lpy VS nbpercent

Also wanted to note another option for python scripts, https://github.com/abo-abo/lpy (that uses lispy). Lispy seems to be actively maintained and lpy is less (abo-abo doesn't seem to be active anymore). There are a few drawbacks, specifically around async evaluation but it's a very interesting option, and with a little more maintenance of the package is even great.

code-cells and ox-ipynb are different tools, what should be compared is jupytext to ox-ipynb. I actually did that when I was writing the post but I did not include it. Basically, the advantage of jupytext is that it offers very solid round trip conversion whereas ox-ipynb only offers org to ipynb conversion. The advantage of ox-ipynb is that it lets you keep the outputs (which for me is also a very important point). Hopefully, outputs will be supported in jupytext at some point (https://github.com/mwouts/nbpercent/). Of course there is also pandoc, which supports round trip conversion and many more formats but like jupytext it removes the outputs.