Jupyter notebooks have begun to be cited on Wikimedia projects (examples: [[ https://en.wikipedia.org/w/index.php?title=Lambert_series&oldid=744731728#cite_note-1 | Wikipedia article ]], [[ File:Parameter estimation process infinite Gaussian mixture model.webm | file page ]], [[ https://en.wikipedia.org/wiki/Wikipedia_talk:Meetup/NYC/Artsy_ArtAndFeminism | meetup page ]] and of course [[ http://paws-public.wmflabs.org/paws-public/ | PAWS ]]), but we currently have no simple way to validate whether these notebooks are actually running, so that readers can benefit from them.
In this session, I propose to build on some earlier work on assessing the reproducibility of Jupyter notebooks [[ https://github.com/sparcopen/open-research-doathon/issues/25 | shared along with scientific publications ]] ([[ https://markwoodbridge.com/2017/03/05/jupyter-reproducible-science.html | initial summary ]]) or on Wikimedia sites ([[ https://meta.wikimedia.org/wiki/WikiCite_2017/Jupyter_notebooks_on_Wikimedia_sites | summary ]]), to improve some first tools to validate Jupyter notebooks and to integrate them into Wikimedia workflows:
- [[ https://github.com/sparcopen/open-research-doathon/issues/25#issuecomment-284736196 | a continuous integration workflow for Jupyter notebooks ]]
- [[ http://paws-public.wmflabs.org/paws-public/995/WikiCite%20notebook%20validator.ipynb | a Jupyter notebook that can run Jupyter notebooks and document the error messages ]]