fix: Mark standalone None as is_optional.

[DanCardin]

Description

This is not generally meaningful, because something is not really ever going to be annotated as foo: None.

However, when writing code that operates over something like for t in TypeView(int | None).inner_types, it can come up.

Perhaps in concert with #10, I might be able to avoid it, but... (it only occurred to me in writing this description :P), it maybe seems like the logical extension of:

assert TypeView(str | int | None).is_optional is True
assert TypeView(int | None).is_optional is True
assert TypeView(None).is_optional is True

[peterschutt]

Thanks, makes sense to me!

I think the collection stuff is unrelated though.

[github-actions]

Documentation preview will be available shortly at https://litestar-org.github.io/type-lens-docs-preview/12

[peterschutt]

I'm having second thoughts on this, Optional is specifically a union that includes None in its members. Would an is_none_type property be better?

[DanCardin]

I have an is_none_type branch pending also. As i mentioned above the strip_optional() branch would also likely handle this case for my particular usecase. Thus, if you want to close this, i think i'll be fine.

But I was just thinking of the logical extension of the types. None is a special type that indicates "optional"ity, whether it's actually required or not. So i dont necessarily see the difference between int | None and None | None and None in terms of the expected result of this function, versus having the presense of None cause it to be True, but only when combined with other types.