Upstream some Diagrams utilities

[georgefst]

In a similar vein to #40, there are various definitions in the diagrams-lib library which seem like they'd be useful here. Personally I've ended up vendoring the following because diagrams-lib can be a bit of a problematic dependency (it has a big footprint, and doesn't work with Wasm for example: diagrams/diagrams-lib#370):

type P2 = Point V2
unitX :: (R1 v, Additive v, Num n) => v n
unitX = zero & lensVL _x .~ 1
unit_X :: (R1 v, Additive v, Num n) => v n
unit_X = zero & lensVL _x .~ -1
unitY :: (R2 v, Additive v, Num n) => v n
unitY = zero & lensVL _y .~ 1
unit_Y :: (R2 v, Additive v, Num n) => v n
unit_Y = zero & lensVL _y .~ -1

Would you consider copying some of these to this library?

Cc @byorgey

[RyanGlScott]

linear does offer the unit combinator, which allows you to write unit _x and unit _y to achieve what you have called unitX and unitY above. Does unit address that need?

Similarly, I wonder if it would make more sense for linear to offer a combinator for defining unit vectors in a negative direction. Perhaps:

-- | Create a unit vector in a negative direction.
--
-- >>> unitNeg _x :: V2 Int
-- V2 (-1) 0
unitNeg :: (Additive t, Num a) => ASetter' (t a) a -> t a
unitNeg l = set' l (-1) zero

[georgefst]

Oh nice, I hadn't noticed unit.

Similarly, I wonder if it would make more sense for linear to offer a combinator for defining unit vectors in a negative direction.

Yeah, I think that'd make sense, although using negate <$> unit _x isn't too bad really.

[Taneb]

You can also write negated (unit _x), with negated in Linear.Vector

[georgefst]

I'd also quite like to see these, although they might be less general than the diagrams versions and thus clash:

reflectX :: (R1 v, Num n) => v n -> v n
reflectX = lensVL _x %~ negate
reflectY :: (R2 v, Num n) => v n -> v n
reflectY = lensVL _y %~ negate