Repository navigation
filewriter: decide where histogramming needs to occur #78
Description
Histogram mode files are looking like they are a requirement for HRPD-X.
We need to decide at what point in the pipeline:
A) TCBs are actually specified for a histogram
B) A histogram file is created
We may be able to do this after and event mode file is created, ie. by the filewriter, using a service like the FIA. Alternatively we could do it in the filewriter.
The streaming hardware does not currently have the capability to stream histograms so this should not be a consideration. It also becomes more of a challenge for propagating TCBs.
Activity
Tom-Willemsen commentedon Jul 14, 2026
I feel this is best considered alongside filewriter requirements gathering/design work.
[-]decide where histogramming needs to occur[/-][+]Filewriter: decide where histogramming needs to occur[/+]Tom-Willemsen commentedon Jul 27, 2026
See also #84 .
Not strictly a duplicate as this issue may still be needed for e.g.
kdaectrlto have some options to specify TCBs and stream them somewhere.[-]Filewriter: decide where histogramming needs to occur[/-][+]`filewriter`: decide where histogramming needs to occur`[/+][-]`filewriter`: decide where histogramming needs to occur`[/-][+]`filewriter`: decide where histogramming needs to occur[/+]Tom-Willemsen commentedon Oct 1, 2026
This needs to happen in the filewriter - as we want to support writing combined histo+event files, separate histo/event files, and all the permutations.
Writing a full event file first and then loading that entire file into external software just to do histogramming is very inefficient (and slow) for instruments who only really want the histogrammed version as the representation for analysis.
Histogramming will be done by a writer-module in the filewriter which will need it's bin boundaries passed to it as a module configuration parameter in the
nexus_structure.