AoC 2022 | Day 3 | Solutions & Spoilers

[sand jungle]

@hoary parcel what is the shortest till now?

[hoary parcel]

pinned

[proud hamlet]

lines = open("../inputs/day3.txt").read().splitlines()

lines = [lines[n:n+3] for n in range(0, len(lines), 3)]

indexes = " abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ"
print(sum(indexes.index(list(set(group[0]).intersection(set(group[1]).intersection(set(group[2]))))[0]) for group in lines))

not great but ¯_(ツ)_/¯

[sand jungle]

Crlf doesn't work for me

[hoary parcel]

are you on windows and did you provide an input using stdin

[sand jungle]

Mac. Input through file

[hoary parcel]

you shouldn't be using CRLF then

use the LF one

[sand jungle]

Tried that as well. P1 works. P2 doesn't falls 10 short of the actual answer

[hoary parcel]

what about the xplatform one

does it give the same wrong answer

[sand jungle]

Xplatform works

Weird...

[hoary parcel]

probably with CR line endings?

[hoary parcel]

added a CR version

[sand jungle]

What does that mean? I mean , how do I do that?

[hoary parcel]

either that's the error or some misprovided input

[hoary parcel]

check the solutions and test the one labeled CR

[sand jungle]

I-

I changed nothing and ran the LF again and it worked

WTH

[hoary parcel]

oh

[hoary parcel]

CRLF did this?

[chilly linden]

todays discovery was for first_elf, second_elf, third_elf in zip(*[iter(data)] * 3): magical

[wintry depot]

indeed ... thanks to stackoverflow I found that as well

[chilly linden]

last night you didn't want to enter the esoteric python territory, what is this today 😭

[lethal spindle]

i did it in lua, I miss Sets

[chilly linden]

i initially used the usual range but wasn't happy with it and found that gem

very neat

when i first saw the question i thought it was gonna be hard, turned out to be surprisingly ez

[wintry depot]

Yeah didn't seem easy at first, but now it looks quite simple. But still took me about 30 minutes the 2nd part

[quiet heath]

I thought part 2 was going to be difficult in particular. Was surprised it was almost exactly the same as part 1, just with slightly altered input

[wintry depot]

I can't wait for the next puzzles. Hope I get further than last year, before I lose motivation

[fallen sinew]

d3a = sum(ord(e.lower())-ord('a')+1+26*e.isupper() for line in in3a.splitlines() for e in set(line[:len(line)//2])&set(line[len(line)//2:]))

in3b = in3a.splitlines()
d3b = sum(ord(e.lower())-ord('a')+1+26*e.isupper() for x,y,z in (in3b[i:i+3] for i in range(0,len(in3b),3)) for e in set(x)&set(y)&set(z))

One liners (for second task, I only did splitlines separately because it would look ugly and repeat it many times) on phone :3

[chilly linden]

im surprised i understood this 100% because its literally a one liner of my program lol

[ornate linden]

import string 
  
 scores = {letter: score for score, letter in enumerate(string.ascii_letters, 1)} 
  
  
 def split_strings(row): 
     lenstr = len(row) // 2 
     return row[:lenstr], row[lenstr:] 
  
  
 def intersections(*tups): 
     tups = iter(tups) 
     first = set(next(tups)) 
     for tup in tups: 
         first &= set(tup) 
     return first.pop() 
  
  
 with open("puzzle.input") as f: 
     print( 
         "part1:", 
         sum(scores[intersections(*tup)] for tup in map(split_strings, f.readlines())), 
     ) 
  
  
 with open("puzzle.input") as f: 
     lines = [line.strip() for line in f.readlines()] 
     zipped = zip(lines[::3], lines[1::3], lines[2::3]) 
     print( 
         "part2:", 
         sum(scores[intersections(*tup)] for tup in zipped), 
     )

[woeful phoenix]

ahh ord() I knew there had to be a function that converted strings to ints. I just used a list.

[chilly linden]

woah

[mighty robin]

with open("inputs/input_3.txt", "r") as input_file:
    inputs = [line.strip() for line in input_file.readlines()]

PRIORITIES = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ"


def part_one(input):
    total_sum = 0
    for i in inputs:
        comp_one, comp_two = set(i[0 : len(i) // 2]), set(i[len(i) // 2 : len(i)])
        intersection = comp_one.intersection(comp_two)
        total_sum += PRIORITIES.index(list(intersection)[0]) + 1
    print(total_sum)


def part_two(input):
    total_sum = 0
    elf_groups = [input[i : i + 3] for i in range(0, len(input), 3)]
    set_elfs = [[set(k) for k in i] for i in elf_groups]
    for i in set_elfs:
        intersection = i[0].intersection(i[1], i[2])
        total_sum += PRIORITIES.index(list(intersection)[0]) + 1
    print(total_sum)


part_one(inputs)
part_two(inputs)

Any suggestions on how to clean part_two?

elf_groups = [input[i : i + 3] for i in range(0, len(input), 3)]
set_elfs = [[set(k) for k in i] for i in elf_groups]

Here I just divide into groups first and then turn each element of the groups into sets, but I feel like there is a better way haha

[frigid hawk]

FYI you can use string.ascii_letters to get the string of all the letters

[mighty robin]

Yeah I just avoid imports in general for AOC

[naive bridge]

""" Day 3 Part 1 Solution """

f = open("puzzle_input.txt", "r")
puzzle_input = f.read()

priority_sum = 0
priority_value = lambda it: ord(it) - (96 if(ord(it) in range(ord('a'), ord('z')+1)) else 38)

for line in puzzle_input.split("\n"):
    middle = int(len(line)/2)
    first = line[:middle]
    second = line[middle:]

    for item in second:
        if item in first:
            priority_sum += priority_value(item)
            break           

print(priority_sum) # 7863


""" Day 3 Part 2 Solution """
import functools

f = open("puzzle_input.txt", "r")
puzzle_input = f.read()
puzzle_input_lines = puzzle_input.split("\n")

priority_sum = 0
team_size = 3
priority_value = lambda it: ord(it) - (96 if(ord(it) in range(ord('a'), ord('z')+1)) else 38)

for lineidx in  range(0, len(puzzle_input_lines), team_size):
    
    elf_team = [puzzle_input_lines[lineidx+i] for i in range(team_size)]
    team_item = functools.reduce(lambda x,y: set(x) & set(y), elf_team)
    priority_sum += priority_value(team_item.pop())

print(priority_sum) # 2488

[frigid hawk]

Fair - I try to avoid importing anything thats not a builtin lib

[regal harness]

🥴 meanwhile me:

[hasty moon]

import os
file_path = os.path.realpath(os.path.dirname(__file__)) + "/input.txt"

def get_priority(char):
    return ord(char) - 64 + 26 if char.isupper() else ord(char) - 96

def part_1(lines):
    def get_points(bag):
        m = len(bag) // 2
        return get_priority(*set(bag[:m]).intersection(set(bag[m:])))
    
    ans = 0
    for line in lines:
        ans += get_points(line)
    return ans

def part_2(lines):
    def get_points(lines):
        return get_priority(*set.intersection(*map(set, lines)))
    
    ans = 0
    for i in range(0, len(lines) - 2, 3):
        ans += get_points(lines[i:i+3])
    return ans

lines = open(file_path).read().strip().split('\n')
print(part_1(lines))
print(part_2(lines))

[mighty robin]

Just why... hahahaha

[naive bridge]

xD

[regal harness]

they take 0 time and importing them if I need to will take a second or two

[frigid hawk]

after importing all that he calculates the answer by hand. The only other line is:" print("The answer is: 12403")

[regal harness]

i ended up not using numpy by default because it actually does take 0.3s to import

[warm oracle]

a&b&c for a,b,c in zip(*[iter(pzl)] * 3) is there a nice way to convert to sets during the zip as well?

[mighty robin]

That sounds about right

[mighty robin]

0.3 just to import really?

[frigid hawk]

I feel like there is a nicer way of doing this but im pretty happy with what I have: ```py
import sys
import os
import string

Get Data

with open(os.path.dirname(sys.argv[0]) + "/input.txt", 'r') as f:
data = f.read()
data = data.split('\n')

Pt 1

priorities = []
for d in data:
splitpoint = int(len(d) / 2)
a = d[:splitpoint]
b = d[splitpoint:]
for i in a:
if i in b:
priorities.append(string.ascii_letters.index(i) + 1)
break

print(sum(priorities))

[hasty moon]

https://tenor.com/view/rickandmorty-showme-gif-7658980

show me the shortest code and the most over engineered code

[regal harness]

yup, you can check yourself via python -X importtime the_script.py 2>importtime_profile, and then showing that file with tuna

[mighty robin]

Huh, will do

[mighty robin]

Any ideas?

[quiet heath]

Here’s my over engineered today: https://github.com/Fiddle-N/advent-of-code-2022/blob/master/day_03/process.py

[hallow vapor]

from pathlib import Path
data = Path("input.txt").read_text().strip().split("\n")

def priority(item: str) -> int:
    return ord(item) - 96 + 58*(ord(item) < 97)

def part1():
    total = 0
    for racksack in data:
        middle = len(racksack) // 2
        common_letter = (set(racksack[:middle]) & set(racksack[middle:])).pop()
        total += priority(common_letter)
    print(total)

def part2():
    total = 0
    for i in range(0, len(data), 3):
        common_letter = (set(data[i]) & set(data[i+1]) & set(data[i+2])).pop()
        total += priority(common_letter)
    print(total)

[frigid hawk]

Is there a reason (other than avoiding imports) to use ord() rather than string.ascii_letters?

[warm oracle]

sum((a & b & c).pop() for a, b, c in zip(*[iter(map(set, pzl))]*3))```
alright I like this

[spring tinsel]

If anything, I like how AoC brings the absolute garbage outta everyone's code

Been long since I last wrote nested listcomp, and fuck yea I'm gonna write more of that stuff

[warm oracle]

both ways, the quick and dirty + the unreadable 1 lines

[hallow vapor]

And then use index? I just grabbed the first tool I had in mind, but that sounds less efficient

[frigid hawk]

I used: string.ascii_letters.index(i)

[hallow vapor]

You need to look through the string every time with index

The time difference is negligible probably but still

[frigid hawk]

Ah right, while ord() + some maths is quicker?

[hallow vapor]

that's my guess, didn't measure

[warm oracle]

ord(literal) should be converted to int on execution

[mighty robin]

It is never worth to do math

Mentally at least

[dark fjord]

There are people that make equations to solve it.

[frigid hawk]

Looks a bit quicker - not much in it though.
for 10 million loops:
ord: 1074 ms
index: 1265 ms

[hallow vapor]

yeah didn't expect it to be significant, the string is pretty small after all

[tropic quarry]

how are u timing ur code?

