DIALS core meeting 2021-09-09

hackmd-github-sync-badge

[Previous meeting: 2021-08-26]

Previous Actions

  • ND: conda-forge pycbf: Make new release to use dials-data directly for tests
  • "Contiguous Nexus" cctbx/dxtbx#356
    • Non-draft state is pending checks against issues @dwpaley raised, and checking for non-contiguous cases [prev]
  • dxtbx src/ layout [prev]
    • Nick: Ask cctbxbb if anyone has knowledge or experience of re-export of dispatchers on windows
    • This is a higher priority now - David feeling pain of missing
  • Removal of DataBlock [prev]
    • Make a longer-term DC for migration over ~6 month period
    • Used in tests, documentation
    • Possibly heavily used by 2020 LS49 paper
  • CCTBX Occasional failure of mirror job
  • DIALS: Remove old environment conda files: dials/dials#1865
  • DXTBX/pycbf [prev]
    • Outstanding: dxtbx#368 for optional cbflib/_adaptbx .
    • Last status: one labelit/labelit_regression test still doesn't pass
    • All except one labelit tests are now working. On NKS todo list to fix.
    • Progress on CBFlib 0.9.7?

Agenda

Datablock

  • Graeme asks how hard to remove and whether we should just allocate a block of time
  • (Possibly heavily used by 2020 LS49 paper)
    • Dan thinks that it's described more than used
  • Discussion on approach of doing this work
    • More visible deprecationwarning
    • Impact assessment and see who/what code is affected
    • Actual removal probably not too difficult

Nonconventional P1 indexing

  • dials/dials#1878
  • XFail test in dials/dials@9b4369
  • GW: Perhaps we need to take what the user gives us, map into standard setting and remember that reindex so that it can invert the reindexing at the end
  • DW: Has encountered this problem, needs to manually reindex back to the setting that he wanted
  • GW: Another example of comparing unit cells in a setting that is more convenient for comparisons
  • Discussion over, should we do this - an argument for always returning in standard setting and relying on user to transform it back if they really want that
  • Discussion seems to be concluding that returning in user-specified order probably makes sense
    • Not doing so has XFEL gui repercussions
    • It already effectively does this for non-P1 cells
    • Action: Richard going to have a look to see how hard this is to do

dxtbx src/ layout

  • dxtbx and xia2 were both fixed for read-only install
  • xia also had the offline issue

Deterministic Experiment Identifiers

  • PR: dials/dials#1864
  • GW: It would be useful to describe the problem that is trying to be solved
  • ND: (again) bring up - would a general metadata field that can have an ast.literal_eval dictionary embedded solve this problem?
    • DP: Yes, this would seem to solve the issue
  • Matching up multiple lattices based on image/lattice count
    • Is this reproduceable - getting the same N lattices doesn't mean that they are in the same order
    • GitHub discussion possibly worthwhile for both of these

Next meeting

Thursday, September 23rd, 4pm UK (BST), 8am PDT.

Select a repo