fix: added branch for value_in to handle Celsius

[dhfwc18]

Added branching logic to handle conversion to Celsius in the value_in method of PySIObject. Celsius now a valid input for the unit argument.

Honestly suprised I haven't caught this one given that I worked in low carbon heating. But the current implementation of value_in that I suggested few days ago does not take into account the seperate implementation of Celsius so results.value_in(si_units.Celsius) would return a TypeError because Celsius is not an instance of SIObject.

The best way I can think of to fix this is to add a branch to handle Celsius as you did with __truediv__.

Think the overhead this fix add should be negligible. Not sure what you guys think?

[prehner]

Thanks for the fix, I didn't foresee this initially. While we're at it, can you also add the value_in method for Angle? Then we have everything consistent.

[prehner]

Sorry, maybe I was unclear. What I meant was basically

impl Angle {
    ...
    fn value_in(&self, unit: &Self) -> f64 {...}

Just to avoid confusion about the value_in function. Most users will not notice that Angle is actually a different class than SIObject, so this is more coherent that way.

[dhfwc18]

Yes sorry I deleted my message

[dhfwc18]

I notice what you meant. Havent read up on Angle when I implemented the fix. Realised what you meant. Added the mirror. Hope this works!