make sure that the recursive reading of .R files doesn't look for the packages listed in a previously formed install.R file. @MilesMcBain we'd just need to make the 'install.R' file an exception to the R_files line in binder_installR, right?

  R_files <- dir(directory, 
                 full.names = TRUE,
                 pattern = "\\.(rmd|r)$" , 
                 recursive = TRUE, 
                 ignore.case = TRUE)

if there are no packages in one of the lists, remove that line from the install.R -- currently it includes the github install step even without any packages and then errors.

Before other packages as these two are required in the calls to load the rest.

👋 @MilesMcBain!

We're in the process of cleaning up ropenscilabs. https://devguide.ropensci.org/curationpolicy.html#incubator-packages

Would you like to transfer this repository to your account, or, seeing it had no recent activity (and that there's holepunch nowadays), should we archive it and transfer it to ropensci-archive?

The function would return the markdown code for the built binder (see binder page) with the details of the user and repo populated:
[![Binder](http://mybinder.org/badge.svg)](http://beta.mybinder.org/v2/gh/<GITHUB_USER>/<REPO>/<BRANCH>?urlpath=rstudio)

Nice work!

Just wondering if it possible to also output a dockerfile, so that there is the option of running the analysis on a docker container instead of on binder?

That would give a backup in case binder doesn't continue at some point

If this is being done retroactively for an older project does binder need to be told to use an older base r version in their docker? If so, how do we add this functionality?

Tried out build_binder for one of my projects, and it got a bit confused with the base packages I had explicitly called. Will need to remove them from the lib_list since they won't need to be in the CRAN or github installs lists.

If a package is not installed locally the error is a bit ugly:

> bindertools::build_binder()
Runtime file created.
Error: $ operator is invalid for atomic vectors
In addition: Warning message:
In utils::packageDescription(package_name) :
  no package 'checkmate' was found

This is unlikely to happen with something you have been actively working on, but could easily happen if you just cloned or pulled changes from a collaborator.