advent-of-code

[earnest forge]

Rip

[normal dragon]

That's why I have a loadfile and a loadlines

[candid berry]

🙂

[rancid creek]

jeeze man, and part 1 took me 5 minutes. I think 4 minutes of that was reading, 1 minute was typing and ~0 minutes thinking.
I'm just here cause I enjoy hanging out with you all.

[normal dragon]

My p1 solve was 5 min too

[tawny sorrel]

Hey i wrote this regex (\w+)([=-])(\d+)? to match these strings rn=1,cm-,qp=3,cm=2,qp-,pc=4,ot=9,ab=5,pc-,pc=6,ot=7 for today's aoc, but I dont match them correctly? What am I doing wrong?

[hollow wharf]

Use #aoc-solution-hints

[jade atlas]

I wrote a Lens class and implemented __hash__ method, hoping that when I will insert the objects in a dict, it will use that function. But surprisingly, it doesn't.

[covert arrow]

dicts use __eq__ to compare items too

[jade atlas]

to compare the keys? I don't think so

[covert arrow]

yes it does

[jade atlas]

then how come when I insert Lens("cm", None) and Lens("cm", 3), it puts them in different keys? I have implemented __eq__ to compare the labels

[covert arrow]

they have to have the same hash and compare equal

#bot-commands message

[tawny sorrel]

Hey where should I send my code @covert arrow ?

[jade atlas]

Ah I see what happened

[covert arrow]

#aoc-solution-hints maybe, but I'm not sure if you'll be able to get decent help if it's in Go

[tawny sorrel]

I just want help with the regex

[covert arrow]

the regex is fine, as demonstrated before

[rancid creek]

which regex ?

[rancid creek]

||\d*|| not \d+

[tawny sorrel]

oh

[lime fulcrum]

what newline character are they talking about?

[modern terrace]

doesn't actually exist

[normal dragon]

The input downloaded straight from AOC's website includes a \n at the end

>>> with open('input.txt') as f:
...     txt = f.read()
...
>>> '\n' in txt
True

@lime fulcrum

[lime fulcrum]

ah yeah i just copied into a file. apparently i lost the \n

[modern terrace]

the what

copy pasting directly from the thing doesn't have it

[normal dragon]

Yeah, it's only if you directly download your puzzle input from Advent of Code's website

But a lot of people do that for automation purposes

[modern terrace]

it's something in the downloading process then

not the input itself

[normal dragon]

The raw file has a new line character

[covert arrow]

doesn't the ? already take care of that?

[normal dragon]

It should

Though it's extra and unneeded

[covert arrow]

I like (\w+)(?:-|=(\d+)) more

[rancid creek]

It might make a difference depending on what they're doing with it.

print(re.match('(\w+)([=-])(\d+)?', 'xx-').groups())
# ('xx', '-', None)
print(re.match('(\w+)([=-])(\d*)', 'xx-').groups())
# ('xx', '-', '')  

[rancid creek]

||re.split('(-|=)', step)||

[modern terrace]

raw file from the downloading process, yeah

anyway whether it exists or not the important thing is actually getting the solution

[covert arrow]

ah, that makes more sense

[normal dragon]

Downloading from curl doesn't magically add a new line character to the end of the file. The file itself has a newline character.

[modern terrace]

something in the html then?

idk

[feral hazel]

the input for 2015-01 doesn't have a newline for me but today's does

[flat sequoia]

Mine doesn't, interesting

but I copy pasted the sample into my code editor manually

and vscode added a newline to that lmao

[modern terrace]

we're going back down now

[hollow wharf]

solved for me, so yeah.

[broken dock]

^

[simple raven]

day 15 part 1 ||was the easiest for me so far, suuurely part 2 will not get too much harder right? 🤪 ||

[ripe meteor]

part2 is comfy as well

[glad tartan]

if you're not reading the HTML you're missing out

[minor cave]

Day 14 was certainly a chatty day

[magic oxide]

interesting

[hollow wharf]

[sleek cave]

today seems like a breather. upon first glance part 2 is also ||easy for a midgame day||. hopefully i didn't miss anything :P

[normal dragon]

Day 14 was a really fun puzzle with some different components that all fit together

[normal dragon]

||I was shocked at how relatively simple day 15 was compared to a typical late day aoc problem. It actually tripped me up. I was like “is that it???”||

[rapid wolf]

It was more of a ||reading comprehension|| day tbh, but even that wasn't too bad.

(definitely did not suffer for 2 hours on p2 because APL really really doesn't like this style of problems

[magic oxide]

my biggest issue today was formatting

[normal dragon]

It was a clear example of the puzzle description being somewhat obfuscated to confuse the LLMs

[normal dragon]

Formatting for what?

[magic oxide]

i decided to do today as a single expression for fun

because p2 was kinda easy

so the formatting was annoying because indentations

[hollow wharf]

You did 1line p2?

[normal dragon]

Nice! Did you end up posting your one liner?

[magic oxide]

day 15 yeah

i've cleaned it up a bit

#1185091269396025344 message

[noble skiff]

Yeah definitely

[noble skiff]

The creator has already said the puzzle descriptions are not being changed at all for the existence of LLMs

There's always been days where you need to read extra carefully

[cyan sleet]

but now people have a scapegoat to blame their reading comprehension skills on :^)

[noble skiff]

Lol

I admit there was one passage I read like 4 times

It was confusing how they said nee entries in the box go in the back, but if it's empty it goes in the front

The added clause has negative value

[cyan sleet]

it goes behind any existing lenses

if no lenses exist, that lens is placed at the front

[noble skiff]

Yes, the second clause adds nothing

[cyan sleet]

what if I add my lens at the back end of the empty box?

there is no room for lenses behind it now

[noble skiff]

Huh

Not sure if being serious or not

[cyan sleet]

serious, that's what I think they want to avoid in their box analogy

[noble skiff]

It's pretty clear that if a line is empty, the front and back of a line are the same place

[cyan sleet]

but this is some physical box in their analogy

(the analogy is bad, don't get me wrong)

[noble skiff]

This doesn't confuse anyone, and what you're saying doesn't really make sense: lenses go at the back of the existing lenses, not at the back of the box

If there's no existing lenses then "behind existing lenses" is the same as the front

[cyan sleet]

without the second clause it's not defined where you should put a lens in an empty box

[noble skiff]

It absolutely is

I mean I guess that's your interpretation, fair enough

[cyan sleet]

you're thinking about a data structure, in their analogy you have this physical box

[noble skiff]

I understand that

I think it's perfectly well defined without that second clause

And clearer

I'm surprised that someone thinks otherwise but it makes sense that different people interpret it differently

[cyan sleet]

so what if I place the lens at the back of the box?

and I want to add one more lens

behind the lens is the wall of the box

I can't move the existing lens per the other rules

[noble skiff]

The box doesn't have finite size so this doesn't even make sense

All I'm saying is that for me the second clause made it less clear, rather than more

[rancid creek]

You can clearly put the lens anywhere, and then perform tilt and rotate operations until the lens goes to the correct spot.

congrats

[cyan sleet]

just be careful when you tilt to not crush the lens with a rolling rock

[lucid narwhal]

&aoc lb

[shut trellis]

Wrong Channel

You are not allowed to use that command here. Please use the #aoc-bot-commands, #sir-lancebot-playground, #bot-commands channel(s) instead.

[rancid creek]

now with more galaxies:

[ripe meteor]

I'd rather rely on a flimsy implementation detail thanks
I just stuffed them in dicts

[autumn kayak]

Sms forword code

[rancid creek]

dicts are flimsy ?

[rapid wolf]

it's not an impl detail since 3.7 I believe (if I figured out what you mean correctly)

[normal dragon]

Yeah, 3.7 (just looked it up)

[ripe meteor]

Flimsy in the sense most other languages wouldnt/dont support it

[normal dragon]

Aha but we're not in other languages! We're in python. :D

[ripe meteor]

For now

What about language roulette peeps

[rapid wolf]

Go randomises iteration order for maps

[normal dragon]

Well language roulette peeps will just have to suffer.

Or they could do what I did, which was use an array/list.

[fluid sierra]

.

[jovial jay]

Chances are some of those languages has their own way of emulating dictionary or can be compiled

It's a lot harder to emulate dictionary or utilize dictionary with languages that don't directly support something akin to dictionary

[normal dragon]

What's an example of a modern language that doesn't support a hashmap?

[jovial jay]

Can't think of any. You could probably make something akin to hashmap too. Even languages that use variables can be used to hash things. I just had to use variables to emulate dictionary, and looking at the source code, yes there is some hashing involved.

