added sigma_G to fits

[damonge]

This is just a copy-paste of @fjaviersanchez 's modifications to avoid having to resolve too many conficts with outdated branches.

The only difference is that sigma_G is now saved in the "background cosmology" HDU, rather than the density skewers. Let me know if this doesn't make sense @fjaviersanchez, @andreufont, @jfarr03

[fjaviersanchez]

Thanks @damonge! Makes sense to me. Let’s see what @andreufont thinks :)

[andreufont]

Looks good!