Convert Nifti to NGFF

nifti2ngff.py

import math
import numpy as np
import nibabel as nb
import zarr
from numcodecs import Blosc
from ome_zarr.writer import write_multiscale
from skimage.transform import pyramid_gaussian

img = nb.load("sub-S01_T2w.nii.gz")
data = img.get_fdata()
data5d = data.swapaxes(0,2)[None, None, :, :, :]

pyramid = [val for val in pyramid_gaussian(data5d, max_layer=4, downscale=2, preserve_range=True)]

storage_options = []
for key in range(len(pyramid)):
    chunkdim = min(64, 2**math.floor(math.log2(min([val for val in pyramid[key].shape if val > 1]))))
    chunks = (1, 1, chunkdim, chunkdim, chunkdim)
    storage_options.append(
        {
            "chunks": chunks,
            "compressor": Blosc(cname="zstd", clevel=5, shuffle=Blosc.SHUFFLE),
        }
    )

for idx, val in enumerate(pyramid):
    pyramid[idx] = np.round(val).astype(np.uint16)

ngff_file = "sub-S01_T2w.ngff"
store = zarr.NestedDirectoryStore(ngff_file)
root = zarr.group(store=store, overwrite=True)
write_multiscale(
    pyramid,
    group=root,
    axes=["t", "c", "z", "y", "x"],
    storage_options=storage_options,
)
multiscales = root.attrs["multiscales"]
for level in range(len(pyramid)):
    multiscales[0]["datasets"][level]["coordinateTransformations"][0]["scale"] = [1.0, 1.0, (2**level)*0.05, (2**level)*0.05, (2**level)*0.05]
root.attrs["multiscales"] = multiscales
for i in range(5):
  root[i].attrs['_ARRAY_DIMENSIONS'] = ['time', 'channel', 'z', 'y', 'x']
zarr.consolidate_metadata(store)