There's a way to fix this?
#algos-and-data-structs
past flare
warm marsh
looks like you switched the encoding
it says UTF16-BE in the lower right corner of the window. Click that and switch back to probably UTF-8
past flare
the problem is i had closed the file before
so i can't even UNDO
I changed to utf-8 again and nothing happened
spiral swallow
hm, it probably read it as utf16 and now it needs to read it again, the question would certainly be better in #tools-and-devops though
fiery cosmos
Can someone explain what's happening here?
stark bridge
someone showing off their LaTeX chops?
spiral swallow
i wish, my linear algebra skills are almost nill 😦
stark bridge
I don't think there was ever a time when I'd have understood that, and I was, ostensibly, a math major
hidden pebble
@hidden pebble I'd reccomend the MIT OpenCourseware introduction to Computer Science
@last blaze huh ty ill check it out
fiery cosmos
Damn does really ML people that are hired at companies earn $120 000 - $500 000 a year?
stark bridge
sure why not?
spiral swallow
it certainly depends on the place
but i don't doubt some people do
that seems more a topic for #career-advice though
fiery cosmos
Hey guys, I'm taking the CSA CollegeBoard course this year, does anyone have tips?
Last year I took a practice, thought I did well and ended up getting a 3...
eager hamlet
hey lepto i though CSA was java
eager hamlet
but anyways, I don't even know how to do this problem:
The cows have not only created their own government but they have chosen to create their own money system. In their own rebellious way, they are curious about values of coinage. Traditionally, coins come in values like 1, 5, 10, 20 or 25, 50, and 100 units, sometimes with a 2 unit coin thrown in for good measure.
The cows want to know how many different ways it is possible to dispense a certain amount of money using various coin systems. For instance, using a system of {1, 2, 5, 10, ...} it is possible to create 18 units several different ways, including: 18x1, 9x2, 8x2+2x1, 3x5+2+1, and many others.
Write a program to compute how many ways to construct a given amount of money using supplied coinage. It is guaranteed that the total will fit into both a signed long long (C/C++) and Int64 (Free Pascal).
PROGRAM NAME: money
INPUT FORMAT
The number of coins in the system is V (1 <= V <= 25).
The amount money to construct is N (1 <= N <= 10,000).
Line 1: Two integers, V and N
Lines 2..: V integers that represent the available coins (no particular number of integers per line)
it probably involves dynamic programming, but I've only done one and that was with a google search, so can someone help me on this one?
orchid palm
ik that i postde this in another channel but this is the place to put it
Any parts worth Salvaging from a printer
that i can use with my raspberry pi 4
half lotus
No, this isn't the place to put it. It's off-topic so #ot0-psvm’s-eternal-disapproval #ot1-perplexing-regexing #ot2-never-nester’s-nightmare
solid cloud
scanners are way more fun to salvage
especially the 10-20kv transformer that'll burn you like you never felt before if misused
fiery cosmos
What is this made with? The 3D stuff and the characters? And how have they been able to do so they move by their self and learn in the 3D world? https://youtu.be/atcKO15YVD8
south parcel
Ooh, 2-min papers! Have you checked the description?
Turns out it's a Ubisoft Montreal paper
fiery cosmos
So it’s like a 3D software they have put in objects and then trained them?
south parcel
At it's very foundation, yes.
The paper covers a wide swath of topics though
Like GPU optimization
Which is fitting considering both training and physics utilise the GPU
Yeah, pretty cool stuff
fiery cosmos
Yeee
How is even 3D software made? I mean what is it that does so it can make 3D stuff?
south parcel
What part of '3D' are you interested in? How rendering works, how to make 3D animations/games, what software's out there, there's a whole lot to explore!
tribal pendant
I think the module panda3d does that
south parcel
If you're talking about python, yeah try what @tribal pendant said
There are lots of 3d graphical libs out there
But for general stuff, you could look into OpenGL? That's real low level stuff though
spiral swallow
panda3d is a high level abstraction on top of OpenGl yeah, to do OpenGl directly it helps a lot to understand matrix math, to understand how the projection from 3d to 2d works, and how to manipulate 3d objects, compute lights and stuff, lots of math involved, but it can be a lot of fun.
south parcel
Woah, thanks @spiral swallow ! Didn't know!
fiery cosmos
@south parcel yeah but I mean from scratch. Ex how to code from scratch from 0 to make a cube in 3D and then later render it out and get I image of a 3D cube. I just want to understand how it’s done from scratch
south parcel
Like tshirtman said, panda3d is a good way to start understanding how 3d works while still writing in python.
Or maybe you'd like to look into graphing in 3d with matplotlib?
gusty grove
opengl is a great way to learn about that stuff
if you want me to i can give you a quick overview of how you get to a image of a 3d cube
gusty grove
does anyone know a fast 3d floodfill approach?
fiery cosmos
Alright panda3d look like something to start with!
fiery cosmos
Is it required to know c++ if you ex are a master at python, linear algebra, creating Neural networks, and can do the most stuff with that, but then if you would like to create a self driving car ex, then i guess that skills isnt enough?
sly patio
@fiery cosmos are you reading Computer Vision by Richard Szeliski?
@fiery cosmos That is the algorithm for DLT (Direct Linear Transformation) that is if you have to sets of known vectors and you want to find a given matrix (e.g. let's say a 3x3 matrix A) where all the points are know scalar factors to equation k. You would first have to get rid of the unknown scalars. That is the weird looking matrix in Section(4) in which he applies the anti-symmetric matrix, in order to obtain the homogenous equations
@fiery cosmos If you wanto know what that means in layman's term. If you use the OpenCV library. the perspective function that you apply prior to the warp function is implementing some of that mathematics under the hood. The OpenCV library is almost entirely derived from that book. FYI
fiery cosmos
I am trying to implement the warpperspective function in python without using the inbuilt opencv function
I have successfully got the homography matrix and applied to the respective x,y values in the source image
But then idk what to do next
@sly patio can you provide some guidance?
@fiery cosmos are you reading Computer Vision by Richard Szeliski?
@sly patio Yes I am
sly patio
I have read it a number of times. What are your questions?
fiery cosmos
Q1. Now that you have the new x'/lambda, y'/lambda, lambda after applying homography to x,y,1 what do I do next
I have read it a number of times. What are your questions?
@sly patio and I am having trouble even going through it for the first time
lol
sly patio
@fiery cosmos what is your actual task? that's where we need to start. that way we can reason about your chosen algorithm
fiery cosmos
I want to project the classic lena image on to a fiducial tag in real time without using the warpperspective function
sly patio
@fiery cosmos That's a start. can you pseudo code your algorithm because fiducial systems can fall into two categories based on the shape of the tags: square tags with a blackand white border on the outside; and tags with other shapes.
