More unit definitions from xclim

[dcherian]

I needed the "degC = celsius" aliases so I copied that and more from xclim.

cc @jthielen

xref #225

[keewis]

looks good to me, although it feels a bit strange that the symbol of degrees_north should be degrees_north, I'd have assumed something like °N or ° (same for degrees_east). I can't find anything on that in the CF conventions (not that I looked very hard), so maybe that's intentional? Why this is important is because something like

u = ureg.Unit("degrees_north ** 2")
f"{u:~cf}"

would return "degrees_north2", which is weird (unless I'm missing something?)

Actually, this is not really the fault of this PR: right now it returns "degrees_N2" which is not that much better.

Edit: now that I'm thinking about it, "°N2" or "°2" also look strange, so maybe we don't have much of a choice?

Edit2: I just realized that this is actually a continuation of #284 (comment), I must have missed that.

[dcherian]

I'd have assumed something like °N or ° (same for degrees_east)

I've realized that there are at least 2 distinct use-cases:

1. units for plotting or reprs where "fancy" stuff like °N would be great (basically emacs' prettify-symbols-mode =) )
2. .pint.dequantify("u:cf").to_netcdf where you want degrees_north to meet the standard.

[dcherian]

The funny exponent syntax is from UDUNITS

I see pint has a "pretty" formatting mode. Is there a way we can define the "pretty" string?

u = units.Unit("degrees_east")
f"{u:~P}"

degrees_east

EDIT: I see even pint doesn't do the pretty ° for "degrees"

u = units.Unit("degrees")
f"{u:P}"

degree

EDIT2: Should we define a Pcf formatter for our own pretty strings?