unexpected behavior: convert delta_degC to degC and vice versa

[Cpp87]

Converting delta_degC to degC implicitly assumes the base value 0 Kelvin. Example:

>>> (3*unyt.delta_degC).to(unyt.degC)
unyt_quantity(-270.15, 'degC')

The offset of 273.15 is suprising to me. I would rather expect an error.

How did I encounter this:
I did not know about the delta_degC. My code calculates in Celcius and I perform x.to(unyt.degC) in case the user inputs Kelvin.
This would have produced terribly wrong results. Hence, I suggest to raise an error and tell the user "provide base temperture for converting delta_degC to degC". Implicit assumption of 0K as base temperature is dangerous.

The same issue happens vice versa:

>>> (3*unyt.degC).to(unyt.delta_degC)
unyt_quantity(276.15, 'delta_degC')

Side note:
Substracting two temperatures in Celcius yields delta_degC as unit.
Substracting two temperatures in Kelving yields Kelvin - seems inconsistent.

[Cpp87]

Unyt version: 3.0.3
Phyton version: 3.10.16
OS: Windows

[neutrinoceros]

Sorry for my delayed reply.
I agree this is a problem. I'm however not sure about your proposed solution, because unyt_array.to doesn't support a base argument.

Side note:
Substracting two temperatures in Celcius yields delta_degC as unit.
Substracting two temperatures in Kelving yields Kelvin - seems inconsistent.

I agree, here too.

[neutrinoceros]

I'm however not sure about your proposed solution

to be more actionable: I do not mean that this is definitely not feasible, but I couldn't convince myself it was feasible by studying the code base for 10min. I would be happy to be persuaded otherwise. A proof-of-concept level PR might do the trick.