advent-of-code

[flat sequoia]

ok

i fixed it

i finished so slowly it didn't even tell me my rank

wow

[magic oxide]

where's enira gone

[naive stone]

i didn't get the joke until you made me give it a second thought lol

[pine tiger]

I'M FREE

lmao how am i still in 2nd

[simple raven]

difficult day 😦 I still don't know how to approach this

[pine tiger]

||dp||

or do you not want hints

[simple raven]

even hearing this doesn't really tell me what to do 😄 need to learn sth new to solve this

[pine tiger]

oh, unfortunate

want me to link some resources for you?

[simple raven]

sure, this is outside of problems I normally encounter so I'll gladly take some help

[pine tiger]

https://usaco.guide/gold/intro-dp

[normal dragon]

This is a problem where you should not be expected to "discover" the algorithm on your own.

So yeah researching how to do it is 10000% valid

[pine tiger]

the general gist here is uh

[atomic cairn]

Aoc is now a math challenge lol

[pine tiger]

||building the solution based on states on step at a time||

||now you can probably think of this as memoized recursion, and it probably is||

[simple raven]

I don't even know how to account for the already existing # vs ? in my program yet 😄 I hope I learn something new and solve this by the end of today 🙂 thanks

[pine tiger]

alright glgl

[magic oxide]

is the brute force not gonna work for p2?

[covert arrow]

I doubt it

[magic oxide]

k

[pine tiger]

you can try

[magic oxide]

I have a ||dp solution|| formulated in my head, just don't have time to code it yet

brute force would be quick and easy tho so I'm asking in advance

[covert arrow]

||there's upwards of 2^50 cases per line, so definitely not||

[pine tiger]

just get a better computer smh

[magic oxide]

oh ||do the cases get that large?||

||they looked relatively small from a glance at my input||

ig ||the dp isn't too bad here||

[pine tiger]

we're talking p2 so

[magic oxide]

oh

I haven't seen p2 then yet

[zenith elbow]

RIP I wrote a ||dp|| sol then realized the '#' must be together and not just nCr

[covert arrow]

p2 is ||repeating each line 5 times joined by ?||

[magic oxide]

ah ic

thx

[lime fulcrum]

Not quite sure how I am going to solve this

[magic oxide]

oh god I think I have a cursed solution involving ||popping from a list while iterating over it||

[flat sequoia]

today was just hard 😧

[flat sequoia]

I actually think it's pretty approachable idk

you don't have to do anything super clever it's just a matter of working thru all the cases

like it's not one that you just

if you don't know you don't know

it'll take a while but i think you can get to the solution on your own

[magic oxide]

I think I agree

but I haven't actually coded it up yet so idk

[cyan sleet]

||for p1 just bruteforcing seems very doable||

[broken dock]

||that's because it is||

||for part 2 though you will have to cache your answers and re-use them||

[broken dock]

I agree 100%, discovering ||dynamic programming|| independently doesn't sound like anything unmanageable

[cyan sleet]

php will probably be my bigger enemy today...

[broken dock]

if you draw out decision trees enough you'll eventually come to realize ||"hey, wait a minute, some of these subtrees are identical and we can save time by caching them"||

[lime fulcrum]

So try to draw it out. I will try that

[hollow belfry]

can someone help me with today? i made recursive solution, it solves p1 under a sec but it fails to do part 2 for some reason, idk why

i have been trying for some time now

[broken dock]

as unhelpful as this is going to sound (and I'm sorry about that), the approach for p1 is too slow

[hollow belfry]

i made it recursive

it solves p2 under a sec

[broken dock]

if you're doing what I think you're doing, then you're on the right track, and as a matter of fact the solution you have now probably looks almost exactly like the intended solution

[hollow belfry]

but wrong

[broken dock]

oh nevermind, then you just set it up wrong

well since someone already posted the link here, i'm going to assume I can do the same

https://usaco.guide/gold/intro-dp

[hollow belfry]

it solves p1 successfulyl

but not p2

same func

[broken dock]

i'm aware, i read your messages

it seems you were lucky

that, or you're setting up your ||dynamic programming|| incorrectly

i'm inclined to think you're probably doing the second

since the chances of the first happening are super low

[broken dock]

there can be tons of reasons why this would produce incorrect answers quickly, can't know exactly what at this point though

[hollow belfry]

its exactly same like p1

but different inputs

[broken dock]

right, so you're dealing with edge cases

[hollow belfry]

probably

[broken dock]

if you're positive that you adjoined the input correctly

[hollow belfry]

it is 100%

[broken dock]

then that can only mean it was wrong from the start

[broken dock]

in which case this is what happened

[hollow belfry]

can u help me with my code, or give me a hint

i have no ideas on what i missed

[broken dock]

sure

[hollow belfry]

do i send it here or dms?

[broken dock]

dms

[covert arrow]

(#aoc-solution-hints exists for aoc-related code help btw)

[stable ivy]

quit yo

[lavish terrace]

I gotta say. I am completely stumped on today's AOC. any hints?

[zenith elbow]

if you're on part 1, ||bruteforce|| is actually feasible, do you have any ideas on how that can be done?

[lavish terrace]

yeah I can see that I probably could do that from the input data, but that feels like it's going to royally **** me over in part 2

[zenith elbow]

well a faster algorithm would also explore all the cases, actually, but due to the ||recursive|| nature some ||memoisation|| could be used

[lavish terrace]

hmmm

well. I may have to rethink my bruteforce then, I wasn't going to use ||recursion||

[zenith elbow]

maybe you're already getting the same benefits ||memoisation|| provides through your approach? you could solve part 1 first and see if it takes too long on part 2

[pine tiger]

no, i’m just dumb

[normal dragon]

You think organically discovering Dynamic Programming is an approachable problem for the average computer programmer?

[lavish terrace]

yeah idk if i'm being stupid but I can't actually figure out how to do even part 1

[zenith elbow]

what was your existing idea that you said didnt need ||recursion||?

[lavish terrace]

bad

lmao

||generate all possible combinations for things with itertools.product or something and then substitute them in and check validity||

[zenith elbow]

I think that could work

maybe you made a mistake somewhere in the implementation?

[lavish terrace]

yeah but it's going to be glacially slow surely

I didn't implement anything yet

[zenith elbow]

never know until you start coding ;)

[lavish terrace]

im 99.9% sure that's going to take a very long time

[pine tiger]

man idt a problem has...angered me since this one

like i coded a brute force solution that like tried to match up substrings to regions

that was taking too long

my first thought wasn't "hm, let's cache the results"

it was "oh this ENTIRE APPROACH" must be wrong

next morning, i code it up in like 15 minutes

and it works wonderfully

even faster than my 2d array approach last night

[dense osprey]

solved yesterdays problem (day 11)
and wasnt part 1 and part 2 the same??
||i just multiplied my prefix sum difference by 1e6-1 instead of 1||

how did other people solve it?

was there some solution for part 1 which didnt work for part 2

[fallow viper]

anyone still working on day 1?

[covert arrow]

depends on your implementation

[dense osprey]

||did people simulate the expansion?||

[covert arrow]

if you ||physically expanded the space|| then it wouldn't work for p2

[dense osprey]

i was thinking about doing it first, but i was too lazy to implement it

[fallow viper]

i can't get the second part of day 1 - my sample input works and gives correct answer and from what i can see my code is calculating correctly. but it's not getting right answer with the actual given input

[woven sable]

i did

[covert arrow]

have you tried the ||oneight|| case?

[dense osprey]

implementing that wouldve taken longer for me

how did you guys solve part 2?

[woven sable]

i got stuck on part 1 tho bc the test input works but not my real input

[dense osprey]

curious if there were other solutions i dont know of

[woven sable]

||are we supposed to use djkstras instead of astar?||

[dense osprey]

today or yesterday?

[normal dragon]

There’s also ||twone||

[dense osprey]

today ig

[woven sable]

yesterday

[simple raven]

I did that for part1 but it was simple to then change that to a solution that works for both parts

[woven sable]

haven't looked at today yet

[dense osprey]

how would you use d in yesterdays problem

it was just manhattan distance

[woven sable]

shortest path?

