float = 1e-9,
    p0: float = 1e9,
    mean: Optional[np.ndarray] = None,
    sigma_mat: Optional[np.ndarray] = None,
):
    """
    Makes Gaussian particle data from a Bmad-style sigma matrix.

    Parameters
    ----------
    n_particle: int, default=100
        Number of particles.
    charge : float, default=1e-9
        Charge in C.
    p0 : float
        Reference momentum in eV/c
    mean : np.ndarray of shape (6,), optional
        Mean positions. Default = None gives zeros
    sigma_mat : np.ndarray of shape (6,6), optional
        Sigma matrix in Bmad units. If default, this is the identity * 1e-3

    Returns
    -------
    dict
        ParticleGroup-compatible data dictionary:
        >>> ParticleGroup(data=gaussian_data())
    """
    if mean is None:
        mean = np.zeros(6)

    if sigma_mat is None:
        cov = np.eye(6) * 1e-3
    else:
        cov = sigma_mat

    dat = np.random.multivariate_normal(mean, cov, size=n_particle)
    x = dat[:, 0]
    px = dat[:, 1]
    y = dat[:, 2]
    py = dat[:, 3]
    z = dat[:, 4]
    pz = dat[:, 5]

    data = {
        "x": x,
        "px": px * p0,
        "y": y,
        "py": py,
        "z": z,
        "pz": (1 + pz) * p0,
        "t": np.zeros(n_particle),
        "weight": charge / n_particle,
        "status": np.ones(n_particle),
        "species": "electron",
    }

    return data

from pmd_beamphysics import ParticleGroup
P = ParticleGroup(data=gaussian_data(100_000))
P.plot('x', 'px')