trjhtr

I had a thought last night on how I could easily write a Pascal's Triangle generator basically just using partition and map. I tried writing it this morning, and it got slightly more complicated than I originally thought it would, although it's still pretty nice.

My main concern is the wrap "hack" that I'm using. I found two similar, but different ways of writing it. I can either wrap the input to the iteration function in 0s, or I can wrap the result of the iteration function in 1s. Both give the same result, but the need for a wrapping helper pollutes an otherwise nice ->> chain. If anyone can think of how I can avoid a separate anonymous function here, and ideally avoid wrapping the input/output in 0s/1s, I'd appreciate it. I'll take any other critique as well, although there isn't much else going on here.

Usage:

(let [pre-t (pascals-triangle-pre) ; Wrapping pre-processing
 post-t (pascals-triangle-post) ; Wrapping post-processing
 n 5]
 (println (take n pre-t))
 (println (take n post-t)))

([1] [1 1] [1 2 1] [1 3 3 1] [1 4 6 4 1])
([1] [1 1] [1 2 1] [1 3 3 1] [1 4 6 4 1])

(ns minesweeper.irrelevant.pas-tri)

(defn pascals-triangle-post 
 (let [wrap #(vec (concat [1] % [1]))]
 (iterate #(->> %
 (partition 2 1)
 (mapv (partial apply +'))
 (wrap))
 [1])))

(defn pascals-triangle-pre 
 (let [wrap #(vec (concat [0] % [0]))]
 (iterate #(->> %
 (wrap)
 (partition 2 1)
 (mapv (partial apply +')))
 [1])))