[covert arrow]

and ||threeighthreeighthreeighthreeight...||

[woven sable]

how would u find it then

j ||brute force every single path||

like ||bfs or sum||

[dense osprey]

|||y_2-y_1| + |x_2-x_1| ||

[woven sable]

oh my god 😭

i did way too much

[pine tiger]

it's a bit more complicated than ^ but yeeah

[dense osprey]

yea

😭

[woven sable]

i spent 2 hours on it 💀

[dense osprey]

oh

[simple raven]

||you only care about absolute change in x and y and how many of the expanded slots you pass||

[dense osprey]

you atleast practiced your d implementation

[woven sable]

true

i couldn't figure out how to do d so i did a*

[covert arrow]

I ||kept track of the rows/columns that get expanded and checked one by one||

prefix sum is very smart

[woven sable]

i still haven't finished d10p2 but i was thinking of j doing ||raycasting||

[simple raven]

I got completely shut down by today's problem though. I tried learning ||dynamic programming approach|| but I can't use it in practice yet 😦 might come back to it eventually on an easier day

[pine tiger]

uh i believe i misrepresented the best method to you last night

[simple raven]

there are at least 4 approaches that work and this was one of them

[pine tiger]

a ||brute force with caching(!!) is just much easier to understand and implement||

see my rant above lol

[dense osprey]

yea but due to the matrix only being 140x140, iterating and checking how many empty rows/columns were between each 2 stars wouldnt take that much longer

[simple raven]

my issue is I can't even figure out how to ||bruteforce|| part1 😦

[dense osprey]

idk if my tenses are correct

[dense osprey]

yes

[simple raven]

hm... I might try to figure it out, just having trouble finding ||the possible combinations while also accounting for what's fixed||

[woven sable]

lowk might give up tho based on what ppl are saying ab d12 bc im like 3 days behind now bc of school

[dense osprey]

ray was defo the best

[pine tiger]

that's unfortunate bro

[simple raven]

||walking the path and then calculating area worked for me because that was the easiest continuation of my part1 solution||

[covert arrow]

https://www.reddit.com/r/adventofcode/comments/18f0s0g/2023_day_10_thank_you_ms_paint/

[dense osprey]

yea but what about the edge cases?

but my ray algo had some weird bug

the final result was 1 above the solution

[simple raven]

I implemented it and it just worked 😄

[dense osprey]

ill do day12 now

but im not good at dp

[magic oxide]

dw me neither

[dense osprey]

always avoided doing any dp

it sounds so difficult

intimidating

[woven sable]

what's ||dp?||

[magic oxide]

oh wait this is #advent-of-code

should prolly spoiler those

[woven sable]

for day12p1 the solution looks like j ||brute forcing every combination and splitting it into their respective groups and then add any that match right?||

[magic oxide]

that would work, sure

||won't work for p2 without modifications tho||

[woven sable]

damn alright

[dense osprey]

dickpic

||dynamic programming||

[woven sable]

💀

[woven sable]

what's that

ohh

||caching?||

[dense osprey]

https://dp-book.com/Dynamic_Programming.pdf

[woven sable]

might have to switch to python atp ||i'm not tryna make a cache and allat for kotlin 😭||

[covert arrow]

yes, ||dp|| is basically just ||@functools.cache||

[dense osprey]

pretty good introduction

[magic oxide]

well ||top down dp is||

[dense osprey]

all numbers in todays input are one digit right?

[rocky spire]

There's a O(galaxies^2) solution and a O(gridsize)=O(galaxies) solution, which one did you implement?

[dense osprey]

first

iterated through the pairs of galaxies

[dense osprey]

||yea i trhink ill just do some recursion to try all combinations and see if they are possible||

for part 1

[woven sable]

question on ab day 12

say u have a group requiring an arrangement of like 1, 2 and 3 broken springs

woild ###.##.# satisfy that or only #.##.###?

oops nvm

didn't read it correctly 😭

[rich delta]

Guys for yesterdays qn part 1 , for sample grid isnt the distance coming out to be 384 instead of 374??

[woven sable]

nah it shld be 374

[rich delta]

for 1 and 7 galaxy they have given dist needed is 15 but shortest is 16

[pine tiger]

ru sure you're expanding the galaxy correctly?

[rancid creek]

   0123456789ABC
 0 ....#........  0
 1 .........#...  1
 2 #............  2
 3 .............  3
 4 .............  4
 5 ........#....  5
 6 .#...........  6
 7 ............#  7
 8 .............  8
 9 .............  9
10 .........#... 10
11 #....#....... 11
   0123456789ABC

10-0 = 10
9 -4 = 5
10+5 = 15

[rancid creek]

You have an extra column before the one with 2 & 7, where did that come from ?

[gloomy kernel]

oh dear lord

we are on that stage of the aoc arent we

[rich delta]

My bad. inserted 2 times

Thankyou

[rich delta]

🙏 🙏

[dense osprey]

||dont simulate the expansion||

[lime fulcrum]

so I don't think I get day 12. bit too much for me

[simple raven]

same 😦

[dense osprey]

||backtracking for part 1||

[lime fulcrum]

that doesn't really tell me anything XD

[dense osprey]

(my algo has a bug but ik that the idea is correct)

[dense osprey]

recursion

[lime fulcrum]

i know what it is but not how to apply it here yet. will give it a day or two

[modern terrace]

naive bruteforce for day 12 part 1 is fine if you can wait a few seconds

[dense osprey]

recusrion is just bruteforcing!

[modern terrace]

the dp thingy is a bit better than bruteforcing

[dense osprey]

dp for part 1?

oh

[modern terrace]

and by bruteforcing i mean simulating each and every change with an ||itertools.product('.#', repeat=springs.count('?'))||

[modern terrace]

wow

[lime fulcrum]

what is the wow?

[rapid wolf]

The first knowledge check IG.

[lime fulcrum]

I don't know about first. 5 and 10 were hard

[rapid wolf]

5 and 10 are about coming up with an algorithm, 12 all but requires some basic knowledge of algorithmization techniques.

[lime fulcrum]

fair. I did not know about ||cash|| for instance. I'll think some more on it

[dense osprey]

I FINALLY FOUND MY BUG

TOOK ME 2HOURS

it was an obvious bug

but i wasnt able to find it

the bug was: ||if all blocks were used, i just added 1 to solution without checking if there were #s later in the string||

[smoky finch]

is day 8 part 2 ||supposed to be a huge number? its taking forever to run||

[ripe meteor]

Yes

[smoky finch]

roughly how big?

[ripe meteor]

14 digits

[smoky finch]

is that a joke

i dont think my 1 million steps every second is gonna cut it then 😭

yeah alright i figured out how to solve it

[dense osprey]

what problem was day 8

ah the lcm one

my bruteforce for part 2 is way too slow

[drifting anchor]

so is mine

[cyan sleet]

Ok, I'll give credit where it's due, this was a good task. Part 1 easy enough to brute-force, part 2 basically infeasible with brute-force and you need to do something clever

[drifting anchor]

what is the clever 👀

[cyan sleet]

A lot of problems in AoC end up not doing the second

some flavor of ||dynamic programming||

I ended up writing the good solution for part 1 since I couldn't be bothered to write a brute force one and I kinda expected the twist 😛

[dense osprey]

dp isnt someting clever

its pretty trivial

im just bad at it

[cyan sleet]

granted, the twist I was actually expecting was "concatenate all inputs"

[drifting anchor]

what part was ||optimised||

the entire solve part?

[cyan sleet]

I mean, something more clever than brute force

[cyan sleet]

wdym?

