How to reproduce the 3D Slicer "resample Image(BRAINS)" action with monai ( or nibabel) ?

[ndzimbong]

With 3D Slicer, I can apply a 4*4 linear transformation matrix (

) to a nifty feature image (), and create a new one with the BRAINSResample module of 3D Slicer (

), which I can then save. The new one has the characteristics:

(note: the spacing and affine have varied). I want to write a python script based on monai to reproduce the same processing, but completely outside of 3D Slicer.
I've tried a few combinations, including:
1- changing the affine, header sform or header qform parameter values of the original nifty image
2- also by trying monai.transforms.SpatialResampling

With the first option, if I pass an affine that is not a diagonal matrix, it is not possible to display the transform on 3D Slicer. Moreover, I don't think the affine changes. The software can no longer read it. But I don't know.
As for the second option, I can't even get it to work

[wyli]

the spatial resample transform should work in this case, script would be something like the one I attached (this is using monai v0.9.1 and it's not tested with 3d slicer) it generates a new nifti file with the updated affine. depending on your linear transform's convention I think you can set lps_xform to true/false.

import torch
import monai
from monai.data.utils import orientation_ras_lps

# download an exmaple
monai.apps.download_url(
    "https://github.com/Project-MONAI/MONAI-extra-test-data/releases/download/0.8.1/avg152T1_RL_nifti.nii.gz"
)
filename = "avg152T1_RL_nifti.nii.gz"
img = monai.transforms.LoadImage(image_only=True, ensure_channel_first=True)(filename)

# linear transform to apply
xform_linear = torch.as_tensor(
    [[0.5, 0, 0, -45], [0, 0.5, 0, -45], [0, 0, 0.5, -50], [0, 0, 0, 1]], dtype=torch.float64
)

lps_xform = True
if lps_xform:
    dst_affine = orientation_ras_lps(xform_linear @ orientation_ras_lps(img.affine))
else:
    dst_affine = xform_linear @ img.affine
new_img = monai.transforms.SpatialResample(mode="nearest", padding_mode="zeros")(img, dst_affine=dst_affine)
monai.transforms.SaveImage(resample=False)(new_img)