dscim.main_recipe.weitzman_min fails for larger values of eta

[JMGilbert]

https://github.com/ClimateImpactLab/dscim/blob/main/src/dscim/menu/main_recipe.py#L902-L912

If very large eta values are used, this function can create very small numbers which can exceed the smallest floats possible. As an example, with eta = 6, and no_cc_consumption and cc_consumption each being of type np.float32, parameter becomes a number usually around 10^13. The w_utility and bottom_utility variables then become numbers around -1*10^(-68) which gets stored as 0 since the float32 closest to 0 is around 10^(-38).

A fix for this may be possible by using an alternative form of the calculation or by taking advantage of the nice properties of isoelastic utility.