CCL compatibility

[humnaawan]

@slosar sorry to bother you again but can you say what is compatibility requirement between LSSLike and CCL? I am using the CCL version that comes with the DESC stack (v1.0.0), and I am getting errors (e.g. there's no Parameters objects in CCL but lss_theory.py calls it).

[slosar]

Ok, humna, no problems at all, but what you are trying to do?
The hsc_fit branch is where all the HSC work is happening and that has been updated to work with the recent CCL (and in fact requires git repo version of CCL).
The master branch of LSSLike is like truly ancient and we'll probably just scrap it as TXPipe functionaliy has superseeded it.
So I think we need to first see what is the project and then pick the best tools for it.
@damonge @anicola

[humnaawan]

thanks for a quick response! i'm trying to set up the code structure to get constraints on cosmological params based on power spectra (for this DESC project). i basically have simulated galaxy density fields with different systematics and i want to quantify their impacts on cosmological params. so far, i have NaMaster working, and I think I have the right SACC objects (following this tutorial notebook).

i can look into the hsc_fit branch and maybe that'll answer my questions. my impression regarding CCL installation is that it might be non-trivial to get a version of it working on NERSC outside of the desc-stack, but i can try.

[slosar]

Then I think you want to use firecrown. As far as I understand it is ready to ingest sacc files and run chains! In fact, I think it would be a great test to see how well this works in practice!
(but then again, you will need to write outputs in sacc2, :) , but believe me, it is going to be worth it.)

[humnaawan]

okay, great. i can try that. thank you!

[joezuntz]

I don't think anyone's written the firecrown-sacc2 interface yet. I know I haven't!