Computation power spectra with folding

[JegerBroxterman]

Hi,

The way the power spectra with folding are computed currently is by changing the positions of particles according to

box_over_fold_x = box_x / fold
inv_box_over_fold_x = 1.0 / box_over_fold_x
x_pos = (x_pos % box_over_fold_x) * inv_box_over_fold_x

with fold = 2.0^folding, i.e. (from swiftsimio readthedocs)

Should this not be
x_pos = (x_pos % box_over_fold_x) * fold ?

For example, without folding (folding = 0 and fold =1) the positions are now converted to
x_pos = (x_pos % (Lbox/1) ) * 1 / Lbox = x_pos / Lbox

[kyleaoman]

At the very least the units seem wrong as-is but correct with the proposed change... I'd be happy to write a patch but I'm a little bit out of my depth on the theory here. @JBorrow if you can confirm that this is a bug as reported, I can probably fix it.

[JBorrow]

Why is this bad? Do you not want x_i' to be dimensionless bounded between 0 and 1? Perhaps the re-use of the variable is not ideal :)

[JegerBroxterman]

I think the re-use of the same variable indeed is the thing that confused me. I think the current version of deposit() in the power_spectrum_py code works, but the descriptions of the input parameters are

    x : np.array[float64]
        array of x-positions of the particles. Must be bounded by [0, 1].

    box_x: float64
        box size in x, in the same rescaled length units as x, y, and z. c

whereas the input in render_to_deposit() is

    positions = data.coordinates

    # Get the box size
    box_size = data.metadata.boxsize

    # Deposit the particles
    arguments = dict(
        x=positions[:, 0].v,
        y=positions[:, 1].v,
        z=positions[:, 2].v,
        m=quantity.v,
        res=resolution,
        fold=folding,
        box_x=box_size[0].v,
        box_y=box_size[1].v,
        box_z=box_size[2].v,
    )

i.e. both box size and positions are not rescaled to [0,1]. I don't think it is an issue as both are not rescaled, but maybe change the description of the parameters in deposit() and deposit_parallel()?