Just more NoneType error bs
#AoC 2022 | Day 7 | Solutions & Spoilers
teal orchid
summer herald
As I said before, when I was using treelib I got around that by just having a counter variable and tacking that on to the directory names then incrementing it so each one was unique
so instead of lmw, lmw, lmw you'd have 45 lmw, 83 lmw, 133 lmw
bronze leaf
nope, i noticed after watching my visualization that each directory is only ever visited once, and in DFS order
proven gyro
Except for when you try to iterate over it
cedar trout
the same what? size?
if there's a chain of nested folders without any files inside of them then they would all have the same size
teal orchid
the same name I guess
teal orchid
so instead of `lmw, lmw, lmw` you'd have `45 lmw, 83 lmw, 133 lmw`
yeah, i was looking at that and I added a uuid attribute to the folder, but im not really sure how i would use that to find the correct folder
cedar trout
nope, i noticed after watching my visualization that each directory is only ever...
they could have been cheeky about it and break the dfs algorithm a couple of times, they don't really tell you that that's what they are doing and you wouldn't notice unless you check the input looking for it
cedar trout
the same name I guess
they can have the same name (their absolute path would be different)
vivid patrol
Except for when you try to iterate over it
!e
a = {0: "a", 1: "b", 2: "c"}
for i in range(len(a)):
print(a[i])
tame groveBOT
@vivid patrol :white_check_mark: Your 3.11 eval job has completed with return code 0.
001 | a
002 | b
003 | c
Yeah but a list doesn't have .values()
vivid patrol
!e
a = {0: "a", 1: "b", 2: "c"}
for i in range(len(a)):
print(a[i])
@vivid patrol :white_check_mark: Your 3.11 eval job has completed with return code 0.
001 | a
002 | b
003 | c
yeah, i was comparing the names of the folders parents but that didnt end up working either
vivid patrol
if it works it works
proven gyro
If you use something the wrong way twice, it cancels out
tough hinge
If you use something the wrong way twice, it cancels out
What if you use it three times?
tough hinge
Can you show your work?
inner ingot
because order of operations
proven gyro
Sigh
(-1) ** 3
tough hinge
!eval print((-1) ** 3)
tame groveBOT
@tough hinge :white_check_mark: Your 3.11 eval job has completed with return code 0.
-1
Checks out
proven gyro
And there you go
vivid patrol
!e print(-1 ** 2)
tame groveBOT
@vivid patrol :white_check_mark: Your 3.11 eval job has completed with return code 0.
-1
tame groveBOT
@proven gyro :white_check_mark: Your 3.11 eval job has completed with return code 0.
1
yeah
pemdas 😔
though usually when you see -1 you assume there's parenthathese around it
proven gyro
Yeah that's quite true
vivid patrol
i seriously can't speak english rn
anyone have something like this ready to copy and paste kek
tough hinge
Isn't that for day 6?
proven gyro
I just have my for line in lines loop ready
vivid patrol
lmao
balmy sable
https://paste.pythondiscord.com/nehahaqaxa could someone look at this code for part 1, it works for the example input but not for mine and I have no idea why
vivid patrol
from a quick look are you assuming each directory has a unique name?
because that's not the case
vivid patrol
$ cd /
$ ls
dir a
14848514 b.txt
8504156 c.dat
dir d
$ cd a
$ ls
dir e
29116 f
2557 g
62596 h.lst
$ cd e
$ ls
584 i
$ cd ..
$ cd ..
