Report attached and loaded packages in reproducibility table

[slager]

Inspired by FredHutch/VISCtemplates#314

Suggested review: Install this branch with remotes::install_github('FredHutch/VISCfunctions', ref = 'repro_tbl_namespace'), then knit a test report with VISCtemplates. The VISCtemplates skeleton will need an adjustment to allow a packages table to carry over onto a second page if needed.

[kelliemac]

tested locally and it works great for me! here is a screenshot for posterity:

note: the way to resolve the associated issue with the VISCtemplates skeleton so that the packages table can be more than one page is just to add option longtable = TRUE to the kable() call.

thanks for tackling this @slager!

[kelliemac]

associated changes for VISCtemplates: FredHutch/VISCtemplates#329

[kelliemac]

@mayerbry do you know if HVTN folks use this function for their reports? if so, we should probably reach out to them.

[slager]

Tagging @bneradil to weigh in for HVTN

[slager]

Per Moni, HVTN SRA reports log get_session_info() to a text file, so the additional lines of output are not an issue. Per Radhika and Emily, this is also not an issue for HVTN STP reports.

[lemireg]

While playing around with doing this for one of my PDF outputs where I wanted to include dependencies in a reproducibility table, I noticed a relatively easy/predictable way we may be able to detect and throw the user a warning (which they can always suppress using warning = FALSE on the r chunk) when they got a package from network drive (usually where our datapackages are located) but then pull from a version stored locally (i.e. risking the have a datapackage stored locally). You could replace "cavd" and "Local" below with other path folder names that catch more cases (if for example we ever have datapackages stored in another folder than N:/cavd)

We could add a warning to this PR if we want to highlight these kinds of cases (even without printing out the local location of every package)

sessinfo <- session_info(pkgs = "loaded", info = "all", dependencies = TRUE)$packages
data.frame(pck = sessinfo$package,
           source = sessinfo$source,
           path = sessinfo$path) %>%
  dplyr::filter(grepl(pattern = "cavd", x = source) & grepl(pattern = "Local", x = path))

[slager]

Note from VISCfunctions meeting:

Path-checking logic like suggested by @lemireg to go in a separate function, or at least in a separate PR

[mayerbry]

thanks!