This was a very challenging and fun kata! It required a lot of thought about which data structure and which algorithm to use to minimize the number of computations. In the end it was worth it and I managed to pass it marginally.

or maybe the shunting yard algorithm.

The goal is to create a function that returns the n-th element of this sequence.

Function must return (n+1)-th element.

u(0) : 1st term of sequence

u(1) : 2nd term of sequence

...

u(n) : n+1 th term of sequence

E -> inner (4) -> outer (4) -> __ (5)

For this case, you don't always end up with 5 nodes in the end.

Here's an counterexample to your case : E -> B -> G -> __(4)

That's why you overcounted.

The subarrays (or tuples or Pairs) will have two elements: first the number the squared divisors of which is a square and then the sum of the squared divisors.

A better way to put it would be :

The subarrays (or tuples or Pairs) will have two elements:

• 1. A number k wheresum_of_squared_divisors(k) is a perfect square.
• 2. sum_of_squared_divisors(k).

Nice! I also made one with Desmos, except my rectangle is rotated 45 deg anti-clockwise when a>b.
https://www.desmos.com/calculator/1okszw7mbp

I already solved the problem by myself and then I stumbled upon your solution which I couldn't understand. No worries if you don't remember the problem. Thanks anyways.

Can you please explain your solution?

my bad I can't count :p

Wrong tests.
For n = 1022 (0011 1111 1110)
Expected: equal to 1279 (0100 1111 1111)
Number of 1s in expected answer and n are not same so clearly expected answer is wrong.
Actual : 1534 (0101 1111 1110) = correct answer

Another wrong test :
For n = 900540478
Expected: equal to 900540495
Actual: 900540510

For those encountering runtime errors on Attempt, don't forget the last sentence of the description : "n being the length of the string array, if n = 0 or k > n or k <= 0 return "" (return Nothing in Elm)."