Re: [Texmacs-dev] TeXmacs and jupyter

[Miguel de Benito Delgado]

There are at least four standard ways of ensuring dependencies are met:

1. python package manager (pip) on top of the OS installation. Not nice since installs are OS wide. 

(pip is itself a python package which we can assume is available as long as there is a python distribution installed ( after 3.something it's installed by default, otherwise it's trivial to install) )

2. virtualenv. Per user, local installs in one directory, including specific python binaries, using pip. AFAIK the "venv" module is also included in python installs by default.

3. conda. Same as virtualenv but allowing for other binary dependencies (libraries). miniconda is lightweight and pretty much a standard

4. containerised solutions (docker + first item in this list). Does not apply to our case.

IMO virtualenv is the way to go in TeXmac's case since it is itself a python package and we won't need any 3rd party libraries (non-python). But maybe someone else has another suggestion? I always do the fourth, so I rarely use virtualenv.

I'm afraid I really cannot help out now, I'm sorry :( Maybe in a couple of months..

--
Miguel de  Benito.

On Wed, 12 Jun 2019 at 11:29, Massimiliano Gubinelli <address@hidden> wrote:

Dear all,

 I was looking at the jupyter protocol, in view of integrating it with TeXmacs. Instead of reimplementing everything (which seems quite complex) we could try to modify the jupyter console here

https://github.com/jupyter/jupyter_console

to make it output the TeXmacs plugin protocol instead that using standard output, in this way we can call it from TeXmacs and have jupyter kernels at our disposal. This will make available many other systems inside TeXmacs.

It should not be very difficult but I need some help since I’m not very familiar with python. In particular I do not undestand how we should do: surely we want to branch the jupyter console to have our own version adapted to TeXmacs protocol, but in order for it to run it would need to have some other packages available (like the one implementing the jupyter protocol, and some other, like for example pygments, to colorise code). How one can be sure that these are available on the client machine? Is there a standard way  to require packages? Can one assume they are always installed? I have a very poor understanding of Python ecosystems.

Are  you interested in join this effort?

Miguel? Darcy?

Best
Max



_______________________________________________
Texmacs-dev mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/texmacs-dev