advent-of-code VS aoc2021

Pretty happy with this one ``` def solve() -> List[int]: seen = [] won = [] scores = [] for n in NUMS: seen.append(n) for board in BOARDS: transpose = list(zip(*board)) for i, line in enumerate(board): if (all(num in seen for num in line) or all(num in seen for num in transpose[i])) and board not in won: won.append(board) scores.append(sum(sum(num for num in line if num not in seen) for line in board) * seen[-1]) return scores

day02 of doing only one-liners and I already failed for the second part, it would have been a huge mess. The first part is pretty messy as well so if anyone can improve on it I would be very grateful ! https://github.com/masmeert/advent-of-code/blob/master/2021/day02/main.py

I actually did the same as you when I first solved it But I saw other people use the l[3] > l[n-3] trick, it didn't make much sense at first but now I get it and it's quite clever. Basically take [1 2 1 3], the windows would be w1=[1 2 1] and w2=[2 1 3], as 2 and 1 appear in both windows what really matters is w1[0] and w2[2] which are unique to each: 1+(2+1) < (2+1)+3 == 1 < 3.

I've been doing similar (in three langs, though Ive fallen behind on go and rust over the past few days) -- https://github.com/A-UNDERSCORE-D/aoc2021/blob/main/aoc/go/util/slice.go is most of my generic go fun, lots of implementations of higher order functions and nice wrappers

Python source

As usual, first implementation. Actually started by doing sum(range(1, steps+1)), and after solving remembered that theres a math thing for that (sum(0..n) = n * (n + 1) / 2). Nothing super fancy other than that. Code Finally starting to see speedup from pypy. That JIT is amazing but has its costs.

https://github.com/A-UNDERSCORE-D/aoc2021/blob/main/aoc/go/six/06.go -- Nothing super fancy. Nice ability to pick arrays over slices here though. Probably contributes a bit of speedup.