[westurner]

> Crafting workflows out of notebooks is a really bad idea; an anti-pattern. If you want to go down the road of "workflows for data scientists"

https://westurner.github.io/hnlog/ Ctrl-F "DVC" ( https://dvc.org/ ) , https://westurner.github.io/hnlog/#comment-24261118 "Ten Simple Rules for Reproducible Computational Research", “Ten Simple Rules for Creating a Good Data Management Plan”, PROV

pygwalker https://github.com/Kanaries/pygwalker :

> PyGWalker: Turn your pandas dataframe into a Tableau-style User Interface for visual analysis

"Generate code from GUI interactions; State restoration & Undo" https://github.com/Kanaries/pygwalker/issues/90

[westurner]

The Scientific Method is testing, so testing (tests, assertions, fixtures) should be core to any scientific workflow system.

- [ ] (It's not possible to run `!pytest` in a Jupyter notebook without installing an extension with JupyterLite in WASM onnly where there's not yet a terminal or even yet a slow-but-usable [cheerpx] webvm bridged to jupyter kernel WASM ~process-space.)

awesome-jupyter#testing: https://github.com/markusschanta/awesome-jupyter#testing

ml-tooling/best-of-jupyter lists papermill/papermill under "Interactive Widgets/Visualization" https://github.com/ml-tooling/best-of-jupyter#interactive-wi...

"Markdown based notebooks" would store files next to the .ipynb.md, which implies a need for an MHTML/ZIP-like archive (for report notebook artifacts produced by scientific workflow systems with provenance metadata); but W3C Web Bundles avoid modifying linked resources with new specs: https://github.com/jupyter/enhancement-proposals/pull/103#is...