search

spacetelescope / poppy

Physical Optics Propagation in Python
https://poppy-optics.readthedocs.io
BSD 3-Clause "New" or "Revised" License
216 stars 72 forks source link

Update intel channel location #627

Closed BradleySappington closed 3 months ago

BradleySappington commented 3 months ago

Github actions are currently failing because the intel channel returns 403 error

The intel channel is used for the mkl_fft library This PR replaces intel channel that no longer exists in coda with the intel self hosted https://software.repos.intel.com/python/conda/

mperrin commented 3 months ago

@BradleySappington Thank you! I have noticed that there was that error from the missing conda channel, but hadn't yet had time to look into it any further. Much appreciated for your taking care of figuring out what to do about it.

BradleySappington commented 3 months ago

@mperrin my pleasure...before I dive too deep, do you have any insight on why this change would affect our coverage test?

mperrin commented 3 months ago

that error in the coverage test is baffling. It's a total non-sequitur, and is from a part of the code that has nothing at all do do with FFTs... 🤷‍♂️

mperrin commented 3 months ago

I'm going to try re-running that one failed job, to check if it was some transient unrelated CI glitch or something like that

mperrin commented 3 months ago

This is definitely weird. The code which is failing, apparently repeatably, has absolutely nothing to do with the FFTs. That portion of the CI is testing the ability of poppy to use synphot for synthetic photometry. The example test cases in poppy do that using a couple of basic standard optical filters, not any of the particular JWST or HST filters, and synphot should automatically retrieve the filter information from some reference data. This works for me locally, but for some reason is failing here in the CI.

For a minimal repeatable example outside of the poppy CI stuff, the relevant tiny bit of code is simply: 

import synphot
synphot.spectrum.SpectralElement.from_filter('johnson_b')

That has absolutely nothing to do with FFT libraries at all...!?!

BradleySappington commented 3 months ago

@mperrin - looks like it just needed some rest, a monday re-run did the trick