$ cd d
$ ls
4060174 j
8033020 d.log
5626152 d.ext
7214296 k
dir a
$ cd a
$ ls
5 n```
TOTAL SIZE: 48381170
SIZE d: 24933647probably better test cases
bleak lantern
worked pre well for me but i dislike the double indentation
oop didn’t check timestamp
bronze leaf
can anyone do this with a single seq[int]?
vivid patrol
lowkey i'm surprised by the amount of people who assumed directories were unique
bronze leaf
i mean, they are unique, but one should consider the full path
vivid patrol
*unique names
mb
but yeah, are people just not that familiar with file systems? (not trying to judge if it's coming off like that idk)
bronze leaf
they could have been cheeky about it and break the dfs algorithm a couple of tim...
yeah, my original solution did not depend on this pattern --- but once i saw it, i had to optimize!
faint lava
can anyone do this with a single `seq[int]`?
is the seq a dictionary?
tough hinge
seq for "sequence"
It's just a list
(I'm assuming)
faint lava
seems like going up and down the tree might be an issue, I don't know if all the instructions were in order such that you don't ever care about remembering subfolders after you're done with them.
rancid heath
seems like going up and down the tree might be an issue, I don't know if all the...
Apparently inputs did a DFS
My code would work even if some things were repeated though because I spent way too long fixing every edge case because of a stupid typo :(
faint lava
I suppose you could abuse that.
cyan wadi
ok so ive finally implemented it with the actual paths in the dictionary instead of the directories names which could be repeated, but its still not giving the right answer. is anything im still missing?
from dataclasses import dataclass, field
@dataclass
class Directory:
size : int = 0
subdirs : list = field(default_factory=list)
def get_next_line(file):
return file.readline().rstrip("\n")
def get_path(current_path, dir):
path = f"{current_path}/{dir}"
return path if current_path == "/" else f"/{path.lstrip('/')}"
max_dir_sum = 100000
f = open("./puzzle_input.txt", "r")
dirs : dict[str, Directory] = {}
current_path = ""
line = get_next_line(f)
while(line != ""):
_, cmd, dir = line.split(" ")
if cmd == "cd":
if dir == "..": # remove the last dir from the path
current_path = current_path.rsplit("/", 1)[0]
else: # add the dir to the path
if current_path == "":
current_path = dir
else:
current_path = get_path(current_path, dir)
line = get_next_line(f)
cmd = line.split(" ")[1]
if cmd == "ls":
full_path = current_path
print(full_path)
dirs[full_path] = Directory()
while((line := get_next_line(f)) != "" and not line.startswith('$')):
val, name = line.split(" ")
if(val == "dir"): # directory
path = get_path(current_path, name)
dirs[full_path].subdirs.append(path)
else: # file
dirs[full_path].size += int(val)
f.close()
result = 0
for dir in dirs.values():
dir_sum = dir.size + sum(dirs[sub].size for sub in dir.subdirs if sub in dirs)
if dir_sum <= max_dir_sum:
result += dir_sum
print(result)
young nacelle
i love how the recent problems allow the usage of patma
proud of my solution yesterday when using match/case
cyan wadi
yesterday's puzzle was pretty easy, im having a hard time with this one 😐
faint lava
Your code confuses me. You get the next line in 3+ different places I'm not sure it's safe, what if it gets a command in the wrong order?
cyan wadi
im assuming the order is always cd dir -> ls
but yeah its kinda confusing
kinda bad for someone who's been doing this for the past 5 hours
xD
cyan wadi
ok so these last 6 directories are wrong since they arent even in the root directory, but i dont know why this is happening
summer herald
I should have gone to sleep five hours ago but I couldn't -_-;
rancid heath
Wrong channel Salt
bronze leaf
or is it
cyan wadi
https://pastebin.com/4davQTL1
you can also use dataclasses in newer versions of python, they're cool 👍
bronze leaf
...or is it
summer herald
you can also use dataclasses in newer versions of python, they're cool 👍
Someone else suggested that. I had a look at them and didn't see the point of them for this
cyan wadi
its just so you dont have to declare the init function in every class
but the regular ones are fine
rancid heath
its just so you dont have to declare the init function in every class
Also typed with defaults
It just looks nicer
cyan wadi
yea
summer herald
They looked more confusing/obfuscated to me looking over
cyan wadi
wrong channel ^^
narrow sail
wrong channel
tender skiff
oh no
narrow sail
also i did day 7 with regular stuff now https://paste.pythondiscord.com/yaqacehabu.py
rancid heath
They looked more confusing/obfuscated to me looking over
If you want a class Foo that contains an int a and a str b, you can write
@dataclass
class Foo:
a: int
b: str
and it will generate an __init__ and a __repr__ for you
I think it also generates rich comparison methods but you might have to ask for that by using @dataclass(...), I don't remember
cyan wadi
@dataclass
class Directory:
size : int = 0
subdirs : list = field(default_factory=list)
i really dont like this thing that i have to use to declare a mutable list in a data class tho
field(default_factory=list) seems so unnecessary
but doesnt work without it
rancid heath
```
@dataclass
class Directory:
size : int = 0
subdirs : list = field(de...
