Trying to setup an environment with an R kernel via conda for use with a Jupyter Notebook I ran into a problem that all but stopped me from progressing. Normally a few permutations of a decent google-ing would suffice but not tonight. Merging a few new technology I was very interested in with one a love dearly, I knew I was just over my head but my fascination with the problem combined with an incredible lack of patience to sit through a drawn out forum conversation or GitHub request - of which there were already multiple making no progress.
If you are half as excited as I am about using these technologies together, then I best get straight to the fix:
Click Here to download the source file. Even better - rather than relying on me to constantly update this link and I wont is to go to the CRAN Package Site and download the most current binary from there.
For me this looked like:
> install.packages('~/Downloads/pbdZMQ_0.2-4.tar', repos = NULL, type="source")
NOTE:
USER$ R CMD INSTALL path_to_file
# and
~/path_to_file USER$ R CMD INSTALL file_name
Using R from the Terminal as shown above should also work - however - between the time I already invested in failing to find a suitable solution, creating this one, writing up this quick draft, and the greater anticipation of playing around with R in a Jupyter notebook where I could simultaneously web scape with Python while generating easy to follow and reproduce journal entries for this same site It was time to move on.
Some more interesting notes that I could look into if I didn’t have more exciting repos of data awaiting my exploration, would be that pbdZMQ is designed to provide high level wrappers for the ZeroMQ library but the Cran site explicitly states:
For convenience, a minimal ‘ZeroMQ’ library (4.1.0 rc1) is shipped with ‘pbdZMQ’, which can be used if no system installation of ‘ZeroMQ’ is available.
So it could also be that linking ZeroMQ as a package dependency could fix all this. The package is only in Version: 0.2-4 at the writing of this piece. And as mentioned, there are plenty of GitHub bug reports, unanswered Stack Overflow queries - which I will be right on tomorrow morning as it’s already well into the morning and I’ve got an interview with a alluring data company in the IoT business in a few hours that I best rest up for.
Interesting academic paper on the pbdZMQ library that inspired a few iterations in my solution.