Therefore, which one did you choose?
@fiery cosmos are you doing just the homography part in Python right?
fiery cosmos
@fiery cosmos are you doing just the homography part in Python right?
@sly patio yes sir
sly patio
@fiery cosmos have you tried singular value decomposition?
using the least squares methos
fiery cosmos
yes, thats how i got my homography matrix
sly patio
@fiery cosmos Ican you post your code or you can just IM it to me if you prefer
@fiery cosmos let me fire up Emacs real quick and see what I can come up with some of my old recycled code
@fiery cosmos did you send it?
fiery cosmos
Hi
haughty sleet
function for disjoint set system
def Find(Set, value):
if parent[value] == value:
return value
else:
parent[value] = Find(Set, parent[value])
return parent[value]
getting error
IndexError: list index out of range
buoyant patrol
*target
lament portal
I need to set up a Venv inside of Visual Studio Code...
Can someone help me to do it?
nvm
earnest basalt
Hii
golden flume
@sly patio are you familiar with computer vision, i need some help. i am trying to set up a single camera + laser system for position measurement, my problem is that i don't know how i could use the knowledge that the laser is esentially just a plane which intersects with the object surface to recover the depth from the image.
long swallow
you want to retrieve distance from the camera to an object using a laser?
just get a sensor for that
rather than a red dot laser
proven folio
spiral swallow
so i was reading on that the other day, on the wikipedia page, and it says the MIT course that used it was replaced to use python instead of lisp, as they considered the thing quite outdated (took them over 10 years to actually replace the course, so they wanted to do it for a long time), and there as been a number of work since designed to provide better books, though i can't judge them or the original effort.
sly patio
@golden flume that's fairly standard liner algebra. What is the actual task that you're trying to accomplish
@snof the little schemer, then read SICP
golden flume
@sly patio i want to get the 3D points of a laser point viewed through a camera
i dont see how it's just algebra? i think there will be some plane fitting involved, but not sure. if it is just algebra i'd like some more information if possible!
sly patio
@golden flume use themapping3d module from the opencv contrib library
Specifically, from lines 193
All of the math exam be found here:
Can not exam (gotta work on my keyboard swiping skills) lol
You have to convert the 3D matrix (in pixels) to the LOS calibrated ray off your camera (which is in meters) if not you'll be off.
tardy forge
north zealot
iPad I'd say
Even if I'm not a fan of Apple products, I think they do the best tablets
golden jacinth
glacial wagon
I got an iPad pro 3rd generation for art, but I use it daily for other stuff too. I fell in love with it
eager hamlet
um kinda off topic
unique ferry
Hi, could someone help me fix my spectrogram?
unique ferry
Can someone at least explain me how to find a dominant frequency in a wav file?
spiral swallow
@unique ferry do you use a FFT (fast fourier transform) on the signal?
unique ferry
G: Yeah.
G: Actually I have to find first and second dominant frequency.
G: It's for car horns, that I have to compare tones to the musical notes.
G; I think I have to use signal peaks, but I'm not sure.
spiral swallow
the FFT will tell you the amount of energy in each frequency, basically, so the strongest one should be dominant, and so on.
burnt wren
Why if priveleged code faults the system must crash, but if user code faults the system can keep going forward?
tribal pendant
What's "priveleged code"
warm marsh
user processes are pretty isolated and controlled by the kernel. kernel code... well, not.
burnt wren
let's say for example a kernel
warm marsh
imagine a school class. if one of the students acts up, they will be thrown out and the lesson continues. If the teacher breaks down, nothing works.
burnt wren
but there's no difference between them, a user program and a kernel is just code in memory, just one has rights to special instructions and is protected
burnt wren
lets say a page fault happens in the kernel, a page fault interrupt gets called by the cpu, the interupt handler is in the kernel and it checks if the faulted instruction is in the kernel, if yes then shutdown the system and display an error message (like a bsos or a kernel panic)
why can't we just jump back to the kernel, it's in memory, we have the context
last blaze
Are you asking how bubble sort works, or for someone to do you homework?
spiral swallow
@burnt wren as i understand it, the teacher/pupil analogy holds, the kernel enforces the rules for other processes, if the kernel code stops working correctly, there is nobody to enforce the rule on itself, all bets are off.
burnt wren
@spiral swallow the kernel handles always handles all faults, all faults that happened in user programs and the kernel, if a fault happened in a process, the handler for the fault just kills the process, but if it happened in the kernel, it shutdowns all disk activity and displays a bsod
why can't the kernel have a routine that would try and fix everything, that the handler could jump to
why can't we just jump past the faulted code or jump back to the code that called the faulted code
spiral swallow
because the kernel code doesn't obey the same restrictions as the non kernel code, so if an issue is discovered in its process, then a lot of assumptions it needs to do to continue working, are not possible to do anymore?
burnt wren
the kernel is still in memory that means we can jump anywhere in it
i just can't understand why don't we do something like that
in the end we can just ignore the fault
spiral swallow
because it wouldn't be safe
sure, the kernel could try to do things that maybe would work, but kernel devs are not the kind to want do to these things
if they can't say what the proper source of action is, they'd rather raise the error and crash
burnt wren
all i can think of is that it's just much more easier to reset the computer and to start everything over
spiral swallow
well when you say "easy" they say "the only acceptable course of action"
if there was a solution, even hard, they would take it, they don't like crashing anymore than you
but if they do that, it's because other ways are unsafe
you don't want a machine to start doing random things when it discovers some error it can't safely recover from, you want it to stop doing things
lone siren
Hey guys, off-topic question here. I'm building a PC for our lead developer, where price is not an issue. I've never done it before. Our company is mostly involved in machine learning and data analytics, and some of the more complex tasks will be carried out on her computer.
Would anyone be able to suggest a suitable CPU/GPU/motherboard combination please?
bronze gazelle
Sounds like a good question for an off-topic channel
hazy topaz
Not sure if this is the right place to ask, but I'm looking for an explanation on what an abstract type is (and maybe it's easier to first explain an abstract class in Python as I know a fair bit of Python and classes already)
My understanding of an abstract class now is that it's basically just a class meant to be a parent of other classes and never gets instantiated itself
spiral swallow
That's technicaly correct, but i would say what's important about an abstract type, is what it's for.
It's a way to declare the properties/aspect/interface of a type, without actually explaining how it works, that's the reason it can't be instaciated, the code to make it work just is not there.