That's why I say that chances are other languages have their own ways.

[normal dragon]

I’m not sure I follow! A python dictionaries, at its core, is a hash map.

[jovial jay]

Yes, that is a hashmap. What I mean is that there are other ways in other languages often. For example, and I'm using a obscure language (G'MIC) as this is the only one I know of that can replicate a different way of hashing: ```
({expr('int(u(0,6,1,0))*12345',123)}) # Create a image with random multiple (0 inclusive - 6 exclusive) of 12345.

repeat w#-1 { # Repeat amount of time corresponding to width of image
v={i[#-1,$>]} # Set v that corresponds to the value of pixel at index $>. $> just means repeat index.

Blah codes here.

H_$v=$> # You see that $v, that correspond with the v variable. Now you can see how this is similar to hashmap. I can access it with ${H_$v}.
}

I use similar technique to solve Advent of Code with this language via hashing like this.

[rapid wolf]

The only hashmap in APL is the one you can get via C# interop, and that's dyalog APL specific.

other than that, you are stuck with arrays

though it does seem like you can use eval and objects

which is definitely sane

[raw wing]

Hello is this the correct channel if I need help with an AOC problem?

[rapid wolf]

#aoc-solution-hints is the better option

[raw wing]

Thank you!

[sleek cave]

yep, didn't miss anything. that puzzle description though ... brought me back to https://adventofrealizingicantread.com

oh shoot that's not the domain

lol i can never remember the url of the joke site

something along those lines, that redirects to aoc

[minor cave]

for some reason it's an http redirect http://www.adventofrealizingicantread.com/

[ripe meteor]

I dont understand this joke

[minor cave]

the redirect? Just a small joke that advent-of-code is also the advent-of-realizing-i-cant-read because so many people miss over some aspect of the puzzle text/instructions

[ripe meteor]

Oh, i didnt think day5 was that difficult to understand tbh

[sleek cave]

but then again today was pretty in your face. i mean, ||HASHMAP|| ... seriously? :P

(this is why i love aoc lol)

[jovial jay]

i think i'm doing aoc 9, now i actually understood it

[rancid creek]

day9 part 1 was nice. day 9 part 2 was || 1 trap||

[silver karma]

have previous years been out of order like this?

[modern terrace]

AoC 2020

[silver karma]

ah yeah

why does Eric do that?

[normal dragon]

It's a picture

[earnest forge]

GLHF everyone

[normal dragon]

Slash story

GOOD LUCK!!!

[silver karma]

glgl

[silver karma]

you could still make the picture with the numbers in order tho

[modern terrace]

that would need a mirror i think

[shut trellis]

<@&518565788744024082> Good morning! Day 16 is ready to be attempted. View it online now at https://adventofcode.com/2023/day/16. Good luck!

[covert arrow]

I'm outside rn 😔

P1 looks pretty simple though

[minor cave]

Language Roulette: Day 16
The language is ... Kotlin

[magic oxide]

definitely don't have the time to do this rn

hopefully not too bad

I have stuff I'd like to do tonight

[spice hemlock]

https://tenor.com/view/chungus-pinged-ben-shapiro-discord-big-gif-21424212

My birthday

[flat sequoia]

Done 3/4

[magic oxide]

ooh today looks fun

||complex multiplication should come in very handy||

[earnest forge]

90/59

I screwed myself by ||forgetting to check for loops||

[earnest forge]

Maybe I'm dumb but I don't see how that would help?

LB full!

[minor cave]

#1185450487009525850

you can blame my cat for the delay, he demanded attention

[earnest forge]

I'll forgive you, if you can show us the culprit 🙃

(cat tax)

[silver karma]

Wow wp

[flat sequoia]

oh yeah that'd be nice

I didn't even think of it

[pine tiger]

seriously what's with me this year

i'm so washed 😭

[flat sequoia]

😧

[minor cave]

A picture of the culprit from a bit ago

[earnest forge]

All is forgiven

He does look very worthy of attention

[flat sequoia]

omggg

[minor cave]

Oh wait, better picture from today

[modern terrace]

how do what

mine works for the test input but fails for the real thing

[earnest forge]

Wow I didn't realise there even were edge cases like that today lol

Is it maybe that ||you're not handling the case where you spawn on a splitter properly||?

[modern terrace]

yes

exactly

that

thanks

[magic oxide]

||not sure if this actually works, but I think mirrors can be represented as a multiplication by 1j/-1j and a conjugate||

||splitters are a real/imag check followed by a multiplication by 1j/-1j||

[earnest forge]

I can't visualise ||a conjugate|| in my head but ||multiplying by i is a rotation||

[magic oxide]

||yeah, your facing vector is one of (-1, 1j, 1, -1j), and you can change the orientation via multiplication to rotate it||

||that's how I typically implement rotation for these types of problems||

||haven't actually done the problem yet, so idk if what I'm thinking of actually works||

||we'll see||

[flat sequoia]

Yeah I think that works

Pretty cool

[modern terrace]

now why's my part 2 not working

[flat sequoia]

######....
.#...#....
.#...#####
.#...##...
.#...##...
.#...##...
.#..####..
########..
.#######..
.#...#.#..

Why is the pic in today's problem kinda sus

[modern terrace]

got it

single letter variable names are not advisable

[pine tiger]

are you talking about amoguses

bc i can't really see it

[flat sequoia]

Maybe I just have brain rot

[rancid creek]

ඞ

[flat sequoia]

today's chat is so inactive 😧

[mossy basin]

it'll continue to drop off probably

[hollow wharf]

Am i not seeing Day 16 spoilers discussion?

[covert arrow]

#1185450487009525850

[hollow wharf]

Discord on mobile so bad i cant go there

[magic oxide]

I kinda see it too

[magic oxide]

a lot of people have dropped off at this point i think

and some people haven't caught up with previous days yet

[flat sequoia]

dang

[magic oxide]

I haven't done today yet since I've been out all day

[trim rain]

😿 i've caught up to day 10 finally doing day 11 rn on my way to catch up, i cannot help but think so far the only challenging question has been day 5 is it just me?

i mean the general difficulty at start was higher than usual but i'm talking more about needing to do find clever solutions instead of just bruteforcing

[magic oxide]

day 8 was challenging until you found the assumption

[trim rain]

yeah if you haven't done alot of graph problems i can see it being tricky but it was still way easier than day 5

it may just be me being stupid tho as i did spend like half a day trying to do day 5 meanwhile no other problem so far has taken more than 30 mins so...

[mossy basin]

was trying to remember what day 5 was

[trim rain]

it was the thing with reallly big ranges

[mossy basin]

that one is easy if you already have an interval library

[magic oxide]

day 5 was on the harder side but I really enjoyed it

[mossy basin]

!pypi mind_the_gaps

[marsh current]

mind-the-gaps v0.3.0

A library for unions, intersections, subtractions, and xors of intervals (gaps).

[trim rain]

i am doing every day in a different language and without any imports (except for needed ones like io.File in java/clojure) so that's kinda hard

[mossy basin]

rip

[magic oxide]

but yeah every other day hasn't been too bad imo

[trim rain]

yeah honestly day 9 was easier than day 1

idk what's going on with the difficulties

[mossy basin]

i think day 12 was the hardest

[magic oxide]

ah right

I didn't find day 12 too hard, but yeah I'd say it was hardest so far

[mossy basin]

but i'm sure we're gonna have some hard ones coming up

[hollow wharf]

dam

<@&831776746206265384>

[hazy ginkgo]

!pban 1182906123657957480 scams

[marsh current]

:incoming_envelope: :ok_hand: applied ban to @stray temple permanently.

[lime fulcrum]

today sounds fun. interesting problem

might have to solve later. it sounds like it would take a while

[glad tartan]

day 1 solution in spasm https://raw.githubusercontent.com/P403n1x87/aoc/main/2023/1/code.pya

targets 3.11

[dense osprey]

shouldnt you spoiler this

[zenith elbow]

I like how I stopped doing it daily just when it went from hard to medium

[lime fulcrum]

how is it a spoiler to say if its hard or not? does it help you solve it faster?

