Expose .cores argument for summary

[stuvet]

Prework

• [x] I understand and agree to the code of conduct and contributing guidelines.
• [x] If there is already a relevant issue, whether open or closed, comment on the existing thread instead of posting a new issue.
• [x] New features take time and effort to create, and they take even more effort to maintain. So if the purpose of the feature is to resolve a struggle you are encountering personally, please consider first posting a "trouble" or "other" issue so we can discuss your use case and search for existing solutions first.

Proposal

posterior::summarise_draws has a .cores option which offers big speedups for summarising larger posteriors. Currently it is not accessible via summary_args.

It'd be very helpful to expose a summary_cores option to tar_stan_summary and tar_stan_mcmc, which could default to 1 to match posterior::summarise_draws defaults. In my fork I've included parallel::detectCores triggered by summary_cores = NULL. I was half way through submitting it as a PR, but saw the prework so raised here first. Currently tested with tar_make_future on slurm (need to set appropriate ncpus in the default resources if called from tar_stan_mcmc).

Happy to submit it as a PR if you'd like, but you may forsee problems that I can't.

Thanks

[wlandau]

Sounds great to me. Would you open a PR? Would be great to have summary_cores argument wherever there is summary_args, which includes the gq, mle, and vb functions as well as the mcmc ones.

[stuvet]

Ah yes, I didn't think about the other options, but I'm just starting to use gq & it'll be very welcome here too. I'll add them all & submit in a sec.

[wlandau]

@stuvet, I actually think you can supply the number of cores to the summary_args argument, e.g. tar_stan_mcmc_rep_summary(..., summary_args = list(.cores = 4)). The summary_args argument populates the ... argument of the summary() method of CmdStanFit objects, which then gets forwarded to posterior::summarize_draws().