Globus at NUANCE

NUScope reads files straight from Globus collections and moves whole runs between them.

ds = ns.load("globus://<collection-id>/instrument/2026-10-13/run1/image_001.dm4")

Setup (once)

1. Register NUScope as a native app at https://app.globus.org/settings/developers and store its client id: nuscope globus setup <client-id> (or set NUSCOPE_GLOBUS_CLIENT_ID).
2. Find the collection id of the instrument or storage collection in the Globus web app (Collections → Overview).
3. nuscope globus login <collection-id> shows a Northwestern login link once; refresh tokens are kept under ~/.globus/app/.

Install the extra first: pip install "nuscope[globus]".

Reading and moving data

nuscope globus ls <collection-id> /path
nuscope globus get globus://<collection-id>/path/file.dm4 [dest]
nuscope globus transfer globus://<src>/run1 globus://<dst>/run1 --recursive --wait

from nuscope.globus import GlobusSession
s = GlobusSession()
s.ls(cid, "/path"); s.find(cid, "/", pattern="*.dm4")
s.download(cid, "/path/file.dm4", "file.dm4")          # (path, sha256, bytes)
task = s.transfer_to(src, "/run1", dst, "/run1", recursive=True)
s.wait(task)

Direct reads need a Globus Connect Server v5 collection with HTTPS enabled. For any other collection, and for large data, use a managed transfer, which Globus verifies with checksums. Files read this way are cached in ~/.nuscope/cache (or $NUSCOPE_CACHE).

The provenance of a Globus read records the collection id, the path, when it was fetched and the file's SHA-256.