Enh/bids units

[smoia]

Closes nothing, but follows #222 .

The code in master is nice, but there are cases that might not be handled.
For instance, if 'mmHg' (Which is a non-BIDS unit we might end up having, since it's the standard unit for some gas applications e.g. CVR) is passed as an argument to bidsify_units, it would return mmhg. The same can be said for BigV or MMLIE (examples adapted from the test). It's not a huge deal, but if it's not a SI unit, at least we should leave it as it is without changing it.
There's still cases that are not supported (e.g. non-SI term plethora would return 'plethorA'), but that's for another day (and I hope no one would measure "0.56 plethora of physiological activity" in the meantime).

Proposed Changes

• Little code refactoring to support more cases
• Add more testing cases, also reorganise weird aliases tests as a dictionary similarly to the current real unit test.
• Quick dictionary reordering to lower the possibility of clashes
• Uppercase all the constants

[codecov]

Codecov Report

Merging #224 into master will not change coverage.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master     #224   +/-   ##
=======================================
  Coverage   93.66%   93.66%           
=======================================
  Files           8        8           
  Lines         679      679           
=======================================
  Hits          636      636           
  Misses         43       43           

Impacted Files	Coverage Δ
phys2bids/bids_units.py	100.00% <100.00%> (ø)

[rmarkello]

These small updates all look + seem reasonable to me! Thanks @smoia 👍