Open Tom-Willemsen opened 3 months ago
Reflectometry scientists, at least, would like axis_date_time
formatted filenames - therefore not using scan_ids.
They would like them in a cycle-number specific folder (the cycle number is available as a PV from the DAE)
As a data administrator (?) I would like to figure out how scans output data gets stored in the long term.
bluesky
output files contain scientifically-useful data, grouping together a series of related measurements into a coherent "dataset". Therefore they should probably be able to be stored long-term in a similar way to the existing.nxs
files.@GRyall and @ChrisM-S have already both told me they have opinions on this topic.
Acceptance criteria
Documented resolution (and implementation tickets) for the following:
scan_id
in various combinations with/without instrument name prepended and so on.bluesky
can write data so that it is treated as long-term scientifically-useful data and gets archived in the right ways for long-term scientifically-useful data.Notes
scan_id
- a numeric ID. By default this is not persisted anywhere and simply starts from 1 and increments in each new python session, but it is easy to configure the run engine to get the "next scan id" from any source (e.g. a file, database, network service, etc etc). If we want a constantly-incrementing scan id on a per-instrument basis that is easy to do.scan_id
at all, as far as I understand it.