Add better ways to work with similarity function

[cvigilv]

Writing the examples I encountered some pain points regarding how we compute similarity.

Currently we only have a single entry point for this operation and it only receives 2
hypervectors as inputs. I think we could do something a little bit better in order to make this
operation be more smooth.

I have this implementation in my mini HDC package:

similarity(u::AbstractHV) = v -> similarity(u, v)
δ(u::AbstractHV) = similarity(u)

This allows to create a function to compare a hypervector via function call and can be used for
in conjunction to the .@ macro to compute one-vs-all, for example.

I will add this probable, but leaving this issue so I don't forget.

[MichielStock]

Yes, that is a good idea, I'll implement it today.

[MichielStock]

Is closed by #34