[dark fjord]

!d timeit

[broken condor]

timeit

Source code: Lib/timeit.py

This module provides a simple way to time small bits of Python code. It has both a Command-Line Interface as well as a callable one. It avoids a number of common traps for measuring execution times. See also Tim Peters’ introduction to the “Algorithms” chapter in the second edition of Python Cookbook, published by O’Reilly.

Similar names: ipython.timeit

[frigid hawk]

time.time

I dont like timeit to be honest

[tropic quarry]

is it possible to use time.time for an entire python file and not a one line

[frigid hawk]

I dont see why not?

[ornate linden]

You could also time it from the commandline. Use time on linux or get-history on powershell

[tropic quarry]

im getting an error when i try subtract t1-t0

[frigid hawk]

You risk timing things you dont want to doing that though

[ornate linden]

True, like interpreter startup

[frigid hawk]

I use this decorator: ```py
def timeit(func):
@functools.wraps(func)
def new_func(*args, **kwargs):
start_time = time.time()
for i in range(100_000_000):
result = func(*args, **kwargs)
elapsed_time = time.time() - start_time
print('function [{}] finished in {} ms'.format(
func.name, int(elapsed_time * 1_000)))
return result
return new_func

Which i named timeit to make it super not confusing

[ornate linden]

Use time.perfcounter instead. Its got better resolution

Time.perfcounter_ns gives it in nanoseconds

[tropic quarry]

time.time_ns() also gives nanoseconds idk what the difference is

[wraith gulch]

#1 
p = lambda x: ord(x) - 64 + 26 if x.isupper() else ord(x) - 96
g = lambda l: p(*set(l[:len(l) // 2]).intersection(set(l[len(l) // 2:])))
s = lambda k: sum(g(o) for o in k)
print(s(open("input.txt").read().split("\n")))

#2
p = lambda x: ord(x) - 64 + 26 if x.isupper() else ord(x) - 96
g = lambda l: p(*set.intersection(*map(set, l)))
s = lambda k: sum(g(k[i:i+3]) for i in range(0, len(k) - 2, 3))
print(s(open("input.txt").read().split("\n")))

[proper barn]

It all depends on your cpu clock anway

[wraith gulch]

[proper barn]

I think the question setting is done in python

with the 'a' and 'A' portions hmm

[ornate linden]

Perf counter is made for checking how long things take. Time.time is meant for lower resolution timing

[frigid hawk]

I switched it out for the perf_counter but its no different in actual performance

[ornate linden]

Probably because you're doing 100mn runs of the function. If you were running it once it would make more of a difference

[frigid hawk]

Just doing it once can get you wildly varying results though

always best to average out over a longer run

[hallow vapor]

what's wrong with timeit?

[frigid hawk]

nothing functionally, just find the interface to it annoying

I find the decorator quicker and easier to understand

[north night]

print("Part 1 : ", sum([ord(min(n))-96 if min(n).islower() else ord(min(n))-38 for n in [set(l[:len(l)//2]).intersection(l[len(l)//2:]) for l in [n.rstrip() for n in open("input.txt").readlines()]]]))
print("Part 2 : ", sum([ord(min(n))-96 if min(n).islower() else ord(min(n))-38 for n in [''.join(set([n.rstrip() for n in open("input.txt").readlines()][i]).intersection([n.rstrip() for n in open("input.txt").readlines()][i+1]).intersection([n.rstrip() for n in open("input.txt").readlines()][i+2])) for i in range(0,len([n.rstrip() for n in open("input.txt").readlines()]), 3)]]))

[proper barn]

unreadable

[nimble pagoda]

these one liners doesn't really optimize your code

[proper barn]

they optimize for unreadability

[nimble pagoda]

naive approach written in one line

[nimble pagoda]

they put so much time in it to write that one line tho

[frigid hawk]

Optimise for smallest screen space

[nimble pagoda]

screen space crisis

[coarse trail]

def part_1(data):
    priority = 0

    for line in data:
        half = len(line) // 2
        r = (set(line[:half]) & set(line[half:])).pop()
        priority += ord(r) - 96 if ord(r) >= 97 else ord(r) - 38

    return priority


def part_2(data):
    priority = 0

    for group in data:
        r = (set(group[0]) & set(group[1]) & set(group[2])).pop()
        priority += ord(r) - 96 if ord(r) >= 97 else ord(r) - 38

    return priority

I use utility functions to load in the data, for part 2 I used a chunking loader, which basically just uses more-itertools.chunked to get the groups of 3.

[chilly linden]

oo i like how you played with ord utilizing ifs, i just went with .isupper() neat solution, pretty much how i did it as well
for part two i found this gem: for first_elf, second_elf, third_elf in zip(*[iter(data)] * 3)

[coarse trail]

I did consider checking for isupper but I quite liked sticking with ord() 😄

I have across this zip(*[iter(data)] * 3) before. I should write it down for times where more-itertools isn't available

[chilly linden]

ya i ended up using that in combination with ord

def part_one(data):
    total = 0
    for sack in data:
        split_index = len(sack) // 2
        first_half, second_half = sack[:split_index], sack[split_index:]
        common = "".join(set(first_half) & (set(second_half)))
        if common.isupper():
            total += ord(common) - 38
        else:
            total += ord(common) - 97 + 1
    return total


def part_two(data):
    total = 0
    for first_elf, second_elf, third_elf in zip(*[iter(data)] * 3):
        common = "".join(set(first_elf) & set(second_elf) & set(third_elf))
        if common.isupper():
            total += ord(common) - 38
        else:
            total += ord(common) - 97 + 1
    return total

i love how python makes these things a breeze, absolutely love it

[zinc wing]

I used string.ascii_letters.index(letter)+1 :P

[chilly linden]

wait what

[zinc wing]

!e

import string
print(string.ascii_letters.index("a")+1)

[broken condor]

@zinc wing :white_check_mark: Your 3.11 eval job has completed with return code 0.

1

[zinc wing]

@chilly linden

[karmic crow]

I deliberately went with alphabet.lower() + alphabet.upper() as I think it's a bit more explicit. It doesn't require any special knowledge about string.ascii_letters to understand. E.g. the fact that lowercase letters come before uppercase (which is the kind of thing I would forget ).

[analog mulch]

I'll tell you what I told Tom

@karmic crow

[karmic crow]

Don't have the energy to write one-liners at the moment 😓

[analog mulch]

Cause you spent it all doing crap

[karmic crow]

¯_(ツ)_/¯

[chilly linden]

🗿

[maiden lava]

I like it 😄 learnt some new things

[hallow vapor]

alternative part2

sum(map(priority, map(set.pop, map(partial(reduce, and_), chunked(map(set, data), 3)))))

[still magnet]

that's a elegant one liner

[naive bridge]

https://github.com/R1c4rdCo5t4/AdventOfCode2022

[quiet heath]

Looks pretty clean. Try not to commit/push puzzle inputs though, the AoC creator would prefer that you refrain from doing so

[naive bridge]

Understood, thanks!

[frail schooner]

Oh, is it okay if I keep the repo private until the finals?

Or even then the inputs shouldn't be pushed to repo?

I wanna make it public later on

[quiet heath]

Probably best to do this. Just don’t push the input files in the first place

[coarse trail]

Why?

[frail schooner]

Okay, gotcha.

[night egret]

Hi

[quiet heath]

The creator is against it. Let me find the link

https://www.reddit.com/r/adventofcode/comments/k99rod/sharing_input_data_were_we_requested_not_to/gf2ukkf/?utm_source=share&utm_medium=ios_app&utm_name=iossmf&context=3

[unique pivot]

In that case, I should probably remove the inputs from my repo

[cursive sail]

it's best to just gitignore them

[coarse trail]

Hmm. Odd reasoning. And if they're bothered they should have this on the site, not some random Reddit post.

[quiet heath]

I think it’s more like they’d prefer you not to post them, rather than decreeing it as a must

[coarse trail]

Well it's sure fire way to make sure everyone doesn't know and commits them. I have. For years. As I was none the wiser.

[full scarab]

A 1-liner for part one but in three lines.```py
def day_3():
f = open("2022/day_3.txt", "r").read().splitlines()
c = [
(((ord(p) - 96) if p.islower() else (ord(p) - 38)) if p.isalpha() else 0) for q in
[[v if v in i[len(i) // 2:] else "0" for v in i[:len(i) // 2]] for i in f]
for p in set(q)
]

print(sum(c)) # PART 1

[random mango]

Oh huh i thought people just pushed them because of a "fuck you" to the author. Honestly I feel like i remembered there being a big label about it on the website before, but i cant find anything.

So i may be completely out of my zone about this 😓

https://github.com/imsofi/advent-of-code/blob/main/2022/day_03.rb

Here is my ruby solution for day 3, challenging myself to try another language proper this time around.

[quiet heath]

I used to do it before completely accidentally. I think that’s the most common situation

[random mango]

Yeah, i just feel like i remember a login screen or similar of importance to be like "hey lets not upload the private inputs, so people cannot revere engineer the algorithmn that generates them, thanks and have fun!"

But its been like 3 years now, and i can find nothing of the sort 😓

[wise rose]

oh dang, I need to update my gitignore then

def parse_group(sacks):
    sack0 = set(sacks[0])
    sack1 = set(sacks[1])
    sack2 = set(sacks[2])
    for item in sack0:
        if item in sack1 and item in sack2:
            print(item)
            return item

def score_item(letter):
    ord_val = ord(letter)
    if ord_val>96:
        return ord_val-96
    return ord_val-64+26

if __name__ == '__main__':
    total_score = 0
    group_size = 3
    group = []
    with open('input.txt', 'r') as sacks:
        for line in sacks:
            sack = line.rstrip()
            group.append(sack)
            if len(group) < group_size:
                continue
            duped_item = parse_group(group)
            total_score+= score_item(duped_item)
            group = []
    print(total_score)

hrmm, code highlighting not happening

[random mango]

Ah shoot. Yeah its no fun that it is so

But i find the reasoning good. I have seen seed crackers in minecraft, so i understand the worry.

[frail schooner]

put py after the three grave symbols

for python colors

[frail schooner]

def parse_group(sacks):
    sack0 = set(sacks[0])
    sack1 = set(sacks[1])
    sack2 = set(sacks[2])
    for item in sack0:
        if item in sack1 and item in sack2:
            print(item)
            return item

def score_item(letter):
    ord_val = ord(letter)
    if ord_val>96:
        return ord_val-96
    return ord_val-64+26

if __name__ == '__main__':
    total_score = 0
    group_size = 3
    group = []
    with open('input.txt', 'r') as sacks:
        for line in sacks:
            sack = line.rstrip()
            group.append(sack)
            if len(group) < group_size:
                continue
            duped_item = parse_group(group)
            total_score+= score_item(duped_item)
            group = []
    print(total_score)

[wise rose]

gotcha

[wise rose]

am I violating some group norms, I'm noticing a lot of people are trying for one liners?

