Use types pervasively

[alvorithm]

Problem

We use types in some function signatures for documentation, but unless one has IDE-level checking, they are not else taken-advantage of.

Solution

• use types pervasively
• activate mypy in the CI pipeline (Github Actions).
• see also Inconsistent indication of types in docstrings #75 for impact on docstrings.

[alvorithm]

Alternatively, use pyright.
getting started
pyrightconfig.json config file to be added and versioned

[alvorithm]

In this context

• Implement a Shape=Union[torch.Size,Tuple[int,...]
• use cast where possible instead of #type: ignore (see https://mypy.readthedocs.io/en/stable/cheat_sheet_py3.html), also see this issue in PyRight for an example.