AoC 2022 | Code Golf Solutions & Spoilers

[sharp scarab]

any path inside the cube would be at most 69

i think

since any "optimal" path could just be shadowed onto the axis

[sullen field]

why does this take a few seconds

[sharp scarab]

because we're floodfilling a lot of space

[orchid charm]

because we don't check bounds at all

[raw frigate]

We're filling from the origin

[sullen field]

i'm gonna wait for 20s again and see if my changes worked

[sharp scarab]

yeah gets the right answer for me

[sullen field]

which they probably do so

[raw frigate]

And if we did 10 steps for example

It would reach 0,0,10

[sullen field]

[Day 18] untested 205 ```py
N=lambda x,y,z:{*zip([x-1,x+1]+[x]*4,[y,y,y-1,y+1,y,y],[z]*4+[z-1,z+1])}
S=lambda A:sum(len(N(*i)-A)for i in A)
C={*map(eval,open(0))}
F={(0,)*3}
exec('for c in{*F}:F|=N(*c)-C\n'*69)
print(S(C),S(F)-57966)

[raw frigate]

But if there was a block in the way

[sharp scarab]

also that yeah

very obvious char save

[raw frigate]

It wouldn't be able to reach 0,0,10 anymore

[orchid charm]

input is a sphere, you are not going to get that kind of point

[sullen field]

confirmed

[orchid charm]

how did i not see that

[sharp scarab]

you would need an input that blocks in all 3 directions

[raw frigate]

Not necessarily

[orchid charm]

like i said, you will never get such an input

[raw frigate]

Just 1 block in the wrong location would affect the result

Idk if it actually happens in any input

But it can

[orchid charm]

because then it wouldn't be a sphere

[sharp scarab]

yeah

if you just look at it from the x,y perspective

once you reach (22, 22) you're basically fine since all points <22

for it to not work it'd need to block all paths to (22,22)

which basically make it very close to a flat plane in that direction

[orchid charm]

nah enthusiast is right, the thing is, i'm also right because such an input won't exist

you get to all locations that can be flooded within 69, but if you have enough points blocking the way, then the corner at the very end will be affected

because it would take longer than 69 to get there

and that affects the constant we subtract

[sharp scarab]

yeah but as you said it'd be more of a swiss cheese square and not a sphere

[orchid charm]

yeah

so we're all right

and the code works

[sharp scarab]

almost sub 200

[raw frigate]

Can we recude the constant by running the floodfill for shorter?

[orchid charm]

we probably can

[raw frigate]

How long does everyone's floodfill take

[orchid charm]

i'll try getting it to 4 digits

[raw frigate]

I think mine was 72

[sharp scarab]

huh

but it still works?

[raw frigate]

Idk

I can't test it on the current version

[orchid charm]

no we cannot reduce the constant to 4 digits

i tried, it's just not happening

[sharp scarab]

yeah 69 is like the minimum

[raw frigate]

How about 5 digits?

[sharp scarab]

it's at 5 digits tho

[orchid charm]

running the golf takes 11.3s for me

[sharp scarab]

takes like half a minute for me

[raw frigate]

Isn't it 118806?

[orchid charm]

hardware diff

[sharp scarab]

that's for 99

[orchid charm]

for 69 it's 57966

[sharp scarab]

I have a 6 year old mac

[raw frigate]

Ah ic

[sharp scarab]

don't expect much from it

[orchid charm]

i have a 1/2 year old mac

[sharp scarab]

actual hardware diff

huh cool 57966 is 6 times a prime

[orchid charm]

yeah

[sullen field]

so is 118806

[orchid charm]

they're only prime because we need to add 6 for the tips

[urban torrent]

does this take like 20s for you? Also, it's 69 for everyone? pog!

[orchid charm]

11.3

[urban torrent]

m2 v m1 😔

[sharp scarab]

that just makes it odd

not prime

[sullen field]

i just wish for some python assignment trick or an efforted rewrite to make a sub-200

[raw frigate]

Also, were counting surface area and not volume

[sharp scarab]

I'd bet on refactoring the neighbours bit

[orchid charm]

57966 is 6 * 9661, 57960 is 6 * a bunch of factors

[sharp scarab]

I was thinking some eval*6 trick but idk where to start with that

[orchid charm]

same with 118806 vs 118800

[sullen field]

wait i think i have an idea

[sharp scarab]

I doubt they're all prime

so we just got lucky

[sullen field]

just gonna wait for the 20 seconds or something that takes to run this

[sharp scarab]

lmao

[orchid charm]

at least this was better than that golf which took 2 mins

[sullen field]

[Day 18] 204 ```py
N=lambda x,y,z:{*zip([x-1,x+1]+[x]*4,[y,y,y-1,y+1,y,y],[z]*4+[z-1,z+1])}
S=lambda A:sum(len(N(*i)-A)for i in A)
C={*map(eval,open(0))}
F={(0,)*3}
exec('for c in{*F}:F|=N(*c)-C\n'*69)
print(S(C),S(F)%9661)

[urban torrent]

works for me. why did you strike it?

[orchid charm]

oh nice

[sharp scarab]

oh very nice

[sullen field]

i striked "untested"

[sullen field]

i didn't strike the whole thing

[orchid charm]

lol

[urban torrent]

i'm lagging in the chat

my bad

[sharp scarab]

we should go back after Christmas and tally up the amount of unexpected 69s this year

[urban torrent]

didn't understand the logic of the numbers that were %

[sullen field]

i think golfing breakthrough should be thoroughly focused on this line ```py
N=lambda x,y,z:{*zip([x-1,x+1]+[x]*4,[y,y,y-1,y+1,y,y],[z]*4+[z-1,z+1])}

[urban torrent]

or golf the programs such that a total of 69 69s show up. kek

[sullen field]

it's too long

[urban torrent]

.

[urban torrent]

what about some matrix logic? maybe too long.idk

[sullen field]

that were %
huh

[urban torrent]

9661,57966

these ones

what do they mean?

[sharp scarab]

57966 is 6*9661, so subtracting 57966 keeps residue the same mod 9661
we know the answer is gonna be probably less than 9661, so taking mod 9661 gives us the unique number with the same residue as the original that is less than 9661, which is gonna be the answer

57966 is just the offset from floodfilling 69 times unbounded

it's the surface area of the diamond thing after floodfilling 69 times

I'm gonna go to bed

and hope that in my dreams the python gods visit me with a shortened formula for the neighbours in 3d space

[sullen field]

imagine ± operator

[sharp scarab]

true

[urban torrent]

I don't have python2 but this most likely would work in it
[Day 18] 201

N=lambda x,y,z:{*zip([x-1,x+1]+[x]*4,[y,y,y-1,y+1,y,y],[z]*4+[z-1,z+1])}
S=lambda A:sum(len(N(*i)-A)for i in A)
C={*map(eval,open(0))}
F={(0,)*3}
exec'\nfor c in{*F}:F|=N(*c)-C'*69
print S(C),S(F)%9661

[orchid charm]

i thought macos came with python2

i must've manually installed it at some point then

[urban torrent]

it defaults to python3.8. for me

[urban torrent]

must be

[urban torrent]

i think 204/201 is the limit

[orchid charm]

what if we use integers instead of tuples?

since we know it's bound between -1 and 22, we can just use base 24

this would make C much longer but would make N much shorter

[urban torrent]

wdym, those integers would need to be added to a collection like a tuple/list tho right?

[orchid charm]

we just convert each tuple to an integer

return (x + 1) + (y + 1) * 24 + (z + 1) * 24 * 24```

something like this?

of course can be golfed

this way we can directly add tuples together

i'm not sure if we'd need bounds for this though

[urban torrent]

this is basically a different integer for every neighbor of x,y,z or one for all 6 neighbors per x,y,z?

[urban torrent]

also, N would still have the N=lambda x,y,z:... which will add to the character. The net difference may be very less

[urban torrent]

@orchid charm any luck with the integer thingy? i'm still confused about what you were trying to explain

[orchid charm]

i was playing chess

[urban torrent]

do you need a challenge?

i just started code.golf and it's amazing!

shortest solution to fizzbuzz, I got. How did someone 58 byte/53 character it?

[sullen field]

probably something to do with slicing FizzBuzz like this? ```py
print("FizzBuzz"[x%-3&4:12&8+x%-5]or x)

or python 2?

[orchid charm]

oh wow i actually recognize some people in their discord

from back when i played tons of clash of code

i was #1 in the us at one point

[sullen field]

how to do this

[orchid charm]

?

[sullen field]

fizzbuzz in 58

[orchid charm]

uh i'll try after dinner

[raw frigate]

[Day 18] 195 ```py
N=lambda x:{x-1,x+1,x-300,x+300,x-9e4,x+9e4}
S=lambda A:sum(len(N(i)-A)for i in A)
C={x+y300+z9e4 for x,y,z in map(eval,open(0))}
F={0}
exec('for c in{*F}:F|=N(c)-C\n'*69)
print(S(C),S(F)%9661)

it turns out you can use negative numbers in the base system as well

otherwise things like balanced ternary wouldn't exist

@urban torrent sub 200!

[orchid charm]

knew it

🙂

[raw frigate]

I don't think the constants can be made shorter

Another 195 which uses base 139 instead ```py
N=lambda x:{x-1,x+1,x-b,x+b,x-bb,x+bb}
S=lambda A:sum(len(N(i)-A)for i in A)
b=139
C={x+yb+zb*b for x,y,z in map(eval,open(0))}
F={0}
exec('for c in{*F}:F|=N(c)-C\n'*69)
print(S(C),S(F)%9661)

[orchid charm]

[Day 18] 193

N=lambda x:{x-1,x+1,x-b,x+b,x-b*b,x+b*b}
S=lambda A:sum(len(N(i)-A)for i in A)
b=9661
C={x+y*b+z*b*b for x,y,z in map(eval,open(0))}
F={0}
exec('for c in{*F}:F|=N(c)-C\n'*69)
print(S(C),S(F)%b)```

[sharp scarab]

[Day 18] 192

N=lambda x:{x-1,x+1,x-b,x+b,x-d,x+d}
S=lambda A:sum(len(N(i)-A)for i in A)
b=9661
F={d:=b*b}
C={x+y*b+z*d for x,y,z in map(eval,open(0))}
exec('for c in{*F}:F|=N(c)-C\n'*69)
print(S(C),S(F)%b)

using d doesn't save any characters btw

i just happened to be testing it

b:=9661 in there works as well

[urban torrent]

That is still 59.

[urban torrent]

Noice!
Haven't had the time to see it. I will in a few hours

[urban torrent]

can you explain the math behind this?looks interesting

[raw frigate]

what math?

it's just base 139

[urban torrent]

base 139?

[raw frigate]

like binary, ternary, etc.

instead of storing x,y,z tuples you store x + y*base + z*base**2

and overflow/underflow doesn't happen if the base is big enough

[urban torrent]

[Day 18] 191 but Code should just take a few months to run I guess

N=lambda x:{x-1,x+1,x-b,x+b,x-d,x+d}
S=lambda A:sum(len(N(i)-A)for i in A)
b=9661
F={d:=b*b}
C={x+y*b+z*d for x,y,z in map(eval,open(0))}
exec('for c in{*F}:F|=N(c)-C\n'*b)
print(S(C),S(F)%b)

[sharp scarab]

no because then %b wouldn't work

9661 is tailored specifically for 69

[urban torrent]

ohh sadge. I thought it was just hella slow

[urban torrent]

it literally seems impossible to me rn. Threw everything I had at it 😔

[wraith crypt]

What's your best? Mine is 61

[urban torrent]

59

Can't get 58 smh

[urban torrent]

anyone golfed today?