[dense osprey]

yea it was obvious that my bruteforce wont work for part 2

but i just wanted part 1 to be done

and worry about the implementation after reading what part 2 wants me to do

[cyan sleet]

in my head either approach was roughly equally annoying to write

[drifting anchor]

the part that ||checks if the combination of damaged and operational records equals the contiguous groups||

[cyan sleet]

I'm just not sure what you're asking about, the computation of the number of ways you can do things is what's being sped up

just the overall way of computing the answer

it's not just a small change to a brute force solution

[dense osprey]

the part i have trouble with is the ? being added between the string when they are added

idk why

[cyan sleet]

(maybe if you ||did it recursive|| you could ||memoize|| and get stuff for free)

[drifting anchor]

whats ||memoize||?

[dense osprey]

memory

memoize

[cyan sleet]

"oh I solved this (sub)problem before, let me re-use that result"

[dense osprey]

nvm there could be records with a "score" of 0 if the ? wasnt added

[cyan sleet]

apparently the DP array for this is too large for PHP's defaults

yeah ok, trying to solve the task I had in my head might actually be quite hard 😅

huh, neat, ignoring the obvious issue of overflowing like crazy, C++ can actually solve that in reasonable time

~5s

so if the question was to count the number of ways modulo some number that would be totally doable

[drifting anchor]

is ||itertools.product|| a suitable way to get all the possible combinations for part 2

[dense osprey]

mod 1e9+7

[dense osprey]

how long does py take

if cpp takes 5s

[cyan sleet]

lol, I'm scared to even try

[magic oxide]

you'll run out of time/memory

[drifting anchor]

what other ways are thbere

[cyan sleet]

be more cleverer

[drifting anchor]

💀

i got 15 mins bro

[magic oxide]

||dp||

[cyan sleet]

we did point to the relevant topic

[drifting anchor]

idk how to ||dp|| it

[dense osprey]

i think you can easily solve this

[magic oxide]

mull over it ig

[cyan sleet]

I'm kinda scared of the memory usage

C++ is already at like 750M I think

just for the array

actually...

you don't even need a full array

[dense osprey]

dp arr?

[cyan sleet]

yeah

(also this might be too spoilery for this channel)

[dense osprey]

nvm idk what i wanted to say

ill try part 2 tmrw

but i have to study for exams too

[cyan sleet]

this was how I defined my thing
||```perl

DP formulation:

How many ways can we put the first i segments into the first j

spaces. Also keep track of whether it's a forced blank from an earlier segment.

dp[i][j][forced]

[drifting anchor]

thx

[cyan sleet]

there might be nicer ways though, I feel like the last thing can be removed with some cleverness

[dense osprey]

3d

yea i wont even try part 2

[cyan sleet]

3d doesn't really mean much

it's not like you're visualizing a cube or anything

and it's probably doable to get rid of the last axis, it just fir how I was thinking about the problem

[dense osprey]

i struggle with doing 1d dp

i wont be able to solve the coin problem in lesd than 30minutes

i cant even solve leetcode dp problems

[cyan sleet]

@hidden musk have you bailed on aoc? 😭

death by fortran

[hidden musk]

I have

from typing import Final
x: Final[int] = 5
y: Final[float] = 3.5

I will prevail! eventually

soon™️

[rancid creek]

do the easy ones first, the easy dp problems sometimes don't even need dp, but kinda guide you into it

[woven sable]

finally finished day 10 after trying to catch up

i used ||shoelace's theorem and picks theorem, pretty nice cuz i only had to add 2 lines of code||

[flat sequoia]

I mean i don't think the basic concept of dynamic programming is difficult to grasp (if anything, the most confusing thing about it is just its name)

It's basically just recursing into subproblems and slapping a cache on it

I think the problem yesterday is approachable in the sense that like

there's not really anything clever in the transition

If you approach it methodically (for example, looking at the first character a time) and think through the possibilities you can pretty reasonably build out a recursive solution, it might just take some time because there are a decent few things to think about

And then from there realizing you can cache intermediate results to avoid recomputing things isn't too hard

[flat sequoia]

I'm not saying it's easy but like, for example, a character-by-character solve where you see ?#?.?#? and think about the ? and how it can be either a # or a . and figure out that the state you need to keep track of is (chars left, current block, blocks left) to be able to recurse is actually pretty discoverable

There's only a tiny bit of additional state you need to keep track of that's not the problem itself, you don't need to do any clever transformation of the state matrix, there is a small fixed set of ways you can transition, you don't need any weird binary searches or math tricks to reduce

It's definitely not one of those things like if you either know it or not (e.g. CRT)

There's a clear sequence of smaller things to think about and solve to build up the solution, it's not like one big revelation

[magic oxide]

I agree with this

I think you could definitely figure something out just by doing cases by hand and a bit of thinking without any background knowledge in dp

[flat sequoia]

Yeah

An example of a DP technique where this isn't really the case is the O(nlogn) solution to Longest Increasing Subsequence

where you have this weird transposition of the state matrix

and you binary search to transition

or even the O(n²) solution to Longest Increasing Subsequence since to transition you loop thru a bunch of previous things and sum them up

In the character by character solution there's only ever one state where you make two recursive calls

Every other state is either as base case or immediately reduces down to a single other state

[flat sequoia]

I actually hate the name Dynamic Programming

it's the worst name in all of CS and Math

[woven sable]

when i first heard it i thought it was something to do with like dynamically programming something itself like a dynamic website or sum

[normal dragon]

Then you were missing what my point was, which is that||, for most people, this isn't a possible day to solve if you don't already know the underlying algorithm.

Do you already have a foundational understanding of Dynamic Programming? Have you done a few problems in it, and understand the general approach to solving it? Great! You can figure out Day 12 Part 2.

If you don't have that, then this is a nearly impossible day to solve--to organically 'discover' the underlying problem-solving technique on your own. Some AoC problems are like that.||

[rancid creek]

You don't need to know the term dynamic programming to know about recursion or the idea of caching the results so you don't have to look them up again.

And if not, it IS perfectly acceptible and expected to google "how to make my code faster" or just ask here.

[minor cave]

Language Roulette: Day 13
The language is ... Elixir

[rancid creek]

This is supposed to be a fund AND educational event.

[zenith elbow]

oh no, only 10 mins left

[rancid creek]

10 minutes to start googling Elixir, and what is elixir. 🙂

https://en.wikipedia.org/wiki/Elixir_(programming_language)
Elixir is a functional, concurrent, high-level general-purpose programming language that runs on the BEAM virtual machine, which is also used to implement the Erlang programming language

[woven sable]

if it looks hard i'm going to sleep 💀

[rancid creek]

should be a medium tonight

[woven sable]

oh yeah cuz it's a weekday now right

[mossy basin]

still a week before the hardest ones

[woven sable]

honestly idk why this year has j been harder for me, maybe it's cuz im doing it in a new language but day 5 & today were awfully hard for me 😭

[mossy basin]

i was so close to top 100 yesterday 😦

i am slow tho

[covert arrow]

I choked so hard on p2 yesterday I won't even call it a choke anymore

[rancid creek]

I didn't even start until 1/2 hour after it came out .

😐

[zenith elbow]

||I think if someone has good problem-solving foundations, it's possible for them to work out the recursive solution and logically think of caching even if they've not properly studied dp
but if the "average computer programmer" does not have solid problem solving experience it might be a bit hard for them to stay on the right track and implement it -- even if they thought about the right solution, without guidance or experience they might easily resort to thinking it wont work any point in time during implementation||

[rancid creek]

Not that I would have made top 100, but top 1000 is a good target.

[magic oxide]

I'm outside again so probably can't

[earnest forge]

sounds like bad planning

[magic oxide]

hope today isn't too bad

[mossy basin]

i try to finish early so i can start visualization early so i dont stay up all night

[magic oxide]

eh I'm on a trip

[woven sable]

i think i gave up too early for part 2 today, bc i just tried to think of a solution in my head instead of trying to come up w a ||dp|| solution on paper

[magic oxide]

would be weird if I stayed inside all the time

[woven sable]

wish i didn't live on the east coast for aoc lowk

[mossy basin]

why are you attacking me

[earnest forge]

Isn't it like 9PM for you there lol

That sounds entirely reasonable

[covert arrow]

I try to finish early so I don't have to do it during my afternoon classes

[mossy basin]

imagine furiously typing during a lecture

[woven sable]

nah it's 12 am rn

[zenith elbow]

it's unbelievable you can finish both the problem and the visualization daily, twice the work of most people

[earnest forge]

Oh yeah, got east/west the wrong way round mb lol

[rancid creek]

I just want to finish asap consume a bottle of nyquil and fall asleep.