But if you get an object who is of this type, you know how you can use it, because it has all the thinge the abstract type defines.
To be able to create an object of this type, you must actually create an object of one of it's subclasses, as the type itself is abstract.
But code manipulating it can completly ignore the exact type it is and how it works, as it knows it has the interface defined by the abstract type.
@hazy topaz
balmy siren
Hey guys ... I'm doing Tree DS and recently done some traversal orders and basic questions related to it and Binary Trees. In short, I've done the intro part and binary trees part and topic next after it are Generic Trees (N-Ary Trees) and Binary Search Trees.. So, i'm not sure what to do next, I mean, most of the people said that you should start with Binary Search Trees once you done with Binary Trees but is that anything special about Generic trees which I should know right now (Also, I'm doing self learning of DS with a book and random blogs, tutorials).. so, is there anything which is related with Generic Trees and binary search trees?
fiery cosmos
@balmy siren you should look at https://gist.github.com/salt-die/ec050420a7750a27bb2d4a7a6fb7d15c
brave oak
can pattern matching (the kind they have in functional languages) fully (and sensibly) replace if-elif-else?
like would it be weird if a languge did that?
spice ravine
as long asyou can match on values and not just types
even if your conditions are completely different (i.e. not based on values of the same variable), you can just fall back on the fact that if-elif-else can be replaced with if-else{if-else}
stable pecan
here's a multiset with choice implementation meant to be fast: (so you don't have to keep making a list to choose items at random)
"""
Setch is a portmanteau of set and choice. A setch is set-like object (containment is O(1)) except
one can pick items at random from the set without converting it to a list. This is a multiset
implementation of a setch.
Note most list operations are O(1) except for indexing in `remove`.
"""
from collections import defaultdict
from random import choice
class MultiSetch:
def __init__(self, *items):
self._item_to_position = defaultdict(list)
self.items = []
for item in items:
self.add(item)
def add(self, item):
self.items.append(item)
self._item_to_position[item].append(len(self.items) - 1)
def remove(self, item):
position = self._item_to_position[item].pop()
last_item = self.items.pop()
if position != len(self.items):
self._item_to_position[last_item].pop()
self.items[position] = last_item
self._item_to_position[last_item].append(position)
def choose(self):
return choice(self.items)
def pop(self):
"""Pop a random item."""
item = self.choose()
self.remove(item)
return item
def __len__(self):
return len(self.items)
def __contains__(self, item):
return item in self._item_to_position
def __iter__(self):
return iter(self.items)
def __repr__(self):
return f'MultiSetch({{{", ".join(map(str, self.items))}}})'
royal kestrel
You can choose items at random from a non-random-access iterator with reservoir sample algorithms
stable pecan
I think that's too slow for me
oblique panther
I'm struggling with how to solve recurrence equations.
My solution is cut off. I wrote T(n) = 2 + 2T(n/2) + 2
I would say that that's O(n), but I don't know how to prove that.
ocean sedge
Slightly off topic but how do I post python code in that nice marked up format (like @stable pecan)
stable pecan
!code-block
halcyon plankBOT
vale parrot
I am wanting to study into computer science can someone describe how their experience was and what they did
last blaze
Are you asking about someone who has graduated from university? As in how they found the course, and their experience after graduating?
vale parrot
yes
last blaze
#career-advice might be a better place
fiery cosmos
Hi. I need some regex help I have the following:
someCategory%someCategory.categoryFoo#value%someCategory.categoryBar#other value%....
As you see it always starts with someCategory, afterwards it repeates (1 to many) "key-value" pairs
the key-values are seperated by hashtag, and each pair is seperated by a percentage symbol
My regex (that doesn't work): ^([\w\s.]+)(?:%\1\.([\w\s]+)#([\w\s]+))+
brisk dawn
gusty grove
Sure
wintry oracle
# Just curious how many iterations there are.
iteration_count = 0
def breadth_first_search(arr):
global iteration_count
# Length of the input array.
arrayLen = len(arr)
# Store traversed index here as proof.
indexArray = []
# Loop through the level as the divide will be exponential.
for i in range(1, arrayLen):
# Break out of loop when indexArray is complete.
if len(indexArray) == arrayLen:
break
# levelArrayLen is the length of array split by the exponent amount.
levelArrayLen = arrayLen // i
# Start at 0 at the beginning of a level loop.
first = 0
# Loop through and start splitting into logarithmic amount.
for x in range(1, i + 1):
# Want to see how many times it ultimately iterated.
iteration_count += 1
# last is the length of array multiplied by how many times splitted.
last = levelArrayLen * x
# mid point of each split within a level.
mid = (last - first) // 2 + first
# Store mid point into indexArray and check its neighbour too.
if mid not in indexArray:
indexArray.append(mid)
elif (mid + 1) not in indexArray and (mid + 1) < arrayLen:
indexArray.append(mid + 1)
elif (mid - 1) not in indexArray and (mid - 1) > -1:
indexArray.append(mid - 1)
# If the value in that index is True, break out.
if arr[indexArray[-1]]:
return indexArray
first = last + 1
return indexArray
testArray = [False, False, False, False, False, False, False, False,
False, False, False, False, False, False, False, False]
print(breadth_first_search(testArray))
print('Iterated %s times.' % iteration_count)
So I made this attempt of breadth-first-search by attempting to divide the testArray down logarithmically. The idea is checking the middle of each logarithmic split of the testArray to see if it is True. The testArray is supposed to represent a sample which should only contain either True or False with the possibility of "clusters". These codes are working about as what I expected and basically it's just cycling through indexes logarithmically and add the middle index value into another array to signify that it has been "checked". The checking part, which is not shown here but may potentially be computationally heavy, doesn't get executed unless it's not part of the indexArray, in my attempt to be efficient with the check. However, if you see the print statement of Iterated X times., you will see that depending on the size of the sample, it can iterate many more times than its actual sample size. For example, if you replace the single True in the testArray and run it, you will see that it iterated 21 times. I am just curious about how bad that could be in a scalable sense and whether you guys can help me think of a way to minimize extra iterations? I am hoping just looping through and manipulating integer (index value) is pretty trivial in and so more iteration isn't the worst thing. Ultimately, given the structure of the array which can have "clusters", logarithmic search makes more sense than linear search. However, given the iteration is more than the sample size, I'm just curious if y'all got an opinion about it.
long wind
Hello, new to the discord world, if I have questions on manipulating large excel sheets using pandas would this be the best channel?
last blaze
Probably #data-science-and-ml would be more like it, but it depends on the question itself
It might be best to ask in #python-discussion where to post the question
or just use a help channel by doing !free in bot commands to find one
long wind
Thanks!
