Closed cschwan closed 1 year ago
In this case I'd really through Python: it is quite a minority of people using Rust in Jupyter. Moreover, in my experience Jupyter works best with interpreted languages, while it provides a poor experience for usually compiled languages, since lacking the compilation step (in a sense, Cling is not C++ and Evcxr is not Rust).
However, concerning the second point, Evcxr might be a better job than what I experienced with other similar interpreters. But the first objection still holds.
If you want to use Jupyter as a development tool, it might be an idea. Otherwise, I'd go for binding as much as possible to Python, and emit Pandas dataframes, that is what people using Jupyter are used to.
If you want to use Jupyter as a development tool, it might be an idea. Otherwise, I'd go for binding as much as possible to Python, and emit Pandas dataframes, that is what people using Jupyter are used to.
That's what I had in mind, or something similar for R, which I think also has a jupyter notebook, or even both.
That's what I had in mind, or something similar for R, which I think also has a jupyter notebook, or even both.
Jupyter literally means "JUlia PYThon & R", but the biggest audience is definitely Python.
I see, I didn't know that - anyway, R is great if you do what I do!
We should evaluate whether it's worth the effort to integrate the PineAPPL into jupyter. For instance the CLI mainly produces tables, through prettyprint-rs, which has support for evcxr.