pace-neutrons / Horace

Horace is a suite of programs for the visualization and analysis of large datasets from time-of-flight neutron inelastic scattering spectrometers.
https://pace-neutrons.github.io/Horace/stable/
GNU General Public License v3.0
7 stars 4 forks source link

POST-CLOSEOUT PLANNING #1719

Open cmarooney-stfc opened 2 weeks ago

cmarooney-stfc commented 2 weeks ago

Format: Looking for customer requirements to be met. We should try to postpone discussion of solutions at this stage.

Agenda:

  1. Toby proposal See circulated document.

    • Ongoing: Maintenance, more comprehensive coverage of the code by the automated tests, distribution of releases etc.
    • Ongoing: Responding to instrument scientist and user requests for additional functionality.
    • Year 1: Creating a GUI.
    • Year 1: Benchmarking and optimisation of CPUtime expensive functions.
    • Year 1: Additional projection types beyond the current linear and spherical (and cylindrical by 31/3/24) e.g. energy focussed cuts.
    • Year 1: Read/Write Mantid MDworkspace and HDF output of DnD objects.
    • Year 1: Implement an sqw object compression methodology (expected reduction of file sizes typically about a factor of 5) to speed up I/O (priority may be increased if problems with CEPH and IDAaaS I/O persist). Essential for continuous sample rotation operation.
    • Year 1:
    • Faster handling of the instrument information (only unique instances of instrument components are retained to keep memory consumption acceptable).
    • Year 2: Making ‘Mushroom ready’
  2. David requirements (decreasing priority)

    • folding spinw and powder fitting in (done except for parallel)
    • rietveld for inelastic finished off with the above line except for
    • speedup of data visualization for larger datasets, or with continuous rotation- may be idaaas/ceph issue
    • resconv - Tobyfit works but not accelerated, mcstas may deal with this (using gpu) plus capture of fine details - other solutions may be available - other groups may also wish mcstas - resconv may also be available to other users as library, requires generic approach - instrument definition files usable for input (though mcstas aalso has such definitions) but this is also calculation - sngle crystal first then powder
    • indirect instruments not yet supported
    • implies GPU McStas done first
    • want powder resconv
    • couple spinw/horace in parallel (=duc fix bug)
    • gui - (a) data visualization e.g. produce cut and export script, reduce button clicks for basic ops, ? include symmetry ops? (noted that data provenance already planned) + uX experience input (b) sqw create priority ultra-low
    • ? crystal field excitations fitting
    • ubmatrix import to pace
    • absorption corrections
    • anharmonic higher order phonons