[earnest forge]

At least you don't live in the UK (5AM)

[woven sable]

damn wtf

[mossy basin]

i'll be properly burned out by the end of this month

[woven sable]

i could never get up that early for aoc

[rancid creek]

You should come hang out in the solutions channel, we tend to help 3-4 people finish completely unique solutions every day.

[covert arrow]

my school timetable lines up perfectly for aoc

lunchtime is 12:45-13:45, aoc opens at 1pm

[earnest forge]

Lol nice

[covert arrow]

I just have to sacrifice my lunch, which is acceptable

GLHF

[earnest forge]

10s left, GLHF everyone

[shut trellis]

<@&518565788744024082> Good morning! Day 13 is ready to be attempted. View it online now at https://adventofcode.com/2023/day/13. Good luck!

[zenith elbow]

I hope today's is ad-hoc

[woven sable]

can there be multiple lines of reflection i'm confused why it says each line of vertical reflection

oh nvm

[magic oxide]

oh p1 doesn't look too bad

[woven sable]

looks like a fun problem today

[magic oxide]

should be a quick leaderboard fill up unless p2 is hell

[woven sable]

yeah this is the first problem in a while i've thought of a solution pretty quick lmao

j too lazy to get on my computer and code it

tmrw ig

[covert arrow]

noooo

i entered a ' while hitting enter

[earnest forge]

I'm getting the wrong answer and I can't tell why :(

[covert arrow]

same

[woven solar]

what does part2 even mean

[hexed axle]

Sub 15 minutes today

[woven solar]

oh. somehow the first dumb guess I tried worked. wild.

maybe after some hours I'll understand why it did.

[minor cave]

#1184363086652637204

[earnest forge]

Wtf, leaderboard is full and I'm still stuck on p1

[covert arrow]

same

[flat sequoia]

I rewrote my code halfway through p1 lol

I should have just done that from the start

[flat sequoia]

after getting 2 wrong answers

[pine tiger]

what was even going on today

i spent like 10 whole minutes just on off by one bullcrappery

[earnest forge]

As far as I can tell my problem isn't even an off-by-one

[fervent helm]

I pressed backspace+enter now i have to wait smh

[earnest forge]

Doing it manually for a grid near the start of my input I get the same answer as my code

Also it passes the test

Like wtf

[pine tiger]

ru on p2?

[earnest forge]

No!

[flat sequoia]

wait you're doing even mirrors and not odd mirrors right

[earnest forge]

?

[flat sequoia]

010 vs 0110

should be the latter

[earnest forge]

Yes, 0110

[flat sequoia]

hm

[pine tiger]

for me i just ||printed out the range i should be going through for each line||

[flat sequoia]

Honestly restarting might be the move

There were a bunch of issues in my original code that i didn't know what was wrong

[pine tiger]

i personally would not recommend

[flat sequoia]

so I just restarted and wrote it cleanly

[earnest forge]

I debug-printed all my ||mirror lines||, and all the ones I checked look right

[pine tiger]

if it's any comfort to you, p2 is trivial this time

[earnest forge]

That would be great, if I could get me p1 working

[pine tiger]

perhaps try larger grid sizes?

[earnest forge]

How would that help

In fact, what do you even mean lol

[pine tiger]

uh sorry i'm like kind of assuming that you're doing the same approach as me

which is really dumb

[flat sequoia]

Are you maybe missing like when the mirror is at the rightmost possible slot

[pine tiger]

||all possible lines -> how far do i have to check for each line -> check that far||

[pine tiger]

!!

[earnest forge]

Wait

I'm stupid

[pine tiger]

massive W

[flat sequoia]

did ya get it

[earnest forge]

I did range(len()-1) because I thought that range(len()) would include a zero-range at the end

🤦

[flat sequoia]

i had this issue too 😧

sobbing

[covert arrow]

ah so that was the issue

yup, part 2 was pretty trivial

[pine tiger]

yk for someone at #3 i've never seen advaya here

perhaps his sole purpose is to take #2 from me \j

[flat sequoia]

Lmaooo

Wait holy shit

I fakesolved

[earnest forge]

?

[flat sequoia]

oh wait jk it works

I thought I wasn't taking care of

the case where the left is longer than the right properly for a sec

[pine tiger]

i feel like the test data would have caught that...

no?

[earnest forge]

Yeah, test data splits at 5

[flat sequoia]

Yeah test data would def have caught it

[hollow wharf]

Hello

[mystic sky]

I'm working on day 5 part 2 after finishing part 1 and I honestly have no clue where to start

are there overlapping seed numbers?

[normal dragon]

We're mostly saying the same thing. I think it's perfectly reasonable to realize "Hey, there's clearly some kind of algorithm or technique I don't know, I'll go look it up."

I just think the statement of "people could discover <insert algorithm or technique>" all on their own is silly, even if you have all the separate components. For some, yes, it will be possible. But for the majority, I don't think so.

[normal dragon]

||Try drawing it out on a piece of paper.||

[mystic sky]

okay, thanks

[broken dock]

incredibly based

so glad i'm not the only person in here who feels that way, and then I see people like captain danger who are actually proponents of leaving certain qualities of the input that make the problem solvable (e.g. would otherwise be extremely difficult to solve or downright impossible) hidden, making that realization part of the puzzle as well

[normal dragon]

Moving my reply here:

I empathize with anyone for whom English is not their native language, because unraveling complex problems in a foreign language is a challenge!

But for native English speakers, I genuinely appreciate how the puzzles are puzzles, where the parameters and assumptions need to be teased out through experimentation or careful reading.

Do I feel dumb when I miss something that was stated? Sure. But that's on me.

[woven solar]

...i don't think it has anything to do with understanding of english at all? Certain info about the input just isn't in the problem statement.

(ah, I guess you're probably responding to some complaint that's slightly different from mine)

[normal dragon]

It's a bunch of stuff in #aoc-solution-hints , but we were clobbering the channel from someone looking for help

[broken dock]

also moving my reply here

^ this is a classic example because dijkstra's only works when you assume the input can't/won't have negative weight cycles

so they are obligated to tell you as much about the input

"assume the graph given has no negative weight cycles"

otherwise you would need to go with bellman-ford

[broken dock]

not morally responsible. obligated.

they don't have a choice, and if they leave it out, people will dispute the question

and send in requests for clarifications

and i feel that aoc should be held to the same standard

and I'm glad at least one other person ( @woven solar ) feels that way

[normal dragon]

Except that the Advent of Code problems are solvable with the information presented.

[broken dock]

subject to debate

[normal dragon]

It's not, though. It's empirically true.

[broken dock]

it's not

but ok

[normal dragon]

You're not making an argument.

[broken dock]

neither are you lol

[zenith elbow]

in CP, your code is (generally) going to be submitted elsewhere and your code would be tested on hidden inputs. in AoC there is only one input and you can see exactly what that input looks like. I don't really see a point in emphasizing the constraints of the inputs in AoC.

except for ConfusedReptile's initial argument on day 8 (say, whether there are internal cycles within the input) which would actually effect the logic of the solution, and would disadvantage those who were trying to solve a general solution that covers edge cases versus those who didn't even think of the edge cases

[normal dragon]

Yes, I am, but it's shorthand. Here--

[woven solar]

I'm still not sure if you're responding to my complaint. Assuming yes - I agree that they are. My complaint was that the AoC problems are simpler than they appear, which means if you optimize for solving them quickly, you'll learn the habit of solving the naivest reading of a problem rather than the general problem, and that habit would be maladaptive in situations other than AoC.

[normal dragon]

Statement: Advent of Code problems are solvable with the information presented in the puzzle.

Proof: Every single advent of code problem published to date has been solved.