[raw frigate]

nope

[urban torrent]

this one is hella short but doesn't work on my p1.

https://topaz.github.io/paste/#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

[urban torrent]

Got from reddit

[urban torrent]

That is to expand the search area

[urban torrent]

stuck with fibonacci on 38 as well when best is 36 😔

[urban torrent]

still no golfs today?

damn

[urban torrent]

Blueprint 28: Each ore robot costs 4 ore. Each clay robot costs 3 ore. Each obsidian robot costs 3 ore and 7 clay. Each geode robot costs 2 ore and 7 obsidian.
doesn't seem to work.
should give 13 but returns 12 instead. rest all are correct

[sullen field]

i have 36 best

anyone who was at #esoteric-python at a certain time probably had 36 best

[urban torrent]

I dm'ed you the best I got.

How can you POSSIBLY shave more chars?

[orchid charm]

do we just not have a day 19 golf

i finally did it because i was finished with exams

[wraith crypt]

I mean given that my code apparently doesn't even work for every input I'm not sure how doable that is

[orchid charm]

💀

[wraith crypt]

Day 19 seems to be a fairly ridiculous problem lol

[orchid charm]

i used z3, so…

[wraith crypt]

Lmao nice

[orchid charm]

can't golf mine either

[raw frigate]

yeah, most of the solutions on reddit aren't guaranteed for all inputs

[orchid charm]

rip

[sullen field]

who's starting with day 20 golf

[raw frigate]

I'm working on golfing mine

[sullen field]

[Day 20] base golf, 250 ```py
A=[int(s)for s in open(0)]
m=0
I,=L=range(f:=5000)
S="i=0\nwhile i<f:x=I.index(i);del I[x];I.insert((x+A[i])%(f-1),i);i+=1\n"
G="for x in 1,2,3:m+=A[I[(I.index(A.index(0))+x1000)%f]]\nprint(m);"
exec(S+G+"A=[x*811589153 for x in A];I,=L\n"+S10+G)

[orchid charm]

[Day 20] 232

A=[*map(int,open(m:=0))]
*I,=L=range(f:=5000)
S="for i in L:del I[x:=I.index(i)];I.insert((x+A[i])%~-f,i)\n"
G=S+"for x in 1,2,3:m+=A[I[(I.index(A.index(0))+x*1000)%f]]\nprint(m);"
exec(G+"A=[x*811589153 for x in A];*I,=L\n"+S*9+G)

[raw frigate]

[Day 20] 229 ```py
A=[map(int,open(m:=0))]
I,=L=range(f:=5000)
G="for i in L:del I[x:=I.index(i)];I.insert((x+A[i])%~-f,i)\nfor x in 1,2,3:m+=A[I[(I.index(A.index(0))+x1000)%f]]\nprint(m);"
exec(G+"A=[x811589153 for x in A];*I,=L\n"+G[:57]*9+G)

[sullen field]

[Day 20] 227 ```py
A,=map(int,open(m:=0))
I,=L=range(f:=5000)
G="for i in L:del I[x:=I.index(i)];I.insert((x+A[i])%~-f,i)\nfor x in 1,2,3:m+=A[I[(I.index(A.index(0))+x1000)%f]]\nprint(m);"
exec(G+"A=[x811589153for x in A];*I,=L\n"+G[:57]*9+G)

[raw frigate]

oh wait

[raw frigate]

p2 doesn't work for me

off by part 1

lmao

[sullen field]

what's weird is that also happened the first time i removed the space between 811589153 and for

‫now it works for some reason (for me)

[raw frigate]

idrk

[orchid charm]

oh wait i didn't check my p2

[raw frigate]

[Day 20] 231 this seems to work ```py
A,=map(int,open(m:=0))
I,=L=range(f:=5000)
G="for i in L:del I[x:=I.index(i)];I.insert((x+A[i])%~-f,i)\nfor x in 1,2,3:m+=A[I[(I.index(A.index(0))+x1000)%f]]\nprint(m);"
exec(G+"A=[x811589153for x in A];*I,=L;m=0\n"+G[:57]*9+G)

[orchid charm]

it's not right

works

[sullen field]

[Day 20] 228 with the space ```py
A,=map(int,open(m:=0))
I,=L=range(f:=5000)
G="for i in L:del I[x:=I.index(i)];I.insert((x+A[i])%~-f,i)\nfor x in 1,2,3:m+=A[I[(I.index(A.index(0))+x1000)%f]]\nprint(m);"
exec(G+"A=[x811589153 for x in A];*I,=L\n"+G[:57]*9+G)

[sullen field]

does this work?

[orchid charm]

ye

[raw frigate]

yes

[sullen field]

that's weird then

‫"also it's supposed to be "with the space

[orchid charm]

also this is 231

[sullen field]

wait

[sullen field]

fixed

[raw frigate]

now it doesn't work anymore

you need to reset m between p1 and p2

[orchid charm]

now it doesn't work

yeah

[sullen field]

ok

[orchid charm]

someone else posted another 228 which did work

pls repost

[sullen field]

works for me for some reason

[raw frigate]

^

Im making the 228 again

[orchid charm]

[Day 20] 228

*A,=map(int,open(m:=0))
*I,=L=range(f:=5000)
G="for i in L:del I[x:=I.index(i)];I.insert((x+A[i])%~-f,i)\nfor x in 1,2,3:m+=A[I[(I.index(A.index(0))+x*1000)%f]]\nprint(m);A=[x*811589153for x in A];*I,=L;m=0\n"
exec(G+G[:57]*9+G)```

sorry idr who did it

[raw frigate]

I did

[orchid charm]

so i'm just gonna repost it

[raw frigate]

smh

[orchid charm]

oh

[sullen field]

why does the 227 work for me

[orchid charm]

because the 228 also works for you

[raw frigate]

it does?

[sullen field]

‫and that was based on the 229

[orchid charm]

it doesn't work for me

[raw frigate]

the base golf doesn't work

[orchid charm]

yeah

but it was fixed

[sullen field]

base golf doesn't even reset m

[raw frigate]

which is why it doesnt work

[sullen field]

it's quite the weird thing that it works for me

why am i the odd one out

[raw frigate]

why does it always have to be cereal lmao

[orchid charm]

[Day 20] 225

*A,=map(int,open(0))
*I,=L=range(f:=5000)
G="m=0\nfor i in L:del I[x:=I.index(i)];I.insert((x+A[i])%~-f,i)\nfor x in 1,2,3:m+=A[I[(I.index(A.index(0))+x*1000)%f]]\nprint(m);A=[x*811589153for x in A];*I,=L;"
exec(G+G[:61]*9+G)```

[raw frigate]

[Day 20] 221 ```py
A,=map(int,open(0))
I,=L=range(f:=5000)
G="\nfor i in L:del I[x:=I.index(i)];I.insert((x+A[i])%~-f,i)\nprint(sum(A[I[(I.index(A.index(0))+x1000)%f]]for x in(1,2,3)));A=[x811589153for x in A];*I,=L;"
exec(G+G[:57]*9+G)

[orchid charm]

[Day 20] 220

*A,=map(int,open(0))
*I,=L=range(f:=5000)
G="for i in L:del I[x:=I.index(i)];I.insert((x+A[i])%~-f,i)\nprint(sum(A[I[(I.index(A.index(0))+x*1000)%f]]for x in(1,2,3)));A=[x*811589153for x in A];*I,=L\n"
exec(G+G[:57]*9+G)```

[raw frigate]

smh how did I miss that

[orchid charm]

[sullen field]

off by part 1 for me

[orchid charm]

💀

what about the 221?

[sullen field]

wait wtf

231 doesn't even work

the 227 does

above is 231 and below is 227 ```py

py test_a.py < aocday20.txt
<string>:3: SyntaxWarning: invalid decimal literal
5498
3390007892081

py test_b.py < aocday20.txt
<string>:3: SyntaxWarning: invalid decimal literal
5498
3390007897579

[orchid charm]

well of course it's off by part 1

you're sure that the 231 is correct?

[urban torrent]

@orchid charm try fixing part 1 of this

[orchid charm]

pretty simple, just don't prune

but that makes it take forever

it might take longer than an hour

[urban torrent]

how will not pruning, fix p1?

[orchid charm]

…

it prunes out the best solution

so obviously not pruning fixes this

you understand what the code does, right?

well whatever

i'll golf day 19 tomorrow

since i made a solution without z3

[urban torrent]

no, I haven't tried understanding it. I just changed the 2000 to a 5000 and p2 started working for me. Haven't tried anything else with that code atm

It was just quite short, so it would be golfable

[orchid charm]

💀

but it uses numpy

[urban torrent]

well third party libs are no problem if the solution is short enough though

[urban torrent]

it works for me

[urban torrent]

and so does this

[urban torrent]

G[:57]. Damnnn

[urban torrent]

what does the %f do?

oh nvm

it's modulo

[urban torrent]

f = lambda x:nums[idx[(idx.index(nums.index(0))+x)%len(idx)]]

I did this in my code. Would that shorten it? 🤔

nah, nvm, that's like 230 chars

[sullen field]

if m is reset it gets off by p1

otherwise it's correct for me

[raw frigate]

Does the 221 work then?

[sullen field]

no

off by p1

[raw frigate]

Nani?

Hmm

[sullen field]

why is it off by p1

[raw frigate]

yeah

[sharp scarab]

do we have a day 19 golf yet?

[sharp scarab]

ok i'm trying to golf my day 19 and i'm running out of variable names

i'm making a variable dictionary

i can't anymore

[urban torrent]

Just use this. This will work if there's no pruning. Just uses numpy if you're okay with that

#1050425859225436241 message

[urban torrent]

[Day 20] 202

*D,=open(0)
for k in 1,811589153:N=[k*int(x)for x in D];I=[*range(L:=len(N))];X=I.index;[I.insert((N[I.pop(j:=X(i))]+j)%~-L,i)for i in I*(k%71)];print(sum(N[I[(X(N.index(0))+p*1000)%L]]for p in[1,2,3]))

[raw frigate]

[Day 20] 200 ```py
D,=open(0)
for k in 1,811589153:N=[kint(x)for x in D];I=[range(L:=5000)];X=I.index;[I.insert((N[I.pop(j:=X(i))]+j)%~-L,i)for i in I(k%71)];print(sum(N[I[(X(N.index(0))+p*1000)%L]]for p in[1,2,3]))

[Day 20] 196 ```py
N=[int(x)for x in open(0)]
for k in 1,811589153:I=[range(L:=5000)];X=I.index;[I.insert((N[I.pop(j:=X(i))]k+j)%~-L,i)for i in I(k%71)];print(sum(kN[I[(X(N.index(0))+p*1000)%L]]for p in[1,2,3]))

[Day 20] 190 ```py
N,=map(int,open(0))
for k in 1,811589153:I=[range(L:=5000)];X=I.index;[I.insert((N[I.pop(j:=X(i))]k+j)%~-L,i)for i in I(k%71)];print(sum(kN[I[(X(N.index(0))+p1000)%L]]for p in[1,2,3]))

[sharp scarab]

good news: i have a golf for day 19 that works for the most part
bad news: it's off by one for part 1 blueprint 1 and works fine for everything else

[sharp scarab]

nvm figured it out

floating point error

[Day 19] 554 (without numpy)

import re,math as p
I=lambda *x:1+M(*map(p.ceil,x),0)
q=n=0
Z=0,0
e=1
M=max
for _,A,B,C,D,E,F in (map(int,re.findall('\d+',l))for l in open(0).read().split('\n')):
 def d(r,l,s,g,b,c,o,t):
  global m
  m=M(m,g)
  if(t-1)*t/2+g<m or t<1:return
  for v,w,x,y,z,h,j,k in((o<M(A,B,C,E)and I((A-r)/o),A,*Z*2,0,1),(D>c and I((B-r)/o),B,*Z*2,1,0),(c*(F>b)and I((C-r)/o,(D-l)/c),C,D,*Z,1,*Z),(b and I((E-r)/o,(F-s)/b),E,0,F,1,*Z,0)):
    if v:d(r+o*v-w,l+c*v-x,s+b*v-y,g+z*(t-v),b+h,c+j,o+k,t-v)
 m=0;d(*Z*3,1,24);q+=_*m
 if n<3:n+=1;d(*Z*3,1,32);e*=m
print(q,e)

can't wait for someone to tell me that it doesn't work on their input

here's the variable dict

var_dict = {
    'q':'quality',
    'n': 'counter for pt2',
    'e': 'pt2 geode product',
    'm': 'max geode',
    'Z': (0,0),

