Have LUT be read bottom-up, and modify the 2x2 LUTs to match true geometry

[diaza]

As title says.
EDIT: Here's the talk where I went over the issues seen. If anyone missed it and would like a more detailed explanation, let me know and we can chat.
EDIT: I'm adding everyone think might have any familiarity with the LUTs. No need to actually review if this isn't within your purview.
To reflect the y-axis in the LUT, I used the following code:

import numpy as np

def swap_y(filename, outfile):
    data = np.load(filename)['arr']
    arr = data.copy()
    arr = arr [:, ::-1, :, :]
    n_op_chan = arr.shape[-1]
    arr[..., :n_op_chan//2] = arr[..., :n_op_chan//2][...,::-1]
    arr[..., n_op_chan//2:] = arr[..., n_op_chan//2:][...,::-1]
    np.savez_compressed(outfile, arr=arr)
    return

filename = "/global/cfs/cdirs/dune/www/data/2x2/simulation/larndsim_data/light_LUT/lightLUT_Mod0_06052024_time_norm.npz"
outfile = filename.replace(".npz", "_swapped_y.npz")
swap_y(filename, outfile)

filename = "/global/cfs/cdirs/dune/www/data/2x2/simulation/larndsim_data/light_LUT/lightLUT_Mod123_06052024_time_norm.npz"
outfile = filename.replace(".npz", "_swapped_y.npz")
swap_y(filename, outfile)

Flipping for FSD shouldn't be necessary.
The optical channel mapping will have to be remapped in flow.

[diaza]

I've opened PRs for op channel remaps for 2x2 and FSD.

[liviocali]

@diaza I did not review the new channel maps in detail. But everything else seems sensible to me. Do you have tools to verify the new channel map? E.g. Angela at some point produced edepsim files that should deposit some charge exactly in front of the SiPMs to verify the mapping.

[diaza]

@diaza I did not review the new channel maps in detail. But everything else seems sensible to me. Do you have tools to verify the new channel map? E.g. Angela at some point produced edepsim files that should deposit some charge exactly in front of the SiPMs to verify the mapping.

If @Angela-White has that ready to send, that would be great! Otherwise I can try to create my own, I've done something similar for 2x2

[diaza]

@liviocali
I've attached two gifs, showing how the light deposition changes as a function of charge deposition position in the FSD. The first shows for the existing code, the other for my modifications. You can see that the depositions in the second gif look more sensible.
EDIT: Replacing with better gifs.

[diaza]

For good measure, I'm showing below two gifs with a similar comparison as above, but in the 2x2. This comparisons should show that the modifications made to larndsim, the LUT, and the op channel mapping in 2x2 result in the same output. The top gif is for the existing code, and the second gif is for the updated.

[cuddandr]

The GIFs are enough validation for me. Everything looks good to me.