Conversation
and updated the notebook about reproducibility
|
Not sure why the tests are failing, but I agree with the changes in the defaults (although we could also delete the options from argparser). |
|
It fails because the functionality is only available in python 3.9 and the tests are in 3.8. @sbailey is there any reason to keep the tests with python 3.8? I see other packages like desispec are already in 3.9... |
|
@alxogm - is this still a relevant PR? Was it addressed elsewhere? |
|
Catching up on ancient PRs — @alxogm @andreufont @abrodze @andreicuceu do we still want this, or is it irrelevant at this point? Go ahead and merge if you want it, or otherwise please close it. Don't worry about github unit tests; we'll clean those up separately since they aren't directly related to these changes (and in some cases are due to python versions we don't support anymore...) |
|
After chatting with @HiramHerrera -- this PR is a harmless update; however, let's not merge it until after unblinding, or the change will need to be propagated through the entire Lya mock production pipeline for DR2. |
4 participants
Changed the defaults for options
--zbest, --bbflux and --desi_footprint to be true as defult. Now in order to set them false we have to give --no-zbest, --no-bbflux, and no-desifootprint.
@andreufont @HiramHerrera @andreicuceu I think those are the only arguments that make sense to change for now... let me know if there are other suggestions.
I also updated a bit the notebook about reproducibility to work with the latest version of the mocks. I simplified the notebook and included some information about the truth files. This can be improved in a future PR is we want to...