Add is_unique for Index & MultiIndex

[itholic]

This PR proposes is_unique for Index & MultiIndex.

Basically it's opposite to has_duplicates

>>> idx = ks.Index([1, 5, 7, 7])
>>> idx.is_unique
False

>>> idx = ks.Index([1, 5, 7])
>>> idx.is_unique
True

>>> idx = ks.Index(["Watermelon", "Orange", "Apple",
...                 "Watermelon"])
>>> idx.is_unique
False

>>> idx = ks.Index(["Orange", "Apple",
...                 "Watermelon"])
>>> idx.is_unique
True

And fixed some tests to sync with pandas'.

[HyukjinKwon]

Looks good otherwise.