Question

Task:
For a given position in 2D array generate list of surrounding positions located in radius.

For example:

input: (1, 1)
radius: 1
output: ( (0, 0), (1, 0), (2, 0), 
          (0, 1),         (2, 1),
          (0, 2), (1, 2), (2, 2) ).

I wrote something like

def getPositions(x:Int, y:Int, r:Int) = {
  for(radius <- 1 to r) yield {
    List(
      for (dx <- -radius to radius) yield Pair(x + dx, y - radius),
      for (dx <- -radius to radius) yield Pair(x + dx, y + radius),
      for (dy <- -radius to radius) yield Pair(x + radius, y + dy),
      for (dy <- -radius to radius) yield Pair(x - radius, y + dy)
    )
  }
}

In this code getPositions returns not a sequance of points, but a sequance of Tuple4 of sequances of points.
How can I “concatenate” 4 generators listed in code? Or is there more concise solution for my task? (I’m pretty new to scala).

P.S.
It’s actually for my starcraft bot.

Answer 1

You need to flatten the List (twice), so this would do:

def getPositions(x:Int, y:Int, r:Int) = {
  for(radius <- 1 to r) yield {
    List(
      for (dx <- -radius to radius) yield Pair(x + dx, y - radius),
      for (dx <- -radius to radius) yield Pair(x + dx, y + radius),
      for (dy <- -radius to radius) yield Pair(x + radius, y + dy),
      for (dy <- -radius to radius) yield Pair(x - radius, y + dy)
    ).flatten
  }
}.flatten

It’s not a ‘lazy’ spiral, though.

Edit

That one is lazy:

def P(i:Int, j:Int) = { print("eval"); Pair(i,j) }

def lazyPositions(x:Int, y:Int, r:Int) = {
  (1 to r).toStream.flatMap{ radius =>

    (-radius to radius).toStream.map(dx => P(x + dx, y - radius)) #:::
    (-radius to radius).toStream.map(dx => P(x + dx, y + radius)) #:::
    (-radius to radius).toStream.map(dy => P(x + radius, y + dy)) #:::
    (-radius to radius).toStream.map(dy => P(x - radius, y + dy))
  }
}


print(lazyPositions(1,1,1).take(3).toList) # prints exactly three times ‘eval’.

I’ve used the def P method to show the real laziness. Everytime, you’d create a Pair, it gets called. In a lazy solution, you’d only want this on demand.