mapX, mapY, mapXY, replace scaleX

[OAGr]

I'm thinking that distributions should have functions for mapX, mapY, and mapXY

dist = normal(5,2)
dist |> mapX({|x| x*3}) // the same as normal(5,2) * 3
dist |> mapY({|y| y*3}) // the same as scaleY(normal(5,2), 3). Only works on pointSet shapes.
dist |> mapXY({|x,y|, x*3, y+2}) // This one is kind of weird. Only works on pointSet. Could be useful for interesting functions. 

We could also have mapXMany for taking an array of elements and mapping all of their X coordinates, and mapYMany and mapXYMany. We could also have mapX2 for two args, if we want.

[OAGr]

One issue with this, (at least as of a few weeks ago), is that map() can be pretty slow, so sometimes hardcoded options are better. Like, yMultiply. Could lead to a ton of functions though so I guess we'll have to decide on trade-offs.

[berekuk]

I checked this, on 0.4 exp with map was 10x slower. On 0.5 it's 2-2.5x slower.

0.4:

$ ./scripts/run.mjs 'd = normal(100,1) -> SampleSet.fromDist; List.upTo(1,100)->map({|i| d->SampleSet.map({|x|exp(x)}) })'
[...]
Time: 2.911 Ok
$ ./scripts/run.mjs 'd = normal(100,1) -> SampleSet.fromDist; List.upTo(1,100)->map({|i| exp(d) })'
[...]
Time: 0.266 Ok

0.5:

$ ./scripts/run.mjs 'd = normal(100,1) -> SampleSet.fromDist; List.upTo(1,100)->map({|i| d->SampleSet.map({|x|exp(x)}) })'
[...]
Time: 0.523 Ok

$ ./scripts/run.mjs 'd = normal(100,1) -> SampleSet.fromDist; List.upTo(1,100)->map({|i| exp(d) })'
[...]
Time: 0.231 Ok

[berekuk]

Related: #973