    '_': 'blueprint id',
    'A': 'orebot ore cost',
    'B': 'claybot ore cost',
    'C': 'obsidianbot ore cost',
    'D': 'obsidianbot clay cost',
    'E': 'geodebot ore cost',
    'F': 'geodebot obsidian cost',

    'r': 'ore amount',
    'l': 'clay amount',
    's': 'obsidian amount',
    'g': 'geode amount',
    'b': 'obsidianbot count',
    'c': 'claybot count',
    'o': 'orebot count',
    't': 'time left',

    'v': 'condition/dt variable',
    'w': 'ore change',
    'x': 'clay change',
    'y': 'obsidian change',
    'z': 'geode change',
    'h': 'orebot change',
    'j': 'claybot change',
    'k': 'obsidianbot change'
}

[Day 19] 538

import re,math as p
I=lambda *x:1+M(*map(p.ceil,x),0)
q=n=0
Z=0,0
e=1
M=max
for _,A,B,C,D,E,F in(map(int,re.findall('\d+',l))for l in open(0).read().split('\n')):
 def d(r,l,s,g,b,c,o,t):
  global m;m=M(m,g)
  if(t-1)*t/2+g>m and t>1:[d(r+o*v-w,l+c*v-x,s+b*v-y,g+z*(t-v),b+h,c+j,o+k,t-v)for v,w,x,y,z,h,j,k in((o<M(A,B,C,E)and I((A-r)/o),A,*Z*2,0,1),(D>c and I((B-r)/o),B,*Z*2,1,0),(c*(F>b)and I((C-r)/o,(D-l)/c),C,D,*Z,1,*Z),(b and I((E-r)/o,(F-s)/b),E,0,F,1,*Z,0))if v]
 m=0;d(*Z*3,1,24);q+=_*m
 if n<3:n+=1;d(*Z*3,1,32);e*=m
print(q,e)

that's all i'm gonna do until someone confirms that this actually works

[urban torrent]

Having dinner. Will check in a bit

[orchid charm]

works for me

[sharp scarab]

how fast is it?

[orchid charm]

within a few seconds

[sharp scarab]

because i did just basically convert my solution

[orchid charm]

like 3 seconds

[sharp scarab]

so we could probably shave a few optimizations

def part_b_solver(blueprints: list):
    geodes = 1
    for _, oo, co, obo, obc, go, gob in blueprints[:1]:
        max_ore = max(oo, co, obo, go)
        current_max = 0
        def dfs(ore, clay, obs, geode, orebot, claybot, obsbot, time):
            nonlocal current_max
            if time <= 0:
                current_max = max(current_max, geode)
                return
            elif geode + time*(time-1)//2 <= current_max:
                return
            # Build robots
            if max_ore > orebot:
                dt = max(ceil((oo-ore)/orebot), 0) + 1
                dfs(ore + orebot * dt - oo, clay + claybot * dt, obs + obsbot * dt, geode,
                    orebot + 1, claybot, obsbot,
                    time - dt)
            if obc > claybot:
                dt = max(ceil((co-ore)/orebot), 0) + 1
                dfs(ore + orebot * dt - co, clay + claybot * dt, obs + obsbot * dt, geode,
                    orebot, claybot + 1, obsbot,
                    time - dt)
            if gob > obsbot and claybot:
                dt = max(ceil((obo-ore)/orebot), ceil((obc-clay)/claybot), 0) + 1
                dfs(ore + orebot * dt - obo, clay + claybot * dt - obc, obs + obsbot * dt, geode,
                    orebot, claybot, obsbot + 1,
                    time - dt)
            if obsbot:
                dt = max(ceil((go-ore)/orebot), ceil((gob-obs)/obsbot), 0) + 1
                if dt < time:
                    dfs(ore + orebot * dt - go, clay + claybot * dt, obs + obsbot * dt - gob, geode + (time - dt),
                        orebot, claybot, obsbot,
                        time - dt)
        dfs(0, 0, 0, 0, 1, 0, 0, 32)
        geodes *= current_max
    return geodes

it's based off this

orebot and obsbot got switched because it allows *Z*3

refer to variable dict for variables

wait lmao

all this time and effort

and Z3 is still here

[orchid charm]

[Day 19] 536

import re,math as p
I=lambda *x:1+max(*map(p.ceil,x),0)
q=n=0
Z=0,0
e=1
for _,A,B,C,D,E,F in(map(int,re.findall('\d+',l))for l in open(0).read().split('\n')):
 def d(r,l,s,g,b,c,o,t):
  global m;m=max(m,g)
  if~-t*t/2+g>m and t>1:[d(r+o*v-w,l+c*v-x,s+b*v-y,g+z*(t-v),b+h,c+j,o+k,t-v)for v,w,x,y,z,h,j,k in((o<max(A,B,C,E)and I((A-r)/o),A,*Z*2,0,1),(D>c and I((B-r)/o),B,*Z*2,1,0),(c*(F>b)and I((C-r)/o,(D-l)/c),C,D,*Z,1,*Z),(b and I((E-r)/o,(F-s)/b),E,0,F,1,*Z,0))if v]
 m=0;d(*Z*3,1,24);q+=_*m
 if n<3:n+=1;d(*Z*3,1,32);e*=m
print(q,e)```

[sharp scarab]

for future reference btw if you do A/o-r/o instead of (A-r)/o it doesn't work

probably because floating point or something idk

[orchid charm]

[Day 20] 188

*N,=map(int,open(0))
for k in 1,811589153:I=[*range(L:=5000)];X=I.index;[I.insert((N[I.pop(j:=X(i))]*k+j)%~-L,i)for i in k%71*I];print(sum(k*N[I[(X(N.index(0))+p*1000)%L]]for p in[1,2,3]))```

[sullen field]

[Day 20] 187 ```py
N,=map(int,open(0))
for k in 1,811589153:I,=range(L:=5000);X=I.index;[I.insert((N[I.pop(j:=X(i))]k+j)%~-L,i)for i in k%71I];print(sum(kN[I[(X(N.index(0))+p1000)%L]]for p in[1,2,3]))

relooked at day 20 and it seems like everything's all correct

[orchid charm]

[sullen field]

it took 17 hours to find that out

[orchid charm]

https://enira.discord.com/channels/267624335836053506/1050425859225436241/1054662563050307585

could've shaved 13 hours off that time if you'd checked when i asked

[sullen field]

[Day 16] 476 ```py
import re,itertools as i
C="for x,b,f,o in*p,:\n A,B=m[x]\n for t in[(u,b,f)for u in A]+(~b&B and[(x,b|B,f+m[B])]or[]):\n if z.get(t,-1)<o+f:z[t]=o+f;p+=(t,o+f),\n"
z,m,p={},{},('AA',0,0,0)
for l in open(0):
a,b,c=re.findall('\d+|[A-Z]{2}',l);m[a]=c,0
if b:=int(b):m[d:=1<<len(m)]=b;m[a]=c,d
exec(C26)
r=[(k,max(x[1]for x in g))for k,g in i.groupby(sorted((k[1],z[k])for k in z),lambda x:x[0])]
exec(C4)
print(max(z.values()),max(b+d(a&c<1)for a,b in r for c,d in r))

[orchid charm]

how did we not see that

[sharp scarab]

I feel bad for whoever has to gather up all the golfs and put them in the github

we're 9 days behind

[orchid charm]

uh oops

that was supposed to be me and tessa

i had exams, will probably do it tomorrow

[sharp scarab]

she had exams so she handed it over to people that also have exams

[sullen field]

i'm 5 days into christmas break

[sharp scarab]

I'm a month into Christmas break

we didn't get github perms tho

afaik

unless u got it as well

actually idk if tessa has exams tbf

[raw frigate]

my christmas break starts in 2 days

[orchid charm]

same

but i'm done with exams

had my last one today

[raw frigate]

my exams are in like may

but I have mocks right after the break

[orchid charm]

would be so much easier if discord had exact text matching or regex search

as well as search within forum posts

[sharp scarab]

we do have the [Day x] thing to search for

just do like in: #1047673173447020564 [Day 12] or something

yeah but discord search sucks so much

[raw frigate]

we need to make lancebot keep track of the golfs

and have them accessible through a command

[sharp scarab]

golfhelper for next year?

[raw frigate]

yeah maybe

[sharp scarab]

if yall are still here next year

[raw frigate]

yeah I probably will

[sharp scarab]

same

[orchid charm]

really depends on how school is

if i forget how to manage my time, probably won't be here next year

[urban torrent]

https://pypi.org/project/golfer-py/0.2.1/
This exists but it's in japanese. I went through a bit of their code and found out that they are using an ast to parse through the entire code to shorten/reduce it. We can build atop it perhaps?

[sharp scarab]

nah I mean like a helper to assist with submitting golfs

not like an actual golf helper

[sullen field]

[Day 21] base golf using @orchid charm's idea, 351 ```py
import re
L=eval
I=int
e=dict(re.findall("(.+): (.+)",open(0).read()))
D='g(%r)'
g=lambda d:L(A+B+D+C+D+E)
A,B,C,E='(len(h:=e[d].split())<2 and I(*h)','or L("','%s','"%(*h,)))'
t=I(g(s:='root'))
A,C,E='d=="humn"and"1j"or'+A[:-5]+'h[0]','+%r+',"+')'"+E
B+="'('+"
a,_,b=e[s].split()
a,b=L(g(a)),L(g(b))
if a.imag==0:a,b=b,a
print(t,I((b-a.real)/a.imag))

[raw frigate]

[Day 21] 264 with rewrite ```py
D=[]
for i in open(0):
a,b=i.split(':')
if'humn'==a:H=int(b)
else:
if'root'==a:R,S=i.split()[1::2]
D+=a+'='+b,
humn=1j
root=0
while root==0:
for i in D:
try:exec(i)
except:0
r=eval(R+'-'+S)
print(f"{root.real+root.imag*H:.0f} {abs(r.real/r.imag):.0f}")

[Day 21] 263 ```py
D=[]
for i in open(0):a,b=i.split(':');a=='humn'and(H:=int(b))or(D:=D+[a+'='+b]);a=='root'and(R:=b[:5])and(S:=b[7:])
humn=1j
root=0
while root==0:
for i in D:
try:exec(i)
except:0
r=eval(R+'-'+S)
print(f"{root.real+root.imag*H:.0f} {abs(r.real/r.imag):.0f}")

[raw frigate]

[Day 21] 259 ```py
D=[]
for i in open(0):
a,b=i.split(':');D+=a+'='+(a=='humn'and(H:=int(b))and'1j'or b),
if'root'==a:R,S=i.split()[1::2]
root=0
while root==0:
for i in D:
try:exec(i)
except:0
r=eval(R+'-'+S)
print(f"{root.real+root.imag*H:.0f} {abs(r.real/r.imag):.0f}")

[sullen field]

[Day 21] 258 ```py
D=[]
for i in open(root:=0):
a,b=i.split(':');D+=a+'='+(a=='humn'and(H:=int(b))and'1j'or b),
if'root'==a:R,S=i.split()[1::2]
while root==0:
for i in D:
try:exec(i)
except:0
r=eval(R+'-'+S)
print(f"{root.real+root.imag*H:.0f} {abs(r.real/r.imag):.0f}")

[sullen field]

[Day 21] 256 ```py
D=[]
for i in open(root:=0):
z=i.split;a,b=z(':');D+=a+'='+(a=='humn'and(H:=int(b))and'1j'or b),
if'root'==a:R,S=z()[1::2]
while root==0:
for i in D:
try:exec(i)
except:0
r=eval(R+'-'+S)
print(f"{root.real+root.imag*H:.0f} {abs(r.real/r.imag):.0f}")

[urban torrent]

how are real+imag and real/imag even get to the result?

