DIALS core meeting 2021-06-03

Previous Actions

(Note from ND: Marking items with further discussion on the agenda below with *)

• MG: organise a typing intro lecture, likely to happen in June.
• dx2 proposal
  • MG to put in a PR to create a proposal in the proposal-space
  • MG+ND: Come up with a proposal to move away from all code being in header files and consolidate into a single library
• *Move to hdf5plugin.
  • GW/MG: Make pull request for moving to hdf5plugin package
  • Wait for the 3.5 release before moving over
• *cbflib conda-forge package
  • ND: Try using old cbflib 0.9.6 (allowing regenerated pycbf) to see if it would be suitable for DIALS/dxtbx to release just the old version
• Merge image projection changes in dials#1716 (prev)
• cbflib conda-forge/pycbf packaging
  • Aaron: Look into labelit regression tests
  • Aaron: Consider, discuss with HJB
• Make mamba the default installation method for dials bootstrap (dials#1730)
• Stop running cctbx testing at diamond
• Aaron: Work out provenance of dials_regression images - what is safe to publish (long term)

Agenda

DIALS deprecations

dials/dials#1731 will need some work in xtc_process for MaskGenerator.

cctbx/dxtbx#371 can go in.

cbflib conda-forge package/pycbf

• Version: 0.9.6.3 - fixes some distribution issues. .4 coming soon to use released dials-data datasets (rather than a fork). Used a Zenodo source for MAR format (no changes to implementation)
• conda install pycbf also works now via conda-forge
• dxtbx#368 still exists for optional usage in dxtbx, no binary dependence on cbflib/_adaptbx if present.
  • Still stray issue with dataset SPring8_ADSC_SN916 being the only one requiring FormatCBFFull.
• dials/pycbf#19 HJB thread
  • 0.9.6 is “Obsolete”
  • 0.9.7 due late June
• Still need to test with labelit/labelit_regression once the test suites actually work
  • GW needs to review https://github.com/cctbx/dxtbx/pull/369 (merged)
• MG Action: Add HJB as pycbf owner. MG: Invited.

Python 3.6 Support

• Previously discussed proposal DC-3 to move DIALS/dxtbx to NEP-29
• We needed an awareness of the Phenix support plan
• Any update on this?
• Phenix is going straight to 3.7
• Okay to commit to this

Delete master branches

We moved to main 3 months ago. Can the master branches be removed?

• Yes

hdf5plugin

Pull requests have been made in dxtbx#373 and dials#1733. Suggest existing developer installs update and remove the hdf-external-filter package manually.

• merge these

Status checks before merging

We had two cases recently where pull requests were merged too eagerly (dials#1699, dials#1729). We now have a status indicating build step. Should we set it so that the pull request build has to complete before merging? We can then also enable automatic merging once the build passes.

scitbx.flex ABI shift problem

cctbx#616

Deferred to next meeting

Next meeting

Thursday, June 17th, 4pm UK (BST), 8am PDT.