trouble with the md:myst output

[parmentelat]

Hi folks

I'm currently running into an issue, which has to do with the md:myst output format

I had filed an issue in jupytext at first, here
mwouts/jupytext#759

digging a little further - i.e. trying to reproduce the issue within the jupytext test context - did not pan out

I could read here
https://myst-nb.readthedocs.io/en/latest/use/markdown.html?highlight=code-cell%20ipython3#syntax-for-code-cells
that the ipython3 addition to code-cell was optional, and I'm curious to know if there are any environment/settings that would cause the myst output to come with that, as all my attempts so far with the jupytext cli would not include that ipython3, while otoh all my attempts at producing a myst from jupyter classic do include that extra ipython3

I'd be really interested if someone from the myst team could take a quick look at that issue
and let me know if I'm not simply missing something obvious here
or redirect me to the right people/channel if relevant

thanks a bunch !

[choldgraf]

Hey 👋 could you clarify your problem a bit? It is a bit hard to understand what is the specific thing you’re expecting to happen, and what is not happening.

[parmentelat]

sure :)

all my notebooks tagged as being 'md:myst' behave the same:
I expect to obtain the same textual representation whether I save under jupyter or if I use jupytext --to md:myst

my problem is I don't get 2 different outputs:

• from jupyter, all my code cells are fenced with

  ```{code-cell} ipython3
  # the python contents
  ```
  

• while from jupytext I more simply get

  ```{code-cell}
  # the python contents
  ```
  

I was not able to reproduce the issue from the jupytext test env, but from there I always get the second form

so I am trying to figure what it is in my environment that causes my installed jupyter to end up producing the first form

hope this clarifies the matter - thanks !

[choldgraf]

Ah - indeed that is strange, because the jupyter interface is just using jupytext under the hood to do the conversion, I think. I don't believe there is any "native" support for myst nb in jupyter

[chrisjsewell]

Yes it's optional and does not affect the execution in any way, it is used for syntax highlighting and is populated from the pymgments/lever key in the the notebook metadata

[parmentelat]

Yes it's optional and does not affect the execution in any way,

right, but the reason it is an impediment is because in some use cases I end up with a lot of meaningless diffs - which was the reason for going text in the first place ;-)

it is used for syntax highlighting and is populated from the pymgments/lever key in the the notebook metadata

ah, this is interesting !
can you please refer to the code that does that ?
should that metadata not impact both workflows ?

thanks !

[choldgraf]

Could you describe more how you're creating the myst-notebook file from Jupyter? Are you just using jupytext's auto-converter extension?

[parmentelat]

in this particular instance, I had a bunch of ipynb files that I mass-converted to myst using
juytext --to md:myst
those files contained the untagged

```{code-cell}

from that point, as soon as I load any of these files under jupyter classic, and save it from there, the ipython3 tag gets added, even if no other change is made in the notebook

I am suspecting some odd configuration in my jupyter custom settings maybe, but before I scrub that thoroughly (which honestly feels scary) I'd really like to figure what on earth can be happening here :)

thanks !