Include the explicit type in ColorLike and RectLike

[aatle]

Add the concrete classes to the ColorLike and RectLike type unions.
They are already implied by the other types, but should be added for clarity.

Since they are only type aliases to unions, their variable name is not always given, so what they represent might not be entirely obvious to users.

Bonus: Sort types in ColorLike by how often they are used.

Example mypy error message, before:

$ mypy test.py
test.py:38: error: Incompatible types in assignment (expression has type "float", variable has type "int | str | SequenceLike[int]")  [assignment]
test.py:39: error: Incompatible types in assignment (expression has type "int", variable has type "SequenceLike[float] | SequenceLike[SequenceLike[float]] | _HasRectAttribute")  [assignment]

After (much more clear!):

$ mypy test.py
test.py:38: error: Incompatible types in assignment (expression has type "float", variable has type "Color | SequenceLike[int] | str | int")  [assignment]
test.py:39: error: Incompatible types in assignment (expression has type "int", variable has type "Rect | FRect | SequenceLike[float] | SequenceLike[SequenceLike[float]] | _HasRectAttribute")  [assignment]

Note:
typing.py now imports things from pygame. I don't think there is any risk of circular imports though.

Note 2:
The _<Shape>Like type unions in geometry.pyi include their concrete classes already.

[damusss]

No, there shouldn't be circular imports.
I have nothing against it, it's redundant but it's clearer, so why not.

[ankith26]

I'm not 100% onboard making verbose unions more verbose, but I can see your argument for these changes. Either way, this PR is not doing anything incorrect so I am approving anyways. Thanks! 🎉

[ankith26]

I'll just leave this PR open for a bit more before merging incase anyone else wants to chime in

[bilhox]

LGTM ! The changes seem fair to me.