[Feature Req] Upcoming Features - Discussion

[camarman]

I want to try to implement these features into the hmf. I thought I could share them here. If they are okay with you, I'll try to implement each of them.

1. Improving docs (adding more equations or maybe creating a special page for units)
2. A function that converts into units that has no reduced hubble parameter (h)
3. A function that converts halo mass into stellar mass.
4. Cumulative stellar mass density.

So far these are my ideas but if more comes I'll share here.

[steven-murray]

Hi @seVenVo1d,

1. This is a good idea
2. One idea here would be to use astropy.units. I can't think of a good interface for it right now. The obvious thing to do would be to make all outputs from hmf astropy.Quantity objects, which can be manipulated. But that does make it a backwards-incompatible change, and would require quite a few changes across the codebase. You could write functions for just some of the most used outputs (like dndm) maybe?
3. This (and 4) is I think outside the scope of hmf. You could write another package that depends on hmf though?

[camarman]

Hello @steven-murray,

I am working on (1) and I'll try to do my best. For 2,3, and 4, I was thinking some sort of a function (tool) that can be used to convert these things, without altering the source code of the hmf. As you have pointed out, doing calculations in terms of natural units could alter the source code a lot..

[camarman]

5. I think we should add matplotlib in the requirements. But I am not exactly sure how to do that...

[steven-murray]

@seVenVo1d I do not think it's a good idea to add matplotlib. None of the hmf modules use matplotlib. While it is a useful extra tool to have if you're making you're own plots, it is not fundamental to the workings of hmf, and every extra dependency is something that has to be managed (i.e. extra work)

[camarman]

I see. That is understandable. In that case I'll work about the docs for now. I was busy for the last couple of weeks so I could not do much, but I'll contribute and open pull request maybe in a week or so. Meantime my research project gave me a lot of insight about the code and structure itself. I am at now much better position in understanding the code and its structure.