Softwaredevelopment (for Machine Learning) with Jupyter Notebooks

Source: Deep Learning on Medium

To run the export, just call the external script. You can do this directly from the notebook: !python notebook2script.py 00_exports.ipynb

This script just loads the notebook in json format. Looks through the cells and exports the marked code cells.

When running the script, it creates a new package exp and puts the .py file into.

run_notebook.py
export cells from notebook

The fire library creates a terminal interface from any function within the script.

Creating your own Unit Test Framework for Jupyter notebooks

To create a simple unit test frame just create a cell like this: