Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Magic cell languages vs magic cell commands

See original GitHub issue

Following #513 and #622, I’d like to think a bit about how magic cells are/should be handled.

Here are my expectations

If a cell has a magic cell that is a language, say e.g. R, that differs from the main notebook language, say e.g. python, then

it is entirely commented out in the .py file
it appears as a ```R code cell in the .md file

Cells with a magic cell that is not a language should appear verbatim in the .md file. The cell magic should always be commented out in the .py file. For some cell magics like %%time or %%capture it may not be necessary to comment out the cell content in the .py file. For some other cell magics like %%script, commenting out the full cell is required.

I think (but would like to check) that we are actually doing that.

Currently the distinction between 1. and 2. is made based on jupytext.languages._JUPYTER_LANGUAGES (script is one of them), plus the configurable custom_cell_magics. Do we cover all the use cases?

Issue Analytics

State:
Created 3 years ago
Reactions:1
Comments:8 (4 by maintainers)

Top GitHub Comments

1reaction

mwoutscommented, Nov 4, 2020

Hi @lc5415 , thanks for reporting the double magic case! I agree with you, the cell should be fully commented out in this case. I’ll try and see if I can fix this in the next release.

0reactions

mwoutscommented, Nov 24, 2020

Hi @fm75 , thanks for asking.

As @chrisjsewell and @choldgraf mentioned in the other discussion, the outputs are never stored in the Jupytext files (and this is not related to magic or not magic commands).

You were probably expecting something closer to the output of jupyter nbconvert ? Indeed this is an area where jupytext is very different from nbconvert - Jupytext focuses on the part of the notebook that the user wrote, not the part that was generated by the machine and which is often thousand times larger. Of course I understand that not having outputs can be troublesome, so depending you what you are trying to achieve, my recommendations are

use jupyter nbconvert if you are looking for a one way conversion .ipynb to .md
use Jupytext’s paired notebooks (synchronized .ipynb + .md files) if you want to edit either the .ipynb (with outputs) or the .md file