The datastorr package by @richfitz provides an framework for simple data retrieval and versioning using GitHub to store data. The package is still under development and the chief developer @richfitz has indicated he'd be happy for additional input. Also @timchurches and @goldingn indicated potential interest.
Some the most useful targets for us would be to
enable datastorr to gracefully handle multiple datafiles (Issue #6 in datastorr) This is really hard to get right because the files might change between versions (so an index is needed) and we'd want to gracefully support the single-file case (i.e., returning the data directly). And then someone might transition from one to the other during the history of a project so there is lots of subtleties there!
For github itself the auth is not done correctly for private repos (Issue #14 in datastorr)
enable a new approach for distributing data whereby interface is specified in json file rather than standalone package. @richfitz - is your mock up of this available somewhere?
The datastorr package by @richfitz provides an framework for simple data retrieval and versioning using GitHub to store data. The package is still under development and the chief developer @richfitz has indicated he'd be happy for additional input. Also @timchurches and @goldingn indicated potential interest.
Some the most useful targets for us would be to