add 'sobol' and "latin hypercube' as distribution types (like 'uniform') and Variance-based sensitivities

[AdityaSavara]

Currently, we have several distribution types, such as 'uniform'. We do not have 'sobol' as an initial distribution type.

It will not be hard to make. Salib returns the parameter vectors for sampling, like in param_values in the example page for Saltelli: https://salib.readthedocs.io/en/latest/basics.html#an-example

there is also latin hypercube: https://salib.readthedocs.io/en/latest/api/SALib.sample.html?highlight=Saltelli#module-SALib.sample.saltelli

This will also allow us to get the first order and second order Variance-based sensitivities to see which parameters are most important for a given response. We can then make spider plots or similar.

https://en.wikipedia.org/wiki/Variance-based_sensitivity_analysis#:~:text=Variance-based%20sensitivity%20analysis%20%28often%20referred%20to%20as%20the,be%20attributed%20to%20inputs%20or%20sets%20of%20inputs.

[AdityaSavara]

We've decided to use scipy Sobol sampling, now that it exists. We could potentially also consider Halton: https://scikit-optimize.github.io/dev/auto_examples/sampler/initial-sampling-method.html