For SIS Course

Jupyter Notebook is a great tool for teaching and learning. Nevertheless, it is not a good tool for version control. In this course, we will use JupyText to sync the notebook with a Python script. Each notebook is paried with a Python script via JupyText. In each cell, we will use pytest-style format to make it compatiable for both notebook and Python script.

1. To Set Up the Sync

jupytext --set-formats ipynb,py:percent notebook.ipynb 

We will see a new file notebook.py in the same directory.

2. Use IDE (nvim or VSCode) to Edit the Python Script as pytest format

# %%

# %%ipytst -qq
def test_task1():
    import pytest
    from transformers import pipeline

    assert 1 + 1 == 2

Note that the # %% is the cell marker for JupyText. The # %%ipytst -qq is the cell marker for ipytest.

To run the test, we can use pytest in the terminal.

pytest notebook.py::test_task1 -s

Note that the -s is to show the print out.

To Sync (py to ipynb)

jupytext --sync notebook.py

To Run the Cells in the Notebook

We have set up ipytest in the notebook.

def in_notebook():
    try:
        from IPython import get_ipython
        if 'IPKernelApp' not in get_ipython().config:  # pragma: no cover
            return False
    except ImportError:
        return False
    except AttributeError:
        return False
    return True

import ipytest
# check if the cell is in jupyter notebook or not
if in_notebook():
    ipytest.autoconfig()

We could not run the cells in the notebook as we do in the regular Jupyter Notebook.