threemc

Installation

To install threemc:

remotes::install_github("mrc-ide/threemc", upgrade = FALSE)

threemc

To see how threemc works for a simple example, please see the relevant vignette.

Development steps

(largely copied from naomi)

• Make changes in a new branch,
• Run checks (requires goodpractice package). Also documents functions with devtools::check(). These checks can be run directly from the script scripts/goodpractice.R or from the command line with make goodpractice (use make goodpractice_vignette if making changes to vignettes, which are very slow to build),
• When branch is ready for merging create a PR and add a reviewer,
• Ensure that the version number has been updated according to semantic versioning and add a news item describing the change, and
• Reviewer should check code and ensure the build passes on Buildkite before merging.

Additionally:

• Wherever possible, add arguments for optional new functionality, with old functionality as default,
• If changing threemc_aggregate or threemc_ppc (the slowest "purely R" functions in this package), a before vs after profiling would be useful, using e.g. profvis,
• Follow the tidverse styleguide as much as possible, but at least have your commits pass the call to lintr from goodpractice,
• Ensure that git commits (and R comments) are in the imperative case (see here for Tim Pope's rational behind this),
• To avoid notes from devtools::check() when using non-standard evaluation (NSE),
  • in dplyr, refer to all columns with rlang::`.data`, e.g. .data$col,
  • in data.table, assign columns referenced in NSE to the value NULL at the beginning of a function (see threemc_ppc for an example).

License

MIT © Imperial College of Science, Technology and Medicine