Jacob B. Socolar
Jacob B. Socolar
@mjskay here's my attempt to factor out the dimension parsing stuff. Mind having a look whenever you get a chance? I've tried to return an object that will be versatile...
Hey @mjskay, The current behavior for `get_indices(c("x[-1,1]", "x[3,3]"))` is: ```r > get_indices(c("x[-1,1]", "x[3,3]")) [[1]] V2 V1 index 1 1 -1 1 5 1 3 NA 3 2 -1 NA 6...
Cool. When you get a chance, check out the new `parse_variable_indices()` over on https://github.com/stan-dev/posterior/pull/152. I think this should do everything we want; hope the code's not too clumsy.
Having a think about how to handle `NA` summaries, and I'm not sure what the best approach is. Some variables will have `NA`s that can be safely ignored. For example,...
@mjskay just mentioning that I have some time to jump back on this now. Once I remember where we're at, I'll keep it moving along.
This is a throwaway comment since it seems like you're all on the same page anyway, but just to mention that in addition to the other concerns, attempting to shoehorn...
Hey guys, I'm still playing with ways to summarize the huge draws array that I'm working with, and realized that this issue is pretty relevant as well. I've approached the...
In thinking about the content of the output of `diagnose_hmc_draws()`, the broad outlines of what to report about divergences and e-FMI seem pretty straightforward, but treedepth less so. * Stan's...
> maybe in the diagnostic files if those are produced But surely the idea here isn't to parse existing diagnostic files in order to... run diagnostics :P
> Yeah I think it's % reaching, but it's been a while since I checked so I could be wrong. Just confirmed this is true. Never realized this! One implication...