Fix Stephan-Boltzmann constant accuracy and add radiation constant

[zingale]

and make it and Stefan-Boltzmann derived from fundamental units

[neutrinoceros]

I'm not against it, though I may have some reserves about using a single greek letter symbol for such a domain specific constant, but I do not understand why it's having an effect on existing doctests. Can you clarify ?

[neutrinoceros]

I'm not sure such a large change is warranted (unless it somehow counts as a bug fix), but why is this output changing at all ?

[zingale]

I didn't change the name of the constant -- it was already σ in the code.

What I did is fix the way the constant is computed. The Stefan-Boltzmann constant is not a fundamental constant -- it is constructed from Planck's constant, Boltzmann's constant and the speed of light (https://en.wikipedia.org/wiki/Stefan%E2%80%93Boltzmann_law#Stefan%E2%80%93Boltzmann_constant). So now its value is consistent with the values of the fundamental constants that are already in unyt.

The reason for the doctest change is that σ's value changed slightly, because it is now defined in terms of the fundamental constants.

[neutrinoceros]

I didn't change the name of the constant -- it was already σ in the code.

Whoops, sorry I misread the diff

Ok thanks for clarifying, I missed that there were actually two changes here: fixing stefan_boltzmann_constant and adding radiation_constant.
Searching for the latter, I find https://en.wikipedia.org/wiki/Radiation_constant, indicating that the name is somewhat ambiguous. Could we use stephan_boltzmann_constant or radiation_density_constant instead ?

[zingale]

radiation density constant is fine, although I've never seen it referred to anything other than the "radiation constant" in textbooks.

[neutrinoceros]

radiation density constant is fine

Alright, let's go with that then.

[zingale]

I've updated the name to radiation_density_constant

[neutrinoceros]

pre-commit.ci autofix

[neutrinoceros]

Thanks @zingale ! FYI I'm planning to release this into unyt 3.0.4 which I'll hopefully be able to prepare sometimes next week.