Add an implementation of unzip

[davidshepherd7]

Here's the unzip implementation I described in #239 cleaned up and with documentation.

I'm still concerned about tee's memory usage as it could result in difficult-to-debug high memory usage. However I don't currently have any other ideas that would allow lazyness.

[eriknw]

I think it will be more convenient if unzip returns a tuple, not an iterator. I also think this will encourage more idiomatic use, since all iterators should be consumed at a similar rate to limit the impact of tee.

I'm less concerned about having ultimate flexibility with being able to handle an infinite number of iterables with infinite length. How would one operate on that effectively? I don't know (but maybe answering that will reveal how to limit the impact of tee). Anyway, there is a balance to be had between flexibility and convenience.

I'd like to merge this. Do you agree about returning a tuple?

[davidshepherd7]

I changed it to return a tuple, and also fixed a bug I came across with empty iterators. Tests pass on my machine (python2 and 3) so it should be ok to merge.

[davidshepherd7]

Any more thoughts on merging this?

[eriknw]

Any more thoughts on merging this?

Yeah, my bad. I think I stretch myself too thinly at times.

This is good to go. Thanks, @davidshepherd7!

[davidshepherd7]

Yeah, my bad. I think I stretch myself too thinly at times.

I completely understand. 👍