Saving and sharing¶

ns.save(ds, "result.h5")       # HDF5, NeXus layout
ns.save(ds, "result.nxs")      # the same, with the NeXus extension
ns.save(ds, "result.zarr")     # Zarr v3, for very large data
ns.save(ds, "result.tif")      # TIFF, calibrated, NUScope document in the description

Every format carries the data, calibrated axes, the normalized and original metadata, the full history and a seal, and every one of them opens again with ns.load() exactly as it was saved. Saving adds a step to the history, and missing folders are created.

Format	Best for	Notes
`.h5` / `.nxs`	most data; sharing with h5py, HDFView, MATLAB, NeXus tools	gzip-compressed; summary of the seal in the root attributes
`.zarr`	very large data, cloud storage	chunked
`.tif`	image viewers, ImageJ/Fiji	resolution tags carry the pixel size

ns.save(..., sign=False) seals without signing; sign=True refuses to save if you have no key. The default signs when you have a key.

RO-Crate packages for a repository¶

from nuscope.crate import write_crate
write_crate([a, b], "deposit", title="Ti alloy BSE series",
            description="...", authors=[{"name": "Your Name", "orcid": "https://orcid.org/..."}],
            include_sources=True)

nuscope crate deposit/ a.h5 b.h5 --title "Ti alloy BSE series" --author "Your Name|https://orcid.org/..."

The folder holds the datasets and an ro-crate-metadata.json (RO-Crate 1.1) describing files, people, instruments, the facility, every processing step, and each file's seal and signers. Zip it and upload it to a repository that issues a DOI.

PROV-JSON¶

import json
json.dump(ds.provenance.to_prov_json(), open("prov.json", "w"))

or nuscope provenance result.h5 --prov-json prov.json.