Fix for Fahrenheit deltas, degree-days and others.

[aulemahal]

Pull Request Checklist:

• This PR addresses an already opened issue (for bug fixes / features)
  • This PR fixes Degree day calculation unit handling errors #1789
• Tests for the changes have been added (for bug fixes / features)
  • (If applicable) Documentation has been added / updated (for bug fixes / features)
• CHANGES.rst has been updated (with summary of main changes)
  • Link to issue (:issue:number) and pull request (:pull:number) has been added

What kind of change does this PR introduce?

• New function ensure_absolute_temperature that converts temperature units to their absolute counterpart assuming they represented a delta. So °C becomes K and °F becomes °R. Anything else goes unmodified (including compound units like °C m (whatever that means)).
• Use this function in to_agg_units to ensure a correct output for op "std", "var" and "integral".

Does this PR introduce a breaking change?

No

Other information:

Pint has "delta_degC" and "delta_degF" for temperature differences, but CF does not. So this new function helps in the specific case where we know the data is delta-like, which happens for the given op.

[Zeitsperre]

Elegant. LGTM.

[coveralls]

coverage: 90.593% (-0.002%) from 90.595%
when pulling 684582e on fix-temp-integral
into a375d78 on main.

[coveralls]

coverage: 90.603% (+0.008%) from 90.595%
when pulling c14e52b on fix-temp-integral
into 23fd681 on main.

[coveralls]

coverage: 90.658% (+0.009%) from 90.649%
when pulling b10e849 on fix-temp-integral
into 67285be on main.

[coveralls]

coverage: 90.658% (+0.009%) from 90.649%
when pulling 9b1889a on fix-temp-integral
into 86b6bc5 on main.