Thanks for taking a look at this, everyone! I look forward to your feedback
on this prototype.
It would be great to get this (or something like it) into the Jupyter
community as a recommendation so that we can build magics that work in any
Kernel. I have a SQL magic implementation that uses Spark
The com.github.jupyter group does come from jitpack. If there's a better
way to publish Jars from the Jupyter organization, I'm all for that.
Otherwise, I think the simplicity of jitpack is worth the trade-off of
having a weird groupId.
On the wrapper class, the choice of a Map was to keep