NB: the notebooks used within create their own modules with nbdev.fast.ai
This includes:
- 00_core.ipynb - some core utility and setup.
- Northern Power Grids data
- 10_npg_data.ipynb for etl.
- 11_npg.ipynb for some plots of the resulting dataset.
After cloning:
pip install -e jupyter_playground
Some features:
print(core.IncrementalPipeline.__doc__)
A class whose instances can dynamically store functions.
When used as a callable passes each stages retunred object as args into each successive function.
Results are wrapped if needed.
print(core.DownloadContent.__doc__)
Masks the repr with it's hash to avoid serialising the content.
This stops libraries like joblib serialising large strings in input reprs