Stravalib 2.0 -> Introduce custom Pydantic types for field conversions

[jsamoocha]

• Introduce custom types that inherit from the generated model's base types and add accessor logic to get the (Pint) Quantity representation of these base types.
• Introduce custom type for time zones
• Introduce custom type for time delta

Depends on #498

[lwasser]

this i think answers my question from the review about what the plan is.

[jsamoocha]

Yes, the plan is to access the "extended" attributes as follows:

d = some_object.distance  # returns float
q = some_object.distance.quantity  # returns pint.Quantity (in meters)
z = unithelper.feet(some_object.distance)  # returns pint.Quantity (in feet)

I found a way to achieve this and maintain a correct type hierarchy.

[lwasser]

@jsamoocha this sounds great.

i have a very basic question. The extender approach sounds like a smooth way to implement these conversions. I"m curious how this compares to returning the distance in quantities via a validator for each attr that forces the conversion when the model is populated. I have a few guesses as to why the extender is better but wanted to get your feedback!

• perhaps involving typing and type overrides or
• perhaps it would make the overall code base more complex having so many validators?
• Or finally perhaps it allows other users who may want to raw response values to convert as they please + the quantity conversions that are so nice to have in place.