Extending the project

[haricash]

Hello All,
I recently came across this package. I feel it can have more functions used in cosmology, such as a quick power spectrum calculator given the density fields. Do you think we can add new stuff to it? Is there an active effort in doing so?

[mileslucas]

Hi @haricash! Sorry for the belated reply 😅

I have very little cosmological expertise, but I did some searching about theoretical power spectra (this package, for example) and that seems like a welcome contribution. However, I'm not sure whether it should be done in a new package or not. @giordano any advice?

[giordano]

Yes, probably it makes sense to have it here.

[haricash]

Cool then! I can't start on it right now as I have exams going on. I'll do it right after!

[JaimeRZP]

Hi All!

My name is Jaime Ruiz. I am cosmology PhD in the Universiy of Oxford working with David Alonso and Pedro Ferreira.
I also recently joined the Turing.jl org.
I recently developed a clone of CCL in Julia which might do what you are proposing, LimberJack:
https://github.com/JaimeRZP/LimberJack.jl

It has a few limitations but we are hoping to publish it soon!
Hope this is useful!

[cgarling]

Hi all, as this package has been seeing some activity lately (e.g., #51) I wanted to pitch the idea of upstreaming some of the development from my fork of Cosmology.jl. I've added a number of methods that I needed for downstream analysis packages (analytic power spectrum models, halo mass functions, etc.) that may be of interest -- check out the "public methods" section of the documentation. There are also built-in cosmological parameter sets (e.g., Planck18) that make doing quick analyses simpler.

I could also look into releasing those downstream packages if there is interest in those applications as well. The power spectrum package, for example, has analytic forms (e.g., Eisenstein & Hu) that take cosmology objects as input, which can then be used to compute halo mass functions and so on (similar to the hmf package in python). It also supports PyCAMB through PyCall.jl, but integrating LimberJack.jl would be nicer. There is also SymBoltz.jl and CLASS.jl as well.

[gwisk]

+1 to @cgarling's suggestion — I think Julia could use a cosmology code, and I think integrating the features you mention would close the gap between the cosmology packages available for Julia and those available for Python.

Personally, I'd most like to see a Boltzmann solver in Julia, so I'd be very interested to see the downstream packages you mention. Personally, I find PyCAMB a little clunky, and since it already wraps the CAMB Fortran code, a 'closer to the metal' approach would be to either integrate LimberJack.jl (which is written in Julia), or to write a Julia wrapper for CAMB, bypassing PyCAMB entirely.