studyGenerics

Code should be readable, and it is the main means of communication for software developers. The objective of studyGenerics() is to centralise common functions (including functions considered trivial!) used in epidemiological analyses performed at the Erasmus MC Department of Medical Informatics.

More than trying to ‘standardise’ how to write a study package, we use this project to save (and, more importantly, test) frequently used functions that make our life easier.

The intention is to break our own silos and learn how to better communicate our ideas and best practices, so we can work more effectively as a team on common ground. Please feel free to contribute.

Installation

You can install the development version of studyGenerics from GitHub with:

# install.packages("remotes")
remotes::install_github("mi-erasmusmc/studyGenerics")

Sharing the same structure

For instance, we have a function that creates the specific structure we need for a study.

studyGenerics::insertStructure(
    path = ".",
    n_obj = 3
  ) 

The file structure of the package will contain most of the elements that we use more frequently.

|-studyPackage
  |-R/
    |-createCohorts.R
    |-runStudy.R
    |-objective1.R
    |-objective2.R
    |-objective3.R
    ...
  |-inst/
    |-concept_set/
    |-cohorts/
  |-LICENSE
  |-README.rmd
  |-extras/
    |-CodeToRun.R
    |-pullFromATLAS.R
    |-shiny/
  |-tests/
  |-renv/

Unifying approaches

Even in the same development team, people may prefer different tools for routine tasks (such as logging or even saving zip files). Here we try to find common ground, test functions, and avoid unnecessary bugs from small and less complicated methods, from asserting data partner names to making sure we are zipping results to the correct folder.

studyGenerics::arrangeCdmNames(
  labels
  )

studyGenerics::createResultsDir(
   outputDir,
   dbname
)

studyGenerics::zipStudyFiles(
  resultsDirName,
  outputDir,
  dbname)

studyGenerics::unZipStudyFiles(
  path,
  pattern,
  negate,
  recursive,
  outputDir
  )

Smoother processes

Furthermore, in studyGenerics() we also develop functions that make our life easier, such as saving an R virtual environment in a cellar on macOS and later using that folder to renv::restore() on a Windows machine without internet access.

studyGenerics::getPkgZips(
  lockfile_path, 
  supplement, 
  override_lock,
  r_rels_vect,
  backupRrel,
  outDir
  )