[sullen field]

no clue

[sullen field]

#1054990914147340288 message this ig

[urban torrent]

and why is root root and not any var. Is it cz of some exec eval shenanigans?

[sullen field]

yes

i've tried changing it to some shorter variable before

[urban torrent]

[Day 21] 252 Does this work for everyone? please check

D=[]
for i in open(root:=0):
 z=i.split;a,b=z(':');D+=a+'='+(a=='humn'and(H:=int(b))and'1j'or b),
 if'root'==a:R,S=z()[1::2]
while root==0:
 for i in D:
  try:exec(i)
  except:0
r=eval(R+'-'+S)
print(f"{root.real+root.imag*H:.0f} {-r.real/r.imag:.0f}")

[wraith crypt]

Invalid syntax

[urban torrent]

huh

how

[urban torrent]

use a different python version?

[wraith crypt]

It seems to be evaling a string with colons in it

What version are you using?

[urban torrent]

I haven't even changed that part. I changed the last line

[urban torrent]

3.10.8

[wraith crypt]

Doesn't work in 3.10.2 either

[urban torrent]

does this work for you?
#1050425859225436241 message

if not, the problem has propagated below

[wraith crypt]

Wait hold on

I messed up the C&P because of how neovim works -_- ignore me

[urban torrent]

[Day 21] 247

D=[]
for i in open(root:=0):
 z=i.split;a,b=z(':');D+=a+'='+(a=='humn'and(H:=int(b))and'1j'or b),
 if'root'==a:R,S=z()[1::2]
while root==0:
 for i in D:
  try:exec(i)
  except:0
r=eval(R+'-'+S)
print(int(root.real+root.imag*H),int(-r.real/r.imag))

[steep swift]

[Day 21] 218

import re
exec(t:=re.sub("\(\):","=lambda:",re.sub("([a-z]+)",r"\1()",open(0).read())))
print(int(root()))
exec(re.sub('[-+*/]','-',re.findall("root.*",t)[0]))
exec("humn=lambda:1j")
x=root()
print(int(-x.real/x.imag))```

[urban torrent]

wtf

[wraith crypt]

Yeah it gives the right answer when I run the right code

🙃

[urban torrent]

218 boisss wth

[wraith crypt]

WA

OBOE

[urban torrent]

works for me. That too 64 ms

[wraith crypt]

Not sure what it's doing to get the wrong answer but it's the wrong answer

[urban torrent]

apparently doesn't work for specific inputs 😔

[steep swift]

[Day 21] 220

import re
exec(t:=re.sub("\(\):","=lambda:",re.sub("([a-z]+)",r"\1()",open(0).read())))
print(int(root()))
exec(re.sub('[-+*/]','-',re.findall("root.*",t)[0]))
exec("humn=lambda:1j")
x=root()
print(round(-x.real/x.imag))

[wraith crypt]

It's an FPE LMAO

[urban torrent]

what's FPE

[wraith crypt]

Floating-point error

[sullen field]

floating point error

[urban torrent]

lol ok

[wraith crypt]

220 works though

[urban torrent]

are we allowing floating solutions? or should they be int-ed

[wraith crypt]

They should be inted yeah

(or in this case rounded)

[urban torrent]

[Day 21] 218

import re
G=print;H=exec
H(t:=re.sub("\(\):","=lambda:",re.sub("([a-z]+)",r"\1()",open(0).read())))
G(int(root()))
H(re.sub('[-+*/]','-',re.findall("root.*",t)[0]))
H("humn=lambda:1j")
x=root()
G(round(-x.real/x.imag))

[wraith crypt]

218, use round

[urban torrent]

wait, they can be rounded up as well? else 216 will not work

[wraith crypt]

(same problem as previous 218 -> 220)

[urban torrent]

ah done

[wraith crypt]

The reason the previous one didn't work (and the same reason yours didn't) is because the real / imag calculation can be just below the right answer instead of just above it

[sullen field]

[Day 21] 217 ```py
import re
G=print;H=exec
H(t:=re.sub("():",A:="=lambda:",re.sub("([a-z]+)",r"\1()",open(0).read())))
G(int(root()))
H(re.sub('[-+/]','-',re.findall("root.",t)[0]))
H("humn%s1j"%A)
x=root()
G(round(-x.real/x.imag))

[urban torrent]

was doing that LMAO

[wraith crypt]

(if the 0000000000003 bit wasn't there, int would work)

[urban torrent]

I still don't understand how the real/imag thingy works though.

seems like magic to me'

[wraith crypt]

It calculates a complex answer a - bi where a / b is the correct answer

[urban torrent]

ohh..i think I get it. damnnn

and why is root root?

and how is root() even working?

[cyan lynx]

a friend did 206 but he is just joining the server now

[wraith crypt]

It appears to be transpiling the input into a set of lambdas

Then running the one for root

[sullen field]

[Day 21] 215 ```py
import re
G=print;H=exec
H(t:=re.sub("():",A:="=lambda:",re.sub("([a-z]+)",r"\1()",open(0).read())))
G(int(root()))
H(re.sub('[-+*/]','-',re.findall("root.",t)))
H("humn%s1j"%A)
x=root()
G(round(-x.real/x.imag))

[echo quartz]

[Day 21] 206

import re
exec(t:=re.sub("..:","=lambda:",re.sub("([a-z]+)","\\1()",open(0).read())))
S=root()
exec(re.sub(r'(root=.*) . (.*)',r'\1 - \2',t))
humn=lambda:1j
z=root()
print(f"{S:.0f}\n{-z.real/z.imag:.0f}")

[sullen field]

nice

[echo quartz]

I tried to replace root with R, but no gain 😦

[wraith crypt]

Hi, welcome

[urban torrent]

Welcome to the server! 😄

..: wot-

[echo quartz]

Welcome, I used to perl golf in good old times, I still claim that perl is much better for golfing 😉

[wraith crypt]

(please use the format though)

[urban torrent]

the regexes are insane

[sullen field]

how does minus work

[Day 21] hehe 203 ```py
import re
exec(t:=re.sub("..:","=lambda:",re.sub("([a-z]+)","\1()",open(0).read())))
S=root()
exec(re.sub(r'(root=.) . (.)',r'\1-\2',t))
humn=lambda:1j
z=root()
print(f"{S:.0f}\n{-z.real/z.imag:.0f}")

[sullen field]

that's only in sets

i mean the minus in the real code

[wraith crypt]

It's comparing a and b

[sullen field]

oh so like assembly cmp

ok

[echo quartz]

a == b -> a - b == 0

[urban torrent]

[Day 21] 199 Sub 200!

import re
exec(t:=re.sub("..:","=lambda:",re.sub("([a-z]+)","\\1()",open(0).read())))
S=root()
exec(re.sub(r'(root=.*) . (.*)',r'\1-\2',t))
humn=lambda:1j
z=root()
print(int(S),round(-z.real/z.imag))

[wraith crypt]

Is using py2-style formatting shorter here? I feel like it might be

Or that

[urban torrent]

would you still need round-ing? or does inting work?

[wraith crypt]

.0f does the rounding

[urban torrent]

ah

[echo quartz]

Isn't the newline required between p1 and p2? Here's just space

[urban torrent]

no need until you get the answer

[sullen field]

as long as it displays the correct input it works in here

[wraith crypt]

Any whitespace is considered valid

[sullen field]

starwort's repo does that rule though ^

[steep swift]

[Day 21] 193

import re
a=re.sub
exec(t:=a("..:","=lambda:",a("([a-z]+)","\\1()",open(0).read())))
S=root()
exec(a(r'(root=.*) . (.*)',r'\1-\2',t))
humn=lambda:1j
z=root()
print(int(S),round(-z.real/z.imag))

[urban torrent]

beat me to it, ahah

[sullen field]

[Day 21] 192 ```py
import re
a=re.sub
exec(t:=a("..:","=lambda:",a("(\w{4})","\1()",open(0).read())))
S=root()
exec(a(r'(root=.) . (.)',r'\1-\2',t))
humn=lambda:1j
z=root()
print(int(S),round(-z.real/z.imag))

[wraith crypt]

Well that's disappointing

[sullen field]

what's your input

[scarlet grove]

different 192, but skipping the int() and assuming that only root starts with ro

[sullen field]

because it better not have numbers in the thousands

[urban torrent]

do you want the file?

[echo quartz]

I thought about it but considered it cheating. Or - gambling 😉

[scarlet grove]

works for my input 😛

[sullen field]

ok

[urban torrent]

https://paste.pythondiscord.com/uhevewagov

[echo quartz]

so does print(31017034894002, 3555057453229) for mine 😛

[urban torrent]

kekw

[steep swift]

@scarlet grove raw is not required for the first string on line 4

[scarlet grove]

doesn't work for me though 😛

[echo quartz]

OK, it was nice to meet you, I have to leave for now, I'll drop later to see your sub 100 😉

[sullen field]

goddamn it ```
humn: 2037

why do you have to be different

[urban torrent]

lol

[wraith crypt]

Lmao

[sullen field]

i have 831 as humn

[Day 21] 192 ```py
import re
a=re.sub
exec(t:=a("..:","=lambda:",a("([a-z]+)","\1()",open(0).read())))
S=root()
exec(a('(root=.) . (.)',r'\1-\2',t))
humn=lambda:1j
z=root()
print(int(S),round(-z.real/z.imag))

[wraith crypt]

2116 here

[sullen field]

ok why do i have to be different

[wraith crypt]

Lmao

[scarlet grove]

189 works for my input, but that's because my humn is 28

exec(t:=re.sub("..:","=lambda:",re.sub("(\w{4})","\\1()",open(0).read())))
S=root()
exec(re.sub('(ro.*). (.*)',r'\1-\2',t))
humn=lambda:1j
z=root()
print(S,round(-z.real/z.imag))```

[sullen field]

[Day 21] 189 assuming no other vars start with ro but humn can be 1000+ ```py
import re
a=re.sub
exec(t:=a("..:","=lambda:",a("([a-z]+)","\1()",open(0).read())))
S=root()
exec(a('(ro.) . (.)',r'\1-\2',t))
humn=lambda:1j
z=root()
print(int(S),round(-z.real/z.imag))

[urban torrent]

so which one works for all inputs?

[sullen field]

test the 190

[wraith crypt]

This one I think

[urban torrent]

int it

for the last line

[sullen field]

o

fixed

[scarlet grove]

works for me

[urban torrent]

works

[sullen field]

nice

[urban torrent]

real/imag is the slope of a complex number. wish there was a function for that 😔

[urban torrent]

what? i don't get it

[sharp scarab]

real/imag is the multiplicative inverse of that

and you do kinda have a formula for that

tan(atan2(y,x))

or tan(arg(z))

[urban torrent]

but you would need to call, numpy or math for it

which will increase char length

[raw frigate]

if only you could just unpack real and imag values like a tuple

[sharp scarab]

it's literally just y/x tbh

atan2 is just arctan with its range modified to fit convention

[sharp scarab]

this, and numpy support for indexing 2d arrays with complex

and complex ints

[scarlet grove]

it's using the fact that python will resolve dependencies between functions by itself.

so if you create a bunch of functions with proper names, it will know in what order to call them.

@urban torrent using @echo quartz code from another discord, it works, because this works: ```python
def f1():
return f2() + f3()
def f2():
return 1
def f3():
return 5

f1()

[echo quartz]

TBH, it's @steep swift 's code, which I only optimized

[sullen field]

python uses dynamic scoping iirc

[echo quartz]

It's most about lazy resolving of object references - there's not much scope here, all is global.

