Add high resolution sources for all ESS instruments

[nvaytet]

Add sources generated with McStas notebook for all 16 instruments.
The new default ess source is created as the average of all sources.

The high-resolution binning will then enable us to use uniform noise during the sample instead of gaussian noise, which would fix issues in sampling optimization in #136

We also add the notebook that was used to generate the sources. It is not ran as part of the docs, as it is too expensive to run in CI.

Updated default ESS source profile:

[nvaytet]

Note that I kept the original odin source as it is already high-resolution and was sampled ~2-3 m from the moderator where cold and thermal neutral are well mixed.

[YooSunYoung]

Are they all that different so that we would need different source per instrument...?
Or will it be generally okay to use the average ess source for most cases...?

[YooSunYoung]

This particular region in the wavelength distribution has a bit of difference between instruments.

[nvaytet]

They are generally close enough that the standard ESS source will be ok for most cases.
I guess your question was: "do we need all these new sources, if they are mostly all identical anyway"?

I guess that's a fair point...

[nvaytet]

I guess for now, it does not hurt to have the files in there, even if they are not used much.
I imagine in the future the sources per instrument might be optimized further, or sampled at a different distance.
Keeping all sources in this PR would mean that the infrastructure is already in place, and future changes would be minimal.
Based on this, I think I will keep them in.