[modern terrace]

how is 15 medium? got it

ok i'm not reading my times correct

hold on

okay

justifiable

[dense osprey]

its a spoiler the same way it is a spoiler when i say what type of problem problem x is

knowing that a problem is easy, makes it obvious that "advanced" algorithms wont be the solution

making the coming up with a solution easier

imagine seeing a leetcode problem marked as easy

would you even consider using dp?

or does the difficulty tell you that the solution probably is some easy algo or just bruteforce

[lime fulcrum]

given that I don't know a lot of these complex algorithms it only tells me can I solve this before I am supposed to go to bed or am I going without sleep

[dense osprey]

which is a spoiler

[lime fulcrum]

fair. but rather small one

[dense osprey]

yea ig

leaderboard times are public thus not a big one ig

[hidden musk]

that's clearly different though. "difficulty" is totally subjective, while saying the type of problem actually gives away information about the problem

[dense osprey]

the difficulty chart being subjective doesnt make it not a spoiler

knowing that a lot of people are struggling with a specific problem gives away the information that the problem wont have an "easy" solution

[hidden musk]

¯_(ツ)_/¯

[lime fulcrum]

It might if you think enough about it

[full zinc]

All problems are easy if you're good enough B)

[dense osprey]

i dont care if ppl spoiler aoc

[woven sable]

yall think it's plausible to catch up if i'm 4 days behind?

or nah

like are days 13, 14, 15, and 16 really hard or nah

[hidden musk]

16 was p easy

[ripe meteor]

Yes

I think day12 was the last hard one

[woven sable]

ima try and push through then

[ripe meteor]

Just be mindful of the little things
Off by one errors and stuff

Conceptually theyre easy but i spent a lot more time than i should hunting down these things

[rancid creek]

yeah should be doable! you got this.

[lime fulcrum]

today is kind of getting long 😅

[hollow wharf]

yo this is my first year doing aoc and today i found out i can still do past years challenges and get stars for it too

it seems i got things to do

[lime fulcrum]

yeah its quite fun but i got enough to do with today XD

[hollow wharf]

xD

[rancid creek]

save the past years for January!

[hollow wharf]

good idea

[rancid creek]

Today's not bad at all. Now tonight's new puzzle might be challenging. 🙂

[lime fulcrum]

i'm just currently doing a lot of if statements. this is going to be very verbose

[rancid creek]

That's what most of coding is. ifs assignments and loops in some variation. 🙂

[hollow wharf]

since there is actually somebody here i will ask what day you find the hardest so far?

[lime fulcrum]

12. haven't done that one yet. 10 was also tricky but in a fun way

[hollow wharf]

yep 12 for me too 😵‍💫

[rancid creek]

Yup 10 and 12 a little harder, but not impossible. Ask for hints or at least hints on what to google, don't waste time spinning when you could be learning. 🙂

[hollow wharf]

thats exactly what im doing

[lime fulcrum]

||doing a check for every possible symbol and then nested if for every direction of movement. check if coordinates are valid and not already a checked state then chuck on tbd pile. probably works. we'll see||

[hollow wharf]

thats p1 i belive

[rancid creek]

Part 12 wasn't ||a lot of code||, but it took ||some time figuring them out||.

[lime fulcrum]

||p2 is a bruteforce?||

[hollow wharf]

thats what i did at least

[rancid creek]

that would be telling. 🙂

[lime fulcrum]

so yes. thank you

[hollow wharf]

but are you on p2

[lime fulcrum]

not yet. just means my p1 can't be too slow

[hollow wharf]

ah if you wanna know p2 is just ||repeating p1 440 times|| 🤫

[lime fulcrum]

yeah that I've stumbled upon

[rancid creek]

I think the first thing I tried for part1 was 15 seconds but worked, and then I optimized it to be instant.

[hollow wharf]

not me ||taking code from day 10 and rewriting a bit||

[lime fulcrum]

||just not sure if its something one does: when there are splits in a path: throw everything in a list with moves yet to be checked and repeat until list is empty.||

[muted aurora]

assignments and loops

Haskell wants to have a chat

[rancid creek]

I mean you can disguise it. But you can do a lot with math, mov and jnz. 🙂

[muted aurora]

Who needs jnz

https://github.com/xoreaxeaxeax/movfuscator

[rancid creek]

Jeeze! /insert scientists not asking why meme 🙂

[agile silo]

im trying to do day8 part 2

im using lcm, but i get a pretty big output

I get too many steps the ghost should travel

[rancid creek]

good.
part1 ||19XXX||
part2 ||13663968XXXXXX ||

[lime fulcrum]

that seems right

[agile silo]

74598175385569379060059

i get

lol

[rancid creek]

That does look too high

[agile silo]

it is

[rancid creek]

||Did you write your own lcm, or use the one in math ?||

[agile silo]

i just multiply all individual steps // gcd(steps)

[rancid creek]

||ok, try using math.lcm and see if you get a different answer||

all of this is too much for this channel.

we should move to hints or the the part 8 channel.

[agile silo]

it is now correct

[rancid creek]

Hurray!

[agile silo]

maybe the overflowed

and that made it not work

i didnt know there was a lcm in math

[rancid creek]

it's only been there a few years, it's a new addition.

v 3.9

[normal dragon]

Love all the batteries in python's standard library

[simple raven]

I would say that they were getting progressively easier since 12

[woven sable]

actually?

ima take a shot at it then

[lusty rover]

Always take the shot

The worst that can happen is you get stuck

[low condor]

you miss 100% of the shots you don't take

[woven sable]

only worry is how much time i have

i've been super low on time this year versus past years

[low condor]

Wayne Gretsky
Michael Scott

[woven sable]

the one thing throwing me off about day 13 is j like

the fact that not all columns will have a reflection

but i'll find a way past that

[rancid creek]

attribution: Walter Gretzky (Wayne's dad)
recorded in interview in 1983 by Wayne recounting the advice.
(had to google the details, but knew it was from a while back)

[rancid creek]

They'll have a reflected row OR column. I'm not sure any have both, and none have neither.

To find the reflection in each pattern, you need to find a perfect reflection across either a horizontal line between two rows or across a vertical line between two columns.

[pine tiger]

lmao i'm tryna to do 2019 again

https://adventofcode.com/2019/day/16
how the hell is this actual fft again

https://people.eecs.berkeley.edu/~vazirani/algorithms/chap2.pdf
i mean my source is here

[jovial jay]

@rancid creek You weren't kidding, part 2 is a trap for Day 9

well, one thing to do is to debug of course

[muted aurora]

oof, good thing I've got an fft function in the standard library for my language LOL

[jovial jay]

same. but, i'm using G'MIC, which has built-in FFT command to process images

not sure if it'll be useful there though, but we'll see

[muted aurora]

probably not, I'm just justifying my decisions to throw everything-but-the-kitchen-sink into my stdlib

[jovial jay]

yeah, i tested it, it's not useful for this case

[rancid creek]

I didn't say a trap, I said ||1 trap||

super spoiler, my solution was posted in the spoiler channel: ||#1182911467717144616 message||

[pine tiger]

i mean the real question for me is how do i link this to fft

i got the matrix multiplication, but it just isn't working out

[rancid creek]

Let me know when you finish which way you did it, there were multiple options. Some cooler than others. I did the boring one. =/

So any predictions? I'm guessing tonight might be a harder one.

n dimensional maze pathfinding while avoiding moving objects.

[pine tiger]

god please no

[rancid creek]

part 1 is pac man, part 2 is pac man in 4d

[rancid creek]

https://adventofcode.com/2019/day/18 comes to mind.

[pine tiger]

bro seriously what was eric on when making 2019

16, that day, 22...

[rancid creek]

the back 1/2 of 2019 was no jokes.

[pine tiger]

wehred u get this list from

[rancid creek]

made it.

[pine tiger]

oh lol

[rancid creek]

using data from leaderboards

[pine tiger]

also why is day 16 put as easy it's literally the one i'm malding on rn

[jovial jay]

i don't wanna know what insane looks like

do i?

[rancid creek]

it's based on when the top 100 finished part 2.

[torn thorn]

totally forgot this was happening this year damn

[jovial jay]

holy 80 mins for top 100 and they're quick guys

[rancid creek]

So it's a relative scale. If it takes you an hour for an easy, you can expect the insanes to be several hours.

[sleek cave]

can we see the full chart? is it on your github

[rancid creek]

No, it's a WIP, but one sec, someone ran with the idea and made a cool web site.

https://aoc.xhyrom.dev/ similar concept

The only reason I made it was to track whether weekends really are harder. They're not, but the last couple of weeks do tend to be harder, regardless of day.

One really cool thing xhyrom did was automate it so it's always up to date. I was updating by hand. lol.

[rancid creek]

2019 day 18 and 22: 2h+ They were fun puzzels. But iirc I was traveling at the time, and was hoping for a 30 minute puzzle. Oops.

[pine tiger]

like i really don't like the ones that require examination of the input

[woven sable]

i don't like the ones where u have to modify the input

[low condor]

How do you mean, modify?

