Generalize mcmcRocPrc for various kinds of input objects

[andybega]

• Closes #5
• mcmcRocPrc() becomes a generic function; methods address different input types, like mcmcRocPrc.rjags() for "rjags" objects; it returns objects of class "mcmcRocPrc", which has the previously added print, etc. methods.
• all methods use the same class constructor to make new "mcmcRocPrc" class objects, new_mcmcRocPrc(); this makes sure that output is consistent accross implementation for different input types, and also avoids having to repeat code over and over again in each method
• the default mcmcRocPrc() method takes a matrix of predicted probabilities (N cases rows, N posterior samples columns) and a vector of observed outcomes (length = N cases) as inputs; this should allow an escape hatch that can handle any type of input even if a method for that particular output type is not available
• methods right now cover input from packages (with class): rstan (stanfit) rstanarm (stanreg), brms (brmsfit), rjags (jags), R2jags (rjags), runjags (runjags), and MCMCpack (mcmc)
• there is no method to accommodate input from R2WinBUGS. The package has been last updated on CRAN in 2015, and it seems clear that STAN is the successor. If someone can add a logit BUGS example I can add a method, though.
• adds tests for most of the new methods, although not 100% for all lines; also for this reason adds several models in tests/testdata. I added these under tests, otherwise if they are in data the package installed size is > 5 MB, which generates a R check NOTE, and also noticeably slows library(BayesPostEst).

No R check ERRORs, WARNINGs, or NOTEs. (Travis-CI on my fork)

[jayrobwilliams]

This looks fantastic! When I have some time I'll probably follow a similar approach with mcmcReg() and mcmcTab() using methods and class constructor functions instead of the giant block of conditionals at the top of each right now.

[andybega]

Thanks! I wasn't sure if using generic + methods ends up making it more complicated than it needs to be.

[jkarreth]

Thank you indeed @andybega , this is great! I'm working on getting a BUGS object, hopefully within 1-2 days.

[jkarreth]

I added a BUGS object representing output from a logit model: data/bugs_logit.rda

This was created by adapting data-raw/jags_logit.R to R2OpenBUGS.

(Note that the actual script to create bugs_logit.rda is not on Github, I asked someone else to run it on their machine. Hope this is OK for future CRAN uploads etc.)

@andybega @jayrobwilliams does this give you enough for the adjusting for BUGS objects as input?