[frail schooner]

no they are doing code golfs

trying to get the shortest solutions possible

in terms of number of characters used in the file

[quiet heath]

You can code however you want to

[wise rose]

:nods: I'm familiar with the concept

[frail schooner]

Its not a group norm, so no worries

[wise rose]

my background values clarity as well as brevity

[frail schooner]

you can solve the stuff however you want

[analog mulch]

OH wat

oops

[wise rose]

I could probably tweak my parse_sacks to use a list comprehension though, i try for generality in the main body, but then it's kind of hard coded to groups of size 3 in the parser

[quiet heath]

I think there’s three main schools of thought for solutions :

1. Get the solution done as quickly as possible and leave it at that
2. Code golf it
3. Overengineer it completely

[frail schooner]

[naive oriole]

Last night a huge group of ~9 ish people all collaborating made the solutions that are pinned

# LF (144)
*z,=open(0,'rb');u=sum;print(u((u({*a[(l:=len(a)//2):]}&{*a[:l]})-96)%58for a in z),u((u({*b}&{*c}&{*d})-106)%58for b,c,d in zip(*3*[iter(z)])))
# CR (144)
*z,=open(0,'rb');u=sum;print(u((u({*a[(l:=len(a)//2):]}&{*a[:l]})-96)%58for a in z),u((u({*b}&{*c}&{*d})-109)%58for b,c,d in zip(*3*[iter(z)])))
# CRLF (146)
*z,=open(0,'rb');u=sum;print(u((u({*a[(l:=len(a)//2-1):]}&{*a[:l]})-96)%58for a in z),u((u({*b}&{*c}&{*d})-119)%58for b,c,d in zip(*3*[iter(z)])))
# xplatform (156)
z=open(0,'rb').read().split();u=sum;print(u((u({*a[(l:=len(a)//2):]}&{*a[:l]})-96)%58for a in z),u((u({*b}&{*c}&{*d})-96)%58for b,c,d in zip(*3*[iter(z)])))
# xplatform one loop (161)
u=sum;print(u(u((u({*a[(l:=len(a)//2):]}&{*a[:l]})-96)%58for a in(b,c,d))+(u({*b}&{*c}&{*d})-96)%58*1j for b,c,d in zip(*3*[iter(open(0,'rb').read().split())])))

[unique pivot]

If i had fluid in my mouth it would have been all over my monitor right now

[frail schooner]

lol

[naive oriole]

What's kind of funny is that it got so golfed that there were different versions for different platforms

because the newline style is different

[unique pivot]

lmao

[wise rose]

https://media4.giphy.com/media/fSBKDj5uNaUpT1HfEs/giphy.gif?cid=790b76115a03f8a60651d36fcabe8780f8a2710f2a1b199b&rid=giphy.gif&ct=ts

[quiet heath]

CR only is quite funny. Isn’t that the old Mac OS before X that used to do that?

[frail schooner]

is there like a full form for CR and LF and CRLF?

[naive oriole]

honestly these solutions have some good principles for golfing:

• ,= operator
• := operator
• *iter unpacking
• {*z} set conversion
• u=sum overused function renaming
• open(0) read from stdin
• uses an overcomplicated (x-96)%58 for the value calculation (a=1, b=2, A=27)
• uses & for lots of stuff

[unique pivot]

,= operator
the WAHT

[naive oriole]

technically it's not an operator

[frail schooner]

We need this pinned bro

I am saving this in my personal Discord for safekeeping

[naive oriole]

*x, = list = x = list[:-1]
iirc

so you just remove the last elem

[unique pivot]

ah, I see, a kind of unpacking

[naive oriole]

It's equivalent to saying

[frail schooner]

wait so *x, means all the elements except the last on in x?

[naive oriole]

yeah

[frail schooner]

lol thats hilarious xD

never knew that

[naive oriole]

*x, _ = list
_ = last element
*x = everything else

[analog mulch]

https://tenor.com/view/im-doing-my-part-serious-stare-gif-11979677
https://github.com/shenanigansd/scratchpad/commit/861ed29821f4feb233745a0edde1e8437f181049

[naive bridge]

really nice

[frail schooner]

wait so to do x = list[:-2] we do *x, , = list? or is there like a different way?

[naive oriole]

you have to do

*x,_,=list but yeah

[analog mulch]

OMD I hate the way Discord plays with embeds

[naive oriole]

,= basically means discard last elem

[frail schooner]

okay, I thought maybe there was a way to do like *x, 2*y = list or something

its basically like regex, ig? like if I think in terms of regex, I get it

[upper tundra]

after just finishing day 3, i now feel really stupid for not just doing python values = 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ' and index() + 1 to get the priority levels

[frail schooner]

xD

in hindsight most of us were being obtuse

[naive oriole]

found this video: https://www.youtube.com/watch?v=KeIMUw22SZ8

unsure if it's good

but hopefully it explains it

[upper tundra]

i made this whole function thinking i was being so smart ```python
def prio(char):
if char.islower():
return ord(char) - 96
else:
return ord(char) - 38

[naive bridge]

you can do

import string
values = string.ascii_letters

[unique pivot]

I got halfway through typing out something like this before I remember string existed

[naive oriole]

The shortest calculation was a hacky math one: (x-96)%58

where x = ord(character)

[upper tundra]

wow thats really nice

good to know about the string.ascii_lettrs as well

[naive oriole]

Artemis, hsop, cereal, cefqrn, tesseract, jenna, salt-die, indie, and xelf

were everyone who helped

with the golfing

I think

[upper tundra]

shoutout to them

[naive oriole]

I'm probably missing some

I was also helping

[upper tundra]

at the very least, im still happy I learned a little bit of new stuff today

with the ord() function

[wise rose]

would the ord function be a bit faster?

[frail schooner]

like the if else?

no, the list would have lesser lines, it just getting the index

[wise rose]

index() is linear, while ord is constant? admittedly small n, but principle of the thing

[naive oriole]

Actually we ended up doing a hacky solution and ignored ord by reading in "rb" (binary mode) to get the characters as binary numbers (what ord would return)

so we skipped the conversion entirely

O(0)

[frail schooner]

true ig, then best solution would be the (ord(x)-96)%58 that these guys found

yeah or just call as rb

[wise rose]

is the else necessary in this case? return breaks out of the function so everything after it doesn't exist

[frail schooner]

I just did everything the hard way, I have no clue how I finished in about 8 minutes considering I was using the wrong index for the second half of the lines while splitting them for like the first 5 minutes..

[wise rose]

I wonder when I'll be writing test cases for my helper functions...

[frail schooner]

helper functions?

[full scarab]

Why the 58%?

[frail schooner]

for the negative values iirc

[naive oriole]

Yeah it's because it ends up being negative

[upper tundra]

yes? because if the character is not a lower case character I want to return something different

[naive oriole]

and then you do the %58 to get it correct

[wise rose]

like my parse_group or score_item functions, when will they be doing complicated enough things I'll want to verify they work on simpler cases

[upper tundra]

the first return statement is not guaranteed to always trigger

oh wait

i think i know what u mean nvm

[wise rose]

but if the if doesn't trigger, then what's after it will always run, so the else provides some semantic feedback, but it's not quite necessary

[full scarab]

Oh, so with %58 you wouldn't have to do -38 on the lowercase ones?

[upper tundra]

yea i see what u mean now 🙂

[wise rose]

I dunno, I just habitually try to early-return to avoid indentation as much as possible

[frail schooner]

Don't laugh at my solutions https://github.com/HeavenEvolved/Advent_Of_Code_2022

[frail schooner]

wait aah I didnt change it back to public while removing the data files

[full scarab]

Uh, it doesn't return negative for me.

[frail schooner]

sec

[naive oriole]

what?

[frail schooner]

can you try now?

[frail schooner]

what do you mean?

[naive oriole]

!e ```py
print(ord("A")-96)
print((ord("A")-96)%58)

[broken condor]

@naive oriole :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | -31
002 | 27

[frail schooner]

it returns negative cuz in ASCII uppercase comes before lowercase

[naive oriole]

the computation is cross-platform

ord("A") is platform-consistent

[frail schooner]

so uppercase has lower value than lowercase

[naive oriole]

!e ```py
print(ord("a")-96)
print((ord("a")-96)%58)

[broken condor]

@naive oriole :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | 1
002 | 1

[frail schooner]

or wait am I saying it the opposite way..

!e print(ord('a'), ord('A'))

[broken condor]

@frail schooner :white_check_mark: Your 3.11 eval job has completed with return code 0.

97 65

[frail schooner]

okay I was right

[dark fjord]

Only thing I can say is why also check for upper? What are you expecting? 👀

[naive oriole]

!e ```py
for a in "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ":
print(ord(a)-96,(ord(a)-96)%58)

[broken condor]

@naive oriole :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | 1 1
002 | 2 2
003 | 3 3
004 | 4 4
005 | 5 5
006 | 6 6
007 | 7 7
008 | 8 8
009 | 9 9
010 | 10 10
011 | 11 11
... (truncated - too many lines)

Full output: https://paste.pythondiscord.com/befifafamu.txt?noredirect

[full scarab]

ord("a")-96 and ord("A")-38 works for me, where should I use the %58?

[frail schooner]

I have no idea bro, my brain was doing donuts while doing the code

my brain just thinks there will be an issue so I did that

[full scarab]

def day_3():
    f = open("2022/day_3.txt", "r").read().splitlines()
    c = [
        (((ord(p) - 96) if p.islower() else (ord(p) - 38)) if p.isalpha() else 0) for q in 
        [[v if v in i[len(i) // 2:] else "0" for v in i[:len(i) // 2]] for i in f]
        for p in set(q)
    ]

    print(sum(c)) # PART 1

[dark fjord]

I'm also using ord-38

[naive oriole]

(ord(character)-96)%58 is the full code where character is any character from "a-zA-z"

[naive oriole]

see this output

and it will demonstrate what is happening

[frail schooner]

(ord(p) - 96) if p.islower() else (ord(p) - 38)) if p.isalpha()

youre using this so you can replace it with (ord(p)-96)%58

wait.. else 0?

[dark fjord]

This my part 1 solution ```py
print(sum([ord(x) - 96 if x.islower() else ord(x) - 38 for r in data for x in set(r[:int(len(r) / 2)]) if x in set(r[int(len(r) / 2): len(r)])]))

[full scarab]

If I remove it, I get invalid syntax.

Doing 0 doesn't effect the sum.