tawdry perch
hello?
need help with my code
`import itertools
def sum_to_n(n):
'Generate the series of +ve integer lists which sum to a +ve integer, n.'
from operator import sub
b, mid, e = [0], list(range(1, n)), [n]
splits = (d for i in range(n) for d in itertools.combinations(mid, i))
return (list(map(sub, itertools.chain(s, e), itertools.chain(b, s))) for s in splits)
print(list(sum_to_n(7)))`
this code gets all the possible ways you can add to get a number using integers
for example, the above code would output this:
[[7], [1, 6], [2, 5], [3, 4], [4, 3], [5, 2], [6, 1], [1, 1, 5], [1, 2, 4], [1, 3, 3], [1, 4, 2], [1, 5, 1], [2, 1, 4], [2, 2, 3], [2, 3, 2], [2, 4, 1], [3, 1, 3], [3, 2, 2], [3, 3, 1], [4, 1, 2], [4, 2, 1], [5, 1, 1], [1, 1, 1, 4], [1, 1, 2, 3], [1, 1, 3, 2], [1, 1, 4, 1], [1, 2, 1, 3], [1, 2, 2, 2], [1, 2, 3, 1], [1, 3, 1, 2], [1, 3, 2, 1], [1, 4, 1, 1], [2, 1, 1, 3], [2, 1, 2, 2], [2, 1, 3, 1], [2, 2, 1, 2], [2, 2, 2, 1], [2, 3, 1, 1], [3, 1, 1, 2], [3, 1, 2, 1], [3, 2, 1, 1], [4, 1, 1, 1], [1, 1, 1, 1, 3], [1, 1, 1, 2, 2], [1, 1, 1, 3, 1], [1, 1, 2, 1, 2], [1, 1, 2, 2, 1], [1, 1, 3, 1, 1], [1, 2, 1, 1, 2], [1, 2, 1, 2, 1], [1, 2, 2, 1, 1], [1, 3, 1, 1, 1], [2, 1, 1, 1, 2], [2, 1, 1, 2, 1], [2, 1, 2, 1, 1], [2, 2, 1, 1, 1], [3, 1, 1, 1, 1], [1, 1, 1, 1, 1, 2], [1, 1, 1, 1, 2, 1], [1, 1, 1, 2, 1, 1], [1, 1, 2, 1, 1, 1], [1, 2, 1, 1, 1, 1], [2, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1]]
how do i change it so that i can limit the size of the addends to be <= y?
for example, if i want to sum to 7 using numbers <= 4, i can do this:
so 3 and 4 is ok
5 and 2 are not
6 and 1 are not either
2 and 2 and 2 and 1 are ok
fierce inlet
could pop a filter on the end
filter(lambda numbers: all(number < some_value for number in numbers), result)
something like that
perhaps not the cleanest, elegant or efficient way
@tawdry perch
stable pecan
def partitions(n, m=None):
if m is None or m >= n:
yield [n]
start = n - 1
else:
start = m
for m0 in range(start, 0, -1):
for partition in partitions(n - m0, m0):
yield [m0] + partition
In [66]: list(partitions(7, 4))
Out[66]:
[[4, 3],
[4, 2, 1],
[4, 1, 1, 1],
[3, 3, 1],
[3, 2, 2],
[3, 2, 1, 1],
[3, 1, 1, 1, 1],
[2, 2, 2, 1],
[2, 2, 1, 1, 1],
[2, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1]]
@tawdry perch
tawdry perch
@stable pecan that's pretty much what i want, but how do you change it so that it gives all the permutations of each partition? ex. 1 + 1 + 2 can also be 1 + 2 + 1 or 2 + 1 + 1
do i put in itertools.permutations() somewhere?
i tried doing so but then it gets 4 duplicates of 1 + 1 + 1 + 1
stable pecan
more_itertools has unique permutations
In [4]: from more_itertools import distinct_permutations
...: for partition in partitions(7, 4):
...: print(list(distinct_permutations(partition)))
[(3, 4), (4, 3)]
[(1, 2, 4), (2, 1, 4), (2, 4, 1), (1, 4, 2), (4, 1, 2), (4, 2, 1)]
[(1, 1, 1, 4), (1, 1, 4, 1), (1, 4, 1, 1), (4, 1, 1, 1)]
[(1, 3, 3), (3, 1, 3), (3, 3, 1)]
[(2, 2, 3), (2, 3, 2), (3, 2, 2)]
[(1, 1, 2, 3), (1, 2, 1, 3), (2, 1, 1, 3), (1, 2, 3, 1), (2, 1, 3, 1), (2, 3, 1, 1), (1, 1, 3, 2), (1, 3, 1, 2), (3, 1, 1, 2), (1, 3, 2, 1), (3, 1, 2, 1), (3, 2, 1, 1)]
[(1, 1, 1, 1, 3), (1, 1, 1, 3, 1), (1, 1, 3, 1, 1), (1, 3, 1, 1, 1), (3, 1, 1, 1, 1)]
[(1, 2, 2, 2), (2, 1, 2, 2), (2, 2, 1, 2), (2, 2, 2, 1)]
[(1, 1, 1, 2, 2), (1, 1, 2, 1, 2), (1, 2, 1, 1, 2), (2, 1, 1, 1, 2), (1, 1, 2, 2, 1), (1, 2, 1, 2, 1), (2, 1, 1, 2, 1), (1, 2, 2, 1, 1), (2, 1, 2, 1, 1), (2, 2, 1, 1, 1)]
[(1, 1, 1, 1, 1, 2), (1, 1, 1, 1, 2, 1), (1, 1, 1, 2, 1, 1), (1, 1, 2, 1, 1, 1), (1, 2, 1, 1, 1, 1), (2, 1, 1, 1, 1, 1)]
[(1, 1, 1, 1, 1, 1, 1)]
@tawdry perch
young plover
It's one of the 2 answers from here: https://stackoverflow.com/a/953097
stable pecan
In [9]: list(chain.from_iterable(distinct_permutations(partition) for partition in partitions(7, 4)))
Out[9]:
[(3, 4),
(4, 3),
(1, 2, 4),
(2, 1, 4),
(2, 4, 1),
(1, 4, 2),
(4, 1, 2),
(4, 2, 1),
(1, 1, 1, 4),
(1, 1, 4, 1),
(1, 4, 1, 1),
(4, 1, 1, 1),
...
thanks!
