Mitzi Morris

> I think we just need to have more examples with more postprocessing examples. this is key. I spent winter break translating a Stan case study (partial pooling binary trials),...

what proportion of processing time is spent in writing a StanFit object to RDS and what proportion of time is spent obtaining a sample from the CmdStan CSV files?

I agree that we need good serialization - but we need to think cross-interface and therefore find serialization formats that can be accessed by Python and C++ as well as...

how are `sampler_diagnostics()` different from `draws()`? is this implemented differently in CmdStanR than in CmdStanPy? aren't the first 7 columns of the `draws` array the source for `sampler_diagnostics`? or is...

in CmdStanPy, everything is stored in a single numpy.ndarray structure - stored column major, so all the sampler diags are contiguous in memory - views on the array will give...

my concern is making sure interfaces aren't too different from one another - we must allow idioms appropriate to the language (i.e., "pythonic" vs "rthonic" - use of quotes quite...

for CmdStanPy, there's Arviz - https://arviz-devs.github.io/arviz/getting_started/InferenceDataCookbook.html#cookbook the underlying storage mechanism doesn't have to be the same - the APIs should have the same features, and as far as serialization goes...

I doubt that CmdStanPy users are interested in the sampler params, so changing CmdStanPy so that draws() returns a view which omits sampler diagnostics would work.

what is the purpose of the serialized object? for reproducibility, you want model, data, inits, seed, chain-ids, and run config info for saved sampler state, you want stepsize and mass...

on my machine, looks like tqdm works in jupyter notebook but not in jupyter lab. something is broken, somewhere.