Ah, mutable defaults. Fun
A) subdirs should probably be hinted with list["Directory"]
B) you could simply not provide a default and make it required on init
cyan wadi
A) `subdirs` should probably be hinted with `list["Directory"]`
B) you could sim...
well didnt know about that 🙂
thanks
wait
@dataclass
class Directory:
size : int = 0
subdirs : list["Directory"] = []
ValueError: mutable default <class 'list'> for field subdirs is not allowed: use default_factory
🥲
rancid heath
Don't provide a default
rancid heath
Using default_factory or override __init__ yourself
rancid heath
Typical choice seems to be field
cyan wadi
yup
rancid heath
FYI even if you override __init__ you can still get a decent amount of value out of a dataclass; you'll get a repr, an eq, a match_args, and a slots by default; by customising the settings (@dataclass(**settings)) you can also get total ordering and kw-only auto-init, frozen instances, and unsafe hashing (allowing hash despite mutability)
narrow sail
also i did day 7 with regular stuff now https://paste.pythondiscord.com/yaqaceha...
feels nice
cyan wadi
FYI even if you override `__init__` you can still get a decent amount of value o...
indeed
cosmic depot
I might have found a bug
if the cwd at the end is a subdir of the directory for part 2 it won't get the correct answer
narrow sail
if the cwd at the end is a subdir of the directory for part 2 it won't get the c...
what input?
cosmic depot
I don't have one, I just know it can happen
I found it while I was writing a readable version of it
narrow sail
I don't have one, I just know it can happen
test input?
narrow sail
test input?
test input looks like it ends in /d/
cosmic depot
oh true
cosmic depot
test input?
yes
unkempt ingot
I like how this guy uses Github issues to document his thoughts and progress especially in using AI assistance to solve day 7. https://github.com/simonw/advent-of-code-2022-in-rust/issues/9
Amazing!
cosmic depot
s=[];d=[]
for r in open(0):
a,*_,b=r.split()
if'/'>b:d+=s.pop(),
elif'$ d'>r:s+=0,
if'/'<a<':':s=[i+int(a)for i in s]
d+=s;print(sum(_*(1e5>_)for _ in d),min(_ for _ in d if _>max(d)-4e7))
``` 192 that actually works for all inputss=[];d=[]
for r in open(e:=0):
a,*_,b=r.split()
if'/'>b:*s,v=s;d+=v,;e+=v*(1e5>v)
elif'$ d'>r:s+=0,
if'/'<a<':':s=[i+int(a)for i in s]
print(e,min(_ for _ in d+s if _>max(s)-4e7))
``` 183@narrow sail @viscid sail @queen breach any more ideas?
narrow sail
idk really
tough hinge
🫵 ▶️ 🛌
unkempt arch
all kinds of blind turns and wrong assumptions yesterday, to the point that I had a minor nightmare last night about there being an entire second independent root directory '\' , and I finish part B this morning before my coffee
narrow sail
🫵 ▶️ 🛌
but it's just 22:06 why should i have 6+ hours of sleep on a day with no school
tough hinge
my first PR ever, hopefully I didn't mess up
Looks fine, except, make sure you create a different branch to work off of
They'll be able to use your PR fine, but it will break your fork after they merge the PR, because GitHub will try to update your fork to have commits that already exist
You can fix it with Git and force pushing and all that, but it's a huge pain
just create branches
main ledge
I finally got day 7! 