[sullen field]

‫so it's literally just loading the variable every time the variable is referenced

[steep swift]

[Day 21] 180, same assumptions as above

import re
a=re.sub
exec(t:=a(":","=lambda:",a("( [a-z]+)","\\1()",open(0).read())))
S=root()
exec(a('(ro.+). ',r'R=\1-',t))
humn=lambda:1j
z=R()
print(int(S),round(-z.real/z.imag))

[sullen field]

[Day 21] this is like the most ridiculous shortening ever but it works, 178 ```py
import re
a=re.sub
exec(t:=a(":","=lambda:",a("( [a-z]+)","\1()",open(0).read())))
S=root()
exec(a('(ro.+). ',r'R=\1-',t))
humn=lambda:1j
print(int(S),round(-R().real/R().imag))

[urban torrent]

this doesn't even make sense anymore. How did you even figure it out?

remind me what \2 was?

[sullen field]

i literally just removed z=R() and its two references and replaced them with R()

‫i'm currently working on a one-exec version

[urban torrent]

but how did you R the root?

[echo quartz]

When replacing operator to minus in root=v1 op v2 he takes a chance to also replace root with R - very clever

[sullen field]

pkubiak's idea

[urban torrent]

why can't we assign S the same way then?

[echo quartz]

Yep, I just tried this but realized in the same point it changes operator to '-' so it would require two replacements, no gain.

[sullen field]

holy moly

[Day 21] 176, same assumptions ```py
import re
a=re.sub
A='=lambda:'
t=a(":",A,a("( [a-z]+)","\1()",open(0).read()))+';'
exec(t+a('ro.+:(.+). ',r'S=root();humn%s1j;R=\1-'%A,t))
print(int(S),round(-R.real/R.imag))

no way i just made a breakthrough at night

[sharp scarab]

night is the best time for golf

in fact it's the only time for golf

you do aoc during the day and golf at night

this channel is like a speakeasy but for golf

we got booted from the public and now we have to gather here in private

[urban torrent]

[Day 21] 174
idk what I changed but seems like this works. I'm on phone and this is 174 chars

import re
a=re.sub
A='=lambda:'
t=a(":",A,a("( [a-z]+)","\\1()",open(0).read()))+';'
exec(t+a('ro.+:(.+). ',r'S=root();humn%s1j;R=\1-'%A,t))
print(int(S),int(-R.real/R.imag))

[sullen field]

wait

that means

[Day 21] 172 ```py
import re
a=re.sub
A='=lambda:'
t=a(":",A,a("( [a-z]+)","\1()",open(0).read()))+';'
exec(t+a('ro.+:(.+). ',r'humn%s1j;R=\1-'%A,t))
print(int(root()),round(-R.real/R.imag))

[echo quartz]

I'm pretty sure we don't need that t anymore, just cannot implement this

[sharp scarab]

you changed round to int

[urban torrent]

ohh, i'm dumb, that won't work

[sullen field]

he changed the wrong thing lmao

[sharp scarab]

lmao

[sullen field]

i thought he was the one that did root() but it turns out i had it partially done on my text editor

[sharp scarab]

I have no idea what the golf is btw I'm still doing my day 21

would golf but hotel has shit internet

[sullen field]

bruh

[urban torrent]

it ain't working what can I do lol

back to 176

[sullen field]

oh i found out why

[urban torrent]

the root() would've have updated value though, right?

[sullen field]

‫no i mean technically updated return value but not updated code

‫humn updated code

[sullen field]

this is crazy short

day 20 didn't even go that far

it's still 11 bytes away

[sharp scarab]

it's shorter than day 3

[sharp scarab]

you should put a space between the Day and 21 just for a consistent format to make it easier to search to put into the github

[sullen field]

i don't think it is

[sharp scarab]

the github says day 3 is 201

[urban torrent]

day 3 was 201

[sullen field]

that's the github

[sharp scarab]

is there a shorter golf floating around?

[urban torrent]

oh it isn't updated my bad

[sullen field]

day 3 pins

[urban torrent]

shortest is 156

[sharp scarab]

ah ok

i only joined the threads like day 9

[sharp scarab]

why doesn't this work again

does it fail on some inputs

[urban torrent]

@wraith crypt's input was rounding up

some FPE

[sharp scarab]

ah

[sharp scarab]

what original operator do you guys have for root?

[sharp scarab]

if everyone has plus we can get rid of a character

also tessa

[Day 21] 175

import re
a=re.sub
t=a(':',A:='=lambda:',a('( [a-z]+)','\\1()',open(0).read()))+';'
exec(t+a('ro.+:(.+). ',fr'S=root();humn{A}1j;R=\1-',t))
print(int(S),round(-R.real/R.imag))

actually lemme just edit a few things to make it cleaner

changed the " to ' for consistency

shame no f string walruses

[orchid charm]

wtf happened while i was asleep

y'all took my idea and ran with it lol

[sharp scarab]

see this is why we need u here next year

u come up with the good golf ideas

[sharp scarab]

[Day 21] Tentative 174, assumes the original root operator is +, ping me if yours isn't

import re
a=re.sub
t=a(':',A:='=lambda:',a('( [a-z]+)','\\1()',open(0).read()))+';'
exec(t+a('ro.+:(.+) ',fr'S=root();humn{A}1j;R=\1-',t))
print(int(S),round(-R.real/R.imag))

[echo quartz]

So you're the one who got the idea of using complex numbers?

[orchid charm]

yeah

[echo quartz]

kudos

[echo quartz]

[Day 21] Tentative 170 (and actually yours doesn't work for me due to this extra semicolon)

import re
a=re.sub
t=a(':',A:='=lambda:',a('( [a-z]+)','\\1()',open(0).read()))
exec(t+a('ro.+:(.+) ',fr'S=root();humn{A}1j;R=\1-',t))
print(int(S),round(-R.real/R.imag))

[orchid charm]

the extra semicolon is from the trailing newline

[sharp scarab]

extra semicolon?

[orchid charm]

i think, at least

[echo quartz]

t=...+';'

[raw frigate]

doesn't work for me

[sharp scarab]

yeah

mine needs the semicolon at the end

[orchid charm]

without the semicolon you need a trailing newline

[echo quartz]

$ python3 golf2 <input-d21.txt 
Traceback (most recent call last):
  File "/home/kompas/git/aoc/2022/golf2", line 4, in <module>
    exec(t+a('ro.+:(.+) ',fr'S=root();humn{A}1j;R=\1-',t))
  File "<string>", line 2284
    ;pmqg=lambda: 2
    ^
SyntaxError: invalid syntax

[sharp scarab]

Traceback (most recent call last):
  File "/Users/stickie/PycharmProjects/AdventOfCode2022/Tools/golf.py", line 4, in <module>
    exec(t+a('ro.+:(.+) ',fr'S=root();humn{A}1j;R=\1-',t))
  File "<string>", line 2621
    rbzv=lambda: gzpt() - spng()sbdn=lambda: bctf() + msbv()
                                ^^^^
SyntaxError: invalid syntax

[orchid charm]

y'all this is because of the trailing newline or lack thereof

[echo quartz]

Well, I'm getting input with wget - so the question is whos input.txt is wrong?

[sharp scarab]

we usually do no trailing newline i think

[orchid charm]

yeah

[sharp scarab]

apparently that's what aoc does

[echo quartz]

INPUT="input-d${DAY}.txt"
curl "https://adventofcode.com/$YEAR/day/$DAY/input" \
  -o $INPUT \
  -H 'authority: adventofcode.com' \
  -H 'cache-control: max-age=0' \
  -H "cookie: ${COOKIE}" \
  --compressed

[orchid charm]

wasn't my decision

[echo quartz]

Is it curl adding missing newline at the end?

[sharp scarab]

idk

i just copy and paste to a txt file

[echo quartz]

Are there any rules for the golf? I think that should be clarified

[orchid charm]

yeah, we have inputs without trailing newlines for golfing

even though the actual inputs do have trailing newlines (at least for me)

[echo quartz]

ok, a bit strange but ACK

[sharp scarab]

we've had this issue before

[orchid charm]

just preaching to the choir lol

[sharp scarab]

so since we're sticking to (int) space (int) format, are we ever gonna go back and fix that one day where the answer was given as (pt1)+(pt2)j

[echo quartz]

How about feature request to python to make complex iterables? 😉 then we could just print(*z)

[sharp scarab]

you'd need complex ints as well

or gaussian ints would be the proper name for it

[echo quartz]

BTW, if the answer is 123 then is 123.0 correct? Would aoc accept it or not?

[sharp scarab]

we've had to convert to ints in the past

honestly i should just write my gaussian integer class one day

i'd have so many uses for it

except it'd be slow af

[echo quartz]

But it wouldn't work for today, as we had non-integer divisions?

[sharp scarab]

yeah

i mean just in general

[urban torrent]

plus for what?

[urban torrent]

works for me

[echo quartz]

$ grep root: input-d21.txt 
root: qntq + qgth

[echo quartz]

Yep, we discussed it - I have input.txt with final newline, as for some reason the rule is it shouldn't be there.

[urban torrent]

yeah, because the input given in AOC doesn't have it

[echo quartz]

orly?

...unless you mean some different "given in AOC"

(or maybe they fixed it lately - IDK)

Actually, any simple (linux) way to strip that last final newline so I can test my script according to the rules?

[wraith crypt]

printf "$(cat INPUT_FILE)" | py my_program.py should do it

I think

[echo quartz]

Yep, works, ty

[sharp scarab]

if your root was originally like
root: abcd + wxyz

[echo quartz]

Well... Nobody even tries today's task?

[sharp scarab]

I haven't done it yet lmao

behind as usual

[echo quartz]

What time zone are you? I'm pretty sick (literally) of getting up before 6am for last 3 weeks

[sharp scarab]

I'm in jst rn

but since I'm on holiday I'm like away for most of the day

[urban torrent]

same here. the question is lacking the fun element. It's just tedious and boring.

I recently discovered code.golf. Really fun website. Have been doing a few of their problems in the last few days.

[orchid charm]

had to travel today

so i couldn't

actually i'll be traveling for the next few days

up until end of advent

[echo quartz]

Strange. Looks like there's a life beyond AoC?

[sharp scarab]

if no one else is doing a golf i can try to turn mine into one

tho it'd probably be huge and not a good golf

[sharp scarab]

yeah even just my parsing is 347 characters lmao

[sullen field]

re?

