Tool to extract raw data from NeXus formatted HDF5 files to "raw" memory mapped data files

extract.py

import os
import sys

import tqdm

import hdf5plugin
import h5py


def extract(nxs, template, start=None, end=None):
    """Extract data from the /entry/data/data dataset in nxs to
    individual frames following template, from start to end (range-style)"""

    with h5py.File(nxs, "r") as f:
        d = f["/entry/data/data"]
        s = d.shape

        if start is None:
            start = 0
        if end is None:
            end = s[0]

        if end > s[0]:
            raise ValueError(f"{end} beyond end of data set ({s[0]})")

        print(f"Source dataset shape is: {s}")

        for j in tqdm.tqdm(range(start, end)):
            i = d[j, :, :]
            i.tofile(template % j)


def main(filename):
    if not filename.endswith(".nxs"):
        raise ValueError(f"{filename} is not NeXus")
    template = os.path.split(filename)[-1].replace(".nxs", "_%06d.raw")
    extract(filename, template)


if __name__ == "__main__":
    main(sys.argv[1])