[frail schooner]

true

just else the other case, ig

[unique pivot]

One thing I was worndering is this is how I split the lines into groups of three, is there a way I could have done this in the forloop without having to have that second list comp?

sacks = input.splitlines()
for i in range(0, len(sacks), 3):
    group = [set(s) for s in sacks[i:i+3]]

[naive oriole]

ord(x) - 96 if x.islower() else ord(x) - 38
# Is equivalent to:
(ord(x)-96)%58

[full scarab]

That's quite cool ngl.

[dark fjord]

[data[i:i + 3] for i in range(0, len(data), 3)]

[unique pivot]

I guess, but then I have to loop over that comp

I was imagining that there was something like

for group in something(sacks):

[naive oriole]

There was some version of that

that we came up with when golfing

I don't know exactly what it was

you can try and find it

[frail schooner]

would for a, b, c in range(0, len(data), 3) work?

[dark fjord]

That's very cool.

[frail schooner]

or do you need the zip thing?

[dark fjord]

Who knew I'd learn something on day =< 3 :3

[frail schooner]

nope doesnt work

[naive oriole]

Honestly I should compile a list of code golfing tips for python

I would just have to find them all

[wise rose]

one thing I've been trying to is to parse the inputs as streams rather than loading it all into memory at once, am I actually doing that?

[frail schooner]

I think they used open(0,'rb') to read directly from stdin

[wise rose]

but like,

if __name__ == '__main__':
    total_score = 0
    group_size = 3
    group = []
    with open('input.txt', 'r') as sacks:
        for line in sacks:
            sack = line.rstrip()
            group.append(sack)
            if len(group) < group_size:
                continue
            duped_item = parse_group(group)
            total_score+= score_item(duped_item)
            group = []
    print(total_score)

is line 5 doing that?

my mental model of how iterators work is lacking

[frail schooner]

line 5 is just opening the input.txt file in read mode

if you want to parse the input from stdin, then I think you use open(0,'r')

[naive oriole]

I'm still finding more code golf tips

[frail schooner]

wait @naive oriole with the stdin open() thing, do you just run the code and then paste the input data in the terminal?

the program is just exiting for me

[naive oriole]

Yeah you paste it I think

[frail schooner]

would it work in any terminal? for example vscode?

[naive oriole]

Not sure actually

Other people were doing more of the testing

I was mainly just looking for more techniques

[frail schooner]

z=open(0,'r')

should this work?

in python?

or *z?

oh .read()..

[north creek]

My attempt to part 1:

sum(ord((c:=({*s[:(l:=len(s)//2)]}&{*s[l:]}).pop()).lower())-96+26*(c<'a')for s in I.strip().split())

101 character

[frail schooner]

aah only part 1

[north creek]

didnt read part 2 yet, wait a bit

[frail schooner]

aah

[naive oriole]

The part 1 section of the code:

*z,=open(0,'rb');u=sum;print(u((u({*a[(l:=len(a)//2):]}&{*a[:l]})-96)%58for a in z))

not sure how many chars

[frail schooner]

85

[naive oriole]

also I might have mismatched parentheses

in that one

[frail schooner]

no, it has no errors

[naive oriole]

good

[frail schooner]

as far as vscode is concerned

[grave vapor]

it took me almost an hour for part 2 (including me eating lunch) bc i didn't read it properly 😭

thought it said that we had to split it in half

not in groups of 3

[unique pivot]

rip

[grave vapor]

here's my solution for part 2 tho

@submit(2, 2022, 3, test=True, submit=False, parser=part_two_parser)
def part_two(daily_input):
    """Solution for AOC Day 3 Part 2."""
    return sum([ 
calculate_priority(set(batch[0]).intersection(set(batch[1])).intersection(set(batch[2])).pop())
        for batch in daily_input
    ])

oh that's formatted weirdly

[dark fjord]

; that's cheating

[north creek]

replace it with \n
anyway it is one char

[naive oriole]

replace the usage of the var

with the command

it would be shorter too

[north creek]

remove .strip -> 95 chars

[dark fjord]

I'm still thinking if my solution could be shorter ```py
print(sum([(ord(x)-96)%58 for r in data for x in set(r[:int(len(r) / 2)]) if x in set(r[int(len(r) / 2): len(r)])]))

[naive oriole]

you would have to also add [:-1] because of the ,= I think

so it would still be ~85 chars

but one line

[north creek]

# Part 1: 93 chars
sum(ord((c:=({*s[:(l:=len(s)//2)]}&{*s[l:]}).pop()).lower())-96+26*(c<'a')for s in I.split())
# Part 2: 97 chars
sum(ord((d:=({*a}&{*b}&{*c}).pop()).lower())-96+26*(d<'a')for a,b,c in zip(*[iter(I.split())]*3))

[lilac atlas]

pandas keeps rocking

m = pd.Index([*pd._testing._random.string.ascii_letters]).to_series().pipe(lambda s: s.groupby(s, sort=0).ngroup().add(1))

# Part 1
print(pd.read_csv("day_3.input", header=None, squeeze=True).pipe(lambda s: s.apply(list).explode().groupby(pd.RangeIndex(len(s)*2).repeat(s.str.len().floordiv(2).repeat(2)))).agg("".join).groupby(lambda idx: idx // 2).agg(list).apply(pd.Series).agg(lambda r: next(iter({*r[0]}.intersection(r[1]))), 1).map(m).sum())

# Part 2
print(pd.read_csv("day_3.input", header=None, squeeze=True).groupby(lambda idx: idx // 3).agg(list).apply(pd.Series).assign(**{"3": lambda fr: fr.iloc[:, :2].agg(lambda r: {*r[0]} & {*r[1]}, 1)}).rename(columns=int).agg(lambda r: next(iter({*r[2]} & r[3])), 1).map(m).sum())

[dark fjord]

Indeed they do.

[lilac atlas]

who needs import string when i can do pd._testing._random.string

[lyric folio]

lol

[broken condor]

Hey @naive oriole!

You either uploaded a .txt file or entered a message that was too long. Please use our paste bin instead.

[naive oriole]

@frail schooner
https://paste.pythondiscord.com/xocamutoce

[frail schooner]

I will save this for later reading! Thank you so much!

[naive oriole]

Yeah I'm probably missing some tips, and just keep in mind that sometimes these will not be shorter (there are always exceptions)

[naive bridge]

it's not much of a difference but you can remove the len(r) in the list splicing in the secod set, so it would be:

this:
print(sum([(ord(x)-96)%58 for r in data for x in set(r[:int(len(r) / 2)]) if x in set(r[int(len(r) / 2):])]))

instead of this:
print(sum([(ord(x)-96)%58 for r in data for x in set(r[:int(len(r) / 2)]) if x in set(r[int(len(r) / 2): len(r)])]))

[naive oriole]

Replace int(len(r) / 2) with len(r)//2, it does the same thing

[naive bridge]

true

so this:

print(sum([(ord(x)-96)%58 for r in data for x in set(r[:len(r) // 2]) if x in set(r[len(r) // 2:])]))

[naive oriole]

yeah

if you want to start golfing that you would start by using :=

[naive bridge]

yup thought about that

[naive oriole]

because you then don't need to calculate len(r) // 2 twice

[naive bridge]

exactly

[frail schooner]

is there a way to do something like print(x:= a for a in [0,1,2])?

like reduce for loop and printing the elements one by one into one line?

[naive oriole]

print(*(0,1,2))

or better yet

print(*"012")

equivalent to doing

print("0", "1", "2")

[frail schooner]

oh shit

[naive oriole]

if you need them each on their own line

it is going to be longer

[frail schooner]

sep='\n'?

[naive oriole]

yeah

but still longer

or

print("0\n1\n2")

is probably shortest

[frail schooner]

print(*data,sep='\n') I just did this?

[naive oriole]

that would prob work

[frail schooner]

aah okay

[naive oriole]

unless '\n'.join(data) is shorter but it looks longer

[frail schooner]

okay so again this * is just taken as a value in data without even having to iterate over it?

hmm..

[dark fjord]

Unpacking

[naive oriole]

print(*data) = print(data[0], data[1] ... data[-1])

each element is an argument

[frail schooner]

this would give error if the values were not str

[naive oriole]

also true

but join is longer anyways I thinl

[frail schooner]

by 1 char

yeah

print(*data,sep='\n')
print('\n'.join(data))

[naive oriole]

the real best answer would be if you shorten your variable name

[frail schooner]

true

print(*data*5)``` is weird af

[naive oriole]

n='\n'
p=print
p(*d,sep=n)
p(n.join(d))

[frail schooner]

just prints the contents 5 times

[naive oriole]

yeah its a multiplication and an unpacking

so it looks really weird

[frail schooner]

how would I do like each entry multiplied by 5?

[naive oriole]

hmm

you mean like
print(5*x for x in l)?

[frail schooner]

without the [x]

just x*5

[naive oriole]

oh ok

hm

it might have to just be that comprehension

and we can't use 5for because it could be 5f or or 5 for

although it might work in previous versions

[oak estuary]

My solution seems to skip a line if there are two consecutive matches 😭

[frail schooner]

two consecutive matches?

[oak estuary]

So my list is 299 when there's 300 characters

For for part A

For if there's a match of B for one line and another match of B on the second line, it skips it

[frail schooner]

did you check the number of lines in your data file?

[oak estuary]

Yes

It's 300

[frail schooner]

and the sum list has 299 entries?

[oak estuary]

Yep

[frail schooner]

just take the intersection of the two halves and use that?

can you send whatever conditional statements youre using?

[oak estuary]

Sure

from string import ascii_lowercase, ascii_uppercase

file = open('day_3.txt', encoding="utf-8")
last_i = None
counter = 0


stuff = []
for line in file.readlines():
    line = line.strip('\n')
    reference = int(len(n)/2)
    part_a = line[:reference]
    part_b = b[reference:]
    print([art_a, part_b)
    for i in part_a:
        for it in part_b:
            if i == it and last_i != i:
                stuff.append(i)
                last_i = i



table = [None]
for digit in ascii_lowercase:
    table.append(digit)

for digit1 in ascii_uppercase:
    table.append(digit1)

print(stuff)

for item in stuff:
    counter += table.index(item)
    if table.index(item) == -1:
        raise TypeError

print(counter)
file.close()

I should have probably used with but oh well (for file handling)

[dark fjord]

Doesn't really matter if your program stops in 20ms anyways

[wise rose]

I'm approaching it as a practice situation, practice makes permanent

[frail schooner]

b[reference:]?

[oak estuary]

Splits it in half

[wise rose]

it's the middle index, odd choice of variable name

[oak estuary]

odd choice yeah

[frail schooner]

no, what is the b list?

it should be line[reference:] no?

[oak estuary]

that's copied weirdly

[dark fjord]

Could someone maybe explain // rq?

[oak estuary]

yes it should be

[eternal cipher]

Floor division

Divide x by y and round down to next whole number

[wise rose]

as opposed to running the result through int?

