refactor: upstreaming improved device classes from erb-thesis

[ErikBjare]

Replaces #92 (switched to branch in the NeuroTechX repo instead of my fork)

Fixes #43

Fixes #77

TODO

• Resolve conflicts caused by changes to master since PR opened (that need to be integrated as part of a merge/rebase): ff5f018...master
  • This was very helpful: https://stackoverflow.com/a/44956845/965332
  • It's mostly changes in the eeg.py file that need attention (and whatever other files conflict)
    • Like the new _muse_get_recent method, and similar (probably best if whoever wrote them resolve the conflict, I feel very unsure...)
  • Fix CI issues following merge with develop
• Test that things work
  • For MuseDevice
  • For BrainflowDevice
• Ensure that tests and typechecking runs correctly in CI
• Check differences to eeg.py between fork and now
  • When I vendored the code into my thesis repo: ErikBjare/thesis@fdaa949
  • First change in eegnb/devices since I vendored the changes: b45201c
    • Full diff between that commit and now: b45201c...master
    • Changes seem to mostly relate to the FreeEEG32
• Figure out how util.py is to be used in the new structure

[ErikBjare]

@JadinTredup I got this PR in an almost mergeable state. The only thing left is to patch the FreeEEG32 stuff that was added after I forked eegnb/devices (see top comment for links to diffs).

If we get the FreeEEG32 stuff fixes soon, I think we might actually be able to merge this soon.

I also added more testing (with pytest), which would be really nice to get merged. (oops, apparently I had similar changes in #88)

[ErikBjare]

Pushed an incomplete merge commit. Needs fixing (and careful review).

[ErikBjare]

@JohnGriffiths @JadinTredup Looks like delaying this made the conflicts a lot worse... I tried to do my best to clean it up, but some issues left and it's 6PM here, so I'm off duty. Will take another look & continue tomorrow.

I would appreciate if you could both have a look and check if the general direction is good, and if the API looks reasonable (or if anything is unclear).

[JohnGriffiths]

Ok so better late than never... @ErikBjare I have looked through the code here and like the ideas.

Will do some tests soon and keen to merge to `develop' soon if no obvious issues.