Strengthening my TS skills by working my way through Advent of Code 2024.
Today was easy. Since JavaScript doesn't have a defaultdict, it's messier than if I had used Python. 🤷 Need to see how other people solved it so I can start writing idiomatic TS.
Surprisingly hard puzzle... it's not really hard but I'm really rusty with my JavaScript Array methods. Array.slice and Array.splice FTW!
Pretty happy with my solution minus the for loop. Interested to see how people do this using pure functional paradigm.
I haven't used Regular Expressions in JavaScript since probably AoC 2021. This was a great puzzle and learned a bit about RegEx. Mozilla Web Docs FTW!
Once I was done the code, I refactored and used a boolean to turn on/off conditionals in my runProgram function. If we keep going down this path ala IntCode, I'll refactor into a class.
I attempted this puzzle right when it came out, but I forgot that JavaScript Maps and Sets store keys as values if passing in an object. Took a week off, came back and wrote up a utility module that has helpers that make JS' data structures function like Python NamedTuples (how I miss thee...).
Once that was done, I was able to write up some hybrid code -- functional shell to perform the calculation, object-oriented core to manage the grid data structure. I'm happy with the code, but I could probably refactor a bit to make part 1 and part 2 flow smoothly.
Did have a bit of a eureka moment when I figured out a better way to do a for loop with a map and then again using filter and keeping the exact data structure I need for the next set of steps. But again, feels like I have this moment everytime I do something complex with functional code vs the standard React DOM mappings. This is my first proper TypeScript job so I'm hoping a lot of the concepts stick.
A+++++ puzzle. Would buy from again.
Part 1 was pretty straight forward. Had to refactor a bit to set things up for Part 2. I was wondering if I could use something like permutations to not have to think, but this was not a brute force solution. Looked at the data a few times before I saw the pattern. Took longer than I'd like to admin.
Update -- saw on reddit that people created a custom comparison function -- which makes sense since we know to sort two numbers.
Completed using an imperative approach. Honestly, I'm not quite sure how I'd make this functional. Part 2 was brute-forced and took 15 seconds. I should be doing better here, but solving a puzzle is solving a puzzle... especially since I'm over a week behind.
Created a wrapper around the native Set to have it work with my custom Point class.
Some optimizations from Reddit:
- put obstacles on the original path vs every single point
- implemented
- store direction in the "seen" set so we can find cycles if something repeats by comparing point and the direction that the guard walks
- do not iterate over every point, you can teleport the guard to the step before the obstacle
- this is probably a good one to do, but honestly I don't really want to implement it
I was thinking about using permutations for this one but 12 choose 12 is a very big number so there definitely must have been another way. Found a recursive solution that worked for part 1, but not part 2 since I recursed backwards vs forwards. Ran into some issues during the part 2 implementation that made me realize I could more easily solve this by recursing forward.
Felt really comfortable to write the solution in TypeScript. Also, learned about structuredClone()
Read through this one almost a couple of weeks ago, but didn't want to spend the time doing yet another grid problem. Came back to it after a while since I have to do some algorithmic thinking for a side project I need to do and this was good practice.
Hacked togethered a solution for Part 1 in a functional way. For Part 2, I had to refactor a bit since my solution was more or less hard coded for part 1. New approach was a bit more functional, but to get things done, I buried a while loop in there as iteration made more sense for the problem at hand. Can clean things up, but 🤷 not worth the effort.
- Day 5 -- create custom comparison function
- Day 6 -- implement some of the optimizations
- Day 6 -- figure out how to do functionally