[sharp scarab]

import re,itertools as i
p,r,e=i.pairwise,range,enumerate
S=re.findall('\d+|L|R',(f:=open(0).read().split('\n\n'))[0])
D=sorted((min(map(lambda z:(z.real,z.imag),d)),d)for n,x in p(r(0,201,50))for m,y in p(r(0,151,50))if(d:={i:v for i,v in((x+y*1j,v=='#')for x,r in e(f[1].split('\n'))for y,v in e(r)if v in'.#')if n<=i.real<x and m<=i.imag<y}))

that's literally just converting the stuff in my parser to golf

and it's already chonk

probably better to use a 3D solution i think

[sharp scarab]

anyone have any idea how to extract a pattern from this for golf

(x,y) refer to complex numbers btw

i just had them written out in vector form for readability

[raw frigate]

[Day 23 part 2] works for example but not actual input ```py
e=enumerate
j=1j
N={x+y*j for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=[-1-j,-j,1-j,1,1+j,j,j-1,-1,-1-j]
D=[A[2:5],A[:3],A[4:7],A[6:]]
while N^E:
E={*N};N=set();D=D[1:]+D[:1];P={}
for e in E:
if len(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),0)):P[l]=*P.get(l,()),e
else:N|={e}
for l,e in P.items():N|=[{l},{*e}][len(e)-1]
if n==9:...#do part 1 here
n+=1
print(n)

[sharp scarab]

ah yes

[raw frigate]

import numpy as np
from collections import defaultdict

data = open(0).read()

grid = np.array([[i == '#' for i in line] for line in data.splitlines()], bool)

new_elves = set(int(j) + 1j * int(i) for i, j in zip(*np.nonzero(grid)))
elves = set()

adj = {
    'NW': -1 - 1j,
    'N': - 1j,
    'NE': + 1 - 1j,
    'W': - 1,
    'E': + 1,
    'SW': + (-1 + 1j),
    'S': + 1j,
    'SE': + (1 + 1j)
}


propose = [
    ['E', {'NE', 'E', 'SE'}],
    ['N', {'NW', 'N', 'NE'}],
    ['S', {'SW', 'S', 'SE'}],
    ['W', {'NW', 'W', 'SW'}],

]

n = 0
while elves ^ new_elves:
    elves = new_elves
    propose.append(propose.pop(0))
    proposals = defaultdict(set)
    new_elves = set()
    for elf in elves:
        adjacent = frozenset(k for k, v in adj.items() if elf+v in elves)
        if len(adjacent) == 0:
            new_elves.add(elf)
            continue
        for k, v in propose:
            if not (v & adjacent):
                break
        else:
            new_elves.add(elf)
            continue
        proposals[elf + adj[k]].add(elf)
    for loc, e in proposals.items():
        if len(e) == 1:
            new_elves.add(loc)
        else:
            new_elves |= e
    n += 1
    if n == 10:
        min_real = float('inf')
        max_real = 0
        min_imag = float('inf')
        max_imag = 0

        for elf in new_elves:
            r, i = int(elf.real), int(elf.imag)
            min_real = min(min_real, r)
            max_real = max(max_real, r)
            min_imag = min(min_imag, i)
            max_imag = max(max_imag, i)

