---
name: DIALS core meeting 2025-05-08
tags: core meeting
---

# DIALS core meeting 2025-05-08

## Previous Actions

- [ ] **ND** Investigate getting `psana` tests running on the DIALS xfel-regression testing
- [ ] **ND**: Move dials.diamond.ac.uk to a redirect to dials.github.io
- [ ] **ND**: Action: Replace dials-support link with slack link/new mailing list address

## Agenda

### Xarray

- Last time welcomed Ian Hunt-Isaak, dials issue was created, https://github.com/dials/dials/issues/2900. Yash Karan looking into this at DLS.

### Dials-support stuff
- **ND:** Ask Charles to apply CCP4 patch to move mailing list notice
- DGW: Charles has been asked. We don't have confirmation that it has been applied yet. Proposed check in ~2 months i.e. start of July. Keep on Agenda for now.

### DIALS Office Hours
- 3pm Thursdays was what we talked about
- ND talk to JBE about running
- Make informal promise for office hours, and a slack channel that people can make known that they have a request to chat/for help about a particular topic
- JBE booked room, not sent out invites/ zoom yet - try to get done in next week.

### DIALS jenkins builds broken since last few days (25/4/25)?

- This is broken in Jenkins
- **ND**: This is broken
- Dan will have a look and try to work out the problem
- dxtbx and dials pull request - egg info changed to dist info https://github.com/dials/dials/pull/2906 and https://github.com/cctbx/dxtbx/pull/807
Dials PR removed the setuptools pin.
- Fixed now, remove from agenda for future.

### Metrics
- Aaron wants to have a dials tag attached to all export, then we can make sure that is always carried through to refined structure. Aaron will handle Phenix, DW will try for CCP4.
- CCP4 GUI, data is lost when packaging mmcif
- Ongoing investigations, optimism that can be fixed
- DGW made some progress, agreement to add history recording into experiments file. Will pick up again after Easter, but WIP.
- Discussion between Aaron/DGW about contents of history/approach across cctbx/dials
- Address when david back
- 08/05/25 update - DGW - PR nearly ready that records history item in experiments file. Top level field outside of experiment list. But will require a number of dials changes as often we make new lists/copies, so will need to copy across history. dxtbx PR first, then follow up with dials once in.At point of export, look through existing history items and extract integrate, scale lines and filter those out.
- Follow up with Aaron again next time.

### Huuuuge Integration
- A discussion about integration memory usage
- Also discussed how this affects usage
- Update (JBE, 04/04/25) - I have made some more progress on this, see draft PR at https://github.com/dials/dials/pull/2890 Putting it here so you can discuss as I am away this week so won't be at this meeting. Although this is a specialisation (photon counting detector, 3d constant background model), I think it is the defaults we use for synchrotron rotation data processing. Might be able to be further generalised in future. Idea would be for some auto logic to use this if the criteria are met in dials.integrate so it would automatically pick up in xia2 as well.
- JBE probably needed to discuss this, defer detail until next time
- discussion about Dan's work measuring and looking at improvement of indexing/refinement compared to XDS
- Also discussion about applicability of James Holton's multi-turn datasets and whether we should focus on single turns
- No further progress, ongoing.

### Big Computer?

- Discussion about a support thread Aaron has about DIALS apparently causing blue screen (!) on windows
- ND/GW Will follow up on forwarded email
- Thread continued
- Format issue now
- DGW doing working
- DGW: turned out to be a user support issue, remove from agenda.

### Orderedset
- Was not a trivial swapover for things in dials
- More changes needed to be made
- Everyone thinks this was resolved?
- Graeme would like dependency update PRs to include instructions on how to update
- This seems reasonable
- Another issue found and fixed by JBE this week.
- PR guidelines for DIALS to include such things? DGW not too in favour of being too restrictive.
- Remove off future agenda unless comes up again.

### Numpy 2 / Python 3.13

- Queries about what it would take to support
- ND is going to look at
- Dials west would like this to be addressed quicker than never
- KD ran into this on install dials and pytorch in same env: ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
nxmx 0.0.4 requires numpy<2, but you have numpy 2.2.5 which is incompatible.
- DGW - is it backward compatible? Worried will hit ccp4. KD says they did for RSShip recently - this PR: https://github.com/rs-station/reciprocalspaceship/pull/261
- To test, let's add to matrix to run both and python 3.13 in branch. DP will do this
- Keep on agenda.

### Retiring `dials_regression`
- https://github.com/dials/dials/pull/2915. Happy with the PR?
- In particular note that `test_filter_overlaps.py` is now skipped, because it does not appear to work as expected
- Not clear why this 'feature' doesn't work, don't understand code or test enough. Already some open issues on this. Crashes if you try to run this with dials.integrate. Good example now in dials data.
- Does CI take longer now running more tests? DGW didn't notice that obviously longer. MacOS timeout.
- Time for a test review? Dedicate some of the DIALS diamond meetings to this?
- Keep on Agenda.

## AOB

### Next meeting

Thursday, May 22nd, 4pm (BST), 8am (PDT), 10am (CDT)