I wasn't taking into account directories possibly having the same name as other directories that are nested somewhere else 🙃
tough hinge
Yeah, a lot of us fell for that 
viscid sail
```py
s=[];d=[]
for r in open(e:=0):
a,*_,b=r.split()
if'/'>b:*s,v=s;d+=v,;e+=...
s,d=[],[]
for r in open(e:=0):
a,*_,b=r.split()
if'/'>b:*s,v=s;d+=v,;e+=v*(v<1e5)
elif'$ d'>r:s+=0,
if'$'<a<'d':s=[i+int(a)for i in s]
print(e,min(_ for _ in d+s if-_<4e7-max(s)))
messed around a bit but couldn't get below 183
here's another version
with things flipped
warped kraken
guys, i made this code for the test data not knowing that for the real data, there would be repeated directory names
and i don't know how to change it so it takes this into account
queen breach
```py
s=[];d=[]
for r in open(e:=0):
a,*_,b=r.split()
if'/'>b:*s,v=s;d+=v,;e+=...
i don't think this works
wrong answer on my input
warped kraken
text = open(r"C:\Users\ollys\Documents\AOC Files\2022 #7.txt", "r").read().splitlines()
curfolder = ""
tree = {}
for line in range(len(text)):
stepfd = 1
if text[line][0] == "$" and text[line][2] == "c":
#cd
if text[line][5] != ".":
#cd directory (not cd ..)
curfolder = text[line][5:]
if text[line][0] == "$" and text[line][2] == "l":
#ls
tree[curfolder] = []
array = []
while text[stepfd + line][0] != "$":
array.append(text[stepfd + line])
stepfd += 1
tree[curfolder] = array
def hasnodirs(dir):
for item in tree[dir]:
if not item[0].isdigit():
return False
return True
#directories that only contain files (no other directories)
def getfoldersize(folder):
total = 0
for file in folder:
for char in range(len(file)):
if file[char] == " ":
endofnum = char
break
total += int(file[:endofnum])
total = str(total) + " "
return total
def searchandreplace():
for dir in tree:
for item in range(len(tree[dir])):
if tree[dir][item].startswith("dir"):
dirname = tree[dir][item][4:]
for finaldir in range(len(finaldirs)):
if dirname == finaldirs[finaldir]:
tree[dir][item] = finaldirsizes[finaldir]
break
foldersizes = {}
while not hasnodirs("/"):
finaldirs = []
finaldirsizes = []
for dir in tree:
if hasnodirs(dir):
finaldirs.append(dir)
for finaldir in finaldirs:
finaldirsizes.append(getfoldersize(tree[finaldir]))
for i in range(len(finaldirs)):
foldersizes[finaldirs[i]] = finaldirsizes[i]
searchandreplace()
part1tot = 0
for dirsize in foldersizes:
if int(foldersizes[dirsize]) <= 100000:
part1tot += int(foldersizes[dirsize])
print(part1tot)
here is the code
queen breach
the 192 did work though
warped kraken
guys, i made this code for the test data not knowing that for the real data, there would be repeated directory names
and i don't know how to change it so it takes this into account
it basically finds all of the folders that contain only files
and then changes the names of them in their parent folders to a file name of the amount of storage they take up
queen breach
use paths instead of folder names
paths will not be repeated, otherwise it's the same folder
tawny creek
for part one:
code returns test input correct, directories and sizes all seem correct.
why does this return the wrong output?
