phsmethods
contains functions for commonly undertaken analytical tasks in Public Health Scotland (PHS):
create_age_groups()
categorises ages into groupschi_check()
assesses the validity of a CHI numberchi_pad()
adds a leading zero to nine-digit CHI numberssex_from_chi()
extracts the sex of a person from a CHI numberfile_size()
returns the names and sizes of files in a directoryextract_fin_year()
assigns a date to a financial year in the format YYYY/YY
match_area()
converts geography codes into area namesformat_postcode()
formats improperly recorded postcodesqtr()
, qtr_end()
, qtr_next()
and qtr_prev()
assign a date to a quarterage_calculate()
calculates age between two datesdob_from_chi()
extracts Date of Birth (DoB) from the CHI numberage_from_chi()
extracts age from the CHI numberphsmethods
can be used on both the PHS server and desktop versions of RStudio.
If you are using the PHS Posit Workbench the default repository is the PHS Posit Package Manager, the benefit of this is that phsmethods
is listed there so you can install it with:
To install phsmethods
directly from GitHub, the package remotes
is required, and can be installed with install.packages("remotes")
.
You can then install phsmethods
from GitHub with:
However, network security settings may prevent remotes::install_github()
from working on RStudio desktop. If this is the case, phsmethods
can be installed by downloading the zip of the repository and running the following code (replacing the section marked <>
, including the arrows themselves):
Load phsmethods
using library()
:
To see the documentation for any phsmethods
’ functions, type ?function_name
into the RStudio console after loading the package:
You can access the full list of functions and their help pages on Reference page of pkgdown website. You will be able to see some examples of each function.
There is also a very useful PHS Methods online training course which gives you an interactive way to learn about this package.
At present, the maintainer of this package is Tina Fu.
This package is intended to be in continuous development and contributions may be made by anyone within PHS. If you would like to contribute, please first create an issue on GitHub and assign both of the package maintainers to it. This is to ensure that no duplication of effort occurs in the case of multiple people having the same idea. The package maintainers will discuss the issue and get back to you as soon as possible.
While the most obvious and eye-catching (as well as intimidating) way of contributing is by writing a function, this isn’t the only way to make a useful contribution. Fixing typos in documentation, for example, isn’t the most glamorous way to contribute, but is of great help to the package maintainers. Please see this blog post by Jim Hester for more information on getting started with contributing to open-source software.
When contributing, please create a branch in this repository and carry out all work on it. Please ensure you have linked RStudio to your GitHub account using usethis::edit_git_config()
prior to making your contribution. When you are ready for a review, please create a pull request and assign both of the package maintainers as reviewers. One or both of them will conduct a review, provide feedback and, if necessary, request changes before merging your branch.
Please be mindful of information governance when contributing to this package. No data files (aside from publicly available and downloadable datasets or unless explicitly approved), server connection details, passwords or person identifiable or otherwise confidential information should be included anywhere within this package or any other repository (whether public or private) used within PHS. This includes within code and code commentary. For more information on security when using git and GitHub, and on using git and GitHub for version control more generally, please see the Transforming Publishing Programme’s Git guide and GitHub guidance.
Please feel free to add yourself to the ‘Authors’ section of the Description
file when contributing. As a rule of thumb, please assign your role as author ("aut"
) when writing an exported function, and as contributor ("ctb"
) for anything else.
phsmethods
will, as much as possible, adhere to the tidyverse style guide and the rOpenSci package development guide. The most pertinent points to take from these are:
<-
, ==
and +
roxygen2
testthat
devtools::check()
It’s not necessary to have experience with GitHub or of building an R package to contribute to phsmethods
. If you wish to contribute code then, as long as you can write an R function, the package maintainers can assist with error handling, writing documentation, testing and other aspects of package development. It is advised, however, to consult Hadley Wickham’s R Packages book prior to making a contribution. It may also be useful to consult the documentation and tests of existing functions within this package as a point of reference.
Please note that this README may fail to ‘Knit’ at times as a result of network security settings. This will likely be due to the badges for the package’s release version, continuous integration status and test coverage at the top of the document. You should only make edits to the .Rmd
version, you can then knit yourself, or a GitHub action should run and knit it for you when you open a pull request.