Final changes for notebook support

[olivialynn]

Some notes:

• Added sphinx building to the testing-and-coverage workflow. Not against taking this out of the unit test workflow and creating a doc-building workflow, if people prefer that.
• Made a number of changes to the Makefile. I'm not terribly up to date on best practices wrt those, so this is an area where I'm especially open to hear feedback.
• This PR didn't end up troubleshooting the jupyter-nb-clear-output pre-commit hook after all (but it solves the sphinx pre-commit check failing with notebooks, though). I had trouble getting its included command to even work from the command line, and was already way past the time I'd budgeted to this. It would definitely be great to resolve that at some point.

[delucchi-cmu]

This is really helpful!

[delucchi-cmu]

Thank you!

[cchris28]

I am used to seeing .PHONY at the end of the Makefile but I do not know if it matters. Its sole purpose is to declare targets that do not represent actual files. So I do not think that Makefile needs to be present (unless you noticed some error?)

[olivialynn]

The initial Makefile that Sphinx generates has the line .PHONY: help Makefile, so I reasoned it was best keep it if I had no specific reason to remove it.

Based on this SO, it seemed to me like it's there to make sure whatever targets are captured by %: Makefile are always run, and--this part goes a little over my head--the presence of Makefile as a dependency of % to avoid weird GNU Make behavior. (But please let me know if you have any thoughts!)

[drewoldag]

This looks great @olivialynn, nice work!