Dear R users,
I would like to ask you for comments on whether you find interesting a
package that would help you generate Makefiles for R analytical projects. I
am developing it for some time mainly for myself and now try to decide
whether it is worth an effort to continue and enhance it for wider
audience. Install the most recent version from github, if you wish:
devtools::install_github("beerda/rmake")
The use of the package is very simple. You write your rules as a pipe
similar to Magrittr's operator:
library(rmake)
job <- 'input.csv' %>>% rRule('preprocess.R')
%>>% 'data.rds' %>>%
markdownRule('report.Rmd') %>>% 'report.pdf'
makefile(job, 'Makefile')
Thats it. This piece of code generates a complete Makefile that runs
preprocess.R, which reads input.csv and writes data.rds, and then compiles
a markdown report.Rmd, which depends on data.rds, and creates a final
report.pdf. (Of course more complicated pipelines are possible as well as
define dependencies programmatically, also some other features are
implemented already, you can imagine...) I ask the community of whether you
find it useful and whether it is good idea to continue in development of
such package. Recently, some other build tools appeared for R, which seem
very interesting, but which I am not so much experienced with (e.g.
"drake"). There are also some other projects with similar purpose, but
they
seem more or less abandoned. Any comments are welcome.
Thanks, in advance.
Best regards,
Michal Burda
[[alternative HTML version deleted]]