%%scala magic is busted

[parente]

I'm working on a sizable PR to add doc and clean up various code paths. Spotted this along the way.

     40             self._interp = get_scala_interpreter()
     41             # Ensure that spark is available in the python session as well.
---> 42             self.kernel.cell_magics['python'].env['spark'] = self._interp.spark_session
     43             self.kernel.cell_magics['python'].env['sc'] = self._interp.sc
     44 

KeyError: 'python'

I'm on the fence about fixing it vs removing support for IPython-to-Scala entirely from this package with the goal of making it more single-purpose: a minimal yet solid Scala+Spark kernel. Pixiedust and https://github.com/maxpoint/spylon can already be used to do scala in ipython.

[mariusvniekerk]

So spylon can't really do scala in Ipython.

[parente]

Clarifying: spylon gives you access to JVM views in IPython. Pixiedust does eval of Scala code.

[parente]

Also, I know @patrick-nicholson is using a simple pattern like:

import spylon_kernel
intp = initialize_scala_interpreter()
intp.interpret(code)
intp.last_result()

which will continue to work without the magic, and can be more easily embedded in Python functions and classes.