[normal dragon]

So - that's a much more interesting argument.

And I think it's a reasonable criticism of Advent of Code if you look at it specifically through the lens of practicing for Competitive Programming.

[broken dock]

in all fairness from my POV this was never about practice

aoc is far from an adequate modality for practicing competitive programming (but this makes sense / is not aoc's fault, it wasn't engineered to be as such), so i was never really expecting it to be anything more than something fun to pass time

but at the same time picking apart the language used in the problem statements to surmise certain things about the input is just not fun

[normal dragon]

Is not fun for you.

[lime fulcrum]

aoc is nice because I'm learning new things

[normal dragon]

@broken dock It's perfectly reasonable to be dissatisfied with aspects about Advent of Code. That's one of the joys of humanity--getting to have opinions about things. Making grand declarative statements about the way it should be or how you would change it if you were in charge... all good things to do.

[broken dock]

there have been a number of instances where I felt like I would like to have a word with the problemsetters just this year alone rofl

at least four i think

wonder if aoc has any official channels to communicate feedback to them

[normal dragon]

The creators are active on reddit.

[broken dock]

do they have certain flairs on their profile

[normal dragon]

They do.

[broken dock]

alright goated, i'll just send them dms

although i'm certain there's gotta be a better way of reaching them

[normal dragon]

I would post an open thread for the discussion.

[broken dock]

i don't want to discuss, i want to share my sentiment with them directly and hear from them, directly

we're already discussing here and i've gauged at least a few different perspectives on this

[normal dragon]

You're much more likely to get a response in a public forum than in a private message, but you do you.

[broken dock]

so as far as reaping benefits from a discussion, that's already been done

[normal dragon]

Three final thoughts:

1. You seem stuck in your mindset and your opinion. You didn't acknowledge my objectively correct statement or offer any kind of rebuttal... it seemed to float past you. That's not so surprising given your age, but it's something you'll want to work on as you grow older.
2. I would challenge to you would be to adjust your mindset. Don't judge Advent of Code for its deficiencies compared to the standards of Competitive Programming. Judge AoC on its own merits and what its creators are trying to accomplish, and ground your suggestions for improvement there.
3. I would strongly encourage you to post in a public forum versus directly messaging the people running AoC, especially given that you haven't even been able to find the creator's email.

[flat sequoia]

Personally I think that the way you've been approaching this conversation has been pretty condescending

There's been a lot of picking apart people's words rather than trying to understand them

Ironically that's sorta related to what a lot of people feel about AoC in general

[broken dock]

1. No arguments have, as of yet, been made, I have established an opinion, and you, yours. Moreover, you know nothing about me, so please do not make any assumptions.
2. I acknowledged this in #aoc-solution-hints
3. No, thank you.

[flat sequoia]

Especially the remark on gian's age I think was wholly unnecessary

[broken dock]

Wildest part is I know he doesn't know how old I am

I think my remarks have so far been pretty tame

so that comment certainly came out of left field

[broken dock]

if i'm wrong and anyone has been in some way offended, by all means, please let me know!

[normal dragon]

The age comment was unnecessary. I was going off your linked website which said you're 18.

[broken dock]

ah, my linkedin must be outdated, then

my birthday is on dec. 1st (when aoc starts :) )

[normal dragon]

Happy (recent) birthday!

[red dragon]

Nice chatgpt

Jk jk

[normal dragon]

o u c h

but. fair...?

[flat sequoia]

ChatGPT is less condescending than that

[broken dock]

In all seriousness, thank you for the engagement @normal dragon, I'm aware judging two different things through one lens usually yields one-sided thought patterns like mine, I've only recently started branching out from competitive programming (that was my introduction to computer science, and my segway into the tech industry), so I know I have to keep an open mind with this stuff

[normal dragon]

You're very welcome. It was a good discussion, and I hope you got as much out of it as I did. ^_^

[lime fulcrum]

you can specify that.

[normal dragon]

To @broken dock and everyone else: I’m sorry for coming off rude and condescending. I was in a bad mood, and I like arguing, and I didn’t temper my words. My goal was to help, but I lost sight of that while trying to win the argument instead.

[dense osprey]

having to look at your input to be able to solve the problem is terrible

[dense osprey]

wjat happened?

[broken dock]

difference of opinion

oldest story in human history

[dense osprey]

who had the correct opinion?

[broken dock]

no one

[woven solar]

looking at your input is a terrible idea; that's way too slow 😛 you just hope.

[broken dock]

i have my reasons for feeling the way I do

danger has his

and that's pretty much it

[dense osprey]

my opinions are always superior

[broken dock]

[dense osprey]

if you disagree with me, you are wrong and probably dont know enough about the topic

[spice hemlock]

Ja

[dense osprey]

Ja

[hollow wharf]

please use #1047673173447020564 for sending solutions or at least send it as spoiler ||code|| like this

if you need help with solution use #aoc-solution-hints instead

[golden python]

@low lodge What Down said. Please use the spoiler channel for posting AOC code. I've removed the message.

[low lodge]

Sry i ve posted in aoc solution hints

[vagrant bloom]

How I solved ?

The niceness of a pair (a, b) is the value you will get by subtracting their product from the difference of their squares. For example, the niceness of the pair (7, 4) is 5, because (7²-4²) - (74) is 5.
Find any pair of numbers with niceness 500, where both numbers have 4 digits. If you find a pair (a, b) then write the number (10000a)+b as the answer.

[rocky spire]

ask in #1035199133436354600

[modern harness]

I recently discovered the library icecream and have been really enjoying using it for aoc problems. "Never use print() to debug again". When you use ic to print, it prints the variable names too, pretty prints structures, and even includes syntax highlighting.

[lime fulcrum]

yeah use debug to debug

what ide are you using?

[jade atlas]

gotta give it a try

[pine tiger]

that sounds awesome what

[modern harness]

Neovim. I do sometimes run a debugger, probably not as often as I should, but it depends on what I'm trying to do. If I want to understand where one specific example is going wrong, yeah, a debugger is the best tool, but if I want to see a summary of values to, say, identify which example I should focus on. I've started using breakpoint() more in my code, which is nice since you can throw it in any code without needing to start it specially with the debugger or anything.

[lime fulcrum]

i've been working with vsc and basically only using the debugger

[plain sequoia]

Ong

[rancid creek]

learning how to use the debugger is a key skill.
(sometimes print is not enough)

debug when I know where I want to stop and look.
print when I need to see what the value on loop ~500 is.

[pine tiger]

i remember spamming debugger when i was starting out

now i barely use it lol

[normal dragon]

Spamming and over using techniques when you first learn them is excellent practice and a great way to learn. Easier to push boundaries and break things and learn to pull back than not knowing where the boundaries are when you need them

[ancient surge]

is advent of code something you can do with pretty much no experience

[pine tiger]

i took a look at that and it is absolutely fantastic

tysm!!

[royal gull]

Kinda depends what you mean by no experience. If you know the basics of Python, and can do basic problems, there’s a few that you could solve… but there’s also a few that are probably too hard. But the easy ones might take a while, just to manipulate the data into a format you can work with. I’d suggest asking for a suggestion for a simple one to try, and asking in the hints channel.

[ancient surge]

so far i only know the basics really

[royal gull]

Might be a little tough, but take a look at today’s problem. Anyone else recommend another easy day?

[ancient surge]

alr

[lime fulcrum]

day 1 part one should be doable

else day 4

[compact pelican]

I still need to start 🥲

[normal dragon]

Check out some of the Day 1 problems from earlier years. You can access them all right here: https://adventofcode.com/2023/events

https://adventofcode.com/2015/day/1 -- the very first puzzle released is a great 'first puzzle' for a Python beginner.

[ancient surge]

Alr

Ty

[normal dragon]

Good luck!!

