Euphonic Calculations: Refactor Abins/Mantid to use Euphonic

pace-neutrons / Pace-Project-Plan

0 stars 0 forks source link

Euphonic Calculations: Refactor Abins/Mantid to use Euphonic #68

Open ajjackson opened 4 years ago

ajjackson commented 4 years ago

Blocking #24

[ ] Refactor Abins to use Euphonic for incoherent powder averaging in the analytic "almost-anisotropic" approximation. This shouldn't be too problematic if the refactoring work to support #67 has already been done.
[x] Arrange for an appropriate release of Euphonic to be included-in/distributed-with Mantid. My understanding is that this will require somebody at ILL to manually do a pip upgrade whenever the packaged Euphonic needs updating, so we should avoid doing that too often. Once per Mantid release, say?

ajjackson commented 3 years ago

The current roadmap is to

integrate Euphonic with the Mantid SimulatedDensityOfStates algorithm (and settle packaging/distribution issues)
Integrate Euphonic with Abins to import force constants data and generate frequencies/eigenvectors on-the-fly

Then it will become possible to develop #67 , which is the first stage of this issue.

So the original check-list should be completed in reverse order!

ajjackson commented 2 years ago

Update: the Euphonic/SimulatedDensityOfStates development has been implemented and is just awaiting a bit of snagging re: reliable installation of Euphonic on Mantid systems. https://github.com/mantidproject/mantid/pull/30760 This could be greatly aided by wheel builds of Euphonic.

I have made a start on the next step, importing force constants into Abins using Euphonic. Packaging issues aside, it seems to already work fine: https://github.com/mantidproject/mantid/pull/32220