FEAT: Extending QuadPrecision input to support arrays, sequences and np.numbers

[SwayamInSync]

closes #198

[juntyr]

LGTM

[ngoldbaum]

Awesome, just a few minor comments.

[SwayamInSync]

@ngoldbaum suggestions applied

[ngoldbaum]

I realized after we merged this that this doesn't update the type stubs. Ping @jorenham. Not sure if there's any testing we can do to ensure the tests themselves don't do things that violate the type stubs?

[jorenham]

Not sure if there's any testing we can do to ensure the tests themselves don't do things that violate the type stubs?

Stubtest is able to detect any api changes that are visible by using the inpsect stdlib. But this PR broadens the types that QuadPrecision.__new__ can accept, so that's not something stubtest can pick up on.
I suppose that you could've detected this if there would've been a hand-written type-test that ensures that type-checkers reject things like QuadPrecision(np.int8()). NumPy has a bunch of tests like these, but those are by no means exhaustive.
And unfortunately there exists hypothesis-like fuzz-testing tool for static typing in Python, as far as I'm aware.

TLDR; theoretically yes, practically no 🤷🏻