make_boot_pmtable design discussions

[barrettk]

A make_boot_pmtable function has been proposed here. It currently feels very similar to the existing make_pmtable, though there are a few questions (see below).

For reference, we are using the tables from the Expo as our guides to develop this function: https://merge.metrumrg.com/expo/expo1-nonmem-foce/posts/parameter-tables.html#create-bootstrap-parameter-table

• Do we want the same . pmtype argument for controlling which parameter type is returned?

• Can we confirm how we want bootstrap CI columns to be named? We have the current handling

  • Regular intervals would go from boot_ci_95 --> 95% CI when displayed (i.e. just displays relevant percentile).
  • There is also the special case where boot_ci_50 is renamed to IQR(rather than a percent).

• Do we need any additional control that make_pmtable does not have?

• As an aside, we wanted to check if we wanted a new footnote added to param_notes for displaying notes like the following:

  st_notes(glue("The confidence interval was determined from the 
                      2.5th and 97.5th percentiles of the non-parametric 
                      bootstrap (n={nrow(boot)}) estimates."))

The caveat is that we would have to provide both a .ci and N number of samples to create this.

[seth127]

There was some discussion in #54 of updating make_pmtable() instead. I think we want to go with a separate function, as discussed here, but let's make sure that's the decision we want.

Either way, that issue can be closed when this one is.

[barrettk]

Something to consider for the future is the idea of centering specific columns as part of make_boot_pmtable. We aren't doing this in make_pmtable currently, so both functions should likely get the same handling.