Repository navigation
filewriter: periods #93
Description
Entries under the raw_data_1/periods group.
Some of these entries are just per-period versions of dynamic metadata that exists elsewhere. These items should be straightforward to do it the same way as other 'dynamic' metadata.
Other items involve details of hardware-period schemes, which need discussion with DSG to work out how the new electronics treat these.
Questions
HRPD-X will likely not use periods, but we need to do research to understand:
- How exactly does the streaming system handle hardware periods, DWELL periods, and suchlike (talk to DSG)
- Which elements of the current format make sense in streaming system?
- Which parts of the
raw_data_1/periodsgroup does mantid minimally require to load a file? (talk to Mantid team) - Which parts of the
raw_data_1/periodscan mantid read for instruments that run in multi-period mode regularly? (talk to Mantid team)
Potential differences from existing files
- Per-period metadata: no differences expected
- Hardware-period related datasets: to be determined, pending discussions with relevant teams about how hardware periods work in new system.
Activity
RichardWaiteSTFC commentedon Oct 15, 2026
I need to look into this further - but quick search shows this (I will discuss with Core team and get back to you)
Possibly just
raw_data_1/periods/proton_chargebut I will check!https://github.com/mantidproject/mantid/blob/91d24dfaf4404fa13f3266c7c12dc066b2a537dc/docs/source/algorithms/LoadISISNexus-v2.rst?plain=1#L63
I know muon instruments use multi-period mode - I found a comment
// For multiperiod datasets, read raw_data_1/periods/good_frameshttps://github.com/mantidproject/mantid/blob/91d24dfaf4404fa13f3266c7c12dc066b2a537dc/Framework/DataHandling/src/LoadMuonNexusV2NexusHelper.cpp#L77
Is it worth asking MNeuData/SuperMUSR data pipeline people about this?
Tom-Willemsen commentedon Oct 15, 2026
Just to note, SuperMuSR is currently separate a separate filewriter from this work - though could potentially end up merging longer term, and this is something we have discussed/are discussing with SuperMuSR team.
Expanding the list from the ticket description, the ones I'm confident we'll be able to keep "the same" are:
periods/numberperiods/good_framesperiods/proton_chargeperiods/raw_proton_chargeperiods/raw_framesperiods/highest_usedThe ones which are a little riskier are:
periods/total_counts(length of this dataset depends on DAQ period vs DWELL period details - this concept will exist but may be implemented differently in streaming system)periods/good_frames_daq- same caveat as aboveperiods/sequences- not impossible to emulate, but could be fiddly/require very careful logic.periods/frames_requested- awaiting details from DSG about exactly how hardware periods work in new system, but in principle should exist and be implementableperiods/type,outputandlabel- as above, awaiting details from DSG - in principle should exist and be implementableSo I'd mostly be curious about how/where mantid uses items in this second list.