[woven sable]

like i remember there was one in 2021 that was a path finding problem that required expanding the map 5 times in each direction

i mean it wasn't that hard it was j annoying lowkey

[lusty rover]

Like how you can't keep your name for more than 24 hours

Gotta change it!!

[woven sable]

and didn't really require much thinking

[low condor]

Gotta switch things up!

[low condor]

Gotcha. The point of those is (usually) to catch out people who’ve not coded the solutions in the best way initially

If you code the best in P1 you are sometimes rewarded in P2

[low condor]

When people actually start calling me “Advent of Verbosity” then it’s time to change lol

[lusty rover]

What's wrong with being Mr. Verbosity?

[woven sable]

yeahh i feel like they're always "make the input super massive so a naive solution won't work"

[jovial jay]

i did the boring way meaning calculating the difference and storing the first/last value only

[low condor]

When you put it that way, nothing wrong at all!

[low condor]

Basically the only way to do it since Eric doesn’t want to run any solution code on the website

[woven sable]

yeah that's fair

i can't imagine how fast his solutions must be to every problem

[lusty rover]

Look at it this way: https://i.imgflip.com/89me0z.jpg

[woven sable]

he must be crazy good at competitive programming

to run thousands of peoples' solutions

[rancid creek]

That's the cool way.

[low condor]

😂

Perfection 👌

[pine tiger]

speaking of day 18, this is gonna sound really weird but i remember you helping me with that day in a reddit thread for a long long time

[low condor]

This was my nightmare day

Needed lots of help just to even solve P1

P2 as yet unsolved

That day and Conway cubes basically are my most hated days

[pine tiger]

what was conway cubes

[rancid creek]

Trying to remember part 2, but I seem to recall some sort of modified bfs where I passed along game state. I'd have to go back and look again.

[low condor]

2020 day 17

Basically could not do that one without looking at a solution

I currently have a solution that is incredibly nicely written by myself to which I have no clue how it works

3D geometry combined with huge ranges is 100% the worst for me

[rancid creek]

My solution to that one included these 3 lines:
||

vecplus   = lambda o,d: tuple(a+b for a,b in zip(o,d))
neighbors = lambda n,l: sum(n.get(vecplus(l,d),0) for d in deltas )
survive   = lambda p,l,n: int(n in (2,3)) if p.get(l,0) else int(n==3)

||

I was more cryptic that year I think. =/

[low condor]

Yeah I have no idea what that code means 😂

Let me post my written code

Actually it’s not Conway Cubes, what am I saying

That one was fine

[pine tiger]

that's not the

switching cubes on and off one, is it?

[low condor]

Yeah it’s another one

[pine tiger]

shudders

[low condor]

That one

Reactor Reboot

2021 day 22

|| https://github.com/Fiddle-N/advent-of-code-2021/blob/develop/day_22/process.py||

If someone can decipher my solution that would be very much appreciated

I wrote comments in order to gaslight myself into believing I understand how this works

[sleek cave]

lmaooo following this thread i was about to say, wasn't fiddle's nightmare reactor reboot?

then i saw your most recent message 😂

[rancid creek]

One of the ones that was harder for me was one I made leaderboard on -- by solving it by hand, and then took forever trying to figure out how to translate what I did by hand into code.

[sleek cave]

this looks absolutely insane

pathfinding with keys and locks - wtf

[jovial jay]

Day 11 map looks a lot like poisson noise

[magic oxide]

did you use negative cuboids?

looks like you did

for explanation, it's far easier imo to start with considering 1d intervals

the inclusion-exclusion principle describes why the method works

when you calculate all the intersections when adding a cube, you're essentially bumping up from inclusion exclusion on n sets (our intervals in this case) to n+1 sets, which you can do by (stuff including n intervals) + (new interval) - (stuff including n intervals intersectioned with the new interval)

makes a lot more sense seeing it that way imo

[earnest forge]

GLHF today everyone

[minor cave]

Language Roulette: Day 17
The language is ... Lua

[mossy basin]

the one non-python language i could do

but won't

[hexed axle]

I can write code in any language--I can't right functional code though, and the roulette is just reminding me of all the ways I can't 😅

[shut trellis]

<@&518565788744024082> Good morning! Day 17 is ready to be attempted. View it online now at https://adventofcode.com/2023/day/17. Good luck!

[shy trench]

finally the path finding day arrives

[tiny tendon]

Fr lmfao

[hexed axle]

nice, congrats to the looking for internship guy, that's fun to see

[jovial jay]

that looks painful to do

[covert arrow]

Looks like ||Dijkstra||

[hexed axle]

||I haven't dug in, but I was thinking something similar, drawing from the Wagner–Fischer algorithm family||

[covert arrow]

I forgot to bring my laptop though, so all I can do is sit and watch

[devout turtle]

Who (or what) is Dijkstra in this context? I believe I’ve seen that name coming by somewhere this month already, but since a it’s a fairly common last name in Dutch, I guess googling for it won’t bring much related results.

[hexed axle]

Pen and paper are a great tool

[covert arrow]

It's an algorithm: ||https://en.m.wikipedia.org/wiki/Dijkstra's_algorithm||

[devout turtle]

Thank you.

[pine tiger]

ok done done

[pine tiger]

😭

[earnest forge]

My pathfinding isn't working why am I stupid

[pine tiger]

yk i have no idea if this is relevant to you

but ||iterating through all directions not the current one|| doesn't work

[hexed axle]

20 on the dot today, that's nifty

[pine tiger]

i had to like ||make a dict for each direction to each other direction|| for crap to work

[minor cave]

#1185813850973540442

[mellow girder]

python language roulette?

[sleek cave]

oooo, pathfinding

time to visit my old friend ||dijks||

forgot how his ||algo|| works lmao

been some time

[magic oxide]

finally

I missed the start because I was eating

[pine tiger]

unfathomably based

[earnest forge]

Ahh I figured out what I did lmao

||I wasn't caching my x and y runs in my visited set||

606/460

[zinc kindle]

Wow. Managed top 1000 for both problems, and took not much time for part2. I am very grateful to the author(s) of the ||pathfinding|| module for making part of the problem a lot easier than it would have been had I implemented the function itself.

[pine tiger]

what's that module?

[zinc kindle]

Sorry, writing in rust, not python. If you're interested, link: ||https://docs.rs/pathfinding/latest/pathfinding/||

[pine tiger]

man, rust

[zinc kindle]

I prefer python (easier input parsing, batteries included, can write recursive inner functions), but I'm trying to learn rust, and that crate is a godsend in AoC.

[pine tiger]

haha gl

i tried rust before and it was so bad i wrote a blog post about it

(golang, which i'm learning now, is worse though)

[zinc kindle]

Could you link the blog post? I'm curious. Maybe in an off-topic channel though.

[pine tiger]

https://sanspapyrus683.github.io/content/other/rustrant/rustrant.html
this was like a year ago

[earnest forge]

You realise that's because char isn't a numeric type, right? Overflows do give RTEs, as well

And this is because a) strings are hard and b) other languages that do string handling properly are also linear-time lookup, they just shove it in the indexing operator and you don't notice because it really isn't a huge deal

(also &String is generally not a useful type, not sure why you mentioned it as well as String)

The one in parse isn't necessary, the one on the left (or alternatively as a turbofish of collect) is necessary to declare the type of container you want, it could also be a HashSet, for example

...they can't? All array indexing is usize

[wet yoke]

Right, a collect could produce many types of collections. You need to tell it which type you're intending to build in some way.

[earnest forge]

This is the good ternary operator btw - since blocks are expressions, why have two ways of writing the same thing
Btw what you want really is <integer_type>::from(ring % 2 == 0) and there's a clippy lint for that

This one does trip me up but there is a good technical reason for it: Inner functions can be monomorphised separately this way (or not at all if they're not generic), which leads to faster codegen and encourages the use of inner functions over private external functions

[pine tiger]

lmao every time i post this

people rag on me for it

[earnest forge]

because your takes are bad /j

[rancid creek]

ironic comment given today's puzzle.

[rancid creek]

hint ||if you know your last dx,dy, then because one of them is guaranteed to be 0, you can get the reverse just by -dx,-dy.||

[rancid creek]

||dijkstra famously said that the reason his algorithm was so simple was he had no pen and paper handy and had to do it in his head||

[pine ravine]

what was the hardest day this year so far?

the almanac one with range splitting?

[covert arrow]

I'd say 12 (hot springs matching)

[hollow wharf]

same

[dense osprey]

day 12 was easier than day 14 imo

[dense osprey]

calc?