        print((max_imag - min_imag + 1) * (max_real - min_imag + 1) - len(elves))
print(n)
```original logic if anyone is interested

[sharp scarab]

good ol' if n==9: ... #do part 1 here

great for golfing

reduces the character count drastically

[raw frigate]

yeah

I made it do part 2 but it breaks on actual input for whatever reason

[orchid charm]

should use if n==9:0

to save chars

[raw frigate]

I'll take a look later to see what the problem is

[sharp scarab]

i'm still debugging my "golfed" day 22

[raw frigate]

but idk what makes it work for example but not actual

[sharp scarab]

sitting at a nice and comfy 1572 characters

someone better come up with a better golf than this monstrosity

it's gonna be a stain on the github

[orchid charm]

lol

[sharp scarab]

bruh

my table was wrong this entire time

[raw frigate]

[Day 23 part 2] 393 ```py
e=enumerate
j=1j
N={x+y*j for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=[-1-j,-j,1-j,1,1+j,j,j-1,-1,-1-j]
D=[A[2:5],A[:3],A[4:7],A[6:]]
while N^E:
E={*N};N=set();D=D[1:]+D[:1];P={}
for e in E:
if len(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),''))!='':P[l]=*P.get(l,()),e
else:N|={e}
for l,e in P.items():N|=[{*e},{l}][len(e)==1]
n+=1
print(n)

[sharp scarab]

so uh

[Day 22] 1212

import re,itertools as i
p,R,e,U=i.pairwise,range,enumerate,complex
J=1j
S=re.findall('\d+|L|R',(F:=open(0).read().split('\n\n'))[1])
*D,=map(lambda x:(U(*x[0]),x[1]),sorted((min(map(lambda z:(z.real,z.imag),d)),d)for n,x in p(R(0,201,50))for m,y in p(R(0,151,50))if(d:={i:v for i,v in((U(x,y),v=='.')for x,r in e(F[0].split('\n'))for y,v in e(r)if v in'.#')if n<=i.real<x and m<=i.imag<y})))
L=-1,J,1,-J
def g(a):
 global S,D
 s,w,f=*D[c:=0],J;o=s
 for v in S:
  if v in'LR':f*={'L':J,'R':-J}[v];continue
  for _ in R(int(v)):
   if(z:=s+f)not in w:C,F=a[c][f];O,W=D[C];r,i=(x:=s-o).real,x.imag;f,o,c,w,s=eval(['f,o,c,w,s','F,O,C,W,z'][W[z:=O+49*U(F==-1,F==-J)+(((1-2*(b:=(F.imag>F.real)^(f.imag>f.real)))*(i if f.real else r))+49*b)*{-1:J,J:1,1:J,-J:1}[F]]])
   elif w[z]:s=z
 return int(sum(x*y for x,y in zip((1000,4,1),(s.real+1,s.imag+1,{J:0,1:1,-J:2,-1:3}[f]))))
print(g([{i:(v,i)for i,v in zip(L,j)}for j in((4,1,2,1),(1,0,1,0),(0,2,4,2),(5,4,5,4),(2,3,0,3),(3,5,3,5))]),g([{-1:(5,J),J:(1,J),1:(2,1),-J:(3,J)},{-1:(5,-1),J:(4,-J),1:(2,-J),-J:(0,-J)},{-1:(0,-1),J:(1,-1),1:(4,1),-J:(3,1)},{-1:(2,J),J:(4,J),1:(5,1),-J:(0,J)},{-1:(2,-1),J:(1,-J),1:(5,-J),-J:(3,-J)},{-1:(3,-1),J:(4,-1),1:(1,1),-J:(0,1)}]))

can someone check that

[raw frigate]

[Day 23] 491 wrong p1 ```py
e=enumerate
j=1j
N={x+yj for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=[-1-j,-j,1-j,1,1+j,j,j-1,-1,-1-j]
D=[A[2:5],A[:3],A[4:7],A[6:]]
while N^E:
if n==9:r,i=zip(((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
n+=1;E={*N};N=set();D=D[1:]+D[:1];P={}
for e in E:
if len(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),''))!='':P[l]=*P.get(l,()),e
else:N|={e}
for l,e in P.items():N|=[{*e},{l}][len(e)==1]
print(S,n)

[sharp scarab]

nice

[raw frigate]

works for me

[sharp scarab]

that's good

[sharp scarab]

I tried to turn the function into a lambda but couldn't get past the if elif

[sullen field]

i might actually be able to take my partially golfed solution and golf it

nvm it's too long

[sullen field]

‫works for me but gives WA for part 1 (fixed by doing n==10)

[Day 23] 484 ```py
e=enumerate
j=1j
N={x+yj for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=[-1-j,-j,1-j,1,1+j,j,j-1,-1,-1-j]
D=[A[2:5],A[:3],A[4:7],A[6:]]
while N^E:
if n==10:r,i=zip(((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
n+=1;E={*N};N=E^E;D=D[1:]+D[:1];P={}
for e in E:
if len(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P:P[l]=*P.get(l,()),e
else:N|={e}
for l in P:N|=[{*P[l]},{l}][len(P[l])==1]
print(S,n)

[Day 23] 483 with a dangerous assumption ```py
e=enumerate
j=1j
N={x+yj for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=[-1-j,-j,1-j,1,1+j,j,j-1,-1,-1-j]
D=[A[2:5],A[:3],A[4:7],A[6:]]
while N^E:
if n==10:r,i=zip(((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
n+=1;E={*N};N=E^E;D=D[1:]+D[:1];P={}
for e in E:
if len(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),j))-j:P[l]=*P.get(l,()),e
else:N|={e}
for l in P:N|=[{*P[l]},{l}][len(P[l])==1]
print(S,n)

[raw frigate]

p2 doesn't work for me

[raw frigate]

[Day 23] 481 ```py
e=enumerate
j=1j
N={x+yj for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
D=[A[4:7],A[2:5],A[6:],A[:3]]
while N^E:
if n==10:r,i=zip(((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
n+=1;E={*N};N=E^E;D=D[1:]+D[:1];P={}
for e in E:
if len(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P:P[l]=*P.get(l,()),e
else:N|={e}
for l in P:N|=[{*P[l]},{l}][len(P[l])==1]
print(S,n)

[orchid charm]

[Day 23] 480

e=enumerate
j=1j
N={x+y*j for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
D=[A[4:7],A[2:5],A[6:],A[:3]]
while N^E:
 if n<11:r,i=zip(*((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
 n+=1;E={*N};N=E^E;D=D[1:]+D[:1];P={}
 for e in E:
  if len(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P:P[l]=*P.get(l,()),e
  else:N|={e}
 for l in P:N|=[{*P[l]},{l}][len(P[l])==1]
print(S,n)```

[sharp scarab]

is E^E not the empty set?

[raw frigate]

N^E?

[sharp scarab]

N=E^E in the first if statement

[raw frigate]

oh

yeah

we need to clear N

[sharp scarab]

ah

[sullen field]

[Day 23] 479 ```py
e=enumerate
j=1j
N={x+yj for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
D=[A[4:7],A[2:5],A[6:],A[:3]]
while N^E:
if n<11:r,i=zip(((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
n+=1;E={*N};N^=N;D=D[1:]+D[:1];P={}
for e in E:
if len(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P:P[l]=*P.get(l,()),e
else:N|={e}
for l in P:N|=[{*P[l]},{l}][len(P[l])==1]
print(S,n)

[Day 23] 478 ```py
e=enumerate
j=1j
N={x+yj for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
D=[A[4:7],A[2:5],A[6:],A[:3]]
while N^E:
if n<11:r,i=zip(((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
n+=1;E={*N};N^=N;D=D[1:]+D[:1];P={}
for e in E:
if len(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P:P[l]={*P.get(l,()),e}
else:N|={e}
for l in P:N|=[P[l],{l}][len(P[l])==1]
print(S,n)

[sharp scarab]

does [{l},P[l]][len(P[l]>1] work?

[raw frigate]

yeah it should

[sharp scarab]

I'm on mobile someone else do it

[raw frigate]

I can't do it rn either

[sharp scarab]

!e

x = [1,2,3]
print(x)
_,*x=x+x[:1]
print(x)

[tawny dawn]

@sharp scarab :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | [1, 2, 3]
002 | [2, 3, 1]

[sharp scarab]

that saves another char

[lapis rover]

surely len(a:={i for i in A if e+i in E}) could just be (a:={i for i in A if e+i in E})?

[sharp scarab]

!e

x,y=39,-7
print(x-y+1)
print(-y-~x)
z,w=65,9
print((x-y+1)*(z-w+1))
print((y+~x)*(w+~z))

[tawny dawn]

@sharp scarab :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | 47
002 | 47
003 | 2679
004 | 2679

[sharp scarab]

probably

[raw frigate]

yeah that too

[sharp scarab]

we really need someone on a computer to get these changes

[raw frigate]

it used to have a >0 and I just missed that

[sullen field]

well what if it's 0

[raw frigate]

it can't be 0

[sharp scarab]

it's for stuff in P

if it was 0 it wouldn't be in P I think

[lapis rover]

btw, it also can't be more than 2, idk if that helps with anything

[raw frigate]

I don't think that's guaranteed

[lapis rover]

only 2 elves can ever propose the same square

[raw frigate]

you can have 3 or 4 in some cases

[lapis rover]

pretty sure you can't

[raw frigate]

oh wait

yeah

[lapis rover]

there are 4 elves around each square. But if there are two at right angles, they will block eachother

they only way you can have two is if they are opposite

[sullen field]

[Day 23] 473 ```py
e=enumerate
j=1j
N={x+yj for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
D=[A[4:7],A[2:5],A[6:],A[:3]]
while N^E:
if n<11:r,i=zip(((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
n+=1;E={*N};N^=N;D=D[1:]+D[:1];P={}
for e in E:
if(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P:P[l]={*P.get(l,()),e}
else:N|={e}
for l in P:N|=[{l},P[l]][len(P[l])>1]
print(S,n)

is there anything else

[sharp scarab]

both of mine

[sharp scarab]

^

[sharp scarab]

^

[sullen field]

ok

[lapis rover]

[{l},P[l]][len(P[l])>1]
[0,P[l],{l}][len(P[l])] only breaks even 😔

[sullen field]

[Day 23] 472 ```py
e=enumerate
j=1j
N={x+yj for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
D=[A[4:7],A[2:5],A[6:],A[:3]]
while N^E:
if n<11:r,i=zip(((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
n+=1;E={*N};N^=N;_,*D=D+D[:1];P={}
for e in E:
if(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P:P[l]={*P.get(l,()),e}
else:N|={e}
for l in P:N|=[{l},P[l]][len(P[l])>1]
print(S,n)

[sullen field]

‫they're floats

[sharp scarab]

oh gdi

stupid floats

well keep that optimization in mind for when we eventually switch from complex to base whatever

[raw frigate]

[Day 23] 465 ```py
e=enumerate
j=1j
N={x+yj for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
D=[A[4:7],A[2:5],A[6:],A[:3]]
while N^E:
if n<11:r,i=zip(((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
n+=1;E={*N};N^=N;_,*D=D+D[:1];P={}
for e in E:
if(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P:
if l in P:N|={e,P.pop(l)}
else:P[l]=e
else:N|={e}
N|={*P}
print(S,n)

[lapis rover]

[Day 23] 459

e=enumerate
j=1j
N={x+y*j for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={(n:=0)}
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
D=[A[4:7],A[2:5],A[6:],A[:3]]
while N^E:
 if n<11:r,i=zip(*((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
 n+=1;E={*N};D=D[1:]+D[:1];P={}
 for e in E:
  if(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P:
   if l in P:del P[l]
   else:P[l]=e
 N|={*P};N^={*P.values()}
print(S,n)

[sharp scarab]

lmao

[raw frigate]

smh

[lapis rover]

seems to be about half the speed

lmao

great minds

I am sad to see the xor-reset thing go

[sharp scarab]

rip xor reset

you will be missed

[lapis rover]

btw it is used in real life assembly to reset a register without loading a constant 0

[sharp scarab]

shame you can't subtract dictionaries

huh

TIL

[raw frigate]

[Day 23] 457 ```py
e=enumerate
j=1j
N={x+yj for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={(n:=0)}
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
D=[A[4:7],A[2:5],A[6:],A[:3]]
while N^E:
if n<11:r,i=zip(((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
n+=1;E={*N};D=D[1:]+D[:1];P={}
for e in E:
if(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P:
if l in P:del P[l]
else:P[l]=e
N={*P}|N^{*P.values()}
print(S,n)

[sharp scarab]

oi what happened to my _,*D=D+D[:1]

[orchid charm]

i ate it

[sharp scarab]

damn

stop eating our golfs

[orchid charm]

sorry

[sharp scarab]

[Day 23] 456 ```py
e=enumerate
j=1j
N={x+yj for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={(n:=0)}
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
D=[A[4:7],A[2:5],A[6:],A[:3]]
while N^E:
if n<11:r,i=zip(((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
n+=1;E={*N};_,*D=D+D[:1];P={}
for e in E:
if(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P:
if l in P:del P[l]
else:P[l]=e
N={*P}|N^{*P.values()}
print(S,n)

should be 456

someone check

[lapis rover]

hahh

wait

I changed E={n:=0} to E={(n:=0)} because I was testing on an old version of python and forgot to remove it

[sharp scarab]

your golf was just so good that despite all that it was still better

so none of us noticed

[lapis rover]

lol

tio.run. A great site but hasn't update python since 3.8 pre-release (which it only includes since the site is built for golfing, and 3.8pre is the first version with walrus)

[sullen field]

so 454?

[Day 23] 454 ```py
e=enumerate
j=1j
N={x+yj for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
D=[A[4:7],A[2:5],A[6:],A[:3]]
while N^E:
if n<11:r,i=zip(((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
n+=1;E={*N};_,*D=D+D[:1];P={}
for e in E:
if(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P:
if l in P:del P[l]
else:P[l]=e
N={*P}|N^{*P.values()}
print(S,n)

[sharp scarab]

if we haven't missed anything then yeah

[sullen field]

i'm surprised ```py
if l in P:del P[l]
else:P[l]=e

[lapis rover]

it's because only two elves can propose the same move

as discussed previously

[sharp scarab]

is there a filler value we could fill P[l] with that wouldn't mess with the xor

[lapis rover]

that would be very cool

tbh i was very surprised it was a byte saving

with that extra level of indentation, .values and everything

[sullen field]

[Day 23] 452 ```py
e=enumerate
j=1j
N={x+yj for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
D=[A[4:7],A[2:5],A[6:],A[:3]]
while N^E:
if n<11:r,i=zip(((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
n+=1;E={*N};_,*D=D+D[:1];P={}
for e in E:
if(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P:exec(("P[l]=e","del P[l]")[l in P])
N={*P}|N^{*P.values()}
print(S,n)

[lapis rover]

hah, right. nice

[sharp scarab]

!e

x={i:v for i,v in zip(range(10),'abcdefghij')}
print(x)
print({*x})

[tawny dawn]

@sharp scarab :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | {0: 'a', 1: 'b', 2: 'c', 3: 'd', 4: 'e', 5: 'f', 6: 'g', 7: 'h', 8: 'i', 9: 'j'}
002 | {0, 1, 2, 3, 4, 5, 6, 7, 8, 9}

[sharp scarab]

k that's what I expected

[orchid charm]

[Day 23] 449

e=enumerate
j=1j
N={x+y*j for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
D=[A[4:7],A[2:5],A[6:],A[:3]]
while N^E:
 if n<11:r,i=zip(*((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
 n+=1;E={*N};_,*D=D+D[:1];P={};[exec(("P[l]=e","del P[l]")[l in P])for e in E if(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P]
 N={*P}|N^{*P.values()}
print(S,n)```

[sullen field]

[Day 23] 445 ```py
e=enumerate
j=1j
N={x+yj for y,l in e(open(0))for x,c in e(l)if'#'==c}
E={n:=0}
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
D=[A[4:7],A[2:5],A[6:],A[:3]]
while N^E:
if n<11:r,i=zip(((e.real,e.imag)for e in N));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(N))