[frail schooner]

// is floor division, it rounds down the result of / to the lowest integer value possible with rounding

so 9.6 result would give 9

instead of 10

[wise rose]

probably a bit more performant since it's staying in the same type the whole time

doesn't it come from bit shifting?

[frail schooner]

it does

[eternal cipher]

Oopsone sec

Mine

[frail schooner]

@oak estuary is it just me or is that another issue?

[oak estuary]

let me send an updated one, the variable names were a little confusing if I was going to share it

from string import ascii_lowercase, ascii_uppercase

file = open('day_3.txt', encoding="utf-8")
last_i = None
counter = 0


stuff = []
for line in file.readlines():
    line = line.strip('\n')
    reference = len(line)//2
    part_a = line[:reference]
    part_b = line[reference:]
    print(part_a, part_b)
    for i in part_a:
        for it in part_b:
            if i == it and last_i != i:
                stuff.append(i)
                last_i = i



table = [None]
for digit in ascii_lowercase:
    table.append(digit)

for digit1 in ascii_uppercase:
    table.append(digit1)

print(stuff)

for item in stuff:
    counter += table.index(item)
    if table.index(item) == -1:
        raise TypeError

print(counter)
file.close()

there

[frail schooner]

also dont use this

[dark fjord]

Is there also a ceil operator then?

[frail schooner]

use ```py
part_a&part_b

[frail schooner]

I think there is a ceil function somewhere? not sure might be in the math or numpy libraries

[frail sparrow]

!d math.ceil

[broken condor]

math.ceil

math.ceil(x)```
Return the ceiling of *x*, the smallest integer greater than or equal to *x*. If *x* is not a float, delegates to [`x.__ceil__`](https://docs.python.org/3/reference/datamodel.html#object.__ceil__ "object.__ceil__"), which should return an [`Integral`](https://docs.python.org/3/library/numbers.html#numbers.Integral "numbers.Integral") value.

[eternal cipher]

math.ceil but dont use that

You'll get floating point errors for some numbers

Use -(a // -b)

[frail schooner]

dont use this

set(part_a) & set(part_b)

[eternal cipher]

!e

a = 21
b = 8
print(-(a // -b))

[frail schooner]

to get the intersection

[broken condor]

@eternal cipher :white_check_mark: Your 3.11 eval job has completed with return code 0.

3

[wise rose]

what does the & operator do?

[eternal cipher]

.

[frail schooner]

!e print(set([1, 2, 3]&set([2,3,4]))

eh?

[oak estuary]

!e print(set([1, 2, 3]&set([2,3,4])))

[broken condor]

@oak estuary :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 1, in <module>
003 | TypeError: unsupported operand type(s) for &: 'list' and 'set'

[oak estuary]

hmm

[frail schooner]

aah

[oak estuary]

oh

[frail schooner]

!e print(set([1, 2, 3])&set([2,3,4]))

ffs

[wise rose]

roller coaster

[frail schooner]

!e print(set([1, 2, 3])&set([2,3,4]))

[broken condor]

@frail schooner :white_check_mark: Your 3.11 eval job has completed with return code 0.

{2, 3}

[frail schooner]

xD

it gives the common elements in both the sets

there's unions, intersections, etc.

[wise rose]

ok, why not use &? or the .intersection() method?

I suppose you can't do an early escape because we know from context that there'll only be one item

[frail schooner]

with .intersection() you would have to write lots of .intersections if you wanted to do more than two at a time

[wise rose]

so there's some performance left on the table

[frail schooner]

but with set(a)&set(b)&set(c)

its a single line to get the intersection between all the sets you want

[broken condor]

Hey @full scarab!

It looks like you tried to attach a Python file - please use a code-pasting service such as https://paste.pythondiscord.com

[wise rose]

you said not to use & but that reads like an argument in it's favor to me

[frail schooner]

no I said use &

[wise rose]

"dont use this"

?

[frail schooner]

I said dont use this cuz I didnt type cast it to set

[wise rose]

eff, am I more tired than I realize?

oh, ok

[frail schooner]

set takes unique values and then you can do the intersection

[full scarab]

part 1:

# priority = letters.index(LETTER) + 1
letters = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ"
total_priority = 0

with open("input.txt", "r", encoding="UTF-8") as f:
    for line in f:
        line = line.replace("\n", "")
        half_len = int(len(line) / 2)
        compartment_1, compartment_2 = set(line[:half_len]), set(line[half_len:])

        total_priority += letters.index(list(compartment_1.intersection(compartment_2))[0]) + 1