[lime fulcrum]

guess i'll be learning path finding then. kind of excited

[pseudo cape]

help me with this code guys so that it could pass the bellow case tests

[covert arrow]

1. please send your code as text (use a pastebin if it's too long) so it's easier to read
2. you're in the wrong channel, see #❓｜how-to-get-help (and don't expect people to spoonfeed the answer either)

[thin blaze]

Today was hard for me... T_T

[scarlet zephyr]

hello everyone, put an asterisk on the repository if it's not difficult for you https://github.com/Ttunay/game
I will be very grateful!!

[ripe meteor]

This is the wrong channel, use offtopic channels

[jade atlas]

What are some other algorithms I should learn for the next few days. I never used dijkastra or a*, I learnt they will be usefull so I studied them a little before today. What are some other well known algorithms that occur frequently in the last few days?

[rapid wolf]

I would look into the chinese remainder theorem

[jade atlas]

so more counting problems?

[rapid wolf]

It's generally more related to N things happening in a cycle and starting to cycle at some step. Find some detail about that

[rocky spire]

It's a theorem that relates a bunch of equations (congruences) like this

x % m1 = a1
x % m2 = a2
x % m3 = a3
...
```To this "merged" equation:

x % (m1m2m3*...) = a
```And along with it you'll usually learn how to calculate a given the x % mk = aks

[jade atlas]

I understand what CRT is. Just wanted to understand more about the kind of problems

[rocky spire]

Oh, nvm then

[hollow wharf]

Hey everybody, can someone tell me where I can learn Python and do practice lessons for free or cheap?

[royal gull]

Ask over in #python-discussion , this channel is for https://adventofcode.com/

[rancid creek]

no calc, ||just throw a minus sign in||
||```py
a,b = -1,0
c,d = -a,-b
c,d == 1,0

you can get the reverse of any direction this way.

[rancid creek]

You could review the problems from last week of previous years, just to get an idea of what is being asked

[rancid creek]

when you're ready, you can look at the solution I posted #1185813850973540442 message to see where I used it to ||remove forward and backwrad moves||

[dense osprey]

havent looked at the problem, jsut saw the dx and dy

[rancid creek]

funny thing is, you remove the braces and ; and it starts to look like python. 🙂

[lusty rover]

Someone introduce this guy to Python

[dense osprey]

this is still better than people who do following:

void func() 
{
  // code
}

writing code like this should affect your credit score

[rancid creek]

agreed 2 space indents are cursed

[earnest forge]

1TBS, 4-space indent is the only correct way to do braced languages

[dense osprey]

oh that was the discord indent

i meant the placement of the {}

i hate seeing it

[dense osprey]

yes

[dense osprey]

idk why not every app has it by default

pressing tab in text editor gives some weird indent

and word too i think

[earnest forge]

That would be the tab character, which is 'move the cursor up to the next multiple of 8 spaces'
Useful for tabular data, not an indent function

[cyan sleet]

:V

[cyan sleet]

believe me, there are worse ones

I bring to you: GNU style

[jovial jay]

Personally, I use 4 spaces/tabs on Python. On other languages: I use 4 spaces/tabs on C++.

[jovial jay]

that's terrible. they should be aligned so that that scope is easier to read

[cyan sleet]

the function calls are also particularly terrible

[earnest forge]

They are aligned if you notice that it's actually 2-space indentation

[jovial jay]

you know that's not what i mean

[normal dragon]

Tab is configurable width, though it’s typically 8 chars

[jovial jay]

imagine deeply nested 8 chars tab if else statements

[earnest forge]

(no but actually it is 2-indent, look at the unbraced if)

[earnest forge]

Yeah all right s/8 spaces/the tab stop/

[normal dragon]

Tangent - Did you know discord actually supports s/search/replace

yo sup

[earnest forge]

Yes I did lol

[cyan sleet]

in a very limited fashion

e.g. this doesn't do anything sensible s/(very)/\1 \1/

and e.g. s/foo/bar/g

bar/g test foo

(starting from foo test foo)

[rancid creek]

any semi colon using language should require the semicolon at the start of the line. 🙂

[cyan sleet]

I mean...you could write it like that

[earnest forge]

pulls out cartoonishly large mallet

[rancid creek]

does it support tr? let's try . qqq

[woven sable]

for day 13 or wtv the reflection one was

[rancid creek]

tr/q/:/

nope

[woven sable]

i think my programs getting confused by head fakes

like it thinks there's a vertical reflection at the end

but in reality there's a horizontal reflection

problem wise, is it better to j use the reflection w more points reflected ig

[earnest forge]

Make sure to check that the reflection isn't empty

Or if this is part 2, ||make sure to check that there actually was a smudge in that reflected region||

[rancid creek]

does it work on the example data?

[woven sable]

nvm i figured it out but part 2 kinda confuses me w the instructions

||in the second pattern, why can't the reflection still be on the 3rd/4th row?||

[earnest forge]

Because ||there is a smudge in the mirror causing a wrong reflection, if the new reflection doesn't contain the smudge it isn't new||

[woven sable]

ohh i see

[ripe meteor]

we escape the filter once more

[woven sable]

i finished day 13 🙏

4 more days to go

[rancid creek]

Nice work! Keep it up!

[broken dock]

i know i'm a bit late, but to add on to this, ||if you defined your directions as an array and going in clockwise or counterclockwise order, and your direction is some integer 0 thru 3, you can add two and take its value modulo 4 and that will also give you the "behind" (this is how I did it just because I was too lazy to multiply the translation vector by -1)||

[woven sable]

is day 14 part 2 ||going to be too slow by j doing a naive solution?||

[rancid creek]

My whole reason for doing this was so I didn't need ||an array or dictionary of positions.|| 🙂
Another thought is, if you use complex numbers than ||getting the rotations is easy as well.|| 🙂

[broken dock]

true

just multiply by i

[covert arrow]

yes, ||look for cycles||

[rancid creek]

Slow yes. It should finish in about 27 hours though, so technically possible.

[broken dock]

@woven sable assuming that the state of the grid after every single cycle was unique, the only valid solution would be to simulate all 10^9 cycles (and this would take a while for anyone)

so my hint is: ||assume that there aren't that many unique states, and that duplicates begin to occur after some number of cycles||

[rancid creek]

That comment reasonably applies to multiple puzzles

[broken dock]

and yet it's something that's very commonly missed

feel like shayaan's question ab part 2 of 14 makes that evident

i would like to note that I was also in the same boat

still can't get used to aoc

i have a habit of operating under the assumption that i'm going to be tested against multiple inputs (most hidden)

[rancid creek]

Another med/heard one last night. Surely tonights will actually be a difficult one then? I keep getting ready. 🙂

At least I was right that last night ended up being maze finding.

[muted aurora]

lets go back under 0.1s

[drowsy oak]

for problem 17, they don't give the size of the 2d array?

or do we assume that the max size is the example array they gave?

[muted aurora]

Yeah they usually don't give you a max size of the 2d array

the full input size is the same for everyone I believe, so I end up just looking at that

[drowsy oak]

what is the max possible operations?

before we time out

I found a brute-forcey dynamic programming sol with 3084588 operations

[muted aurora]

there's no time out

you're not submitting the code, only the answer

as long as you're ok waiting for it lol

[drowsy oak]

w-what

LMFAO

[drowsy oak]

how long does 10^7 operations generally take?

like how many operations before I wait a day

[muted aurora]

that's what 10 million? shouldn't be that long. Depends on your definition of "operation" though

let's not discuss solutions here

[drowsy oak]

like idk if i'm being dumb or not

[pine tiger]

what language is this

[muted aurora]

my own lang, it compiles to C

[pine tiger]

damn

[jovial jay]

I wish I can make my own lang, but that's too much time and effort

[normal dragon]

If I had $1500, I'd take Dave Beasley's Write A Compiler course https://www.dabeaz.com/compiler.html

[jovial jay]

In my case, I want my own interpretative language with JIT support

But, for now, I can get by current options

[normal dragon]

https://craftinginterpreters.com/

[hexed axle]

Language Roulette: Day 18
The language is ... Ruby

[covert arrow]

GLHF

[earnest forge]

GLHF

[shut trellis]

<@&518565788744024082> Good morning! Day 18 is ready to be attempted. View it online now at https://adventofcode.com/2023/day/18. Good luck!

[magic oxide]

oof

haven't even had time to do yesterday's yet

[covert arrow]

pycharm is not working for some reason

another ||optimization|| puzzle today

I don't even know where to start with this one

[rocky spire]

Part 1 doesn't seem to be using the colors and I'm scared

[devout tusk]

Is P1 just regular|| floodfill||?