commands = []
x = input()
while x != "":
commands.append(x.split())
x = input()
current_directory = "/"
directory_dict = {"/": []}
size_dic = {}
for i in commands:
if i[0] == "$":
if i[1] == "cd":
if i[2] == "..":
for key, values in directory_dict.items():
if current_directory in values:
current_directory = key
else:
current_directory = i[2]
if i[0] == "dir":
directory_dict[current_directory].append(i[1])
directory_dict[directory_dict[current_directory][-1]] = []
if i[0].isnumeric():
size_dic[i[1]] = int(i[0])
directory_dict[current_directory].append(i[1])
def size(node):
if node in size_dic:
return size_dic[node]
else:
count = 0
for i in directory_dict[node]:
count += size(i)
size_dic[node] = count
return size_dic[node]
total_size = 0
for i in directory_dict:
x = size(i)
if x <= 100000:
total_size+=x
print(total_size)
royal axle
who's gonna solve day 7 in bash
summer herald
for part one:
code returns test input correct, directories and sizes all seem co...
in the whole puzzle input, two directories can have the same name
It looks like your code uses the directory names as dictionary keys. That won't work unless you add something to them to make them unique
storm epoch
who's gonna solve day 7 in bash
that would actually be hilarious, create all the files and then just query using du
rancid heath
that would actually be hilarious, create all the files and then just query using...
You can't do it with du, you have to use find
du rounds up to the next block I think
limber shell
OKay - here is my code so far - hold on a sec while I get it up.
vivid patrol
a meh input to test it against is:
$ cd /
$ ls
dir a
14848514 b.txt
8504156 c.dat
dir d
$ cd a
$ ls
dir e
29116 f
2557 g
62596 h.lst
$ cd e
$ ls
584 i
$ cd ..
$ cd ..
$ cd d
$ ls
4060174 j
8033020 d.log
5626152 d.ext
7214296 k
dir a
$ cd a
$ ls
5 n```
TOTAL SIZE: 48381170
SIZE d: 24933647Where do I post code snippets if there are too long for Discord?
vivid patrol
!paste
tame groveBOT
vivid patrol
wow
vivid patrol
i'm just impressed that you wanted to write a nice looking solution
are you sure you're doing directory shennanings right
it's kinda weird how you did it but maybe fine? 🤷♂️
are you assuming directory names are unique? because it isn't
limber shell
I was, but I wrote the whole lookup table thing to account for that.
Each file and directory goes into the tree with a unique id as the name.
vivid patrol
looks like your tree is working and fine
probably an issue of where you do number calculations
limber shell
Thats what I suspect - but I have no idea what it might be
That all happens right at the bottom.
vivid patrol
not sure how anytree works
are you sure you're recursively getting files
i.e. in directory a
it sums up files i, f, g, h.lst
limber shell
Yes, and then adds them to the directory attribute length of the parent node.
to the length of a, in your example.
vivid patrol
weird
i don't have anytree installed, can't test personally 🤷♂️
you can try to replace with paths and test against input
tawny creek
in the whole puzzle input, two directories can have the same name
oh thats such a silly reason -_-
summer herald
oh thats such a silly reason -_-
It's a pretty simple fix though tbh
My first go at day 7 I used a module called treelib that threw errors immediately when I tried to create multiple nodes with the same name
rancid heath
Lmao rip
summer herald
What I did was just put some code in my main for loop that incremented a counter each time it went through and then appended that as a string to the name of each directory node when I made it.
Thus ensuring that every directory had a unique name
I eventually had to go back to the drawing board because it was too difficult to extract the information I needed from the output, but that at least was a good idea lol
summer herald
That's another solution. It didn't occur to me for some reason, even though I was recording the paths in that version of the script.
tawny creek
wait if they can have the same name when you call "$ cd x" how would you know which to choose
summer herald
because when you call $ cd it can only go to subfolders of the current folder
just like when you use it on a regular shell
(which is one reason why I was tracking the path)
hazy gust
I kept track of it with a cursor stack, appending a directory on every cd foo and popping it with each cd ... With a function that could convert that stack to a single string to use as a dict key when adding files and sizes.