print(total_priority)```
part 2:
```py
# priority = letters.index(LETTER) + 1
letters = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ"
total_priority = 0
buffer = []

with open("input.txt", "r", encoding="UTF-8") as f:
    for count, line in enumerate(f):
        line = line.replace("\n", "")
        count += 1

        buffer.append(line)

        if count % 3 == 0:
            total_priority += letters.index(list(set(buffer[0]) & set(buffer[1]) & set(buffer[2]))[0]) + 1
            buffer = []

print(total_priority)    

anything that could be shortened/simplified?

[frail schooner]

why are you doing encoding?

just as a precaution?

[full scarab]

just used to it, pylint complains otherwise (i use vscode)

[frail schooner]

no worries

[frail schooner]

doesnt for me?

[hallow vapor]

int(len(line) / 2) -> len(line) // 2

[wise rose]

I personally would put list(compartment_1.intersection(compartment_2))[0] on it's own line just for readability

[hallow vapor]

intersection can be done with &

[full scarab]

is pylint activated? lol

[frail schooner]

yep

[full scarab]

yeah i was using that before but decided to use &

[frail schooner]

also you can ignore pylint errors per line if you want

[full scarab]

is there a difference or preference between & or .intersection apart from readability?

[frail schooner]

& can be used with multiple sets

and is shorter

[full scarab]

yeah had to use that for part 2

well it's probably good practice to specify encoding + i'm used to specifying it anyway 🤷‍♂️

[hallow vapor]

Also assigning to two variables two unrelated values is pretty meh

I would assign to two variables in the same line only if it was something like unpacking

[frail schooner]

@full scarab use to ignore the error on the line

[full scarab]

i thought i was going crazy, len(line) / 2 returns a float so i just cast it to int, i checked if all of the lines were even to determine to use // or not

[eternal cipher]

My message just sent like 20 mins late

[frail schooner]

xD

Man I really should go to sleep.. its 1:20AM..

but I really wanna sit here and try new things xD

[full scarab]

i'm aware, but if pep says it's better to specify encoding then i may aswell do it 🤷‍♂️
i'm doing this during the day anyway, if i was sweaty i'd probably not specify it

but sadly it releases at 4 am for me so can't compete

[frail schooner]

rip

thankfully its like 10:30AM release for me so its all good

[full scarab]

!e

a = [1,2]
print(len(a) / 2)

[broken condor]

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

1.0

[full scarab]

like what

[oak estuary]

I found out what happened, it's skipping a "B"

[frail schooner]

and why is it doing that?

[wanton arrow]

it's more about clarity than shortness. Also as we are indexing the list the extra chars would ruin the indexes.

[oak estuary]

I think its how I coded it

[frail schooner]

ah okay

[oak estuary]

the B matches are consecutive

but obviously my logic was to ensure that each line has a unique match - I obviously coded it wrong the first time

[frail schooner]

I think you made everything a little more complicated than was needed xD

[oak estuary]

Yeah 😂

thank you!

[frail schooner]

since you already have the uppercase and lowercase lists

[oak estuary]

Hoepfully part 2 is more straightforward

[hallow vapor]

you can also pop from the set, you don't need to convert it to a list

[frail schooner]

you could've just used the indexes to get the values

[full scarab]

oh nice, i did not think of that

[full scarab]

on a set?

[frail schooner]

not you, was talking to mesub

I will ping you when I reply to you xD

wrong emoji..

[hallow vapor]

line.strip() is a bit more elegant than line.replace("\n", ""), although for aoc I just read the whole input and then split by \n

[full scarab]

i also did from string import ascii_lowercase, ascii_uppercase but realized the string is literally hardcoded into the module so i copy pasted it into my open program

[frail schooner]

isnt there like a string callable for the whole list of alphabets?

thats seems like a missed opportunity

[real musk]

!e

from string import ascii_letters
print(ascii_letters)```

wait

[frail schooner]

I think I saw someone import something from string library in the morning.. dont remember it now

[hallow vapor]

in part two you do count += 1, but you can do enumerate(f, start=1)

[frail schooner]

or ascii_letters

[broken condor]

@real musk :white_check_mark: Your 3.11 eval job has completed with return code 0.

abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ

[frail schooner]

wait why didnt you just import that instead of those two separate ones xD

[full scarab]

.

[frail schooner]

meh, it would have taken less characters

but okay

[full scarab]

ah didn't know you could do that, thanks

[real musk]

you also could use ord

[full scarab]

well i don't think start= is required

[real musk]

wait

[full scarab]

since start is the 2nd arg

[dark fjord]

With all the new knowledge this is my final result for part 1

print(sum(((l:=len(r)//2,ord(*(set(r[:l])&set(r[l:])))-96)[1]%58for r in data)))

[real musk]

chr*

[full scarab]

what

[real musk]

talking about generating the alphabet

[full scarab]

for what?

[hallow vapor]

in general though I would just iterate over indices with a step of 3, or user more_itertools.chunked. I think that's better than appending to a list and then checking if it's modulo 3

[real musk]

generating the alphabet (but you have to use chr)

[full scarab]

i'm not too familiar with itertools, i'll check it out

[frail schooner]

!e print(sum(((l:=len(r)//2,ord(*(set(r[:l])&set(r[l:])))-96)[1]%58for r in [[1,2,3,4,5,6,7,8]])))

[broken condor]

@frail schooner :x: Your 3.11 eval job has completed with return code 1.

001 | <string>:1: SyntaxWarning: invalid decimal literal
002 | Traceback (most recent call last):
003 |   File "<string>", line 1, in <module>
004 |   File "<string>", line 1, in <genexpr>
005 | TypeError: ord() takes exactly one argument (0 given)

[frail schooner]

[dark fjord]

print(sum(((l:=len(r)//2,ord(*(set(r[:l])&set(r[l:])))-96)[1]%58for r in [[1,2,3,4,5,6,7,8]])))

[frail schooner]

!e print(sum(((l:=len(r)//2,ord(*(set(r[:l])&set(r[l:])))-96)[1]%58for r in [[1,2,3,4,5,6,7,8]])))

[broken condor]

@frail schooner :x: Your 3.11 eval job has completed with return code 1.

001 | <string>:1: SyntaxWarning: invalid decimal literal
002 | Traceback (most recent call last):
003 |   File "<string>", line 1, in <module>
004 |   File "<string>", line 1, in <genexpr>
005 | TypeError: ord() takes exactly one argument (0 given)

[frail schooner]

I am just copying your code

[dark fjord]

That isn't a valid input tough

[full scarab]

i was trying to simplify ```py
set(buffer[0]) & set(buffer[1]) & set(buffer[2])

e.g can't do the first two then do the last one (or am i wrong?)

[frail schooner]

oh right

!e print(sum(((l:=len(r)//2,ord(*(set(r[:l])&set(r[l:])))-96)[1]%58for r in [['xAbchasjhHgsdG']])))

[broken condor]

@frail schooner :x: Your 3.11 eval job has completed with return code 1.

001 | <string>:1: SyntaxWarning: invalid decimal literal
002 | Traceback (most recent call last):
003 |   File "<string>", line 1, in <module>
004 |   File "<string>", line 1, in <genexpr>
005 | TypeError: ord() takes exactly one argument (0 given)

[frail schooner]

!e print(sum(((l:=len(r)//2,ord(*(set(r[:l])&set(r[l:])))-96)[1]%58for r in ['xAbYHu'])))

[broken condor]

@frail schooner :x: Your 3.10 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 1, in <module>
003 |   File "<string>", line 1, in <genexpr>
004 | TypeError: ord() takes exactly one argument (2 given)

[sand jungle]

check pinned for best golf till now. current 144 chars

[frail schooner]

!e print(sum(((l:=len(r)//2,ord(*(set(r[:l])&set(r[l:])))-96)[1]%58for r in ['xAbYHu'])))

[broken condor]

@frail schooner :x: Your 3.11 eval job has completed with return code 1.

001 | <string>:1: SyntaxWarning: invalid decimal literal
002 | Traceback (most recent call last):
003 |   File "<string>", line 1, in <module>
004 |   File "<string>", line 1, in <genexpr>
005 | TypeError: ord() takes exactly one argument (0 given)

[frail schooner]

!e print(sum(((l:=len(r)//2,ord(*(set(r[:l])&set(r[l:])))-96)[1]%58for r in ['xAbYAu'])))

[broken condor]

@frail schooner :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | <string>:1: SyntaxWarning: invalid decimal literal
002 | 27

[frail schooner]

okay finally

[oak estuary]

okay I found the actual solution to what I had 😭

[oak estuary]

basically for every new line, I had to stop comparing to the last line

which fixed it

!!!!

[frail schooner]

xD

[dark fjord]

part 2 :3 ```py
print(sum([(ord(*(set(x[0])&set(x[1])&set(x[2])))-96)%58 for x in [data[i:i+3] for i in range(0, len(data), 3)]]))

[sand jungle]

check pinned for best golf till now. current 144 chars

[dark fjord]

I saw but I don't like that "short" one. It's good but using ; is kinda cheating imo

[cursive sail]

you can shave off some whitespace there

print(sum([(ord(*(set(x[0])&set(x[1])&set(x[2])))-96)%58for x in[data[i:i+3]for i in range(0,len(data),3)]]))

[woeful phoenix]

what do the different codes mean in this? (lf, cr...)

[cursive sail]

that's the newline sequence

windows uses carriage return (cr) line feed (lf), unix uses line feed, old macos uses carriage return

[dark fjord]

Actually I updated it to ```py
print(sum([(ord(*(set(x)&set(y)&set(z)))-96)%58 for x,y,z in zip(3[iter(data)])]))

[cursive sail]

the space before the for is unneeded

[dark fjord]

lmao yeah but my ide is on drugs if I do that.

And... I like it like this. I'm gonna stop now... ||maybe if I kept going it will be 10 characters long.||

[wanton arrow]

does that get the right answer?
ah the iter pops 1 each time? so the zip does work?

[dark fjord]

Yep

[wanton arrow]

what does your data = line look like? You could move the ord part there.

[azure cargo]

sets are cool py def day3(inp: str) -> tuple[int, int]: sco = 0 slc = inp.strip().split("\n") for i in slc: num = int(len(i) / 2) a, b = i[:num], i[num:] ins = list(set(a) & set(b))[0] sco += list(" abcdefghijklmnopqrstuvwxyz" + "abcdefghijklmnopqrstuvwxyz".upper()).index(ins) gsco = 0 for i, j, k in zip(slc[::3], slc[1::3], slc[2::3]): gsco += list(" abcdefghijklmnopqrstuvwxyz" + "abcdefghijklmnopqrstuvwxyz".upper()).index(list(set(i) & set(j) & set(k))[0]) return (sco, gsco)

[wanton arrow]

add a .splitlines() to the text test data

[dark fjord]

f.read().split("\n")

[azure cargo]

nonono BAD

use f.read().strip().splitlines()

trailing newline

[dark fjord]

!e ```py
data = """dWlhclDHdFvDCCDfFq
mGdZBZBwRGjZMFgvTvgtvv
jwwJrzdzGdSbGGnNlzWczHzPHPhn
cczcbMBszhzzDBTBPPPGjtvtlt
LqJLfpwdLnvQLRGQjGtj
gSgnSJJCGSGpGSrwgfhchmmmHzcrHDmbrmMm""".split("\n")

print(sum(((l:=len(r)//2,ord(*(set(r[:l])&set(r[l:])))-96)[1]%58 for r in data)))

[broken condor]

@dark fjord :white_check_mark: Your 3.11 eval job has completed with return code 0.

192

[dark fjord]

Doesn't matter since it won't have any matches.

[azure cargo]

but for the actual data...

[dark fjord]

There aren't any trailing whitespaces so it doesn't matter.

[azure cargo]

there are

trailing newlines

[fleet ivy]

#part 1
def dupe(sttr):
    '''finds dupe characters in the string but needs to make sure
    :param sttr: string
    :return string: character that is in twice, but only once in each half of the string
    '''
    for char in sttr:
        if sttr[len(sttr) // 2:].count(char) >= 1 and sttr[:len(sttr) // 2].count(char) >= 1:
            return char
def summer(lll):
    '''sums up the priority score for each duplicate character
    :param lll: list of string
    :return int: sum of priority scores
    '''
    import string as str
    PRIORITIES = dict(zip(str.ascii_lowercase, range(1,28)))
    UPPERCASE = dict(zip(str.ascii_uppercase, range(27,53)))
    PRIORITIES.update(UPPERCASE)

    sum = 0
    for line in lll:
        sum += PRIORITIES[dupe(line)]
    return sum

with open('advent3.txt', 'r') as file:
    text = file.read().splitlines()
print(summer(text))

#part2
def dupe2(sttr1, sttr2, sttr3):
    ''' finds the character in three given strings
    :param sttr1: string
    :param sttr2: string
    :param sttr3: string
    :return string: character that is in twice, but only once in each half of the string
    '''
    for char in sttr1:
        if char in sttr2 and char in sttr3:
            return char

def summer2(lll):
    '''sums up the priority score for each duplicate character in the three string group
    :param lll: list of list of string
    :return int: sum of priority scores
    '''
    import string as str
    PRIORITIES = dict(zip(str.ascii_lowercase, range(1,28)))
    UPPERCASE = dict(zip(str.ascii_uppercase, range(27,53)))
    PRIORITIES.update(UPPERCASE)

    sum = 0
    for list in lll:
        sum += PRIORITIES[dupe2(list[0], list[1], list[2])]
    return sum

with open('advent3.txt', 'r') as file:
    text = file.read().splitlines()
lolos = [[text[index], text[index + 1], text[index + 2]] for index in range(0, len(text), 3)]
print(summer2(lolos))

I didn't even know about sets. 🤡 🤡 🤡 🤡

[dark fjord]

Not in my input.

[dark fjord]

Then you're going to find & hilarious.

[fleet ivy]

is that the same thing as set?

[dark fjord]

No it shows if both sets have the same values.

[fleet ivy]

bloody hell

[dark fjord]

!e ```py
print({1, 2, 3} & {3, 4, 5})

[broken condor]

@dark fjord :white_check_mark: Your 3.11 eval job has completed with return code 0.

{3}

[fleet ivy]

Thank you

[cursive sail]

set intersection is great today

i'm very proud of how concise my solution was because of it

[dark fjord]

What did you come up with?

[cursive sail]

# part 1
s = 0
for r in data.splitlines():
    a, b = r[:int(len(r)/2)], r[int(len(r)/2):]
    c = list(set(a) & set(b))[0]
    s += string.ascii_letters.index(c) + 1
print(s)

# part 2
s = 0
for g in chunked(data.splitlines(), 3):
    c = list(set(g[0]) & set(g[1]) & set(g[2]))[0]
    s += string.ascii_letters.index(c) + 1
print(s)

that was not particularly clean though

i was trying to get a solution as fast as possible

the int(len(r)/2) should really have been floor div

[proper barn]

What's chunked?

[cursive sail]

!d more_itertools.chunked

[broken condor]

more\_itertools.chunked

more_itertools.chunked(iterable, n, strict=False)```
Break *iterable* into lists of length *n*:

```py
>>> list(chunked([1, 2, 3, 4, 5, 6], 3))
[[1, 2, 3], [4, 5, 6]]
```  By the default, the last yielded list will have fewer than *n* elements if the length of *iterable* is not divisible by *n*:

```py
>>> list(chunked([1, 2, 3, 4, 5, 6, 7, 8], 3))
[[1, 2, 3], [4, 5, 6], [7, 8]]
```  To use a fill-in value instead, see the [`grouper()`](https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.grouper "more_itertools.grouper") recipe...

[cursive sail]

it's coming as itertools.batched in 3.12

[earnest sail]

you can also find something equivalent named grouper in https://docs.python.org/3/library/itertools.html#itertools-recipes

[stray spear]

curious why they decided on batched as a name

[cursive sail]

i dunno

that's the name of it they use in the recipe in 3.11

[full bane]

I thought they use grouper

[cursive sail]

so maybe they just mostly copied that

oh they have both

def batched(iterable, n):
    "Batch data into lists of length n. The last batch may be shorter."
    # batched('ABCDEFG', 3) --> ABC DEF G
    if n < 1:
        raise ValueError('n must be at least one')
    it = iter(iterable)
    while (batch := list(islice(it, n))):
        yield batch
```is in there

[stray spear]

"chunks" or "chunked" is the kind of term I see a lot

batched...not so much

[cursive sail]

same, i like chunked

notably they don't have the strict from more_itertools

[stray spear]

I also have issues with pairwise

[solid estuary]

Rust Structs for the win

[stray spear]

because from math that feels like it should be about all pairs

not adjacent pairs

rust structs aren't that special

rust's enums are fun though

[solid estuary]

just go line by line with a counter, adding each line to a list, when the counter reaches 3 go to the next list index

[stray spear]

or nicer https://doc.rust-lang.org/std/primitive.slice.html#method.chunks

[solid estuary]

neato

[hallow vapor]

nothing like the trick with zip

[stray spear]

sadly array_chunks https://github.com/rust-lang/rust/issues/74985 isn't in stable

