Added scan()

[414owen]

Scanning is like reducing but keeping the intermediate accumulator states, haskell docs

[colbyr]

Hey @414owen sorry I took so long getting back to you here. This looks pretty useful!

My question for you is can scan be implemented in terms of one of the existing protocols? It seems to me that you might be able to use _map (since the length of the input is the same as the length of the output) or failing that _reduce, but I definitely haven't through all the semantics.

The benefit of using an existing protocol is that if a project depends on transmute and has implemented the TransmuteCollection protocols for any project-specific types, they'll be able to use scan with those types without doing any extra work. If we add a new scan protocol, they'll have to implement it for their types before they can use it.

For examples of using the existing protocols, see pluck, which uses _map, and merge, which uses _reduce and _set. Anything of the functions in src/internal are fair game.