[earnest forge]

Yeah that's what I did

No idea what the optimisation for p2 is

[zenith elbow]

what's p2 about?

[covert arrow]

||same thing with really big numbers|| basically

[somber crow]

i just leaderboarded for the first time :)

[rancid creek]

well done.

[covert arrow]

how do you do p2

[somber crow]

i used ||shapely||

[rancid creek]

||think about past puzzles||

[covert arrow]

||d10?||

[rancid creek]

omg. my code stalled on the 623rd line...

[rancid creek]

You achieved rank 80 on this star's leaderboard and gained 21 points!

heart stopping moment there when it paused on the last 2 lines.

[zenith elbow]

leaderboard is full

[hexed axle]

https://discord.com/channels/267624335836053506/1186176336734203954

20:55 today, we're late in the month and still getting leaderboards filled up early on

[flat sequoia]

I had it so fast man

but then I spent 8 minutes debugging correct code

LMAOOO

sobbing

[pine tiger]

wait wtf

what's wrong with my ||shoelace||

[rapid wolf]

spoilers ty

[flat sequoia]

probably ||area ≠ lattice points||

[pine tiger]

ok done

not sure why my answer is off by one tbh

[flat sequoia]

😧

[pine tiger]

ok wait i think i realized

honestly kinda elegant tbh

[earnest forge]

I think I know why that happens

[pine tiger]

smth smth ||4 corners||

[earnest forge]

Yeah

[pine tiger]

really, as soon as problems require some amount of neuron activation

i crumble

[earnest forge]

If you imagine ||the coordinates to be the centre of each cell||, when you ||halve the walls|| you're effectively ||undercounting ¼ of each corner||

Wait no that would be ||overcounting|| why is the OBOE in the other direction

[flat sequoia]

oh noooo

[pine tiger]

OBOE?

[earnest forge]

Off By One Error

[pine tiger]

no i'm pretty sure that's ||undercounting||

i took the ||shoelace area and added the distance traversed / 2||

[earnest forge]

I'm figuring it out in a text editor lol

[zenith elbow]

okay but what's with the extra off by one?

[earnest forge]

Oh wait I have a better way to understand it
||if you treat the coordinates as the top left of each cell, the shoelace gives you the internal area + half the walls - 1 cell (the top right and bottom left corners are both omitted)||
||so when you add half the walls rounded down you need an additional cell to make up for the corner that wasn't counted||

[pine tiger]

||corners exist||
a simple input that makes the square should make it obvious

[devout tusk]

Why are the colors even provided? Is p2 horrendous?

[pine tiger]

they're not colors.

[earnest forge]

||p2 tells you that the colours and instructions are the wrong way round||

[devout tusk]

Is p2 basically using the hex value of colors as lengths?

[pine tiger]

thousand yard stare

[devout tusk]

Lol, I was right then lmao

||shapely|| ftw, i guess

[silver karma]

it depends what the operation is, adding 1 to an integer 10^7 times will take a lot less time than sorting a list 10^7 times, for example

[drowsy oak]

idk, but for competitive coding 10^8 is generally how many operations you can do

[tall wasp]

Can I see your solution? I tried that but got wrong answer so ended up just implementing it myself

[modern terrace]

if i had remembered how ||pick's|| and ||shoelace|| worked maybe i could've been in the leaderboard again 😔

[surreal sapphire]

wat is advent of code

[thin blaze]

Day 18 was trivial comparatively to yesterday. x)

[rapid wolf]

meh, ||knowledge check like knowledge check, not much sense in talking difficulty IMO||

[simple raven]

a coding Christmas challenge where every day you are given a coding problem to solve https://adventofcode.com/2023

[jade atlas]

I feel ashamed for forgetting about ||pick's theorem|| from ||day 10||

[somber crow]

||https://github.com/TheSuperGamer20578/Advent-of-Code/blob/master/2023%2F18.py|| you need to ||add half the Trench plus one||

[tall wasp]

ahh that makes sense, thanks! ||I did end up doing that for my eventual solution without shapely but for some reason didn't think about doing it there 🙃 i only tried adding the whole path and none of the path||

[silver karma]

Is it just me or is the puzzles this year much easier than they were in 2022? I had to stop around ~11 last year as it was just beyond my comprehension but this year I've stumbled my way to 18, with a couple of the later days such as 15 and 16 being quite easy to do on my own

[modern terrace]

yeah they are

or there was just more knowledge accumulated throughout the year

[simple raven]

I must say that my decision to seek help this year instead of just trying to solve everything on my own has made my AoC experience waaay better. I still managed to solve half of the problems without any help at all which makes my non-CS heart happy. I do feel like the later days seem quite manageable, especially after seeing all sorts of approaches to problems from previous days. To me that is a sign of a well-made set of puzzles where earlier solutions help you with the rest.

[simple raven]

||first we found the bits inside a 2D loop (day 10), then we traced a ray across mirrors where some mirrors has special behaviour based on direction of travel (day 16), then we had to solve day 10 in 3D with a much larger step size, except all motion only happened in a 2D plane (day 18, we were removing cubic metres). Are we about to do these steps in 3D? O.o||

[vale fossil]

the instrument oboe

||stop using spoilers pls i am going to jump offa brig 💀||

[misty wolf]

|||hi|||

[simple raven]

just two |

[misty wolf]

||||||oh yeah ||||||

[simple raven]

||lol||

[sleek cave]

i procrastinated learning ||shoelace and pick|| on day 10, looks like it's time to do it now :P

i have no idea what those two magic words entail

but it's time to Find Out ™️

[ripe meteor]

You dont need those two if youre giga brained
Apparently a friend ||computed all the corners and fed em to autocad to build the polygon shape and get its area||

[simple raven]

you can do the same in python

[rocky spire]

Or yknow, ||sweepline||
I honestly find it surprising I still haven't seen anyone solve today with it

[ripe meteor]

Well yes but you need to write code for that

[simple raven]

there is a library that lets you do exactly the same thing autocad does

[ripe meteor]

Which one

[simple raven]

there are many geographical libraries in python and they use it as a baseline for calculations

[simple raven]

||shapely||

[ripe meteor]

Oh ok i've heard of that one

[pine tiger]

wonder how far i'm going to fall these days

[magic oxide]

my day 17 works for the test cases but not for my input

ok changing a number worked

idk why

but it works

[gloomy kernel]

ok fuck fuck fuck

i missed 2 days

[cyan sleet]

no, patrick, off by one error is not an instrumemt

[feral hazel]

considering adding a bunch of imports to my template
it makes a pretty even staircase lol```py
from collections import Counter, defaultdict, deque
from contextlib import suppress
from functools import cache, partial
from itertools import batched, chain, count, combinations, groupby, islice, pairwise, permutations, product, starmap, tee
from operator import eq, gt, lt, ge, le, add, sub, mul
from pathlib import Path
from bisect import bisect, bisect_left, insort, insort_left
from heapq import heappop, heappush
from math import ceil, dist, floor, gcd, inf, lcm, prod, sqrt
from sys import setrecursionlimit
from re import findall, finditer, search

[ripe meteor]

What template is this for

[feral hazel]

advent of code
it's the file i start with on every day
currently it looks like this
https://github.com/cefqrn/advent-of-code/blob/main/2023/template.py

[ripe meteor]

Oh i didnt realise i was in this channel

[feral hazel]

actually i could probably just paste it in here it's about the same size as the previous code block lol```py
from contextlib import suppress
from itertools import groupby
from pathlib import Path

with (Path(file).parent / "input").open() as f:
data = f.read().rstrip()

lines = data.splitlines()
blocks = [tuple(group) for has_content, group in groupby(lines, bool) if has_content]
with suppress(ValueError):
ints = [int(x) for x in data.split()]

[rancid creek]

functools reduce! also itertools cycle and repeat, and for me I use from time import perf_counter a lot as well. 🙂

for re I have to use import re as I use the flags a fair bit. Not to mention the occasional use of compile and match

[feral hazel]

ooh yeah
i just grepped the folder to see what i tended to use
in the problem where cycle would've been useful i just multipled the string by a big number lol

[normal dragon]

I’m not as fast as everyone else so I like using dataclasses/namedtuples to give my fields semantic meaning

[normal dragon]

Also, I often use sys.exit() to dump myself out of code when I’m debugging things

[feral hazel]

oh yeah last year i used namedtuples a bunch
working in 2d with classes was so much easier

[feral hazel]

ah at that point i usually plop it all into a function and return lol

[normal dragon]

Yeah I’m too lazy to indent things sometimes

[feral hazel]

fair

[normal dragon]

Or properly debug using a debugger

[feral hazel]

print debugging ftw

i've been dabbling in haskell and it's a miracle i haven't passed in io to print debug a function yet

[simple raven]

icecream debugging > print debugging

[agile silo]

I was trying to solve day 10 part 2, but the squeezing part is complicated...

otherwise I would do a simple coloring algorithm

like, travel from the inside, all tiles that are connected

but if there are pipes that are squeezed together, I wouldnt be able to travel through those

unless I did something...

any hint?

or any other different approach that I could take (just hint)

[rancid creek]

That's always fun.