[cursive sail]

eh who doesn't use nightly anyway /hj

if you're writing a binary, nightly is fine

it's not like it breaks all that often

[stray spear]

I usually do

this time I ended up using tuples from itertools

I'm assuming I could drop the 53 from the type side

or maybe something better still

oh _ for array lengths is not in stable

sad

ok, I'm dumb, I can do

let mut counts = [0u8; 53];

[stray spear]

wow that is so long

||pls don't make the obvious joke||

[mystic kindle]

day three 🎉

here was my solution ```py
priorities = dict(zip(string.ascii_letters, range(1, 53)))

def part_1(data: list[str]) -> int:
halves = [
next(iter(set(i[: int(len(i) / 2)]).intersection(set(i[int(len(i) / 2) :]))))
for i in data
]
priorities_sum = sum(priorities[s] for s in halves)

return priorities_sum

def part_2(data: list[str]) -> int:
chunks = chunked(data, 3)
badges = [
next(iter(set(a).intersection(set(b)).intersection(set(c))))
for a, b, c in chunks
]
priorities_sum = sum(priorities[s] for s in badges)

return priorities_sum

sets are nice

oh, you can do intersection with &, can't you

[cursive sail]

mhm

[mystic kindle]

ah, my solution would've been more concise then

yeah that's better ```py
priorities = dict(zip(string.ascii_letters, range(1, 53)))

def part_1(data: list[str]) -> int:
halves = [
next(iter(set(i[: int(len(i) / 2)]) & set(i[int(len(i) / 2) :]))) for i in data
]
priorities_sum = sum(priorities[s] for s in halves)

return priorities_sum

def part_2(data: list[str]) -> int:
chunks = chunked(data, 3)
badges = [next(iter(set(a) & set(b) & set(c))) for a, b, c in chunks]
priorities_sum = sum(priorities[s] for s in badges)

return priorities_sum

[solid estuary]

still blazing fast

[eternal cipher]

https://cdn.discordapp.com/attachments/1048465205916598353/1048684077127061524/image.png

This was mine

They're one liners it's just formatted

[fallen sinew]

||never gonna give you up||

[full scarab]

with open("the.txt","r") as f:
    inpt = f.read().split("\n")
    letters = "abcdefghijklmnopqrstuvwxyz"
    letters += letters.upper()
    x = []
    total = 0
    dic = {}
    for i in range(1,53):
        x.append(i)
    the_list = list(zip(letters,x))
    for i in the_list:
         dic[i[0]] = i[1]
    print(dic)
    print(dic)
    for i in inpt:
        number_of_elements = len(i)
        half_indx = (number_of_elements/2) -1
        first_half = i[0:int(half_indx)]
        second_half = i[int(half_indx):]
        common_element = set(first_half).intersection(second_half)
        for n in common_element:
            total += dic[n]
    print(total)```

It says too low

Why?

[eternal cipher]

You're taking away 1 when you divide, why?

[stray spear]

I mean the task is super tiny, basically any solution will be fast, mightnas well write something sebsible
https://github.com/algmyr/aoc-2022/blob/master/src/solutions/day03.rs

[mystic kindle]

time to do it in rust

[solid estuary]

used no outside info, the only thing I looked at was the aoc prompt

[stray spear]

the time taken is...insignificant

I wonder what the first day with some meat to it will be

[cursive sail]

what benchmarking lib is that?

i've only used criterion

[stray spear]

none

[cursive sail]

ah

[full bane]

i hand rolled my benchmark thing also. your percentage progress bar thing looks nice

[stray spear]

I have my own code for timing runs

and yeah, the display code is probably where the biggest chunk of work went 😛

[full bane]

i wrote a macro to generate a match case for all the days, which was fun

[wanton arrow]

what does it do?

[stray spear]

I wrote a macro too, but nothing clever

[full scarab]

Index star from 0

While the len starts at 1

[full bane]

so in rust you can't dynamically use things, but i wanted to do that. i wrote a proc macro that generated a function with a match case that went

match day {
  0 => {
    // load day 0
  }
  ...
}
``` basically

[full scarab]

Oh wait

[cursive sail]

mind just does the tiniest bit more

macro_rules! days {
    ($($day:literal)*) => {
        paste! {
            $(mod [<day $day>];)*

            pub fn parts(day: u8) -> (fn(&str) -> String, fn(&str) -> String) {
                match day {
                    $($day => ([<day $day>]::part_one, [<day $day>]::part_two),)*
                    _ => panic!("invalid day"),
                }
            }
        }
    };
}

days! {
    01
}

the machinery is very very yucky

[mystic kindle]

man I'm just manually cargo running all my solutions

[cursive sail]

i might make it into a trait

[mystic kindle]

I was thinking of writing some small helper to automatically fetch and submit stuff, though

[full bane]

your solutions are forced into returning strings?

[cursive sail]

right now. i'll planning to change that later.

whenever i stop being lazy

will probably do some trait nonsense or something to make it cleaner

[full bane]

currently my thing just runs the functions. i want to return it also, but then i'd either have to make the functions all return strings which is sad or something else

[cursive sail]

i'll probably just make it generics

maybe a trait with associated types

[stray spear]

I have a run result struct that holds Box<dyn Display> for the answers

dumb but works

[cursive sail]

Box bad /hj

*const () + TypeId >>>

[stray spear]

I wouldn't use it for anything performance critical

[cursive sail]

i mean Box isn't actually horrible

but i hate that it's magic

since it has the aliasing rules of &mut T or something like that

[stray spear]

it's a very good type for what's actually needed, just something that I can display

though I guess just wrapping stuff in format! works as well

and just deal with strings

but you can at least move that out of solutions

[cursive sail]

(don't actually do this, but it's fun to when you can)

[wanton arrow]

it still feels like python should have an easier way of splitting a string in 1/2 the three methods I've seen so far all feel clunky, and 2 of them involve downloading other libraries.

[mystic kindle]

I just sliced by dividing the len in half

and they're all even, so it'll work fine

[stray spear]

yeah, I would say this is quite nice already

n = len(s)
s[:n//2]
s[n//2:]

[full bane]

but in rust 👀 let (l, r) = s.split_at(s.len() / 2); 👀

[mystic kindle]

more-itertools has split_at too >.>

[wanton arrow]

more itertools has divide

that was one of the 2

numpy reshape was the other.

map(set,np.fromiter(map((' '+items).index,d),int).reshape(2,-1))

[mystic kindle]

oh right

guess I'll refactor to use that

[stray spear]

eww

[wanton arrow]

yeah

[wanton arrow]

import numpy as np
from string import ascii_letters as items
data = [[*map((' '+items).index,line.strip())] for line in open(filename)]

λ=lambda d:set.intersection(*map(set,np.array(d).reshape(2,-1))).pop()
print(sum(map(λ,data)))

[mystic kindle]

oh, popping is probably a nicer way to get the first item

I did next(iter(...)) >.>

[wanton arrow]

yeah .pop() seems simple.

[stray spear]

!e

s = 'abcdefgh'
a, b = zip(*zip(*len(s)//2*[iter(s)]))
print(a, b)

[broken condor]

@stray spear :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 2, in <module>
003 | ValueError: too many values to unpack (expected 2)

[stray spear]

err, maybe I should have test run this

oh, one zip too many

!e

s = 'abcdefgh'
a, b = zip(*len(s)//2*[iter(s)])
print(a, b)

[broken condor]

@stray spear :white_check_mark: Your 3.11 eval job has completed with return code 0.

('a', 'b', 'c', 'd') ('e', 'f', 'g', 'h')

[stray spear]

it's terrible, but it works

[dark fjord]

!e ```py
s = 'abcdefgh'
print(*zip(4[iter(s)]))

[wanton arrow]

*zip

[broken condor]

@dark fjord :white_check_mark: Your 3.10 eval job has completed with return code 0.

('a', 'b', 'c', 'd') ('e', 'f', 'g', 'h')

[mystic kindle]

oh hey, this works ```py
priorities_sum = sum(
[priorities[(set(a) & set(b)).pop()] for a, b in [divide(2, i) for i in data]]
)

[wanton arrow]

l=len(s:="abcdef")//2
print(*zip(*l*[iter(s)]))

[mystic kindle]

that's part 1 one-lined, pretty much

[wanton arrow]

yeah that's a lot cleaner to read too.

except I hate seeing i be a non-int. 🙂

[mystic kindle]

lol, I read it as item

maybe l for line?

[wanton arrow]

d ?

r?

[mystic kindle]

sure

[wanton arrow]

map

[stray spear]

remove the []

[mystic kindle]

I was doing some fancy str.maketrans stuff before I realized I could just skip it and do ```py
priorities = dict(zip(string.ascii_letters, range(1, 53)))

[mystic kindle]

ah right, sum will take the generator

[full bane]

the chad (string.ascii_lowercase + string.ascii_uppercase).index

[wise rose]

I dunno, left, right = string[mid:], string[:mid] is pretty legible

ok, I got those backwards, but still

[mystic kindle]

it would've been if I calculated mid on its own line >.>

[stray spear]

I did

fn priority(n: u8) -> u8 {
  if n & 32 == 0 {
    (n & 31) + 26
  } else {
    n & 31
  }
}

because knowing the logic of ascii is neat

[wanton arrow]

halves = partial(divide,2)
for a,b in map(halves,data)

[mystic kindle]

a map would be more concise here, yeah

[wanton arrow]

fn priority(n: u8) -> u8 {
    n & 31 + 26 * !(n & 32)
}

[cursive sail]

not that it matters for this challenge

[wanton arrow]

can you do ! for not?

[full bane]

yeah

[wanton arrow]

n & 31 + 26 * !(n & 32)

[stray spear]

err 1 - (n&32) is for sure wrong

[wanton arrow]

yeah, at a min it needs the ==0 in there

[mystic kindle]

    priorities_sum = sum(
        priorities[(set(a) & set(b) & set(c)).pop()] for a, b, c in chunked(data, 3)
    )
``` part 2 one-lined too

[wanton arrow]

I'll cut myself some slack for using a language I don't know.

is it rust ?

[stray spear]

it is

[dark fjord]

chunked hmmm

[mystic kindle]

more-itertools!

[stray spear]

fn is a giveaway

[wanton arrow]

chunked is good

[mystic kindle]

actually, I think salt's aoc_lube has a chunked function too, but whatever

[clever dune]

yeah, it's just chunk though

[wanton arrow]

yeah, I'm sure there are other languages that use it too.

