Dear list,

Our current R interface uses pexpect. It is functional, at least from the 
command line.

However, the "%r" magic, working from the command line and the Sage 
notebook, cannot be (usefully) used in the Jupyter notebook. This has led 
me to explore what could be done with rpy2 (already part of sage) : see 
Trac#19638 <http://trac.sagemath.org/ticket/19638> for an initial step in 
this direction.

It turns out that this interface, quite useful for the notebook, has some 
serious drawbacks :

   - It starts a new R instance, different from the one started by our 
   current interface
   - It can exchange data with Sage, but in a format different from the one 
   used by our current interface

This interface needs enhancements.

I see at least two paths to enhance it :


   - Tweak Rpy2 to use the same R instance as we do, to use the same data 
   conversions as we do, etc...
   - Re-implement a new interface, re-processing the exchanged data as 
   needed.

I am seeking this list's advice about how to do it.

Note : the problem with %r in the Jupyter notebook is shared by the other 
interpreter-specific magics, such as %maxima. A more general solution 
should be studied...

Note that, whatever path you advise, this is not a small task, and won't 
happen overnight.

Note also that, whereas, the Sage notebook has its own documentation, the 
Jupyter notebook and its extensions currently do not. Shouldn't that be 
addressed ?

Ladies and gentlemen, to your keyboards...

--
Emmanuel Charpentier

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Reply via email to