Allow arithmetic of jlwrap in Python

[marius311]

If you're in PyJulia and grabbed some Julia object which has custom +/-/*/etc.. defined, it'd be nice if you can do that directly in Python. E.g.

In [1]: %load_ext julia.magic 
Initializing Julia interpreter. This may take some time...

In [2]: %julia using LinearAlgebra 

In [3]: x = %julia I 

In [4]: 2*x 
   ...:                                                                                                                             
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-4-cac53a56d728> in <module>
----> 1 2*x

TypeError: unsupported operand type(s) for *: 'int' and 'PyCall.jlwrap'

where that error is the current behavior but it would be nice if it instead returned:

<PyCall.jlwrap UniformScaling{Int64}
2*I>

I think I can probably implement this if you think you'd accept this functionality.

From a quick glance it looks like the way to proceed is to give the jlwrap object __add__, __mul__, etc... via some mechanism or another somewhere near https://github.com/JuliaPy/PyCall.jl/blob/master/src/pytype.jl#L379, either by adding to that if-clause, or something else. Am I on the right track?

[tkf]

If you want to try this right now, you can use one of my prior attempts:

• https://github.com/tkf/ipyjulia_hacks
• https://github.com/tkf/PyBase.jl

I think the direction of PyBase.jl is the way to go. That is to say, implement it as a Julia package; or rather, the current plan is to put it in PyCall (yay!): #617. It's in the PyCall 2.0 roadmap #629 so I think we will eventually have this out-of-the-box. If you want to make it happen sooner, see my comment here: #617 (comment)

Let me close this and let's continue the discussion in #617