tawdry perch
wait
stable pecan
that means we can map it
tawdry perch
it got this error
NameError: name 'distinct_permutations' is not defined
and it doesn't work if i type itertools. before it
stable pecan
distinct_permutations is from more_itertools
tawdry perch
oh
and i liked the above functions, but i don't like tuples within list
how do i make it give me lists within lists?
stable pecan
well, tuples take up a lot less memory but if you want lists we can do that
In [17]: list_of_tuples = list(chain.from_iterable(map(distinct_permutations, partitions(7,4))))
In [18]: list_of_lists = list(map(list, list_of_tuples))
In [19]: list_of_lists
Out[19]:
[[3, 4],
[4, 3],
[1, 2, 4],
[2, 1, 4],
[2, 4, 1],
[1, 4, 2],
[4, 1, 2],
thanks!
stable pecan
np
tawdry perch
here was the reason i wanted lists
i want to convert the numbers to words
using this:
for curWordNum, curWord in enumerate(myList):
for curLtrNum, curLtr in enumerate(curWord):
curLtr += 64
myList[curWordNum][curLtrNum] = chr(curLtr)
myList[curWordNum] = "".join(curWord)
print(myList)
and i wanted to limit it at 26 because 26 is Z
i am trying to do a little puzzle i like to call the A1Z26 checksum
so assume A is 1, B is 2, ..... Z is 26
now take the word you want and add up all its values
and we will call the sum the "A1Z26 checksum"
now i want to see if it's possible to get the original word back from the A1Z26 checksum
stable pecan
In [25]: import string
...: translate = {letter:value for value, letter in enumerate(string.ascii_lowercase, start=1)}
...: my_string = 'hi there'
...: sum(translate.get(char, 0) for char in my_string)
Out[25]: 73
you want to figure out the word from 73
tawdry perch
well, here is my code, which i somehow broke
i'm confused
our com-sci classes are far too basic
(this is not for a class)
(i just wanted to make a fun puzzle)
tawdry perch
i'm confused by my own code
stable pecan
oh
tawdry perch
your code encodes the checksum
i'm doing the opposite
so if someone gives you the checksum 7, here are the possibilities
stable pecan
it just seems like there's going to be a lot of words that sum to 73
tawdry perch
yes
stable pecan
or whatever
halcyon plankBOT
Hey @tawdry perch!
tawdry perch
this is a sample output
*** English words that have an A1Z26 checksum of 24 ***
X
AW
CU
ES
FR
HP
IO
KM
LL
MK
NJ
OI
PH
RF
SE
TD
VB
WA
ACT
ADS
ARE
AVA
BOG
CAT
CBS
CFO
CHM
DOE
EAR
EMF
ENE
ERA
FAQ
FEM
GAP
GEL
GOB
GPA
IAN
IBM
INA
JAM
KID
LBJ
LEG
MAJ
NEE
ODE
OED
RAE
SAD
ADAR
ADEN
BAIL
BALI
BANG
CRAB
DADO
DANE
DEAN
DENA
EDNA
ELBE
FLEA
GLAD
IBID
IEEE
LEAF
AGANA
BALED
BLADE
CAKED
CECAL
CHAFF
JADED
ACHEBE
BABBLE
BEDDED
CACHED
CADGED
CANADA
DEFACE
JABBED
you trying to find all the possibilities? that is a partition problem so you're on the right track
tawdry perch
i figured that out
stable pecan
don't do this though len(list(partitions(73, 26)))
or do do it, if you want, but it's incredibly large
tawdry perch
is there a way to make it save RAM though @stable pecan ?
when i put in a large number like 27 it uses 10GB+ of RAM
and my computer only has 8GB
stable pecan
partitions, like most combinatorial objects explode violently in number of possibilities as you increase size
you can chunk it and dump it to hard drive
but you could run out of that too
tawdry perch
how would i save the list to my hard drive?
stable pecan
partitions is a generator, so you can generate the first, say, 1000, elements and save them to disk, and then generate the next 1000 and so on
tawdry perch
how do you do that?
tawdry perch
but how would i type it in my code?
i didn't learn much about all these because our school uses http://cscircles.cemc.uwaterloo.ca to teach it
stable pecan
In [44]: parts = partitions(73, 26)
In [45]: list(islice(parts, 100))
Out[45]: [[26, 26, 21], [26, 26, 20, 1], [26, 26, 19, 2], [26, 26, 19, 1, 1], [26, 26, 18, 3], [26, 26, 18, 2, 1], [26, 26, 18, 1, 1, 1], [26, 26, 17, 4], [26, 26, 17, 3, 1], [26, 26, 17, 2, 2], [26, 26, 17, 2, 1, 1], [26, 26, 17, 1, 1, 1, 1], [26, 26, 16, 5], [26, 26, 16, 4, 1], [26, 26, 16, 3, 2], [26, 26, 16, 3, 1, 1], [26, 26, 16, 2, 2, 1], [26, 26, 16, 2, 1, 1, 1], [26, 26, 16, 1, 1, 1, 1, 1], [26, 26, 15, 6], [26, 26, 15, 5, 1]
...
that gives me the first 100 partitions, if you use islice again with the same exact code, it will give the next 100
stable pecan
In [46]: list(islice(parts, 100))
Out[46]: [[26, 26, 11, 8, 1, 1], [26, 26, 11, 7, 3], [26, 26, 11, 7, 2, 1], [26, 26, 11, 7, 1, 1, 1], [26, 26, 11, 6, 4], [26, 26, 11, 6, 3, 1], [26, 26, 11, 6, 2, 2], [26, 26, 11, 6, 2, 1, 1], [26, 26, 11, 6, 1, 1, 1, 1],...
txt or jsn or i dunno, some pickler
tawdry perch
how should i change the code to make it save to hdd?
because i just did this to make it in ram
allCombos = list(map(list, list(itertools.chain.from_iterable(more_itertools.distinct_permutations(partition) for partition in partitions(userChecksum, 26)))))
can i use the pickle module like so: https://stackoverflow.com/a/27745539/12946280 ?
stable pecan
yep, pickle or jsn, but they aren't very memory efficient
fiery cosmos
Hi
Does big oh notation equal the number of steps needed for the code?
like O(n^2) does that mean if i have 10 numbers, for example, there would be 100 steps in the algorithm?
serene helm
basically yes
bronze gazelle
It's generally not that literal, but it does explain to you the trend of how your algorithm performs as the number of inputs increases
tawdry perch
@stable pecan hello?
ionic glacier
hi, im trying to learn algorithims and i don't understand this vid at 34:40. How come he groups it all the log 2 n times. https://www.youtube.com/watch?v=HtSuA80QTyo&t=1673s
last blaze
an O(1) operation happens O(log n) many times
ionic glacier
and is log n the concept of just splitting in half?
fiery cosmos
I think spliting is log 2
ionic glacier
ah ok. i just didnt understand where he got this "log2" or "log n " stuff from. prob missing some background
sleek hare
if you have a list of N things, and you split that into N/2 things, and then again into N/4, things and so on until the result is of length 1
the number of splits is O(Log N)
fiery cosmos
Is quicksort's space complexity O(log n) or O(n)?
