Add support for many more MIRA probonto distributions

[SamWitty]

See #590 for the original PR from @liunelson .

[SamWitty]

@liunelson , why are some of the types for the parameters Dict[str, Union[float, list]] and not just Dict[str, float]?

See e.g. mira_betabinomial_to_pyro

[SamWitty]

@liunelson , it looks like you used Gamma where you meant to use InverseGamma.

Also, I'd like to add some tests to ensure these mappings are completed correctly. Do you have any example AMRs that use some of these new distributions?

[liunelson]

@liunelson , why are some of the types for the parameters Dict[str, Union[float, list]] and not just Dict[str, float]?

See e.g. mira_betabinomial_to_pyro

The Pyro Binomial distribution expects a list for either the probs or logs arguments
https://docs.pyro.ai/en/stable/distributions.html#binomial

You're right about the BetaBinomial distribution; it just wants Dict[str, float]; I copy-and-pasted the type hints from the Binomial one.

[liunelson]

I don't actually have any example of AMRs handy that use all these different probability distributions.

However, MIRA has this test but it just directly invokes a LogNormal1 distribution out of the blue.
https://github.com/gyorilab/mira/blob/7d98f519f06a471284009b56e7af9b0d8f4201e9/tests/test_modeling/test_petri.py#L38

[liunelson]

Besides the type hinting of the BetaBinomial function, everything else looks fine to me.