[normal dragon]

It’s not a huge amount of work either

@dataclass
def Point:
    r: int
    c: int

[rancid creek]

Point = namedtuple("Point", "x y")

[feral hazel]

ah i've always preferred using the one in typing
why is it in typing actually

[normal dragon]

That’s what I use 99% of the time, but I use row / column so I don’t have to think about swapping the y and x variable on lookups

[simple raven]

"x y" just doesn't sit well with me, I know the namedtuple can parse it nicely, but it's such an odd way of passing in individual names

[rancid creek]

Also nice to add some dunders. For the Beacons puzzle a couple years ago where I needed 3 dim points:

class Beacon(tuple):
    def __add__(p1,p2): return Beacon((p1[0]+p2[0],p1[1]+p2[1],p1[2]+p2[2]))
    def __sub__(p1,p2): return Beacon((p1[0]-p2[0],p1[1]-p2[1],p1[2]-p2[2]))

Should add x,y,z to that and clean it up.

[normal dragon]

Also __iadd__ is nice

a = Point(1,2)
a += (1,1)

Work sreally nice if you have named directions

Up = Point(1,0)
A += Up

[rancid creek]

a = Point(1,2)
a.NorthEast()

[normal dragon]

Well you could do that too ^_^

[simple raven]

this was my Range class for ||day 5||, it made writing the solution much much easier
||```py
class Range:
def init(self, start: int, end: int = None, steps: int = None):
self.start = start
if steps:
self.end = start + steps - 1
else:
self.end = end
self.seen = False
def contains(self, item: int):
return self.start <= item <= self.end
def repr(self):
return f"Range({self.start}, {self.end})"
def lt(self, other):
return self.start < other.start
def gt(self, other):
return self.end > other.end

[rancid creek]

Probably not, I'd go with a move, and some enums

[normal dragon]

Same

[rancid creek]

nice. I just used the actual Range class. but thought about extending it to support adding and subtraction of ranges.

[normal dragon]

I don’t think you can extend the range type?

!e

class Range(range):
    pass

[marsh current]

@normal dragon :x: Your 3.12 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "/home/main.py", line 1, in <module>
003 |     class Range(range):
004 | TypeError: type 'range' is not an acceptable base type

[normal dragon]

You could still make a custom Range() class and use the range item under the hood (though that’ll be slow for some operations, like the in operator)

[feral hazel]

why in specifically?

[normal dragon]

Oh I’m thinking of something else

There was something where range was expanding out into the actual numbers of it

I don’t remember I was probably doing something stupid

[feral hazel]

ah ok

[rancid creek]

ah right, that's one of the reasons I abandoned it. Also it was getting silly with the permutations.

def __sub__(self, other):
    b0,b1 = self.start,min(self.stop,other.start)
    a0,a1 = max(other.stop,self.start),self.stop
    return (ORange(b0,b1) if b1>b0 else (),ORange(a0,a1) if a1>a0 else ())

pretty sure that spoils nothing at all.

[normal dragon]

Yeah I have no clue what that represents

[rancid creek]

compared to just using actual ranges which was fine.
||```py
for r in seed_ranges:
# Split range into 3 possible ranges using r1, r2 as the potential breakpoints
B = range(r.start,min(r.stop,o))
A = range(max(a+o,r.start),r.stop)
M = range(max(r.start,o)+d-o,min(a+o,r.stop)+d-o)
if B: new_ranges.append(B)
if A: new_ranges.append(A)
if M: converted_rules.append(M)

[normal dragon]

Built-ins where possible is usually a good move. That looks clean

[rancid creek]

BAM.

[normal dragon]

Though…

I don’t know, self, I feel like you could compress that down into a single line

[rancid creek]

It would be less clean. If you want to do that move it to a separate function. 🙂

[normal dragon]

(I was just trying to bait you into doing absurd golf)

Also, now I recognize the code

Or rather which problem it was

[lime fulcrum]

bit stumped today. i liked day 10 more

[normal dragon]

I haven’t read the puzzle yet, but ||I saw someone mention it requires knowledge of a specific algorithm||

[lime fulcrum]

i hope it doesn't. its basically day 10 but with directions instead of a given grid

[rancid creek]

I was the only person I saw using range() so maybe it was not the optimal approach. But the code felt clean.

[normal dragon]

Super clean

[rancid creek]

Your solution from ||day 10|| should be applicable here. But it might be ||slow if you do it 1 point at a time|| but should still finish within a ||couple minutes||

[lime fulcrum]

||might not be. I relied on knowing when there is a corner. guess i could do where I came from and where i'm going at that point||

[rancid creek]

No, there's lots of ways to do it. But any of the ons that passed ||day 10|| should probably work. But ||there is a subtle difference in what they're counting||.

[normal dragon]

Cool!

I’m very far behind. It always gets challenging keeping up in the later days with my social obligations

[rancid creek]

I posted links to 2 different wikipedia pages in the spoiler channel. ( spoilers: #1186176336734203954 message )

[normal dragon]

Noice

[lime fulcrum]

||ok idea i go basically the same way with tracking in and out by toggling but i'm doing it it in x and y seperately so it doesn't matter if there is a lot of space inbetween. might have to mark in all intersections though on both lists||

[rancid creek]

Best to just get help and keep plodding on. The exposure to the problems and techniques used alone is worth it, even if you get a lot of help.

[normal dragon]

Oh I more meant from finding time to do it

I usually can figure out when I need to start looking up algorithms or asking for help

[rancid creek]

I liked how one guy figured out a neat trick for one of the days and explained it, and someone else replied with a link to the wiki page explaining the same theorem.

[normal dragon]

Like the last three days have all been filled with holiday social/family things

[normal dragon]

I love that! Reading the solves is always such a joy

Especially the super fast solvers with the their one line solutions

[rancid creek]

I've got a personal battle with one guy. We think alike and post similar code, but his is so much cleaner especially when he uses numpy. Curses.

[normal dragon]

4HbQ?

[rancid creek]

me: I did a cool solve!
him: like this (link to better code)

[rancid creek]

curse him. 🙂

[normal dragon]

Haha

[rancid creek]

jokes aside, he's great. especially when he does numpy stuff, and we've bounced some ideas off each other in the past.

[normal dragon]

I almost never use numpy

[rancid creek]

4HbQ and jonathonpoulson are prob my 2 favs to follow on reddit and anything @mossy basin posts on discord. always a good read.
jp's code is generally a mess, but I love the speed he cranks it out at and the way he explains what he's doing after.

[normal dragon]

I love watching Jp’s solves

I always go “gfdi how is he so fast”

[rancid creek]

I did one last year (or 2021?) I finished but wasn't happy with how slow my code ran, so I watched his vid, and 1/2 way through the solve he says "this won't work" and swapped to c for the rest of it.

[normal dragon]

I also really enjoy the early isolang days

I’ve hit the leaderboard all of once. That was early day when it wasn’t nearly as competitive as it is now

[rancid creek]

Yesterday was my first leaderboard this year. I've been going casually and didn't expect to hit it at all this year.

[normal dragon]

That’s awesome! Way to go

[rancid creek]

Thanks. But I was happier with rewriting it to finish in 1ms. the Original version the one that got on the leaderboard took 2 minutes to finish. lol

[normal dragon]

Nice. Excited to peek at your code later. I’ve enjoyed the discussions so far this year

[rancid creek]

In short there's an ||O(n) solution|| but my original ||was O(n*m)|| (<--- you can look at this later)

ok, brb. Being taken shopping. My favorite. 🙂

[normal dragon]

I wish more engineers understood algorithmic complexity. Like you don’t need a super in depth knowledge of it, but having some knowledge is good

[normal dragon]

🫡

[rancid creek]

I wish more engineers just had common sense. 🙂

[lime fulcrum]

common sense is surprisingly uncommon?

[rancid creek]

I had a team working on a web front end to showcase how amazingly fast and real time our search engine was. But it took ~5 seconds to LOAD the web page. The Engineer working on it added a cute spinning logo to mitigate it but I was still annoyed it made us seem slow.