I am confused since some websites say O(log n) or O(n)
stable pecan
Mathematical analysis of quicksort shows that, on average, the algorithm takes O(n log n) comparisons to sort n items. In the worst case, it makes O(n2) comparisons, though this behavior is rare.
so it's O(n log n)
fiery cosmos
space complexity
Is quicksort's space complexity O(log n) or O(n)?
I am confused since some websites say O(log n) or O(n)
Not time complexity
I guess it depends on the implementation
Using tail-recursion or iteration the largest sub-array can only ever be half of the input size
so in that case the space complexity is O(log n)
royal kestrel
The space complexity is O(log n) because you either
- need to store
log nstack frames for a recursive implementation - need to store
log nelements in a stack if you implemented it iteratively
Though one could argue in the worst possible case the space complexity could reach O(n) in certain implementations under very specific conditions
tribal pendant
you wrote time, but you meant space @royal kestrel , no?
royal kestrel
yes
fiery cosmos
@royal kestrel thanks
thick flicker
Is there a relationship between logical addresses and Virtual Memory?
north zealot
Depends on the implementation
thick flicker
It seems to me that we have to go through several layers before accessing the real memory. I mean, first we have logical addresses for the processes, and then the Virtual Memory Space, and only after is the physical memory accessed.
Am I right?
north zealot
Well, the logical adresses are inside the Virtual Memory Space, which are just address range of the physical memory
And the Virtual Memory is the technique capable of mapping the logical adress to the physical address, and assign the memory space
thick flicker
So the Memory Management Unit is the responsible for Virtual Memory
stable pecan
anyone ever played with generating sets of non-intersecting circles?
i spent some time today playing with it
class Word:
def __init__(self, word1=None, word2=None, concat=True):
if word1 is None:
self.word = '()'
elif concat:
self.word = f'{word1.word}{word2.word}'
else:
self.word = f'{word1.word[0]}{word2.word}{word1.word[1:]}'
def __repr__(self):
return self.word
def __add__(self, other):
return Word(self, other)
def __rshift__(self, other):
return Word(self, other, concat=False)
def __eq__(self, other):
return self.word == other.word
def __hash__(self):
return hash(self.word)
def all_words_length(n):
if n == 1:
yield Word()
return
for word in all_words_length(n - 1):
yield word + Word()
yield Word() + word
yield word >> Word()
yield Word() >> word
you can generate them with a simple grammar:
In [152]: set(all_words_length(4))
Out[152]:
{(((()))),
((()())),
((())()),
((()))(),
(()(())),
(()()()),
(()())(),
(())(()),
(())()(),
()((())),
()(()()),
()(())(),
()()(()),
()()()()}
where this set represents the non-intersecting circles
but there's a lot of redundant words using this method
can anyone think of a way to get all the non-intersecting circles without the redundancy
thick flicker
How far should I learn CS to become a better programmer.
I have some basic knowledge of Computer Architecture and Operating System.
candid osprey
Depends. Cs knowledge usually only helps, so more the merrier. You can focus on the things in your critical path first though, such as data structures, logical operators and io
torn pine
Does anyone know any good learning programs? Other then codecademy, ping me if you respond please
compact vine
@torn pine sololearn
torn pine
Thanks
woven wind
What's a good book for data structures and algorithms? Just preparing before data science
royal kestrel
Algorithms + Data Structures = Programs by Niklaus Wirth
woven wind
I have a question, is it necessary for me to be able to code an algorithm if I can understand it completely? for example, if I know how bubble sort works, but I can't write code that does bubble sorting, is it ok?
woven wind
I'm having problems writing code like that, what's the first something I should do, make a flowchart? pseudocode?
candid osprey
If you know how it works, why can't you code it?
runic tinsel
writing the code is what you're good for
woven wind
Coding it seems more complicated
candid osprey
Pseudocode isn't a bad idea. But point being, you might need practice then
Because if you can understand something end to end, you should be able to code it up. It might take a bit, but it's a very important skill to solidify
You might think it's not important for algos. But think of it this way, What's the use of coming up with a solution if you can't give it life?
woven wind
Because most of these algorithms are built in
like libraries or built-in functions, methods
candid osprey
That's what I meant, The skill applies to solutioning beyond just algos.
woven wind
So how should I start trying to code one of them
candid osprey
Tricky, because usually understanding the logic is the hard part.
Well i suppose, if you say you understand an algo.. Can you write it as a series of steps in plain English?
woven wind
yes
candid osprey
Well defined instructions. One instruction per sentence. In order.
OK. That's a good start. Then take one sentence at a time, and try to translate that to code
Don't proceed till you make sure one sentence is translated and works
woven wind
TYSM, I'm gonna try that, I hope it helps
candid osprey
Fingers crossed
fiery cosmos
Hi. As homework, I've got to calculate the millionth prime number in under 1 second. What's the best algorithm to go about this, and should I bother doing it in Python? Thanks!
Does this sound sensible? I'll estimate the millionth prime using pi(n) = n/log(n), and then use the Sieve of Eratosthenes to calculate all the primes up to there?
serene helm
What's the point of estimating first?
fiery cosmos
The sieve doesn't calculate primes, but only finds primes upto a certain number. Making it sieve specifically upto 15,485,863 (the millionth prime) would be cheating.
serene helm
Pretty sure there's a way to implement it without having to stop at any specific number
fiery cosmos
That's great! What's the way?
serene helm
I did it by nesting filters, but that's way too slow
def primes():
numbers = itertools.count(2)
while True:
prime = next(numbers)
yield prime
numbers = filter(lambda num, prime=prime: num % prime, numbers)
Thank you!
gusty grove
I would estimate the range that the nth prime is in and check for it in there
vital quiver
So I have a slightly headache-inducing problem. Let's say I have a number of jobs, denoted with an ID. Each of those jobs has parents, and I need to map out all job chains given the parents so given the data on runtime/memory for those jobs, assign priority to them.
gusty grove
No need to sieve all the way up to the nth prime
vital quiver
Jobs can have multiple parents, and jobs can have multiple children, for reference.
gusty grove
@fiery cosmos until when do you need this done?
swift breach
@fiery cosmos https://en.wikipedia.org/wiki/Sieve_of_Atkin
If you're doing it in python, make sure the startup cost of the interpreter doesn't count against your timing. Other than that, any modern method of generating primes should be fine
fiery cosmos
@gusty grove There's no urgency, but the sooner the better obviously. Thanks for the help
@swift breach Thanks. I've tried it with the Sieve of Eratosthenes, and got about 5-6 seconds.
