Feature request: Typings interop

[warrenronsiek]

It would be really cool if you could use defined schemas as types. E.g.:

s = Schema({'x': int, 'y': str})
xy: s.type = {'x': 4, 'y':'a'}

I find myself defining Schemas to validate http responses in tests, but once they are defined, I would like to use them to provide type hints to my IDE. However, ATM I have to define a different type using either Typings or a dataclass in order to do so. This leads to code duplication where I have to define datatypes for type hints that are exact clones of the Schemas used in validation. This is bad.

It would be much better to have a schema.to_dataclass or schema.type method to extract a dataclass or type. Even better if these methods would be called implicity when schemas are used as types.

[cl-parsons]

+1 on that feature. Type-hints are a good practice in python, as it aims to write documented code.

[kohtala]

One should also be able to use types as schemas.

Python 3.8 also has TypedDict:

class S(TypedDict):
    x: int
    y: str

xy: S = {'x': 4, 'y':'a'}

Schema(S).validate(xy)

Or some function to use get_type_hints on a value and validate from that.

pydantic seems close.

[cglacet]

I just tried the library for the first time, it works as expected but I agree that it makes no sense to have both a type and a schema that describe the exact same thing.

In my opinion @kohtala 's proposition looks more interesting because people already have types declared in their backend and they could use your validation tool without having to rewrite any code.