[clever dune]

i don't live in the past

[mystic kindle]

lolol

[stray spear]

I don't know of any off hand

[clever dune]

actually found that often the easiest way to parse input is with a combination of extract_ints and chunk

[stray spear]

fun, func, function, defun, def

those I have seen around

not much fn

[wanton arrow]

I just did index + 1 or added a space at the start and then indexed.

map((' '+items).index,line.strip())

[clever dune]

like, 2018 day 3 input looked like:

#1 @ 596,731: 11x27
#2 @ 20,473: 23x22
#3 @ 730,802: 23x23

and you can just

chunk(extract_ints(raw_input), 5)

[mystic kindle]

gleam uses fn, but its a lesser known thing

[mystic kindle]

mm, yeah that's more concise

[cursive sail]

elixir uses fn

[mystic kindle]

I thought elixir used def?

oh, anonymous functions?

[stray spear]

lol, if there is any critique I can give to AoC, many of the input formats kinda suck

[cursive sail]

i think it uses fn for lambdas? i'd have to check.

[clever dune]

there's a library for reverse format strings that can help parse if you have issues

[mystic kindle]

yeah it uses fn for anonymous functions

[stray spear]

if I really need it I'll probably just bash things with regex or do something hacky

[grave vapor]

omg i should've used binary and

i did. intersection

[mystic kindle]

I did at first too >.>

I switched when I went and tried to shorten everything

[grave vapor]

i really wanna do aoc in kotlin but there aren't many good helpers 😭

[clever dune]

nim uses proc

[mystic kindle]

nim seems cool

[clever dune]

func is reserved for procs with no side effects

it's pretty cool

[wanton arrow]

nim does seem cool, and iirc can use python libraries and has an option to compile to python code or something like that?

I just don't have time at the moment for yet another language. 🙂

[clever dune]

yes, i use aoc_lube to fetch input in nim

import nimpy

proc fetch*(year, day: int): string =
  ## Get AoC input.
  "aoc_lube".pyImport.callMethod(string, "fetch", year, day)

that star after the proc name means it's public function

[wanton arrow]

it just feels like nim should be a natural progression from python, there's a lot of coupling there. I hope it does well in the future.

[mystic kindle]

I wanted to do AOC in haskell, but I forgot to properly learn it 😔

[clever dune]

my day 3 solution:

include aoc

type Sacks = seq[string] | seq[seq[char]]

let sacks = fetch(2022, 3).splitLines()

proc priority(c: char): int =
  1 + c.ord - (if c >= 'a': 'a'.ord else: 'A'.ord - 26)

proc sumPriorities(groups: seq[Sacks]): int =
  sum groups.mapIt(
    it.mapIt(it.toSet).foldl(a * b).chooseOne.priority
  )

part 1: sumPriorities sacks.mapIt(it.distribute(2))
part 2: sumPriorities sacks.chunk(3)

[wanton arrow]

that doesn't seem like a great choice for syntax though, like one of the devs was being clever instead of having a "public" keyword.

[mystic kindle]

yeah, I'm not a fan of the star

[full bane]

feels like latex

[naive oriole]

does nim support macros

[clever dune]

i prefer it, it makes it very easy to see public functions at a glance with little boilerplate

yes, nim has templates and macros

[mystic kindle]

templates, like, C++ templates?

[naive oriole]

so technically you could just make a macro replacing public with *

[clever dune]

templates are like compile-time code replacements

[naive oriole]

templates would probably work better

[wanton arrow]

like pub proc fetch( seems better. but I'm complaining about a language I don't use so I have little room to stand on.

[naive oriole]

so you could put the public first

[mystic kindle]

variable names can have spaces?

[stray spear]

AoC in LaTeX 👀

[clever dune]

did you know that snake_case and snakeCase are the same name in nim

[stray spear]

(lualatex is cheating)

[clever dune]

yeah, but you need to use ```

[mystic kindle]

I see

[clever dune]

actually i don't know if spaces work

[mystic kindle]

what's up with part 1, then?

[cursive sail]

what the fuck i hate it

[stray spear]

Writing an explanatory pdf with latex code computing the answer would be wack

[clever dune]

but operators are done with:

proc `+`(a: MyClass): MyClass =
  ...

[mystic kindle]

does it ignore cases and special characters?

[naive oriole]

the devs were really trying to be clever with nim

[clever dune]

no, the first letter case matters

[cursive sail]

impure operator definitions 👀

[full bane]

i think it's turing complete

[mystic kindle]

tfw arithmetic has side effects >.>

[stray spear]

there is clever, and there is that

[clever dune]

the reason they chose to do that is not to force a style on anyone -- similarly you can take a c-library and bind the functions to nim and keep your libraries style

[naive oriole]

still less "clever" than JS (==, var, {} + {})

[naive oriole]

I mean the naming thing isn't a huge issue

[stray spear]

JS isn't clever, it's a clusterfuck

[naive oriole]

you don't really need two variables with a nearly exact same name

[stray spear]

so I could alternate between camel_case and snakeCase between every usage? sounds like fun

[mystic kindle]

aLMOST_SCREAMING_SNAKE_CASE

[clever dune]

the reason i like nim though is the uniform function call syntax:

type Vector = tuple[x, y: int]
 
proc add(a, b: Vector): Vector =
  (a.x + b.x, a.y + b.y)
 
let
  v1 = (x: -1, y: 4)
  v2 = (x: 5, y: -2)
 
  # all the following are equivalent
  v3 = add(v1, v2)
  v4 = v1.add(v2)
  v5 = v1.add v2

[full bane]

core dev sponsored obfuscation

[cursive sail]

i can get on board with those first two, but the last scares me

[lunar lance]

You don't need, but they are telling me I can't.
I think it is my decision as a programmer to make, not theirs as a language designer.

[cursive sail]

i don't mind parenless function application, but i've never seen them optional

[clever dune]

you whitespace hater

[naive oriole]

https://en.wikipedia.org/wiki/Whitespace_(programming_language)

[clever dune]

but it makes dot auto-complete really good

[cursive sail]

i like that

i use haskell sometimes, which does function application without parens too

but having both is weird to me

[clever dune]

i really hate parens, mostly because of the difficulty in typing them

shift + 9/0 is the most awkward combination

[naive oriole]

Shift + 9 and Shift + 0 are so far

Though none of !@#$%^&*() are particularly simple, yet are used very often

[clever dune]

the curly and square brackets are much better

[naive oriole]

[]{} are fine

also

-=_+ are annoying

particularly +_

[mystic kindle]

That's why real programmers swap the top letter row with the number row

[cursive sail]

i usually keep my right hand a little further right on my keyboard than normal

so it's strange to hear that these are difficult for some

[mystic kindle]

I don't have much problem typing special characters, yeah. But I guess it does vary person to person

[naive oriole]

not difficult per se but difficult compared to asdfhjlk

[clever dune]

i do shift combinations one-handed --- pinky on right shift + middle-finger on parens is so weird

[cursive sail]

i've been tempted to completely remap certain key combos on my keyboard to arbitrary symbols, but i can't figure out how to do it nicely on windows

[naive oriole]

Razer supports complete keyboard remapping

[stray spear]

[](){} is good to practice of you need to type C++ lambdas

[mystic kindle]

are you on a laptop?

[dark fjord]

Maybe something for u

[naive oriole]

i hate this

[mystic kindle]

Or do you use an external keyboard?

[naive oriole]

I use a laptop

that I literally never move

[dark fjord]

I got the keychron k8 pro recently.

[mystic kindle]

mechanical keyboards >>

[stray spear]

sadly []()<>{} isn't quite valid C++

[clever dune]

if you want to see some crazy macros -- this is me trying to implement multiple inheritance in nim with macros:
https://github.com/salt-die/nimheritance/blob/main/nimheritance/classes.nim

[stray spear]

but very close

[mystic kindle]

I've been putting together my first custom board for the past few weeks

[dark fjord]

@clever dune Use <url> to suppress embeds.

[naive oriole]

wow that is unique although tbf I'm not particularly familiar with nim

[stray spear]

oh god, vtables

[naive oriole]

the macro *class looks insane for a macro

[stray spear]

I know they exist and what they are for, but I don't want to think about them...

[clever dune]

for this, they only exist at compile time

spits out some normal nim objects and functions at the end

i've seen some aoc helpers in nim that use macros that turn something like:

day 1:
    part 1:
        some_code()
    
    part 2:
        some_more_code()

into auto-submissions

[stray spear]

idk, I'm all for automating tedious things

[naive oriole]

I mean that sounds good

[stray spear]

but I've never seen a need for that

submitting is close enough to being a one-off thing that I don't feel the need to automate it

[naive oriole]

I've just been too lazy to automate it honestly

[stray spear]

so you save a few seconds 25 times a year

let me find the relevant xkcd...

[wanton arrow]

I'm not fast enough that automating the submission would help, worse it might get the wrong answer. lol

[clever dune]

it's not just saving time --- it's very easy to read

[stray spear]

https://xkcd.com/1205/

xkcd: Is It Worth the Time?

Don't forget the time you spend finding the chart to look up what you save. And the time spent reading this reminder about the time spent. And the time trying to figure out if either of those actually make sense. Remember, every second counts toward your life total, including these right now.

(and don't get me wrong, I love automating stuff, I use vim and write a lot of small scripts)

(thinks just have to start annoy me by being tedious for me to automate them)

[quiet heath]

whilst others code golf, I've been adding type hints to my code instead XD

all my solutions pass mypy with strict mode

[plush crater]

uppercase = string.ascii_uppercase
lowercase = string.ascii_lowercase
letters =   dict(enumerate((lowercase + uppercase),1))

common_letters = ''
total = 0
def clean(first_half,second_half):
    global common_letters 
    for letter in first_half:
        if letter in second_half:
            common_letters += letter
            return

with open('zzzinp.text','r') as file:
    compartment = file.readlines()
    for line in compartment:
        half = len(line)// 2
        clean(line[:half],line[half:])

# 
def get_key(val,my_dict):
    for key, value in my_dict.items():
        if val == value:
            return int(key)
    return 0
print(common_letters)
for common in common_letters:
    total += get_key(common,letters)
print(total) ```

onto 2nd

[wanton arrow]

!global

[broken condor]

When adding functions or classes to a program, it can be tempting to reference inaccessible variables by declaring them as global. Doing this can result in code that is harder to read, debug and test. Instead of using globals, pass variables or objects as parameters and receive return values.

Instead of writing

def update_score():
    global score, roll
    score = score + roll
update_score()

do this instead

def update_score(score, roll):
    return score + roll
score = update_score(score, roll)

For in-depth explanations on why global variables are bad news in a variety of situations, see this Stack Overflow answer.

[wanton arrow]

also, you can reverse the indexes of a dict by swapping the key/value.

letters = {v:k for k,v in enumerate((string.ascii_letters),1)}

oh and ascii_letters is lowercase + uppercase, so you can use that too. 🙂

[plush crater]

thanks man

i spent so much time on y it isen't possible to get key from value

😭