[magic oxide]

day 6

[hidden musk]

4 i think was pretty straightforward

i know of a cabal of ~beginner programmers that mostly got it

2 too

day 7 part 1

[ripe meteor]

Day 8 part1
Day 9 was ok

What did xelf do with that chart of difficulties

[modern harness]

Yeah, I also like the feature where you can just leave an empty ic() and it'll print the location... a much quicker way to do "reached this point" type printing

[shadow musk]

hello 🙂

[shadow musk]

icecream is awesome for debugging and logging

[flat sequoia]

Bruh I just realized

I'm going to be on a plane on the 20th, 21st, and 23rd

gonna have to buy the in-flight wifi and try my best to do aoc 😭

[zenith elbow]

aoc in the air ✈️ ✨

[hexed axle]

Language Roulette: Day 14
The language is ... Dart

[earnest forge]

GLHF everyone

[covert arrow]

GLHF

[shut trellis]

<@&518565788744024082> Good morning! Day 14 is ready to be attempted. View it online now at https://adventofcode.com/2023/day/14. Good luck!

[magic oxide]

oh

I forgor

again

[potent pumice]

man

discord has a weird push notification system

since this server is big enough and whatnot

I viewed that message on mobile and then closed the app

then got a notification for it

[magic oxide]

today looks fun

have a 40 min train ride soon so hopefully I can finish in that time

[flat sequoia]

I keep getting 69

idk why

[hexed axle]

https://discord.com/channels/267624335836053506/1184725793826422844

[flat sequoia]

sobbing

I was missing this for 10 minutes

[earnest forge]

I'm getting 87 for the sample for p2 lol

[hexed axle]

|| while 👀 What on earth is part 2? ||

[flat sequoia]

sobbing

[earnest forge]

||do things lots of times||

[hexed axle]

||ohno||

[earnest forge]

Uh oh this isn't good

[devout turtle]

Whut? 😮
How do you do that. Can you give me tips on that? I mean, I quickly looked at the puzzel for today and my first thought was: another bunch of periods and hashtags, how will I have to handle them this time??
I need those 40 minutes just to get an understanding of what I need to do.
What skills am I “missing” ??

[flat sequoia]

Are you adding to the dict before checking the dict?

[devout turtle]

Don’t get me wrong: I’m definitely not asking for the solution to be handed over, as I want to solve them myself, but the last few days, I seem to have a lot more trouble getting the input into a form I can do something with it.

I’ve been working with JavaScript so far, as I’m a bit more familiar with that now, but those hints I’m looking for are not language specific.

[earnest forge]

Oh my god I'm dumb

I was doing it the wrong way round

🤦

[flat sequoia]

what was the issue

[hexed axle]

Honestly a lot of it is just experience. A lot of puzzles have some common parts, and a lot of real life programming problems have similar common parts. Binary searches are commonly taught for a reason--a lot of time you need to search something. Some folks are really really good at flexible thinking, and noticing how a puzzle needs a common component. But you don't get there until you play with a wide variety of problems and puzzles. It's just going to take a while of figuring out how to view a problem, break it down to smaller, known problems, and tackling those so the big issue is simpler

[earnest forge]

Hey, this is on-topic smh

[earnest forge]

||literally just running the spin cycle the wrong way round, clockwise instead of anticlockwise||

[flat sequoia]

oh nooo 😭😭😭

[earnest forge]

90% of my time in part 2 was that stupid bug lol

(explains why I wasn't ever getting 64 for the sample lmao)

[pine tiger]

I'M GOING TO COMMIT MULTIPLE FELONIES

I READ IT WRONG

I DIDN'T KNOW

A CYCLE WAS ALL 4 DIRECTIONS

I THOUGHT A STEP WAS A SINGLE DIRECTION

LIKE NORTH ONE, EAST ONE

ykw i'll settle for just being on lb at this point

[magic oxide]

I've done 2 years of aoc before this, so I'm kinda used to this input style

[earnest forge]

Another useful thing is lots of people use utility libraries

For example, ||aoc_helper.Grid|| is quite well-suited for today's puzzle

[flat sequoia]

Personally I don't think it provides that much utility if you've worked with the language a lot

parsing is hard sometimes though

[magic oxide]

idk what part 2 is, but from the comments here ||is it about detecting a cycle from shifting north east south west north ...?||

[earnest forge]

Idk, ||rotate_anticlockwise|| and ||rotate_clockwise|| were nice, in addition to the ||from_string constructor||

[earnest forge]

Kinda

(you're doing it ||a large number of times||, so you de-facto need to do that)

[woven solar]

today is pretty much a slight modification of that tetris task from a year or two ago.

[flat sequoia]

😭😭😭😭

[magic oxide]

yeah I thought of the tetris task as well

[flat sequoia]

Today's part 2 is classic aoc

[pine tiger]

bro i spent way too long on checking modulus off by one etc. etc.

perhaps i should make some notes related to that

[earnest forge]

🤝 stupid bugs from not reading the question properly

At least you didn't also have my bug

:P

[flat sequoia]

Meanwhile I just had stupid bug from being bad

😧

[pine tiger]

man as soon as day > 10 my placing just 📉

[earnest forge]

Tbh I kinda find the opposite, in that lb spots usually take longer after day 10 so my being a little bit slow to write the code usually doesn't matter as much

OTOH stupid bugs might occur more like yesterday oh god

[flat sequoia]

I'm def faster on faster days compared to slower days

[pine tiger]

i guess you're just better cuz

[earnest forge]

OOF

That's a huge dropoff for p2

[pine tiger]

bro i'm washed as hell it's not even funny

[earnest forge]

Yeah, me too

[pine tiger]

how do i get my brain to tell me funny elves are not the same as a life or death situation

[earnest forge]

My dropoff for p2 was also pretty bad because of that bug ._.

[devout turtle]

I have to get ready to go to the office, but can you tell me more about “binary searches” in this regard (or point to a good source?, I don’t mind reading in)?

[devout turtle]

Last year was my very first attempt at some after I heard about it for the first time.

[pine tiger]

uh searching for # in sorted array

you can do it by looking in the middle

if too low, that means everything on the left is not worth searching

[hexed axle]

(I haven't read todays problem, it probably doesn't apply to it, but was a generic example)

[pine tiger]

if too high -> everything on the right

you keep on doing that till you narrow your range down to a single #

[devout turtle]

Are those libraries “ready made and available” in Python, or is it something people created themselves?

[earnest forge]

!pip aoc_helper

[marsh current]

aoc-helper v1.8.17

A helper package for Advent of Code

[pine tiger]

both

starwort suggested that library, i just code some stuff myself

[earnest forge]

(but yeah people have offline ones too)

[devout turtle]

Thanks so far. I’ll catch up later.

[rancid creek]

asked chatgpt to help me reduce some voodoo math, and told it the correct answer.

start = 114
r = 156
print((start%(r-start)) + (start-1))
# 143 (it said 119)
print(r+(start-1))
# 269 (which it got right at first, and then bizzarly just said it's 143)

[jade atlas]

today was fun!

[normal dragon]

I liked it a lot

[edgy briar]

Vim

[normal dragon]

[edgy briar]

:vim:

[lime fulcrum]

Today sounds fun

[stable ivy]

I still am stuck on cosmic expansion

[simple raven]

big hint if you want: ||there is no need to do the actual expansion||

[fleet fox]

how do u solve puzzles in less than 10 mins

daheck

[ripe meteor]

Practice and experience

[rancid creek]

Last nights solution is one of the reasons I enjoy aoc: I found something new. the second param to re.sub can be a function/lambda.

spoiler thread for the curious: #1184725793826422844 message

[rancid creek]

It takes me more time than that to read and type. 🙂 Never mind think.

[earnest forge]

You ignore 90% of the puzzle text :P

[fleet fox]

Ohh

I see

thank

[dense osprey]

i hate chatgpt

[dense osprey]

i hate regex

[rancid creek]

It's a pretty simple regex though. '[.O]+'
You can get a lot of utility out of very simple regex without it turning into a giant mess.

[normal dragon]

It’s a slippery slope from that to RFC 822: https://pdw.ex-parrot.com/Mail-RFC822-Address.html

/s

Regex’s are super helpful, though I don’t find them particularly fun to use in Python.