Assigning total size per directory, I would just recursively add the file size to the total for every key of the current cursor back to the root.
Once that was all set up, totalling everything was pretty straightforward.
tawny creek
would something like this work?
starting in my code at line 20
else:
for dire in directory_dict[current_directory]:
if dire == current_directory + "_" + i[2]:
current_directory = dire
if i[0] == "dir":
directory_dict[current_directory].append(current_directory + "_" + i[1])
directory_dict[directory_dict[current_directory][-1]] = []
I struggle so much with recursion, it took me more than half a day to solve Day 7, and I ended up getting a lot of help with people. I barely understand my own code for it 😂
vivid patrol
lmao rip
limber shell
I am STILL working on Day 7 - if I don't figure this out soon I'm giving up this year cause I can't take this level of frustration during my christmas vacation lol
neat tulip
Hello
Am bad at recursion
Does this seem like it'd do what I expect it to? (Recursively sum the sizes of a directory)
def _recursive_size(entry: FileSystemEntry, sum) -> int:
if entry.type == EntryType.FILE:
return entry.size
for e in entry.entries:
sum += _recursive_size(e)
return sum
FileSystemEntry is a custom class meant to mockup a file or folder. Only files have sizes. Folder sizes need to be calculated from summing all their nested file sizes
tough hinge
Hello
Am bad at recursion
Does this seem like it'd do what I expect it to? (Recu...
I'm too sleepy for recursion right now. But... You're defining the total as a parameter, and then you're not actually passing it into the recursive call. You'll only get the result of the first return, AFAIU. Have you tried it yet?
Also you're overriding the built-in sum
neat tulip
I'm too sleepy for recursion right now. But... You're defining the total as a pa...
Right. Got an error that I was missing the required parameter. Passed it in and it worked like a charm
As for overriding sum, I'll have to on an epic journey through the mountains and valleys to find out where I asked
tough hinge
Right. Got an error that I was missing the required parameter. Passed it in and ...
Well 🖕you too then birdo, 
neat tulip
(I'll probably be back later asking why int is not a callable)
rancid heath
(I'll probably be back later asking why `int is not a callable`)
Only if you try to use it in a shadowed scope, as long as you constrain it to that function you're fine
tough hinge
Only if you try to use it in a shadowed scope, as long as you constrain it to th...
Don't encourage him!!!
finite tundra
Anyone know whats wrong with this? Works for test input but not actual input (part one)
from collections import defaultdict
with open('day7\input.txt', 'r') as f:
data = f.read().split('\n')
path = []
sizes = defaultdict(int)
sums = 0
for line in data:
words = line.split()
if words[0] == '$':
if words[1] == 'cd':
if words[2] == '..':
path.pop()
else:
path.append(words[2])
elif words[0].isnumeric():
for i in range(len(path)):
sizes[path[i]] += int(words[0])
print(sizes, i, path)
for keys, values in sizes.items():
if values <= 100000:
sums += values
print(sums)```
brave raptor
Anyone know whats wrong with this? Works for test input but not actual input (pa...
This doesn't produce enough folders, maybe duplicates?
yeah that's probably it
finite tundra
wdym?
finite tundra
for example
oh, so there could be file sphbzn inside of a folder called sphbzn? or something along those lines? Like there are duplicate file/folder names?
brave raptor
exactly and you'd be counting them as the same
finite tundra
Hm, do you see an easy way to fix that with my code?
brave raptor
Hm, do you see an easy way to fix that with my code?
sorry for the late reply
for path in accumulate(current_path): # current_path = ['/', 'folder1/', 'folder2/']
sizes[path] += size``` this is what a lot of people did!docs itertools.accumulate
tame groveBOT
finite tundra
sorry for the late reply
```python
for path in accumulate(current_path): # curre...
Ah interesting, ty!
keen bronze