Any ideas to make it go down to 1s, or should I try a new algorithm entirely? Any supercomputers I can rent out for 10 seconds? :p
gusty grove
im working on something, but its not quite working yet 😉
swift breach
It's certainly doable
In [7]: m.nth_prime(1000000)
Out[7]: 15485863
In [8]: %timeit m.nth_prime(1000000)
217 ms ± 7.52 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
(this is timing from some old code)
I'd suggest looking for a faster means of generating primes. Sieve of atkin and some wheel based options both work much faster.
swift breach
There are a handful of way to improve the speed at which you're checking in python as well.
Division across a numpy array is faster than divison across each element of a python list.
There's also little ways to only check specific numbers along the way.
Observations like:
All primes > 3 are of the form 0 = p mod 6 ± 1
let you reduce your search space.
fiery cosmos
tysm!
stable pecan
i've got a numpy sieve that's quite fast, but not so readable
In [95]: %timeit primesfrom2to(6000000)
43.9 ms ± 718 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)
though i have a pure python version that's really close to the limit:
In [98]: %timeit sieve(6000000)
890 ms ± 7.25 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
urban gust
I have a question dealing with Scraper API and Python. Can someone help?
how do i insert code in here? I haven't used the site in a while and forget
topaz pulsar
use ``` to open code block
fiery cosmos
@stable pecan Thank you so much! Do you think if I run my sieve in numpy it'll become faster?
stable pecan
yes, numpy is written in c, so you get to replace slow python loops with fast c-loops
fiery cosmos
This is my current code. How would I go about sticking numpy in here? An array of zeroes perhaps?
start = time.time()
#limit = pass + 1 # Prime number theorem
limit = 16000000
flags = [True] * (limit) # Start with all numbers as potential primes
flags[0] = flags[1] = False # 0 and 1 are not primes
for i in range(2, int(limit ** 0.5)):
if flags[i] is True:
for j in range(i * i, limit, i): #Start at prime squared, step by prime
flags[j] = False
print([i for i, n in enumerate(flags) if n == True][999999])
end = time.time()
print(end - start)
Thanks - that sounds promising about numpy
stable pecan
you can speed that version up with slice assignments
but numpy lets you flip flags to false even faster
fiery cosmos
Excellent! What function should I use in numpy to do that?
I tried using slice, but I got some error about integer iteratives
stable pecan
In [76]: import numpy as np
...: arr = np.zeros(20, dtype=bool)
...: arr
Out[76]:
array([False, False, False, False, False, False, False, False, False,
False, False, False, False, False, False, False, False, False,
False, False])
In [77]: arr[::3] = True
...: arr
Out[77]:
array([ True, False, False, True, False, False, True, False, False,
True, False, False, True, False, False, True, False, False,
True, False])
you can use slice assignment like so for a numpy array
fiery cosmos
thank you very much!
fiery cosmos
Is there any difference if I use an array of ones?
Just to make it easier, cos it represents true
stable pecan
no difference, but boolean arrays will take less memory so be slightly faster to traverse
fiery cosmos
I'm going with np.ones(n, dtype=bool). is that the same?
stable pecan
In [78]: import numpy as np
...: arr = np.zeros(20, dtype=np.uint8)
...: arr
Out[78]:
array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
dtype=uint8)
you uint8 is the smallest integer data type i think
fiery cosmos
tysm! even faster than bool?
Also, what's the fastest way to change those 1s to 0s?
stable pecan
np.ones will just start you with all Trues, so whichever one works
stable pecan
use the numpy slice assignments as above arr[::3] = True will make every third element True
fiery cosmos
ah, thanks a lot!
fiery cosmos
yep, thanks a lot
stable pecan
you can put the starting number at the beginning as well
fiery cosmos
And if I just need to set flags[0] and flags[1] to 0, what's the quickest way to do that?
stable pecan
arr[n**2::3] = True
fiery cosmos
thanks
stable pecan
maybe this is more illustrative of slice-assignments:
In [79]: arr = np.arange(20)
In [80]: arr
Out[80]:
array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,
17, 18, 19])
In [81]: arr[::3]
Out[81]: array([ 0, 3, 6, 9, 12, 15, 18])
In [82]: arr[::3] = 100
In [83]: arr
Out[83]:
array([100, 1, 2, 100, 4, 5, 100, 7, 8, 100, 10, 11, 100,
13, 14, 100, 16, 17, 100, 19])
tysm you've been so helpful
stable pecan
np
fiery cosmos
np - nice pun lol
just one q - what's the quickest way to set flags[0] and [1] to 0? with python alone?
stable pecan
you can do it like normal, but that's slow, you want to replace this whole loop:
for j in range(i * i, limit, i): #Start at prime squared, step by prime
flags[j] = False
with the single slice assignment:
flags[i * i:limit: i] = False
if you need to flip a single flag you can index like before
fiery cosmos
index? soz
stable pecan
flags[i] = False
stable pecan
whichever you want
fiery cosmos
if im using dtype uint8, whats better?
stable pecan
False, True will automatically get converted to 1,0 for a uint array and vice-versa for a boolean array
fiery cosmos
ah tysm!
stable pecan
boolean array just uses less memory
stable pecan
boolean
fiery cosmos
tysm!
got it in 0.2 seconds. tysm!
my compsci teacher has a bit of a prize waiting :D
so yeah, thanks for your invaluable help - it was super useful
stable pecan
no worries, numpy should blow other solutions out of the water
fiery cosmos
thats so helpful, seriously tysm
stable pecan
oh, there's a faster method than enumerate for recovering the integers
stable pecan
that works
In [84]: arr = np.random.randint(0, 2, size=20, dtype=bool)
In [85]: arr
Out[85]:
array([ True, True, False, False, True, True, True, True, True,
False, False, False, True, True, False, True, True, False,
True, False])
In [86]: np.nonzero(arr)
Out[86]: (array([ 0, 1, 4, 5, 6, 7, 8, 12, 13, 15, 16, 18]),)
you don't have to flatten it though
fiery cosmos
oh wow, thanks
i get an error without flat, but it blows the 1s mark out of the water, so im happy
tysm gtg
stable pecan
take care
sand coral
I want to start a project. The project is a chat app like this, with voice and everything. I would like to make it lik skype and completly free. So, to start, I need some guidance on what to learn. I need to know some libraries, some pattern and everything, but first I need to know if I can start a free software project. I'm one person team by now, but I would like to get people envolved.
