This repository has been archived on 2024-12-22. You can view files and clone it, but cannot push or open issues or pull requests.
aoc2023/README.md

54 lines
2.9 KiB
Markdown

My solutions to the Advent of Code 2023 programming challenges.
Thanks to Eric Wastl for creating this enjoyable event.
- Requires `lib.py` from [aocpy](https://git.felixm.de/felixm/aocpy) repository.
- Requires `sympy` for day 24.
- Requires `matplotlib` and `networkx` for hands-on day 25.
# Times
- Day 1: 40:00 (I don't know what I am doing.)
- Day 2: 14:15 (Okay, but far way from leaderboard.)
- Day 3: 1st 20:00, 2nd 70:00... (I had a logic error that took me a while to
find.)
- Day 4: 1st 9:06, 2nd 22:31; it wasn't hard but I didn't think quick enough :/
- Day 5: 1st 25:00, 2nd 1:55:00; Required patience and accuracy
- Day 6: 13:54; I was slow because I thought it is much harder?
- Day 7: 75:00; leaderboard 16:00... that was just bad; no excuse
- Day 8: 25:00; I was doing pretty decent here.
- Day 9: 57:00; my input parse function did not consider negative values...
- Day 10: 180:00; this one was hard for me.
- Day 11: 68:00; okay but not elegant and way too slow ofc; x-ray solution
would have been neat
- Day 12: 52:00 and 22:00 for leaderboard; had the right idea and I am good at
this type of problem
- Day 13: 90:00; pretty straightforward but way too slow
- Day 14: 5:55 for first and then 48:00; straightforward but slow, ofc
- Day 15: 4:30 and 31:20; more reading comprehension than programming
- Day 16: 00:27:30 745; best placement so far, of course still horribly slow
- Day 17: a couple of hours; I realized that I need A* after a while; reused
implementation from Project Euler but improved with heapq which was super fun
- Day 18: a couple of hours; I realized that I need shoelace algo for part two
but didn't realize that I have to compute the outer edges for a while and
after I did, I still got clockwise/counter-clockwise issues. They could have
made it meaner by using different clock directions for example and input.
- Day 19: This one was pretty straightforward and required the interval
technique we applied earlier.
- Day 20: Part 2 was tough. I had the right idea of printing out the periods of
the input conjunction gate pretty early, but then messed up the
implementation and thought it wasn't gonna work. Spent a half day thinking up
something else before returning to the idea and it worked flawlessly.
- Day 21: Part 1 was straightforward, but part 2 maybe the hardest problem this
year.
- Day 22: Not too hard, but definitely way too slow for leaderboard.
- Day 23: I found this fun because it required some creativity for part 2. Slow
af, of course.
- Day 24: Solve problem with sympy. I first used numpy to solve part 1 and it
was much faster than using sympy, but I lost that solution when switching to
sympy. Takes about three minutes to run for part 1 and then part 2 is under a
second.
- Day 25: I cheeky solved this by plotting the graph and manually removing the
nodes. I should probably try to write an algorith that does that, but meh.
Manually plotting requires matplotlib and networkx packages.