[hollow wharf]

chatgpt finished all assigmnents

if u cant u will be replaced by chat

[warm berry]

Hello, please don't spread misinformation and spam in channels. Thanks

[hollow wharf]

how can it be misinformation when he solved all days for me for now 😭

[ripe meteor]

We dont care, move along

[hollow wharf]

rude @ripe meteor

[jovial jay]

i'm way behind. day 8 seems to be a doozy for me. but i think i can do it

[normal dragon]

You 100% can do it!!!!

[jovial jay]

where's visualization by the way?

[minor cave]

#1180191057498083418

[jovial jay]

yeah, found that. it seems so hidden, but i understand why

[rancid creek]

It's great if you've been able to ask chatgpt for help and get useful information, you're essentially doing enahnced googling. But chatgpt absolutely has not been able to solve all of the aoc problems. If any.

[rancid creek]

day 8 should be easier than day 7, so you can do this. ask if you get stuck, or check the solutions channel for help/hints.

[jovial jay]

i'm not using a generic language for these challenges though. but, still confident i can solve day 8, and if i want, all the days after that

[rancid creek]

if you got part1 done you're in good shape. after that it's a matter of realizing there's a technique to shortcut part 2. You won't want to brute force it.

That comment probably applies to many of these problems.

[ripe meteor]

What are you using?

[simple raven]

I doubt chatgpt could solve any part to any of these problems just from the input text

[jovial jay]

G'MIC. A Domain-Specific Language designed from bottom-up for creating/editing raster graphics images with some 3D capability.

sometimes, the only trouble is parsing string to easily parsable images. at the end of the day, everything so far can be broken down to numbers

[rancid creek]

chatgpt also seems to be under heavy load lately, I've been getting really crap answers from it on simple things like "how do I simplify this math expression".

[simple raven]

I haven't used chatgpt much tbh, I'm satisfied with copilot and even that is often more distracting than helpful

[rancid creek]

microsoft's cheeky labeling for chatgpt4 enhanced help "Bing is your AI-powered copilot"

that can't have been an accident

[simple raven]

lol

[jovial jay]

i wish chatgpt and others just say to people "learn to code" next time

[rancid creek]

They should at least have a warning "you'll need to know how to code to figure out this untested crap I'm giving you" 🙂

[lime fulcrum]

it does explain code that does work somewhat well. for that it is useful

[cosmic tartan]

I haven't started aoc and should but I don't even know how to make a proper function without getting a syntax error 💀

[normal dragon]

This year has been pretty hard for newbies. Other years have been more gentle!

The first AOC Problem from the first year is very doable for newbies to python.

https://adventofcode.com/2015/day/1

[cosmic tartan]

thanks

[normal dragon]

Feel free to ping for help if you need any! AoC is great :D

[cosmic tartan]

I've tried aoc with rust but not with python so I don't really have the pythonic wisdom

[simple raven]

I wanted to do this and it turns out github copilot remembers the whole solution 😄

[normal dragon]

Well you could just ... turn off github copilot XD

[simple raven]

I know, I was just surprised 😄

[rancid creek]

Good time to start learning! Ask as you go and learn as you go!

[rancid creek]

say 1 this year felt a little harder compared to day 1 last year, but it was still in the range of "easy"

[cosmic tartan]

thank you :)

[rancid creek]

You'll find there are a lot of people here that are happy to help anyone that is motivated to learn. 🙂

[cosmic tartan]

that's very encouraging, thank you :)

[sleek cave]

school made me miss three days 😩

good ol' aoc with "great! now for part 2, ||repeat what you did in part 1 one billion times. what is the answer?||"

this year's day 8 has taught me to ||not overthink - if you see a cycle, go for it|| :P

so i hypothesized ||there will be a cycle||

i tested, and ||there was indeed a cycle||

now it's just a matter of if it works, it works :P

[jovial jay]

day 8 taught me to read a little harder

it takes forever to do Part 1 if you don't start at the appropriate place

[sleek cave]

... and solved ||using a cycle||

just like day 8, feels like cheating :P

if it works, it works

[magic oxide]

||there has to be a cycle by the pigeonhole principle||

[jovial jay]

i'm close to wrapping on Day 8 😄

[feral hazel]

||ah thank you i was wondering how to prove it but i didn't know where to start||

[magic oxide]

||proof would go like
each tilt of north west east south can be modeled as a function f that takes in one board state and outputs another board state
there are a finite number of possible board states (for a loose upper bound take N= 2^(board size))
define g(n) = f(g(n-1)), and g(0)=starting board state. Then, in the sequence g(0),g(1),...,g(N), there are N+1 values, so by the pigeonhole principle there has to be two values s and t with s<t s.t. g(s)=g(t), which means that after s applications of f on the initial board state, or g(s), t-s applications of f will return back to g(s), which is a cycle, so there is always a cycle
||

[sleek cave]

let's gooo! excited to see your solution :)🎄

[sleek cave]

||if there is a much larger number of possible board states than N + 1, where N = 10**9, how can you apply the pigeonhole principle?||

||nothing constrains the sequence g(0),g(1),...,g(N) - they can very well be all different game states, no?||

[magic oxide]

||N is defined to be 2^(board size), which is an upper bound for the number of possible board states||

||there are looser upper bounds you can have for this problem, but just any upper bound works||

[sleek cave]

oh okay, i misunderstood ... but even in this case ||who's to say the cycle period isn't MUCH larger than 10**9?||

[magic oxide]

||wdym by 10**9?||

[steady yew]

^ (i would assume mathematically)

[magic oxide]

||the proof shows that the cycle length is bounded by N as well, since it is at most N-0=N||

[sleek cave]

yeah, and so ||the cycle length can be much larger than the number of times we need to spin the platform||

[magic oxide]

oh you mean for this specific problem

[sleek cave]

yep

[magic oxide]

maybe

[sleek cave]

OHHH you meant in general

[jovial jay]

well, i had to use a mixture of languages, the day 8 part 2 answer does not fit in a double too well

[sleek cave]

yeah of course in general ||there has to be a cycle bc there's a finite number of game states||

[magic oxide]

||2^board size is a very loose upper bound, it's possible the number of board states can be reduced||

[sleek cave]

definitely, but through whatever reduction ||it's still >> 10**9||

actually i'm unsure

[magic oxide]

yeah

[sleek cave]

maybe it can be significantly reduced

[magic oxide]

possible reductions are:
only consider spaces without #
only consider states where all rocks are touching something to the north

2nd one is fairly large

1e9 is quite small tho so I still don't think that's guaranteed

there might be other possible reductions

either way ||assuming the existence of a small cycle is a feasibility assumption imo, which I'm personally fine with||

[rancid creek]

My logic went like this: at 10,000 searches per second it would take over 24 hours to search a billion iterations. Ergo per AOC there must be a better way. ||Cycle seemed a possible choice so I tried that first.||

[sleek cave]

yeah same

(and maybe spoiler that lolol)

[magic oxide]

my logic was ||this reminds me of that tetris problem last year I'll just do the same thing||

[noble skiff]

It would be interesting to see the distribution of cycle sizes over all possible boards

[rancid creek]

Ha valid, Now I have to go find that one.

[noble skiff]

My wild guess is that in fact most boards enter a limit pretty quickly

It's pretty easy to test actually

[magic oxide]

i feel the same intuitively yeah

[normal dragon]

I’m sure Eric designed it that way

[candid berry]

Heh, I thought the same thing

but I didn't do the same solution. I thought: ||a billion is not such a large number. I'm guessing that it must cycle, but rather than make any attempt to determine the cycle length, I'll just throw a @functools.cache on my function that does a single spin cycle, and see how long it takes to do a million iterations. Turns out: "fast enough". That basically reduces it into a billion-ish dict lookups, which only takes ~2 minutes on my laptop||

[magic oxide]

lmao

[candid berry]

took less time to run than it would have taken me to code something better 😄

[jovial jay]

what kind of prompt is day 9?

