drat Tutorial: First Steps towards Lightweight R Repositories
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.
Now that drat is on CRAN and I got a bit of feedback (or typo corrections) in three issue tickets, I thought I could show how to quickly post such an interim version in a drat repository.
Now, I obviously already have a checkout of drat. If you, dear reader, wanted to play along and create your own drat repository, one rather simple way would be to simply clone my repo as this gets you the desired gh-pages
branch with the required src/contrib/
directories. Otherwise just do it by hand.
Back to a new interim version. I just pushed commit fd06293 which bumps the version and date for the new interim release based mostly on the three tickets addresses right after the initial release 0.0.1. So by building it we get a new version 0.0.1.1:
edd@max:~/git$ R CMD build drat * checking for file ‘drat/DESCRIPTION’ ... OK * preparing ‘drat’: * checking DESCRIPTION meta-information ... OK * checking for LF line-endings in source and make files * checking for empty or unneeded directories * building ‘drat_0.0.1.1.tar.gz’ edd@max:~/git$
Because I want to use the drat
repo next, I need to now switch from master
to gh-pages
; a step I am omitting as we can assume that your drat
repo will already be on its gh-pages
branch.
Next we simply call the drat
function to add the release:
edd@max:~/git$ r -e 'drat:::insert("drat_0.0.1.1.tar.gz")' edd@max:~/git$
As expected, now have two updated PACKAGES
files (compressed and plain) and a new tarball:
edd@max:~/git/drat(gh-pages)$ git status On branch gh-pages Your branch is up-to-date with 'origin/gh-pages'. Changes not staged for commit: (use "git add <file>..." to update what will be committed) (use "git checkout -- <file>..." to discard changes in working directory) modified: src/contrib/PACKAGES modified: src/contrib/PACKAGES.gz Untracked files: (use "git add <file>..." to include in what will be committed) src/contrib/drat_0.0.1.1.tar.gz no changes added to commit (use "git add" and/or "git commit -a") edd@max:~/git/drat(gh-pages)$
All is left to is to add, commit and push—either as I usually via the spectacularly useful editor mode, or on the command-line, or by simply adding commit=TRUE
in the call to insert()
or insertPackage()
.
I prefer to use littler’s r for command-line work, so I am setting the desired repos in ~/.littler.r
which is read on startup (since the recent littler release 0.2.2) with these two lines:
## add RStudio CRAN mirror drat:::add("CRAN", "http://cran.rstudio.com") ## add Dirk's drat drat:::add("eddelbuettel")
After that, repos are set as I like them (at home at least):
edd@max:~/git$ r -e'print(options("repos"))' $repos CRAN eddelbuettel "http://cran.rstudio.com" "http://eddelbuettel.github.io/drat/" edd@max:~/git$
And with that, we can just call update.packages()
specifying the package directory to update:
edd@max:~/git$ r -e 'update.packages(ask=FALSE, lib.loc="/usr/local/lib/R/site-library")' trying URL 'http://eddelbuettel.github.io/drat/src/contrib/drat_0.0.1.1.tar.gz' Content type 'application/octet-stream' length 5829 bytes opened URL ================================================== downloaded 5829 bytes * installing *source* package ‘drat’ ... ** R ** inst ** preparing package for lazy loading ** help *** installing help indices ** building package indices ** testing if installed package can be loaded * DONE (drat) The downloaded source packages are in ‘/tmp/downloaded_packages’ edd@max:~/git$
and presto, a new version of a package we have installed (here the very drat interim release we just pushed above) is updated.
Writing this up made me realize I need to update the handy update.r
script (see e.g. the littler examples page for more) and it hard-wires just one repo which needs to be relaxed for drat. Maybe in install2.r
which already has docopt support…
R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.