Fit EpiNow2 model

[zsusswein]

Take in-memory data, parameters, and sampler settings and fit an EpiNow2 model.

This model recapitulates our existing approach. It continues to be careful to avoid several of our key bugs from the last year:

• We strip the dataframe down to the minimal needed columns with the appropriate names before calling epinow()
• It doesn't allow a right-truncation PMF longer than the data
• It forces the random seed to be used both within R and by Stan, so initialization and the MCMC sampler are deterministic

I also downgrade model fitting failure to a warning here from an error. This is to enable metadata/output writing despite model failure.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Additional details and impacted files

📢 Thoughts on this report? Let us know!

[kgostic]

Minor comments mostly on documentation, but this looks good and the code is clean! Excited to see this go!

[natemcintosh]

I think this is great. My only question was about what happens if a user forgets to specify all the necessary fields in one of the input lists to fit_model()

[zsusswein]

/document

[athowes]

LGTM!