So I assigned another engineer to help see if he could speed up the loading code. Around lunch time I see him and asked if he'd found anything yet.

him: yeah, what do you want the load time set to?
me: well 0 ideally but do your best
(pause)
me: what do you mean "set to"

Turns out that the first engineer had added a sleep(5) to the page to make it look like it was a professional site and was doing a lot of setup.

[simple raven]

LOL

[lime fulcrum]

oof

yeah I aspire to never do that

[rancid creek]

ok, brb, have fun all

[simple raven]

every now and then just regex find sleep(%d) and replace with '' in the codebase

[normal dragon]

That sounds like the old speedup loop

https://thedailywtf.com/articles/The-Speedup-Loop

I can’t believe you ran into it in the wild

[normal dragon]

man grep

[simple raven]

who and what fruit?

[normal dragon]

grep is a command line utility that searches through text files

[strange zinc]

Cap, you a vim user?

[ripe meteor]

I can actually imagine people complaining that an app's loading time is so quick it doesnt appear to be working "properly" to them
Sometimes slowing things down isnt such a bad idea

[rancid creek]

I've seen some things over the years. 🙂

[normal dragon]

What is vim ??? /s

[normal dragon]

There is a legit UX experience need for slowing down a response to convince people the app is doing something. TurboTax is a well known/studied example. But it depends on the use case and should be an actual documented need before you slow down the experience for an end user

[normal dragon]

Same.

I just hadn’t seen that one in the wild. I’ve seen unintentional examples. Like once some dev wrote some silverlight code that had an animation that ate up huge amounts of cpu and eventually crashed the computer from a memory leak

[lime fulcrum]

but lets not pretend its isolated to programming. screw up at work: we have o ring seals for different temperatures. the 180 degrees ones are a lot more expensive. someone put in 160 degree seals in the 180 degree seals box. which work for the 30min of testing we do in house but start leaking after 5 hours when its with the client.... not ideal. also 180 was like 20.- and 160 about 5.- visually identical

[royal gull]

.... ... like tears in the rain

[rancid creek]

I actually looked up the exact quote thinking I'd paste it, but the last line is a bit morbid. 🙂

[royal gull]

lol, I did same too...

[rancid creek]

...1999 Mars Climate Orbiter

An investigation indicated that the failure resulted from a navigational error due to commands from Earth being sent in English units (in this case, pound-seconds) without being converted into the metric standard (Newton-seconds). The error caused the orbiter to miss its intended orbit (about 90 miles, or 145 kilometers) and to fall into the Martian atmosphere at approximately 35 miles (56 kilometers) in altitude. The spacecraft disintegrated due to atmospheric stresses.

[normal dragon]

http://www.cs.unc.edu/~smp/COMP205/LECTURES/ERROR/lec23/node4.html

Floating point rounding errors

[lime fulcrum]

its the last words of a character of course its morbid

[simple raven]

it's funny to see in the stats the drop on day12 and then everyone who managed to get through day 12 stayed for the next 3 days 😄

[noble skiff]

i must admit I may be close to my limit 😛

didn't have time to do much Friday-Sunday because of real life; caught up 16 and 17 now, went relatively smoothly.
18 isn't hard but I'm just pretty sick of such questions; not much motivation to od it

[simple raven]

fair, I'm waiting for the next dijkstra so I can skip it again and be done 😄

[noble skiff]

lol

I didn't mind the djikstra as much tbh

I think largely because at least that felt a little more fresh

and the actual grid-like nature was not nearly as forefront. for day 18 most people are just using a formula they barely understand which I'm not that interested in doing 😛

[simple raven]

I have never learned path solving algorithms so it was either spending all day on the problem or watching movies with family...

[noble skiff]

djikstra is fun and worth learning fwiw

[simple raven]

I want to come back to it after christmas celebrations are over and redo day 12 for dp and 17 for dijkstra

[rancid creek]

You might still be working on day22 then ....

[rancid creek]

to be fair, using established tools and algorithns is a lot of what makes python great
You don't need to know how timsort works to use sorted. You don't have to implement a priority queue to use hyeapq, etc.

the puzzle is figuring out you need something that counts the ||area of a polygon||, googling algorithms to do that is part of the solve.

and maybe you learn enough, that next time you still know the options you have.

[noble skiff]

I'm honestly not sure how the convo was related to python in any way 🤔

I think as always for AoC, people draw the line in different places for what tools they'll use. Some people will use basically anything, some will use the standard lib, some will do it in C with zero dependencies, etc

personally I try to mostly only use things that I understand relatively precisely how they work. I got a binary heap implementation for day 17 because I didn't want to write one out from scratch, but I know how to do it.

i don't know how that formula works, and to judge from some of the discussion around it, it seems like most people have only the vaguest idea. some people will be okay with using that regardless, some will not.

Though honestly, I might sit down and learn it... it would probably be a lot less painful than dealing with all the edge cases in alternative ways of doing it

[rancid creek]

The point is you don't need picks or greens theorem here, You can use them. But they're not required. Using them just makes sense in the same way using timsort instead of writing your own sort makes sense. Sometimes knowing how to use a tool is all you need. For example the vast number of people that use something like excel, vs the small number of people that actually worked on it.

[noble skiff]

like I said, the difference is that I know how sort works - whether it's timsort or not is a red herring because it doesn't specifically need to be timsort

If the difference doesn't matter to you - that's cool. But there is a difference, and for me it matters 🙂

[spark dust]

Looking for someone in here trying to make billions

On a business idea

Starting from nothing

[hollow wharf]

Lmao

[spark dust]

Capital B

[civic dune]

can someone give me a medium sized project that is for beginners? please?

[spark dust]

Billions

Money sign &

$

Billion$

[hollow wharf]

Make an app that displays weather/forecast for a user's specified city/current city

[civic dune]

"beginners"

[spark dust]

You will have to scrape data

[civic dune]

::))))

[hollow wharf]

:')

[spark dust]

And move work with api

That is beginners

Extremely beginner

Simple request and parsing

[still delta]

If you're in the US the NOAA is where to look for all your weather needs. Believe it or not most all weather apps or websites get their raw info from NOAA for free and then repackage it.

[minor cave]

@spark dust@hollow wharf@civic dune this is better suited for #python-discussion it's off topic for this channel

[spark dust]

Shhh

[hexed axle]

Language Roulette: Day 19
The language is ... C#

[earnest forge]

You need glasses?

[hexed axle]

No no, I've got them, that's how I C#

[earnest forge]

GLHF everyone

[shut trellis]

<@&518565788744024082> Good morning! Day 19 is ready to be attempted. View it online now at https://adventofcode.com/2023/day/19. Good luck!

[broken bloom]

yay!

I'm not gonna know how to solve it anyway

I've been stuck since day 10

[hexed axle]

oh part 2 must be a doozy today

[magic oxide]

haven't even looked at day 18 yet

[devout tusk]

Today looks tedious to implement

[earnest forge]

239/14 let's go

[hexed axle]

I'm purely judging off of the leaderboard's very slow trickle in completion

[earnest forge]

It's actually not that bad

[hexed axle]

I haven't even read today's question yet

[earnest forge]

Won't give hints until LB is full though

I keep forgetting to put 'return total' in my functions and I swear it's costing me several spots each day

[pine tiger]

REDEEMED

[pine tiger]

sheesh

starwort sweep!

[earnest forge]

no I got 239 for p1

Tbf writing better code for p1 is probably why I did so well at p2

[covert arrow]

oh shit i forgot

[earnest forge]

But I think I literally cost myself 2 spots on p2 by forgetting the return total first time for each part

[pine tiger]

😭 😭

[earnest forge]

FULL

[feral hazel]

40/70 :D

[earnest forge]

Nice

[hexed axle]

https://discord.com/channels/267624335836053506/1186540774872059995

[flat sequoia]

idk why my p2 doesn't work tbh

maybe it's just because the code is so bad

I should rewrite it..

[pine tiger]

yk i hate when aoc relies on leaps of faith

source: just trust me bro

[earnest forge]

Leaps of faith? Where lol

[pine tiger]

||i'm probably wrong somewhere, but why don't the xmas ranges ever intersect? i think you can construct a workflow such that they do||

[earnest forge]

I genuinely have no idea what you mean by this lol

Talk to me in spoilers

[modern terrace]

||i'm getting 177152000000000 for p2 test cases||

;-;

wait what

[earnest forge]

LB is full so I'll give the hint now: ||it's not that bad if you remember day 5||

[modern terrace]

86/855