n+=1;E={*N};_,*D=D+D[:1];P={};[exec(("P[l]=e","del P[l]")[l in P])for e in E if(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P];N^={*P,*P.values()}
print(S,n)

[lapis rover]

if'$'>c would work if not for newlines 😔

[Day 23] 433

e=enumerate
j=1j
E={x+y*j for y,l in e(open(0))for x,c in e(l)if'#'==c}
n=0
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
P=D=[A[4:7],A[2:5],A[6:],A[:3]]
while P:
 if n<11:r,i=zip(*((e.real,e.imag)for e in E));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(E))
 n+=1;_,*D=D+D[:1];P={};[exec(("P[l]=e","del P[l]")[l in P])for e in E if(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P];E^={*P,*P.values()}
print(S,n)```

[sharp scarab]

rip N

[lapis rover]

if we're going to do mutability we may as well do it properly

[sharp scarab]

does P[l]=[l,e][l in P] work

[lapis rover]

ooo good point

[orchid charm]

not terminating when i run

with P[l]=[l,e][l in P]

[sharp scarab]

huh

[lapis rover]

should be e,l i think

oh nah i don't think it's valid

[sharp scarab]

oh yeah true

[lapis rover]

since deleting actually removes the key

[sharp scarab]

yeah that too

[lapis rover]

buuut

maybe if we separate the add and remove again

since its quite a big saving

[orchid charm]

[Day 23] 431

e=enumerate
j=1j
E={x+y*j for y,l in e(open(0))for x,c in e(l)if'#'==c}
n=0
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
P=D=A[4:7],A[2:5],A[6:],A[:3]
while P:
 if n<11:r,i=zip(*((e.real,e.imag)for e in E));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(E))
 n+=1;_,*D=D+D[:1];P={};[exec(("P[l]=e","del P[l]")[l in P])for e in E if(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P];E^={*P,*P.values()}
print(S,n)```

idk why you added the square brackets

[lapis rover]

nice

e=enumerate
j=1j
E={x+y*j for y,l in e(open(0))for x,c in e(l)if'#'==c}
n=0
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
P=D=A[4:7],A[2:5],A[6:],A[:3]
while P:
 if n<11:r,i=zip(*((e.real,e.imag)for e in E));S=int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(E))
 n+=1;_,*D=D+D[:1];P={};[exec("P[l]=[e,l][l in P]")for e in E if(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D if not{*i}&a),P))!=P];E|={*P};E^={*P.values()}
print(S,n)```this doesn't work but I'm not sure why

[orchid charm]

don't the keys need to be removed

[lapis rover]

Yes, but if something is both a key and a value then E|={*P};E^={*P.values()} should add it and remove it again, the same as if it was not there at all

[orchid charm]

but it's while P

[lapis rover]

ahhh

right

thanks

[sharp scarab]

surely there's a better way to do A

[orchid charm]

tried range, too many chars

[sharp scarab]

I was thinking using powers of (1+i) but normalizing is a pain

[orchid charm]

how do you do this without range?

[sharp scarab]

well that does use range

[orchid charm]

yeah

i tried the same

but too many chars

[lapis rover]

wait no

i'm editing the wrong version

[sharp scarab]

lmao

[orchid charm]

worked for me

[lapis rover]

what did

[orchid charm]

the 427 you sent

[lapis rover]

wait what do you still have it?

[orchid charm]

yeah

i made it better too

[lapis rover]

bcs it was based off of this

iirc

[sharp scarab]

huh

[orchid charm]

[Day 23] 423

e=enumerate
j=1j
E={x+y*j for y,l in e(open(0))for x,c in e(l)if'#'==c}
n=0
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
P=D=A[4:7],A[2:5],A[6:],A[:3]
while P:
 if n<11:r,i=zip(*((e.real,e.imag)for e in E))
 n+=1;P={};[exec("P[l]=[e,l][l in P]")for e in E if(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D[n%4:]+D[:n%4]if not{*i}&a),P))!=P];E|={*P};E^={*P.values()}
print(int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(E)),n)```

[lapis rover]

and when i applied the D[n%4:]+D[:n%4] to the actual version it only broke even

[sharp scarab]

I've been staring at the wrong code block this entire time lmao

[lapis rover]

yeah sorry i shouldn't post broken stuff i even confuse myself

[orchid charm]

either way i have a -4 optimization

[lapis rover]

sooo... why does this work

[orchid charm]

idk

[lapis rover]

oh actually no

it should work

[orchid charm]

great, i didn't look at what you did

[lapis rover]

because there won't be anyone making any propositions when the simulation stabilises

so idk why it didn't work before

maybe I just didn't wait long enough

[orchid charm]

[Day 23] 419

e=enumerate
j=1j
E={x+y*j for y,l in e(open(0))for x,c in e(l)if'#'==c}
n=0
A=j-1,-1,-1-j,-j,1-j,1,1+j,j,j-1
P=D=A[4:7],A[2:5],A[6:],A[:3]
while P:
 if n<11:r,i=zip(*((e.real,e.imag)for e in E))
 n+=1;P={}
 for e in E:
  if(a:={i for i in A if e+i in E})and(l:=next((e+i[1]for i in D[n%4:]+D[:n%4]if not{*i}&a),P))!=P:P[l]=[e,l][l in P]
 E|={*P};E^={*P.values()}
print(int((max(r)-min(r)+1)*(max(i)-min(i)+1)-len(E)),n)```

[sharp scarab]

shame setting n=-1 and shifting everything to compensate is worse by 1 char

[sharp scarab]

[Day 22] 1109

import re,itertools as i
p,R,e,U,Q,H=i.pairwise,range,enumerate,complex,lambda z:(z.real,z.imag),zip
J=1j
S=re.findall('\d+|L|R',(F:=open(0).read().split('\n\n'))[1])
*D,=map(lambda x:(U(*x[0]),x[1]),sorted((min(map(Q,d)),d)for n,x in p(R(0,201,50))for m,y in p(R(0,151,50))if(d:={i:v for i,v in((U(x,y),v=='.')for x,r in e(F[0].split('\n'))for y,v in e(r)if v in'.#')if(n<=i.real<x)*(m<=i.imag<y)})))
L=-1,J,1,-J
def g(a):
 global S,D
 s,w,f=*D[c:=0],J;o=s
 for v in S:
  if v in'LR':f*={'L':J,'R':-J}[v];continue
  for _ in R(int(v)):
   if(z:=s+f)not in w:C,F=a[c][f];O,W=D[C];r,i=Q(s-o);f,o,c,w,s=eval('fF,,oO,,cC,,wW,,sz'[W[z:=O+49*U(F==-1,F==-J)+(((1-2*(b:=(F.imag>F.real)^(f.imag>f.real)))*(i if f.real else r))+49*b)*{-1:J,J:1,1:J,-J:1}[F]]::2])
   elif w[z]:s=z
 return int(sum(x*(y+1)for x,y in H((1000,4,1),(*Q(s),{J:-1,1:0,-J:1,-1:2}[f]))))
print(g([{i:(v,i)for i,v in H(L,j)}for j in((4,1,2,1),(1,0,1,0),(0,2,4,2),(5,4,5,4),(2,3,0,3),(3,5,3,5))]),g([{i:(v,f) for i,v,f in H(L,*j)}for j in H(((5,1,2,3),(5,4,2,0),(0,1,4,3),(2,4,5,0),(2,1,5,3),(3,4,1,0)),((J,J,1,J),(-1,-J,-J,-J),(-1,-1,1,1))*2)]))

100c removed, yay?

[sharp scarab]

[Day 22] 1083

import re,itertools as i
p,R,e,U,Q,Z=i.pairwise,range,enumerate,complex,lambda z:(z.real,z.imag),zip
J,N=1j,-1
S=re.findall('\d+|L|R',(F:=open(0).read().split('\n\n'))[1])
*D,=map(lambda x:(U(*x[0]),x[1]),sorted((min(map(Q,d)),d)for n,x in p(R(0,201,50))for m,y in p(R(0,151,50))if(d:={i:v for i,v in((U(x,y),v=='.')for x,r in e(F[0].split('\n'))for y,v in e(r)if v in'.#')if(n<=i.real<x)*(m<=i.imag<y)})))
L=N,J,1,-J
def g(a):
 global S,D
 s,w,f=*D[c:=0],J;o=s
 for v in S:
  if v in'LR':f*=[J,-J][v>'M'];continue
  for _ in R(int(v)):
   if(z:=s+f)not in w:C,F=a[c][f];O,W=D[C];f,o,c,w,s=eval('fF,,oO,,cC,,wW,,sz'[W[z:=O+49*U(F==N,F==-J)+(((1-2*(b:=(F.imag>F.real)^(f.imag>f.real)))*Q(s-o)[f.real!=0])+49*b)*[1,J][F.real!=0]]::2])
   elif w[z]:s=z
 return int(sum(x*(y+1)for x,y in Z((1000,4,1),(*Q(s),{J:N,1:0,-J:1,N:2}[f]))))
print(g([{i:(v,i)for i,v in Z(L,j)}for j in((4,1,2,1),(1,0,1,0),(0,2,4,2),(5,4,5,4),(2,3,0,3),(3,5,3,5))]),g([{i:(v,f) for i,v,f in Z(L,*j)}for j in Z(((5,1,2,3),(5,4,2,0),(0,1,4,3),(2,4,5,0),(2,1,5,3),(3,4,1,0)),((J,J,1,J),(N,-J,-J,-J),(N,N,1,1))*2)]))

that's all i'm doing

at least until after aoc

is anyone doing a day 24 golf?

[orchid charm]

haven't even done day 24 yet

it's just dfs with pruning, isn't it?

[sharp scarab]

i tried dfs

it sucked

bfs is better

[orchid charm]

oh

[sharp scarab]

or maybe i'm just bad at dfs

that's definitely possible as well

[orchid charm]

what's bad about dfs

[sharp scarab]

i couldn't store states properly with dfs because branch reasons

[orchid charm]

i didn't think about bfs because i thought the search space would grow too rapidly before finding solutions

[sharp scarab]

valid states got wiped off

[orchid charm]

are there enough storms so bfs is viable?

[sharp scarab]

you store (pos, time%blizzard_cycle_time)

bfs takes a bit to run but it finishes

idk try dfs

you know more about dfs and dp than me

[orchid charm]

maybe after i get to the next hotel

am on the road rn

[sharp scarab]

just code in the car

that's what i did for day 15

also if you can

can you check the dimensions of your input

[orchid charm]

not logged in on phone

i also can't use a hotspot

[sharp scarab]

ah that's a shame

[sharp scarab]

my input is about 90% blizzards at the start

[cold dagger]

my queue size for bfs doesn't go past 3000

total ~120,000 states searched

[sharp scarab]

i was gonna golf today

but it's a real pain

[orchid charm]

alright, i'll probably do bfs then

[sharp scarab]

for anyone else that's gonna do a golf, the board size apparently comes in either 102*35 or 152*21

maybe more idk

[orchid charm]

i'll check mine after i get to the hotel

[sharp scarab]

also if you're gonna store blizzards in a dict be careful of duplicate blizzards removing each other

[orchid charm]

that's why you store blizzard direction in the key

obviously

[sharp scarab]

i just had 4 sets in my original solution

[orchid charm]

oh

[sharp scarab]

for the golf i was thinking that a floodfill would be pretty short

just floodfill to all possible spots until you get to the end

then drain the pool and flood again

and do that again

[sharp scarab]

also thinking about it my last two golfs were 500c and 1200c

probably shouldn't be doing first golfs anymore

mine are cursed

[raw frigate]

I started on one earlier

should I keep going with it?

[raw frigate]

122*27 for me (both accounts but that might be by chance)

[sharp scarab]

yeah I gave up on mine

I'll assist in golfing it tmr

[raw frigate]

ok I feel like base 200 will be better than complex for this

but I'll just use complex anyway for now

[sharp scarab]

we can always convert to base 200

complex is easier to debug tho

[raw frigate]

yeah

that's why I'm sticking with it for now

[raw frigate]

[Day 24 part 1] 320 ```py
j=1j
e=enumerate
B={(m:=~-x+~-yj,{'<':-1,'>':1,'^':-j,'v':j,'#':0,'\n':0}[c])for y,l in e(open(0))for x,c in e(l)if'.'!=c}|{(j+(E:=m-1),0),(-2j,0)}
S=-j
n=0
s={S}
while{E}-s:B={((i:=d+p).real%m.real+j(i.imag%m.imag)if d else p,d)for p,d in B};s={n for l in s for n in{l,l-1,l+1,l-j,l+j}}-{i for i,_ in B};n+=1
print(n)

does this work for everyone else?

[Day 24] 351 ```py
j=1j
e=enumerate
B={(m:=~-x+~-yj,{'<':-1,'>':1,'^':-j,'v':j,'#':0,'\n':0}[c])for y,l in e(open(0))for x,c in e(l)if'.'!=c}|{(j+m-1,0),(-2j,0)}
S=m-1,-j
n=0
for D in 1,0,1:
s={S[D]}
while{S[1-D]}-s:B={((i:=d+p).real%m.real+j(i.imag%m.imag)if d else p,d)for p,d in B};s={n for l in s for n in{l,l-1,l+1,l-j,l+j}}-{i[0]for i in B};n+=1
if D:print(n)

I'm not bothered to convert this into base-n

I think I'll go to sleep now

apparently this is also 9x faster than my original solution btw

[lapis rover]

{'<':-1,'>':1,'^':-j,'v':j,'#':0,'\n':0}[c]
[-1,1,-j,j,0,0]["<>^v#\n".index(c)]

~-x+~-y*j
x-1+y*j-j is not shorter but...

[orchid charm]

less cursed, so it's worse

[raw frigate]

[Day 24] 343 ```py
j=1j
e=enumerate
B={(m:=~-x+~-yj,[-1,1,-j,j,0,0]["<>^v#\n".index(c)])for y,l in e(open(0))for x,c in e(l)if'.'!=c}|{(j+m-1,0),(-2j,0)}
S=m-1,-j
n=0
for D in 1,0,1:
s={S[D]}
while{S[1-D]}-s:B={((i:=d+p).real%m.real+j(i.imag%m.imag)if d else p,d)for p,d in B};s={n for l in s for n in{l,l-1,l+1,l-j,l+j}}-{i[0]for i in B};n+=1
if D:print(n)

[sullen field]

[Day 24] 333 ```py
j=1j
e=enumerate
B={(m:=~-x+~-yj,[-1,1,-j,j,0]["<>^v".find(c)])for y,l in e(open(0))for x,c in e(l)if'.'!=c}|{(j+m-1,0),(-2j,0)}
S=m-1,-j
n=0
for D in 1,0,1:
s={S[D]}
while{S[1-D]}-s:B={((i:=d+p).real%m.real+j(i.imag%m.imag)if d else p,d)for p,d in B};s={l+n for l in s for n in{0,-1,1,-j,j}}-{i[0]for i in B};n+=1
if D:print(n)

[sullen field]

[Day 24] 324 ```py
j=1j
e=enumerate
z=-1,1,-j,j,0
B={(m:=~-x+~-yj,z["<>^v".find(c)])for y,l in e(open(0))for x,c in e(l)if'.'!=c}|{(j+m-1,0),(-2j,0)}
S=m-1,-j
n=0
for D in 1,0,1:
s={S[D]}
while{S[1-D]}-s:B={((i:=d+p).real%m.real+j(i.imag%m.imag)if d else p,d)for p,d in B};s={l+n for l in s for n in z}-{i[0]for i in B};n+=1
if D:print(n)

[sharp scarab]

wtf

!e

print(*map(ord,'^>v<.'))
print([(i,i>'2') for i in '^>v<.'])

[tawny dawn]

@sharp scarab :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | 94 62 118 60 46
002 | [('^', True), ('>', True), ('v', True), ('<', True), ('.', False)]

[sharp scarab]

there we go

replace '.'!=c with '2'>c

actually I could just do it

[sullen field]

have you tested it

because it doesn't work

[sharp scarab]

it doesn't?

[sullen field]

[Day 24] 323 ```py
j=1j
e=enumerate
z=-1,1,-j,j,0
B={(m:=~-x+~-yj,z["<>^v".find(c)])for y,l in e(open(0))for x,c in e(l)if'.'!=c}|{(j+m-1,0),(-2j,0)}
S=m-1,-j
n=0
for D in 1,0,1:
s={S[D]}
while{S[1-D]}-s:B={([(i:=d+p).real%m.real+j(i.imag%m.imag),p][d==0],d)for p,d in B};s={l+n for l in s for n in z}-{i[0]for i in B};n+=1
if D:print(n)

[sullen field]

we need #

[sharp scarab]

ohh

[sullen field]

which translates to 0

[sharp scarab]

fair enough then

[orchid charm]

[Day 24] 322

j=1j
e=enumerate
z=-1,1,-j,j,0
B={(m:=~-x+~-y*j,z["<>^v".find(c)])for y,l in e(open(0))for x,c in e(l)if'.'!=c}|{(j+m-1,n:=0),(-2j,0)}
S=m-1,-j
for D in 1,0,1:
 s={S[D]}
 while{S[1-D]}-s:B={([(i:=d+p).real%m.real+j*(i.imag%m.imag),p][d==0],d)for p,d in B};s={l+n for l in s for n in z}-{i[0]for i in B};n+=1
 if D:print(n)```

lol

[sharp scarab]

why is there a space

[orchid charm]

btw is pt 2 600 greater than pt 1 for y'all?

it was 600 more on both of my accounts

not sure if coincidence or…?

[sharp scarab]

mine was 518 more

probably a coincidence

[orchid charm]

alright

[sharp scarab]

is the space just a discord thing

[orchid charm]

yeah

the while should be under the for

so it should all be indented

[sharp scarab]

ah ok

[orchid charm]

308 908 and 334 934 for both accounts

i thought there was a thing going on there

😔

[sharp scarab]

even if it did hold

we'd lose like 10-20c

I think

not that much

[orchid charm]

22x152 and 27x122 for input size if you were curious