Add typing check

[PhilippeMoussalli]

This PR adds type checking into the CICD pipelines of Fondant. We're leveraging the latest version of mypy.
#46

[NielsRogge]

Tbh I find these typing hints pretty hard to read and I'm not sure they're helpful for us. They seem useful for computers but not for humans

t.List[t.Union[str, t.Dict[str, str]]] => I mean just look at this

Just my personal opinion :D

[PhilippeMoussalli]

I twas getting errors from mypy as follows
Incompatible return value type (got "List[str]", expected "List[Union[str, Dict[str, str]]]") which went away when I explicitly defined the type.
However, the args of kubeflow does required to have a list that can contains either a string or a dictionary so in that sense I think the type hint in the function's return is valid.

I'll just add a mypy exception for that one

[RobbeSneyders]

The typehints are meant to be useful for computers though, that's why we are adding mypy :)

If we have complex types that are reused across Fondant, we can create aliases for them. Eg:
https://github.com/encode/starlette/blob/master/starlette/types.py

I would rather have a correct complex type than ignore the issue. Defining the types helps catch bugs.

[PhilippeMoussalli]

alright I think that's a better approach indeed :) thanks for the suggestion. Added

[RobbeSneyders]

Nit, but the following is a bit easier to follow:

spec_data = pkgutil.get_data("fondant", "schemas/manifest.json")

if spec_data is None:
    raise FileNotFoundError("component_spec.json not found in fondant schema")

spec_str = spec_data.decode("utf-8")
spec_schema = json.loads(spec_str)

[RobbeSneyders]

Thanks @PhilippeMoussalli!

I would revert ignoring the args type, but otherwise LGTM.