Haskell

This was quite fun! I got a bit distracted trying to rewrite safe in point-free style, but I think this version is the most readable. There's probably a more monadic way of writing lessOne as well, but I can't immediately see it.

safe xs = any gradual [diffs, negate <$> diffs]
  where
    diffs = zipWith (-) (drop 1 xs) xs
    gradual = all (`elem` [1 .. 3])

lessOne [] = []
lessOne (x : xs) = xs : map (x :) (lessOne xs)

main = do
  input :: [[Int]] <- map (map read . words) . lines <$> readFile "input02"
  print . length $ filter safe input
  print . length $ filter (any safe . lessOne) input

Uiua

Uiua is still developing very quickly, and this code uses the experimental tuples function, hence the initial directive.

Try it Live!

# Experimental!
"7 6 4 2 1\n1 2 7 8 9\n9 7 6 2 1\n1 3 2 4 5\n8 6 4 4 1\n1 3 6 7 9"
⊜(⊜⋕⊸≠@\s)⊸≠@\n # Partition at \n, then at space, parse ints.

IsSorted ← +⊃(≍⇌⍆.|≍⍆.)        # Compare with sorted array.
IsSmall  ← /××⊃(>0|<4)⌵↘¯1-↻1. # Copy offset by 1, check diffs.
IsSafe   ← ×⊃IsSmall IsSorted  # Safe if Small steps and Ordered.
IsSafer  ← ±/+≡IsSafe ⧅<-1⧻.   # Choose 4 from 5, check again.

&p/+≡IsSafe .            # Part1 : Is each row safe?
&p/+≡(±+⊃IsSafe IsSafer) # Part2 : Is it safe or safer?

C

First went through the input in one pass, number by number, but unfortunately that wouldn't fly for part 2.

#include "common.h"

static int
issafe(int *lvs, int n, int skip)
{
	int safe=1, asc=0,prev=0, ns=0,i;

	for (i=0; safe && i<n; i++) {
		if (i == skip)
			{ ns = 1; continue; }
		if (i-ns > 0)
			safe = safe && lvs[i] != prev &&
			    lvs[i] > prev-4 && lvs[i] < prev+4;
		if (i-ns == 1)
			asc = lvs[i] > prev;
		if (i-ns > 1)
			safe = safe && (lvs[i] > prev) == asc;

		prev = lvs[i];
	}

	return safe;
}

int
main(int argc, const char **argv)
{
	char buf[64], *rest, *tok;
	int p1=0,p2=0, lvs[16],n=0, i;

	if (argc > 1)
		DISCARD(freopen(argv[1], "r", stdin));

	while ((rest = fgets(buf, sizeof(buf), stdin))) {
		for (n=0; (tok = strsep(&rest, " ")); n++) {
			assert(n < (int)LEN(lvs));
			lvs[n] = (int)strtol(tok, NULL, 10);
		}

		for (i=-1; i<n; i++)
			if (issafe(lvs, n, i))
				{ p1 += i == -1; p2++; break; }
	}

	printf("02: %d %d\n", p1, p2);
}

https://github.com/sjmulder/aoc/blob/master/2024/c/day02.c

Haskell

runningDifference :: [Int] -> [Int]
runningDifference (a:[]) = []
runningDifference (a:b:cs) = a - b : (runningDifference (b:cs))

isSafe :: [Int] -> Bool
isSafe ds = (all (> 0) ds || all (< 0) ds) && (all (flip elem [1, 2, 3] . abs) ds) 

isSafe2 :: [Int] -> Bool
isSafe2 ds = any (isSafe2') (zip [0..length ds] (cycle [ds]))

isSafe2' (i, ls) = isSafe . runningDifference $ list
        where
                list = dropIndex i ls

dropIndex _ []     = []
dropIndex 0 (a:as) = dropIndex (-1) as
dropIndex i (a:as) = a : dropIndex (i - 1) as

main = do
        c <- getContents
        let reports = init . lines $ c
        let levels  = map (map read . words) reports :: [[Int]]
        let differences = map runningDifference levels
        let safety = map isSafe differences
        let safety2 = map isSafe2 levels

        putStrLn . show . length . filter (id) $ safety
        putStrLn . show . length . filter (id) $ safety2

        return ()

Took me way too long to figure out that I didn't have to drop one of them differences but the initial Number

Rust

The function is_sorted_by on Iterators turned out helpful for compactly finding if a report is safe. In part 2 I simply tried the same with each element removed, since all reports are very short.

fn parse(input: String) -> Vec<Vec<i32>> {
    input.lines()
        .map(|l| l.split_whitespace().map(|w| w.parse().unwrap()).collect())
        .collect()
}

fn is_safe(report: impl DoubleEndedIterator<Item=i32> + Clone) -> bool {
    let safety = |a: &i32, b: &i32| (1..=3).contains(&(b - a));
    report.clone().is_sorted_by(safety) || report.rev().is_sorted_by(safety)
}

fn part1(input: String) {
    let reports = parse(input);
    let safe = reports.iter().filter(|r| is_safe(r.iter().copied())).count();
    println!("{safe}");
}

fn is_safe2(report: &[i32]) -> bool {
    (0..report.len()).any(|i| {  // Try with each element removed
        is_safe(report.iter().enumerate().filter(|(j, _)| *j != i).map(|(_, n)| *n))
    })
}

fn part2(input: String) {
    let reports = parse(input);
    let safe = reports.iter().filter(|r| is_safe2(r)).count();
    println!("{safe}");
}

util::aoc_main!();

#Rust

initially, for part two I was trying to ignore a bad pair not a bad value - read the question!

Only installed Rust on Sunday, day 1 was a mess, today was more controlled. Need to look at some of the rust solutions for std library methods I don't know about.

very focussed on getting it to actually compile/work over making it short or nice!

pub fn task_1(input: &str) -> i32{
    let mut valid_count = 0;

    let reports = process_input(input);

    for report in reports{
        let valid = is_report_valid(report);

        if valid{
            valid_count += 1;
        }
    }

    println!("Valid count: {}", valid_count);
    valid_count
}

pub fn task_2(input: &str) -> i32{
    let mut valid_count = 0;

    let reports = process_input(input);

    for report in reports{
        let mut valid = is_report_valid(report.clone());

        if !valid
        {
            for position_to_delete in 0..report.len()
            {
                let mut updated_report = report.clone();
                updated_report.remove(position_to_delete);
                valid = is_report_valid(updated_report);

                if valid { break; }
            }
        }

        if valid{
            valid_count += 1;
        }
    }

    println!("Valid count: {}", valid_count);
    valid_count
}

fn is_report_valid(report:Vec<i32>) -> bool{
    let mut increasing = false;
    let mut decreasing = false;
    let mut valid = true;

    for position in 1..report.len(){
        if report[position-1] > report[position]
        {
            decreasing = true;
        }
        else if report[position-1] < report[position]
        {
            increasing = true;
        }
        else
        {
            valid = false;
            break;
        }

        if (report[position-1] - report[position]).abs() > 3
        {
            valid = false;
            break;
        }

        if increasing && decreasing
        {
            valid = false;
            break;
        }
    }

    return valid;
}

pub fn process_input(input: &str) -> Vec<Vec<i32>>{
    let mut reports: Vec<Vec<i32>> = Vec::new();
    for report_string in input.split("\n"){
        let mut report: Vec<i32> = Vec::new();
        for value in report_string.split_whitespace() {
            report.push(value.parse::<i32>().unwrap());
        }
        reports.push(report);
    }

    return reports;
}

Nim

Got correct answer for part 1 on first try, but website rejected it. Wasted some time debugging and trying different methods. Only to have the same answer accepted minutes later. =(

proc isSafe(report: seq[int]): bool =
  let diffs = collect:
    for i, n in report.toOpenArray(1, report.high): n - report[i]
  (diffs.allIt(it > 0) or diffs.allIt(it < 0)) and diffs.allIt(it.abs in 1..3)

proc solve(input: string): AOCSolution[int, int] =
  let lines = input.splitLines()
  var reports: seq[seq[int]]
  for line in lines:
    reports.add line.split(' ').map(parseInt)

  for report in reports:
    if report.isSafe():
      inc result.part1
      inc result.part2
    else:
      for t in 0..report.high:
        var mReport = report
        mReport.delete t
        if mReport.isSafe():
          inc result.part2
          break

Codeberg repo

Of course I ended up with a off-by-one error for the second part, so things took a bit longer than they really should've.

But either way, behold, messy C#:

int[][] reports = new int[0][];

public void Input(IEnumerable<string> lines)
{
  reports = lines.Select(l => l.Split(' ').Select(p => int.Parse(p)).ToArray()).ToArray();
}

public void Part1()
{
  int safeCount = reports.Where(report => CheckReport(report)).Count();
  Console.WriteLine($"Safe: {safeCount}");
}
public void Part2()
{
  int safeCount = reports.Where(report => {
    if (CheckReport(report))
      return true;

    for (int i = 0; i < report.Length; ++i)
      if (CheckReport(report.Where((_, j) => j != i)))
        return true;

    return false;
  }).Count();

  Console.WriteLine($"Safe: {safeCount}");
}

bool CheckReport(IEnumerable<int> report)
{
  var diffs = report.SkipLast(1).Zip(report.Skip(1)).Select(v => v.Second - v.First);
  return diffs.All(v => Math.Abs(v) <= 3) && (diffs.All(v => v > 0) || diffs.All(v => v < 0));
}

Factor

: get-input ( -- reports )
  "vocab:aoc-2024/02/input.txt" utf8 file-lines
  [ split-words [ string>number ] map ] map ;

: slanted? ( report -- ? )
  { [ [ > ] monotonic? ] [ [ < ] monotonic? ] } || ;

: gradual? ( report -- ? )
  [ - abs 1 3 between? ] monotonic? ;

: safe? ( report -- ? )
  { [ slanted? ] [ gradual? ] } && ;

: part1 ( -- n )
  get-input [ safe? ] count ;

: fuzzy-reports ( report -- reports )
  dup length <iota> [ remove-nth-of ] with map ;

: tolerable? ( report -- ? )
  { [ safe? ] [ fuzzy-reports [ safe? ] any? ] } || ;

: part2 ( -- n )
  get-input [ tolerable? ] count ;

Rust

use crate::utils::read_lines;

pub fn solution1() {
    let reports = get_reports();
    let safe_reports = reports
        .filter(|report| report.windows(3).all(window_is_valid))
        .count();

    println!("Number of safe reports = {safe_reports}");
}

pub fn solution2() {
    let reports = get_reports();
    let safe_reports = reports
        .filter(|report| {
            (0..report.len()).any(|i| {
                [&report[0..i], &report[i + 1..]]
                    .concat()
                    .windows(3)
                    .all(window_is_valid)
            })
        })
        .count();

    println!("Number of safe reports = {safe_reports}");
}

fn window_is_valid(window: &[usize]) -> bool {
    matches!(window[0].abs_diff(window[1]), 1..=3)
        && matches!(window[1].abs_diff(window[2]), 1..=3)
        && ((window[0] > window[1] && window[1] > window[2])
            || (window[0] < window[1] && window[1] < window[2]))
}

fn get_reports() -> impl Iterator<Item = Vec<usize>> {
    read_lines("src/day2/input.txt").map(|line| {
        line.split_ascii_whitespace()
            .map(|level| {
                level
                    .parse()
                    .expect("Reactor level is always valid integer")
            })
            .collect()
    })
}

Definitely trickier than yesterday's. I feel like the windows solution isn't the best, but it was what came to mind and ended up working for me.

TypeScript

import { AdventOfCodeSolutionFunction } from "./solutions";


/**
 * this function evaluates the 
 * @param levels a list to check
 * @returns -1 if there is no errors, or the index of where there's an unsafe event
 */
export function EvaluateLineSafe(levels: Array<number>) {
    // this loop is the checking every number in the line
    let isIncreasing: boolean | null = null;
    for (let levelIndex = 1; levelIndex < levels.length; levelIndex++) {
        const prevLevel = levels[levelIndex - 1]; // previous
        const level = levels[levelIndex]; // current
        const diff = level - prevLevel; // difference
        const absDiff = Math.abs(diff); // absolute difference

        // check if increasing too much or not at all
        if (absDiff == 0 || absDiff > 3)
            return levelIndex; // go to the next report

        // set increasing if needed
        if (isIncreasing === null) {
            isIncreasing = diff > 0;
            continue; // compare the next numbers
        }

        //  check if increasing then decreasing 
        if (!(isIncreasing && diff > 0 || !isIncreasing && diff < 0))
            return levelIndex; // go to the next report
    }

    return -1;
}


export const solution_2: AdventOfCodeSolutionFunction = (input) => {
    const reports = input.split("\n");

    let safe = 0;
    let safe_damp = 0;

    // this loop is for every line
    main: for (let i = 0; i < reports.length; i++) {
        const report = reports[i].trim();
        if (!report)
            continue; // report is empty

        const levels = report.split(" ").map((v) => Number(v));

        const evaluation = EvaluateLineSafe(levels);
        if(evaluation == -1) {
            safe++;
            continue;
        }
        
        // search around where it failed
        for (let offset = evaluation - 2; offset <= evaluation + 2; offset++) {
            // delete an evaluation in accordance to the offset
            let newLevels = [...levels];
            newLevels.splice(offset, 1);
            const newEval = EvaluateLineSafe(newLevels);
            if(newEval == -1) {
                safe_damp++;
                continue main;
            }
        }
    }

    return `Part 1: ${safe} Part 2: ${safe + safe_damp}`;
}

God, I really wish my solutions weren't so convoluted. Also, this is an O(N^3) solution....

J

There is probably a way to write this more point-free. You can definitely see here the friction involved in the way J wants to regard lists as arrays: short rows of the input matrix are zero padded, so you have to snip off the padding before you process each row, and that means you can't lift some of the operations back up to the parent matrix because it will re-introduce the padding as it reshapes the result; this accounts for a lot of the "1 everywhere (you can interpret v"1 as "force the verb v to operate on rank 1 subarrays of the argument").

data_file_name =: '2.data'
data =: > 0 ". each cutopen toJ fread data_file_name

NB. {. take, i. index of; this removes trailing zeros
remove_padding =: {.~ i.&0

NB. }. behead, }: curtail; this computes successive differences
diff =: }. - }:

NB. a b in_range y == a <: y <: b
in_range =: 4 : '(((0 { x) & <:) * (<: & (1 { x))) y'

NB. a row is safe if either all successive differences are in [1..3] or all in [_3.._1]
NB. +. or
ranges =: 2 2 $ 1 3 _3 _1
row_safe =: (+./"1) @: (*/"1) @: (ranges & (in_range"1 _)) @: diff @: remove_padding

result1 =: +/ safe"1 data

NB. x delete y is y without the xth element
delete =: 4 : '(x {. y) , ((>: x) }. y)'"0 _
modified_row =: 3 : 'y , (i.#y) delete y'

modified_row_safe =: 3 : '+./"1 row_safe"1 modified_row"1 y'
result2 =: +/ modified_row_safe data

Lisp

(defun p1-process-line (line)
  (mapcar #'parse-integer (str:words line)))

(defun line-direction-p (line)
  "make sure the line always goes in the same direction"
  (loop for x in line
        for y in (cdr line)
        count (> x y) into dec
        count (< x y) into inc
        when (and (> dec 0 ) (> inc 0)) return nil
        when (= x y) return nil
        finally (return t)))

(defun line-in-range-p (line)
  "makes sure the delta is within 3"
  (loop for x in line
        for y in (cdr line)
        for delta = (abs (- x y))
        when (or (> delta 3) )
          return nil 
        finally (return t)))

(defun test-line-p (line)
  (and (line-in-range-p line) (line-direction-p line)))

(defun run-p1 (file) 
  (let ((data (read-file file #'p1-process-line)))
    (apply #'+ (mapcar (lambda (line) (if (test-line-p line) 1 0)) data))))

(defun test-line-p2 (line)
  (or (test-line-p (cdr line))
      (test-line-p (cdr (reverse line)))
  (loop for back on line
        collect (car back) into front
        when (test-line-p (concatenate 'list front (cddr back)))
          return t
        finally (return nil)
  )))

(defun run-p2 (file) 
  (let ((data (read-file file #'p1-process-line)))
    (loop for line in data
          count (test-line-p2 line))))

I forgot that this started yesterday, so I'm already behind. I quite like my solution for part one, ~~but part two will have to wait~~ edit: part 2 was a lot simpler than I thought after a night's sleep.

Rust

use color_eyre::eyre;
use std::{fs, num, str::FromStr};

#[derive(Debug, PartialEq, Eq)]
struct Report(Vec<isize>);

impl FromStr for Report {
    type Err = num::ParseIntError;

    fn from_str(s: &str) -> Result<Self, Self::Err> {
        let v: Result<Vec<isize>, _> = s
            .split_whitespace()
            .map(|num| num.parse::<isize>())
            .collect();
        Ok(Report(v?))
    }
}

impl Report {
    fn is_safe(&self) -> bool {
        let ascending = self.0[1] > self.0[0];
        let (low, high) = if ascending { (1, 3) } else { (-3, -1) };
        self.0.windows(2).all(|w| {
            let a = w[0];
            let b = w[1];
            b >= a + low && b <= a + high
        })
    }

    fn is_dampsafe(&self) -> bool {
        if self.is_safe() {
            return true;
        }
        for i in 0..self.0.len() {
            let damped = {
                let mut v = self.0.clone();
                v.remove(i);
                Self(v)
            };
            if damped.is_safe() {
                return true;
            }
        }
        false
    }
}

fn main() -> eyre::Result<()> {
    color_eyre::install()?;

    let part1 = part1("d02/input.txt")?;
    let part2 = part2("d02/input.txt")?;
    println!("Part 1: {part1}\nPart 2: {part2}");
    Ok(())
}

fn part1(filepath: &str) -> eyre::Result<isize> {
    let mut num_safe = 0;
    for l in fs::read_to_string(filepath)?.lines() {
        if Report::from_str(l)?.is_safe() {
            num_safe += 1;
        }
    }
    Ok(num_safe)
}

fn part2(filepath: &str) -> eyre::Result<isize> {
    let mut num_safe = 0;
    for l in fs::read_to_string(filepath)?.lines() {
        if Report::from_str(l)?.is_dampsafe() {
            num_safe += 1;
        }
    }
    Ok(num_safe)
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn sample_part1() {
        assert_eq!(part1("test.txt").unwrap(), 2);
    }

    #[test]
    fn sample_part2() {
        assert_eq!(part2("test.txt").unwrap(), 4);
    }
}

Haskell

Had some fun with arrows.

import Control.Arrow
import Control.Monad

main = getContents >>= print . (part1 &&& part2) . fmap (fmap read . words) . lines

part1 = length . filter isSafe
part2 = length . filter (any isSafe . removeOne)

isSafe = ap (zipWith (-)) tail >>> (all (between 1 3) &&& all (between (-3) (-1))) >>> uncurry (||)
 where
  between a b = (a <=) &&& (<= b) >>> uncurry (&&)

removeOne [] = []
removeOne (x : xs) = xs : fmap (x :) (removeOne xs)

Nim

import strutils, times, sequtils, sugar

# check if level transition in record is safe
proc isSafe*(sign:bool, d:int): bool =
  sign == (d>0) and d.abs in 1..3;

#check if record is valid
proc validate*(record:seq[int]): bool =
  let sign = record[0] > record[1];
  return (0..record.len-2).allIt(isSafe(sign, record[it] - record[it+1]))

# check if record is valid as-is
# or if removing any item makes the record valid
proc validate2*(record:seq[int]): bool =
  return record.validate or (0..<record.len).anyIt(record.dup(delete(it)).validate)

proc solve*(input:string): array[2,int] =
  let lines = input.readFile.strip.splitLines;
  let records = lines.mapIt(it.splitWhitespace.map(parseInt));
  result[0] = records.countIt(it.validate);
  result[1] = records.countIt(it.validate2);

I got stuck on part 2 trying to check everything inside a single loop, which kept getting more ugly. So then I switched to just deleting one item at a time and re-checking the record.

Reworked it after first finding the solution to compress the code a bit, though the range iterators don't really help with readability.

I did learn about the sugar import, which I used to make the sequence duplication more compact: record.dup(delete(it).

This is my very naive rust solution, part 2 is mostly just an extra function, so they’re bother covered in this one.

AoC day 2

JavaScript

Also wrote a solution in JavaScript to play around with list comprehension. Wrote some utility functions for expressiveness (and lazy evaluation).

const fs = require("fs");
const U = require("./util");

const isSafe = xs =>
    U.pairwise(xs).every(([a,b]) => a!==b && a-b > -4 && a-b < 4) &&
    new Set(U.pairwise(xs).map(([a,b]) => a < b)).size === 1;

const rows = fs
    .readFileSync(process.argv[2] || process.stdin.fd, "utf8")
    .split("\n")
    .filter(x => x != "")
    .map(x => x.split(/ +/).map(Number));

const p1 = U.countBy(rows, isSafe);
const p2 = U.countBy(rows, row =>
    isSafe(row) || U.someBy(U.indices(row),
        i => isSafe([...row.slice(0, i), ...row.slice(i+1)])));

console.log("02:", p1, p2);

https://github.com/sjmulder/aoc/blob/master/2024/js/day02.js

#Zig

const std = @import("std");
const List = std.ArrayList;

const splitScalar = std.mem.splitScalar;
const parseInt = std.fmt.parseInt;
const print = std.debug.print;
const concat = std.mem.concat;

var gpa = std.heap.GeneralPurposeAllocator(.{}){};
const alloc = gpa.allocator();

const Answer = struct {
    safe: u32,
    tolerated: u32,
};

pub fn isSafe(levels: []i32) bool {
    if (levels.len == 0) {
        return false;
    }
    // slide window in pairs, advancing by one
    var it = std.mem.window(i32, levels, 2, 1);
    const first = it.first();
    const decreasing = first[0] - first[1] > 0;
    it.reset(); // rewind the iterator

    while (it.next()) |slice| {
        const lhs: i32 = slice[0];
        const rhs: i32 = slice[1];
        if (decreasing) {
            if (lhs <= rhs) return false;
            if (lhs - rhs < 1 or lhs - rhs > 3) return false;
        } else {
            if (rhs <= lhs) return false;
            if (rhs - lhs < 1 or rhs - lhs > 3) return false;
        }
    }
    return true;
}

pub fn solve(input: []const u8) !Answer {
    var rows = splitScalar(u8, input, '\n');

    // PART 1

    // determine how many reports are safe
    var safe_reports: u32 = 0;
    var tolerated_reports: u32 = 0;
    var unsafe_reports = List([]i32).init(alloc);
    defer unsafe_reports.deinit();

    while (rows.next()) |row| {
        var levels = splitScalar(u8, row, ' ');

        var report = List(i32).init(alloc);
        defer report.deinit();

        while (levels.next()) |level| {
            const value = parseInt(i32, level, 10) catch continue;
            report.append(value) catch continue;
        }

        if (isSafe(report.items)) {
            safe_reports += 1;
        } else {
            try unsafe_reports.append(try alloc.dupe(i32, report.items));
        }
    }

    // PART 2

    // determine how many unsafe reports can be tolerated
    for (unsafe_reports.items) |report| {
        var index: usize = 0;
        while (index < report.len) : (index += 1) {
            // mutate report by removing one level
            const mutated_report = concat(
                alloc,
                i32,
                &[_][]const i32{ report[0..index], report[index + 1 ..] },
            ) catch report;
            defer alloc.free(mutated_report);

            if (isSafe(mutated_report)) {
                tolerated_reports += 1;
                break;
            }
        }
    }

    return Answer{ .safe = safe_reports, .tolerated = safe_reports + tolerated_reports };
}

pub fn main() !void {
    const answer = try solve(@embedFile("input.txt"));
    print("Part 1: {d}\n", .{answer.safe});
    print("Part 2: {d}\n", .{answer.tolerated});
}

test "test input" {
    const answer = try solve(@embedFile("test.txt"));
    try std.testing.expectEqual(2, answer.safe);
    try std.testing.expectEqual(4, answer.tolerated);

Smalltalk

Discovered a number of frustrations with this supposedly small and elegant language

1. Smalltalk's block based iteration has NO control flow
2. blocks are very dissimilar to functions
3. You cannot early return from blocks (leading to a lot of horrible nested ifs or boolean operations)
4. Smalltalk's messages (~functions) cannot take multiple arguments, instead it has these sort of combined messages, so instead of a function with three arguments, you would send 3 combined messages with one argument. This is fine until you try to chain messages with arguments, as smalltalk will interpret them as a combined message and fail, forcing you to either break into lots of temp variables, or do lisp-like parenthesis nesting, both of which I hate
5. Smalltalk's order of operations, while nice and simple, is also quite frustrating at times, similar to #4, forcing you to break into lots of temp variables, or do lisp-like parenthesis nesting. For instance (nums at: i) - (nums at: i+1) which would be nums[i] - nums[i+1] in most languages

Part 1

day2p1: input
    ^ (input lines 
        collect: [ :l | l substrings collect: [ :s | s asInteger ]])
        
        count: [ :nums |
            (nums = nums sorted or: nums = nums sorted reverse) 
                and: [
                    (1 to: nums size-1) allSatisfy: [ :i | 
                        ((nums at: i) - (nums at: i+1)) abs between: 1 and: 3
        ]    ]    ]

Part 2

day2p2: input    
    | temp |
    
    ^ (input lines 
        collect: [ :l | (l substrings collect: [ :s | s asInteger ]) asOrderedCollection ])
         
        count: [ :nums |
            
            (self day2p2helper: nums)
            or: [ 
                ((1 to: nums size) anySatisfy: [ :i |
                    temp := nums copy.
                    temp removeAt: i.
                    self day2p2helper: temp
                ])
            
            
            or: [(self day2p2helper: nums reversed)
            or: [
                (1 to: nums size) anySatisfy: [ :i |
                    temp := nums reversed.
                    temp removeAt: i.
                    self day2p2helper: temp
                ]
            ]]] .
        ]

day2p2helper: nums
    ^ (1 to: nums size - 1) allSatisfy: [ :i | 
        ((nums at: i+1) - (nums at: i)) between: 1 and: 3    
    ].

Kotlin

A bit late to the party, but here you go.

import kotlin.math.abs

fun part1(input: String): Int {
    return solve(input, ::isSafe)
}

fun part2(input: String): Int {
    return solve(input, ::isDampSafe)
}

private fun solve(input: String, condition: (List<Int>) -> Boolean): Int {
    var safeCount = 0
    input.lines().forEach { line ->
        if (line.isNotBlank()) {
            val nums = line.split("\\s+".toRegex()).map { it.toInt() }
            safeCount += if (condition(nums)) 1 else 0
        }
    }
    return safeCount
}

private fun isSafe(list: List<Int>): Boolean {
    val safeDiffs = setOf(1, 2, 3)
    var incCount = 0
    var decCount = 0
    for (idx in 0..<list.lastIndex) {
        if (!safeDiffs.contains(abs(list[idx] - list[idx + 1]))) {
            return false
        }
        if (list[idx] <= list[idx + 1]) incCount++
        if (list[idx] >= list[idx + 1]) decCount++
    }
    return incCount == 0 || decCount == 0
}

private fun isDampSafe(list: List<Int>): Boolean {
    if (isSafe(list)) {
        return true
    } else {
        for (idx in 0..list.lastIndex) {
            val shortened = list.toMutableList()
            shortened.removeAt(idx)
            if (isSafe(shortened)) {
                return true
            }
        }
    }
    return false
}

G'MIC solution

it day2
crop. 0,0,0,{h#-1-2}
split. -,{_'\n'}
foreach { replace_str. " ",";" ({t}) rm.. }

safe_0,safe_1=0
foreach {
	({h}) a[-2,-1] y
	num_of_attempts:=da_size(#-1)+1
	store temp

	repeat $num_of_attempts {

		$temp

		if $> eval da_remove(#-1,$>-1) fi

		eval "
			safe=1;
			i[#-1,1]>i[#-1,0]?(
				for(p=1,p<da_size(#-1),++p,
					if(!inrange(i[#-1,p]-i[#-1,p-1],1,3,1,1),safe=0;break(););
				);
			):(
				for(p=1,p<da_size(#-1),++p,
					if(!inrange(i[#-1,p-1]-i[#-1,p],1,3,1,1),safe=0;break(););
				);
			);
			safe;"

		rm

		if $>
			if ${} safe_1+=1 break fi
		else
			if ${} safe_0,safe_1+=1 break fi
		fi

	}

}

echo Day" "2:" "${safe_0}" :: "${safe_1}

Elixir

defmodule Day02 do
  defp part1(reports) do
    reports
    |> Enum.map(fn report ->
      levels =
        report
        |> String.split()
        |> Enum.map(&String.to_integer/1)

      cond do
        sequence_is_safe?(levels) ->
          :safe

        true ->
          :unsafe
      end
    end)
    |> Enum.count(fn x -> x == :safe end)
  end

  defp part2(reports) do
    reports
    |> Enum.map(fn report ->
      levels =
        report
        |> String.split()
        |> Enum.map(&String.to_integer/1)

      sequences =
        0..(length(levels) - 1)
        |> Enum.map(fn i ->
          List.delete_at(levels, i)
        end)

      cond do
        sequence_is_safe?(levels) ->
          :safe

        Enum.any?(sequences, &sequence_is_safe?/1) ->
          :safe

        true ->
          :unsafe
      end
    end)
    |> Enum.count(fn x -> x == :safe end)
  end

  defp all_gaps_within_max_diff?(numbers) do
    numbers
    |> Enum.chunk_every(2, 1, :discard)
    |> Enum.all?(fn [a, b] -> abs(b - a) <= 3 end)
  end

  defp is_strictly_increasing?(numbers) do
    numbers
    |> Enum.chunk_every(2, 1, :discard)
    |> Enum.all?(fn [a, b] -> a < b end)
  end

  defp is_strictly_decreasing?(numbers) do
    numbers
    |> Enum.chunk_every(2, 1, :discard)
    |> Enum.all?(fn [a, b] -> a > b end)
  end

  defp sequence_is_safe?(numbers) do
    (is_strictly_increasing?(numbers) or
       is_strictly_decreasing?(numbers)) and all_gaps_within_max_diff?(numbers)
  end

  def run(data) do
    reports = data |> String.split("\n", trim: true)
    p1 = part1(reports)
    p2 = part2(reports)
    IO.puts(p1)
    IO.puts(p2)
  end
end

data = File.read!("input.in")
Day02.run(data)

Elixir

defmodule AdventOfCode.Solution.Year2024.Day02 do
  use AdventOfCode.Solution.SharedParse

  @impl true
  def parse(input) do
    for line <- String.split(input, "\n", trim: true),
        do: String.split(line) |> Enum.map(&String.to_integer/1)
  end

  def part1(input) do
    Enum.count(input, &is_safe(&1, false))
  end

  def part2(input) do
    Enum.count(input, &(is_safe(&1, true) or is_safe(tl(&1), false)))
  end

  def is_safe([a, b, c | rest], can_fix) do
    cond do
      (b - a) * (c - b) > 0 and abs(b - a) in 1..3 and abs(c - b) in 1..3 ->
        is_safe([b, c | rest], can_fix)

      can_fix ->
        is_safe([a, c | rest], false) or is_safe([a, b | rest], false)

      true ->
        false
    end
  end

  def is_safe(_, _), do: true
end

python

import re
import aoc

def setup():
    return (aoc.get_lines(2), 0)

def safe(data):
    order = 0 if data[0] < data[1] else 1
    for i in range(0, len(data) - 1):
        h = data[i]
        t = data[i + 1]
        d = abs(h - t)
        if d not in [1, 2, 3] or (order == 0 and h >= t) or (
                order == 1 and h <= t):
            return False
    return True

def one():
    lines, acc = setup()
    for l in lines:
        if safe([int(x) for x in re.findall(r'\d+', l)]):
            acc += 1
    print(acc)

def two():
    lines, acc = setup()
    for l in lines:
        data = [int(x) for x in re.findall(r'\d+', l)]
        for i in range(len(data)):
            if safe(data[:i] + data[i + 1:]):
                acc += 1
                break
    print(acc)

one()
two()

def is_safe(report: list[int]) -> bool:
    global removed
    acceptable_range = [_ for _ in range(-3,4) if _ != 0]
    diffs = []
    if any([report.count(x) > 2 for x in report]):
        return False
    for i, num in enumerate(report[:-1]):
        cur = num
        next = report[i+1]
        difference = cur - next
        diffs.append(difference)
        if difference not in acceptable_range:
            return False
        if len(diffs) > 1:
            if diffs[-1] * diffs[-2] <= 0:
                return False
    return True

with open('input') as reports:
    list_of_reports = reports.readlines()[:-1]


count = 0

failed_first_pass = []
failed_twice = []

for reportsub in list_of_reports:
    levels = [int(l) for l in reportsub.split()]
    original = levels.copy()
    if is_safe(levels):
        safe = True
        count += 1
    else:
        failed_first_pass.append(levels)

for report in failed_first_pass:
    print(report)
    working_copy = report.copy()
    for i in range(len(report)):
        safe = False
        working_copy.pop(i)
        print("checking", working_copy)
        if is_safe(working_copy):
            count += 1
            safe = True
            break
        else:
            working_copy = report.copy()

print(count)