[FEAT] Add support for logit-normal distributions for parameters

[liunelson]

Problem 5 of the monthly demo scenario has three models A, B, and C.

Model C is extracted from this paper:
https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1012263

Its configuration (Table 1) requires one of the parameters \epsilon^{DC} to be a logit-normal distribution (with \mu = -1.46 and \sigma = 0.71).

Logit-normal distributions are supported by MIRA (present in its probonto.py) but it's missing from PyCIEMSS:
https://github.com/ciemss/pyciemss/blob/main/pyciemss/mira_integration/distributions.py

[SamWitty]

@liunelson , do you have an example AMR we can use as a test case for this?