"Module not found" error when autograding

[jnak12]

Operating system

RedHat

nbgrader --version

0.9.1

jupyterhub --version (if used with JupyterHub)

4.1

jupyter notebook --version

7.1.2

Expected behavior

We have an assignment that was put in a folder and it has the following files:
CGastTest.ipynb
CGwat.ipynb
CGwatTest.ipynb
P0.ipynb
P0ParsingTest.ipynb
P0TypeCheckingTest.ipynb
SC.ipynb
SCTest.ipynb
ST.ipynb
STTest.ipynb
'01 Extending P0 with Bitwise Set Operations (WASM) [24 points].ipynb'

Inside some of the files (e.g. P0ParsingTest.ipynb) are the lines:

import nbimporter; nbimporter.options["only_defs"] = False
from P0 import compileString

Expected behaviour: file is auto graded.

Actual behavior

If the file is opened, each cell runs without issue.
But once submitted and autograded, the following error shows up:

ModuleNotFoundError: No module named 'P0'

From what I have been told, nbimporter converts a notebook to a module.

[perllaghu]

I know nothing about nbimporter, so can't help with that, however I'm also surprised to get to autograde '01 Extending P0 with Bitwise Set Operations (WASM) [24 points].ipynb' (I would have expected the braces to give you problems)

[jnak12]

That seems to be okay. I don't know much about nbimporter either, but if it has to create a new file from the notebook conversion, where would it put that file? In the autograded directory? Can files get created in the autograded directory?

[brichet]

Thanks for opening this @jnak12.

Except if there is another repo for nbimporter, I would recommend not to use this package:
https://github.com/grst/nbimporter?tab=readme-ov-file#update-2019-06-i-do-not-recommend-any-more-to-use-nbimporter