[oak flax]

am i allowed to post paste links in #aoc-solution-hints ? im kinda stuck on the second part of todays problem for a few hours already, just cant get it the way i want, and im not sure if there something im missing or everything is just a totally mess

[magic oxide]

go ahead

we have a pastebin btw

!paste

[marsh current]

Pasting large amounts of code

If your code is too long to fit in a codeblock in Discord, you can paste your code here:
https://paste.pythondiscord.com/

After pasting your code, save it by clicking the Paste! button in the bottom left, or by pressing CTRL + S. After doing that, you will be navigated to the new paste's page. Copy the URL and post it here so others can see it.

[normal dragon]

WHO'S READY FOR DAY 15?!

[magic oxide]

anytime now we're gonna get a really hard problem

[covert arrow]

GLHF

[earnest forge]

GLHF!

[earnest forge]

In -3 days :P

[modern terrace]

just in time

[jovial jay]

Now Ill just read and laugh.

[shut trellis]

<@&518565788744024082> Good morning! Day 15 is ready to be attempted. View it online now at https://adventofcode.com/2023/day/15. Good luck!

[loud lintel]

[jovial jay]

It's not bad

Was expecting something painful

[modern terrace]

YES

[flat sequoia]

37/20

why was that so hard to understand wtf

i had to read part 2 so many times

[earnest forge]

121/35 let's go

I spent time in p1 checking that the input was really one line because the way the puzzle was worded made me think p2 was going to be like 'now there's 2 lines, don't ignore the newline'

LB full

[modern terrace]

83/158

:D

i got in the leaderboard for once

[earnest forge]

do we ping kat?

Nice!

And for p1 rather than p2 which is a little surprising

[modern terrace]

i had to do a test with the test input which probably caused me to have a much lower rank

[earnest forge]

Wait I just realised, we didn't get a language roulette post today

Do you reckon kat is asleep

[modern terrace]

maybe

[pine tiger]

lmao today's reading was ass

[covert arrow]

I threw today because of reading comprehension

[pine tiger]

and yet you still got #20, curious

[covert arrow]

511/342

[earnest forge]

People were slow on p2

I gained 90 places lol

[modern terrace]

i had to do an unnecessary from collections import defaultdict

[shy trench]

wait so how do you know which box to put the new lens in

[pine tiger]

that's why you have it preimported : )

[modern terrace]

and faced a problem with mutability
||(b = [{}]*256)||

[pine tiger]

||hash it||

[covert arrow]

||hash function from p1||

[earnest forge]

Hash it

[mossy basin]

jeeeez the reading

i can't read this much this late at night

[shy trench]

oh wait thanks

im dumbass

[modern terrace]

i had to reread the part for = so many times

[flat sequoia]

I can't have been the only one struggling to comprehend LMAO

😭😭😭

[mossy basin]

i did pretty well considering

but i don't see well this is not my strong suit

[covert arrow]

I kept thinking the symbols meant something for p1

[mossy basin]

gotta wait for the screen reader

[modern terrace]

||If there is not already a lens in the box with the same label, add the lens to the box immediately behind any lenses already in the box.||
this confused me so much

[covert arrow]

When really you have to ||throw the whole thing into hash||

[earnest forge]

I was getting worried about the 'ignore newlines, it's only one line' comments lmao

[mossy basin]

i was just confirming that || it's just normal dict operations||

[covert arrow]

I imagine it's because trailing newlines would change the answer?

[mossy basin]

like, you're just describing a ||dict|| right? right?!?

[pine tiger]

i mean a rogue \n could easily change the hash

[covert arrow]

But the whole thing was one line

[mossy basin]

i did input.replace("\n", "") cause i got scared

[earnest forge]

I mean probably but like, if someone's copying the input wrong that's on them?

[modern terrace]

was there a newline

[earnest forge]

No

[modern terrace]

oh

[mossy basin]

the problem made me paranoid

[feral hazel]

ah i rstrip my input

[modern terrace]

i literally wrote ||if x != '\n':|| just in case

[covert arrow]

How about the ide auto-inserting newlines

[earnest forge]

Yeah same here, I actually checked with shell commands - I really shouldn't have done lmao

[modern terrace]

someone else can open the spoilers for day 15 right..?

[earnest forge]

This is why you use an automation library

[covert arrow]

Fair enough

[mossy basin]

we can try to ping a mod to open a channel

admins asleep

[pine tiger]

yk salt i'm a bit surprised you aren't mod

[mossy basin]

i've never applied, i don't want the responsibility

[pine tiger]

haha relatable

[feral hazel]

oh i just download it from the input page
so far they all end with a newline
iirc they didn't always in the earlier years

[earnest forge]

They need to get the language roulette as well

[earnest forge]

That's just how Eric's server software works

[pine tiger]

Language Roulette is ... Brainfuck

[mossy basin]

let me bug them in the aoc channel for us helpers

[earnest forge]

Automation libraries will remove the extra nl btw

[modern terrace]

there's a special aoc channel for helpers?

[covert arrow]

We have a aoc prep channel

Language roulette today is Go btw

[modern terrace]

ok

[mossy basin]

oh yeah, i did know the language for today

[modern terrace]

my part 2 took 10 minutes longer than part 1 but still somehow landed me on 158

[earnest forge]

I didn't know the hit board game Go was Turing-complete

[mossy basin]

i got like the same place on both

427/410

never 420

[modern terrace]

i guess this is how important milliseconds are in the leaderboard

[mossy basin]

wtf

insane

[earnest forge]

yeah and I wasted 20 seconds checking the input like a moron

(2:16 finish p1)

[mossy basin]

i do a lot of really slow things, i'm not really meant for competitive programming

but i try

[flat sequoia]

Holy shit lmao

[mossy basin]

now let me rename all my variables to be lore friendly

[earnest forge]

So like, any progress? Lmao

[normal dragon]

This was a nice easy puzzle today

[mossy basin]

no one's responded

[magic oxide]

we have an aoc channel for helpers?

[mossy basin]

i know kat said she was watching a movie earlier and couldn't do it, and dawn couldn't

[magic oxide]

yo aboo to the rescue

[mossy basin]

i pinged xith cause it looked like they're online

[normal dragon]

I'm surprised it doesn't just get posted automatically

[mossy basin]

if aboo can do it that'd be awesome

[magic oxide]

it opens after leaderboard fills up

[mossy basin]

i was gonna ping aboo, but then they went offline

[earnest forge]

It's hard to automate because it's supposed to be 'once leaderboard fills' and we don't want to DoS AoC lol

[tepid cairn]

Language Roulette: Day 15!
Sorry for the delay ~ today's language is......Go!

[modern terrace]

30 minutes woo

[mossy basin]

thank you aboo!

[modern terrace]

can someone make the solutions post

[magic oxide]

the heck does it mean by ||ignore newlines||

[feral hazel]

||the trailing newline at the end of the input||

[candid berry]

yeah, I got that wrong at first

I also thought "what kind of red herring is this?", ignored it, and promptly got the wrong result

[magic oxide]

I always remove mine so

[earnest forge]

I thought 'what kind of red herring is this', didn't ignore it, and wasted like 20 seconds checking the input only for the answer to be to do literally nothing

[normal dragon]

I always ||strip the newline of my aoc inputs||

[candid berry]

me too, I just missed it this time and couldn't figure out what I'd done wrong for 30 seconds or so

I even threw in an explicit loop, and then I was ||confused by the blank line|| that showed up in the output before it finally clicked what dumb fucking thing I'd done

🤦

[jovial jay]

If there's one thing that AoC taught me, I can't read

[earnest forge]

http://adventofrealizingicantread.com/

[normal dragon]

I copy paste the same AOC file for every day... I've got all of my convenience functions just pasted at the top

[earnest forge]

I'll do you one better: aoc template all

[normal dragon]

Sure I could do that, but that's more effort

[earnest forge]

(it's literally one command you run at the beginning of the event and it generates the files for each day)

[candid berry]

I copy paste from one day to the next, but yesterday I had a ||.read().splitlines()|| that became a ||.read()|| today since there was only one... 😠

[earnest forge]

((you can also say, e.g. 16-25))