Clever Geek Handbook

How do I check and compare multiple versions of R functions for a package?

I use unit tests through testthat

for many of my simpler functions in R, but I have more complex functions that combine simple functions and business logic. I would like to test the cumulative impact through the before and after view for a sample of input data. Ideally I would like to do this for various candidate changes.

At the moment I:

  • Using Rmarkdown Documents
  • Downloading the package as is
  • Getting my sample
  • Running my sample through the as-is package and outputting the results table
  • source

    new copies of functions
  • Running my sample again and outputting the results table
  • Reload the package and source

    execute different copies of functions as needed

This proved to be difficult due to some functions that are in the package namespace still executing versions of the function packages, which makes the results unreliable unless I completely reload all dependent functions downstream. In addition, the mechanism is complex to manage and difficult to reuse.

Is there a better strategy for testing candidate changes in R packages?

+3


source to share


2 answers


I mitigated this issue by creating a subpackage in my impact analysis folder that contains the changed functionality.

Then I use devtools::load_all

to download these new versions of the functions.



I can then compare and contrast the results by referring to the originals through a namespace, eg. myoriginalpackage:::testfunction

looking for new ones withtestfunction

+1


source


Perhaps you can replace steps 5 and 7 with "Rcmd build YourPackage" for each version. Then with

install.packages("Path/To/MyPackage_1.1.tar.gz", type="source")


check old version and then replace 1.1 with 1.2

0


source






More articles:


All Articles