Add binder examples

[KirstieJane]

In #133 @tsalo added some really lovely jupyter notebooks to explain some aspects of the tedana🎉 processing.

It would be really cool if these were accessible through mybinder.

I don't have time to work on this so any help would be greatly appreciated! ✨ 😻 👾

I think this requires:

• an update to the info.py file to include (for example) matplotlib, seaborn and nilearn.
  • Side note, thank you to Minn in the binder team for helping understand where binder goes looking for its requirements 😸
  • @emdupre would prefer this to be in the "DOCS REQUIRES" section so these packages aren't added when a regular user installs tedana.
  • An alternate idea is to make a binder/requirements.txt file that contains everything needed for the notebook. Note that this option won't run setup.py so everything needs to be added into that file, not just additional packages on top of those defined for tedana.
• moving the jupyter notebooks to somewhere more accessible, eg an examples folder
• uploading some data to OpenNeuro so it's easy to access
• creating the binder link
• updating the documentation to make it clear that this information exists!!

[souravsingh]

@tsalo @KirstieJane I would like to take a stab at the issue. Where do I start?

[tsalo]

@souravsingh I don't know anything about binder, unfortunately, but the notebooks we want to binderize are in this folder. The five-echo data used for the notebooks is at https://osf.io/ea5v3, although the paths in the notebooks will have to be changed to fit whatever filesystem binder allows us to access.

[emdupre]

Hi @souravsingh ! If you're not yet familiar with Binder, they have some really great resources to get started ! These tutorials might be the best first place to look.

As @tsalo said, the notebooks and data are already hosted online, so the issue would be pulling them together in the right environment. Which Binder is perfect for ! Let us know if you have any specific questions about next steps :)

Edited to add: @KirstieJane has done a great job of laying out the overall picture, so let us know which of those steps you have questions on !

[jbteves]

I see that the original PRs were closed for this. @KirstieJane do you mind if I edit your comment so that your task list is at the bottom and we can track progress?

[jbteves]

Note: we already have data hosted on OSF for our CI. Relevant links:

• five-echo task
• three-echo rest

[KirstieJane]

Totally fine to edit!! Go for it 🚀🌟

[jbteves]

Thanks! Done, now we can track progress directly on the issue title/tracker

[jbteves]

@KirstieJane @emdupre @tsalo I have an example of it working for optimal combination here. It's in binder/optimal_combination_workflow_plots.ipynb. I haven't adjusted the workflow one yet but I'll replace the hardcoded paths similarly. The big thing is whether we want to use kundu, kundu-stabilize or revise it to be mle.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions to tedana:tada: !

[emdupre]

I'd like to suggest we use Jupytext for this, so that we're including .py rather than .ipynb files in our repository. It will keep the version history much more manageable !

And if we add Jupytext to the dev-requirements and put the correct YAML front matter in the examples, Binder will open the python scripts as Jupyter Notebooks ! ✨

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions to tedana:tada: !