jupyter-book and textual formats like py:percent

[parmentelat]

Hi

my understanding is that jupyter-book can be used on natively on notebooks stored as ipynb, or md:myst

turns out I have a lot of py:percent notebooks; of course, as the jupyter-book doc mentions, it's easy to use jupytext to translate py:percent into md:myst, but my feeling is that a smooth tooling should ideally be able to take care of that (in particular in order to keep an incremental workflow where only changed notebooks trigger a conversion + execution)

where is the team standing on this ?
is it that it's actually supported and I failed to spot it ?
or is it that nobody has yet the time to tackle that ?
or is it intentionally left out b/c deemed as outside of the project scope ?

thanks for your kind feedback

[chrisjsewell]

Heya

in particular in order to keep an incremental workflow where only changed notebooks trigger a conversion + execution

Yeh this should already work, have you tried it?

[chrisjsewell]

As in by using https://jupyterbook.org/file-types/jupytext.html

[parmentelat]

all righty - sorry for not spotting that earlier - thanks !!

[chrisjsewell]

no worries 😄

[parmentelat]

so just for the record, it appears that

• I need to have at least one file entry at the toplevel of the _toc (I would even assume, exactly one)
• and this one file, apparently again, must be in either ipynb, md or rst format; that's what was suggested by the error message that I was getting, when in my case that landing page was in py:percent format; the error vanished when I translated it manually into `md:myst'

no big deal of course, was just a little surprising, given that all my notebooks were in py:percent

[chrisjsewell]

thanks for the feedback yes I seem to recall maybe someone else having this issue that @choldgraf was talking to

[choldgraf]

Yeah I think we need to improve the documentation around text based formats, and i think there were some assumptions in the code that we need to loosen. IIRC there's a section that checks the extension in the toc code and if it isn't a subset of allowed ones, displays a misleading error. I just haven't had the time to dig into it and figure it out