Just want to know If it's doable, and what stuff I need to work on first. I would like to create it from scratch, and get developing this app through time. Like 2 years or more. I really would like a speaking foreing language chat app available for free for Linux.
!subscribe
fiery cosmos
@stable pecan I've worked it to use np.bool and got it to under 10 milliseconds. Tysm!
stable pecan
anytime
past lance
If you're going to make a chat application I would strongly recommend learning socket... requires very basic networking knowledge but fairly potent. Here's a good place to start: https://docs.python.org/3.8/howto/sockets.html
sand coral
Thank you!
minor chasm
just a quick question ... before starting to learn path finding algorithm do u need to have knowledge of other algorithm and concepts ??
minor chasm
oh sweet thnx bud
tribal pendant
depends on the complexity of the algorithm, and the time complexity you want to achieve.
royal kestrel
You most likely need to be familiar with graphs, and some form of priority queue
solid umbra
I am planning to use python for competitive programming.. is it good idea?
ripe flame
What are you going to be competing against?
solid umbra
I am planning to compete in Hackerrank Hackathons or CodeChef
radiant vault
I don't think so
fiery cosmos
In competitive programming it doesn't really matter what language you use. Its about you knowledge in algorithms, datastructures, maths, etc
Some languages make it easer to implement a solution to a given problem, but you need to know what to do
For instance in ProjectEuler, just brute forcing problems won't work
that being said,Python is very handy since you do have numpy, and scipy
tribal pendant
python is one of the best language then, with its huge stlib
last blaze
Speed does matter - there are some languages where bruteforcing will work, and then there are some algos that you can implement in a less optimised way if you don't do it in Python
most just default to C++ afaik - or Java or another compiled lang
tribal pendant
ah ok
fiery cosmos
does anyone know how to show enum values in a combo box c# UWP?
tribal pendant
I didn't even understand the question, so afaik I can't help, sorry
fiery cosmos
does anyone know c# here
I mean numpy, and scipy are written in C so they will run fast as well. And again as I mentioned, it comes down to knowledge not what language you use
last blaze
You don't want to be calling in and out of C for your logic
Also, judging environments don't even necessarily allow libraries
fiery cosmos
Don't sure I'm asking in right channel, but why there's a numpy.array([...])? Why do not just a = [...]?
last blaze
This isn't really the right channel. But basically, numpy arrays are faster than python lists
for more info #python-discussion would probably be better
tribal pendant
If it’s about speed it’s the right channel
umbral shuttle
New to python trying to figure out how to read a file into an array
I have a textfile with a sudoku in it
0 0 0 0 0 0 2 0 0
0 8 0 0 0 7 0 9 0
6 0 2 0 0 0 5 0 0
0 7 0 0 6 0 0 0 0
0 0 0 9 0 1 0 0 0
0 0 0 0 2 0 0 4 0
0 0 5 0 0 0 6 0 3
0 9 0 4 0 0 0 7 0
0 0 6 0 0 0 0 0 0
what will the array have in it?
I'm guessing you want integers in rows and columns
@umbral shuttle ^^
umbral shuttle
yeah
I found a way to do it
Not very elegant though @stark bridge
def readSudoku(directory):
# {
file = open(directory)
sudoku = []
for i in range(9):
row = []
for i in range(9):
row.append(int(file.read(1)))
file.read(1)
sudoku.append(row)
# file.read(1)
file.close()
return sudoku
# }
def printSudoku(sudoku):
# {
for y in range(9):
row = ""
for x in range(9):
row = row + str(sudoku[y][x])
row += " "
print(row)
# }
printSudoku(readSudoku("puzzle.txt"))
heh
that's not terrible but I see what you mean
the one change I would make, is: instead of reading one byte, turning it into an int, and then reading another byte and just disacarding it ...
... I'd read the entire line, then use .split to get a list of integers
umbral shuttle
yeah thats what I was looking for
what the method was called
I actually do know js funnily enough
stark bridge
is this homework?
umbral shuttle
no
I am doing a python course but it is very slow
I want to brute force some puzzles for funzies because it is raining out and I have some free time
Trying to quit league of legends actually haha
stark bridge
I can show you how I'd do it -- there's a lot of little differences, but some might inspire you
def readSudoku(directory):
with open(directory) as inf:
sudoku = []
for line in inf:
row = [int(field) for field in line.split()]
sudoku.append(row)
return sudoku
def printSudoku(sudoku):
for row in sudoku:
for index, col in enumerate(row):
if index > 0:
print(' ', end='')
print(col, end='')
print()
printSudoku(readSudoku("puzzle.txt"))
it's more general -- you could have two- or three-digit numbers in your puzzle if you wanted
and it could have any number of rows or columns
umbral shuttle
what does map(int, ['1', '2']) do?
def readGrid(directory):
with open(directory) as file:
grid = []
for line in file:
grid.append(list(map(int, line.split())))
return grid
def printSudoku(sudoku):
# {
for y in range(9):
row = ""
for x in range(9):
row = row + str(sudoku[y][x])
row += " "
print(row)
# }
printSudoku(readGrid("puzzle.txt"))
this is what I have so far
I am going to do a validation step
and a generic printing method
as well as a brute force solver after lunch
thanks for the help!
stark bridge
map(int, ['1', '2']) returns [1, 2]
it runs the int function on each string in the list, and returns a list of the results
stark bridge
I used to know that 🙂
they might be the same thing. it's been so long since I did scheme that I can't remember
reduce certainly takes a sequence and returns a single value; "accumulate" sounds like it'd do the same thing
stark bridge
dunno
umbral shuttle
oh
stark bridge
I wonder if you're looking at docs for numpy
umbral shuttle
The summation of list using accumulate is :[1, 4, 8, 18, 22]
The summation of list using reduce is :22
umbral shuttle
stark bridge
accumulate() returns a list containing the intermediate results
that sounds like the crucial difference. Use whichever you like, as long as it works
umbral shuttle
reduce works exactly the same as it does in js
which is what I was looking for
any way to neaten this up?
import functools
def readGrid(directory):
with open(directory) as file:
grid = []
for line in file:
grid.append(list(map(int, line.split())))
return grid
def printGrid(grid):
for row in grid:
print(functools.reduce(lambda x,y:x+" "+y, map(str, row)))
printGrid(readGrid("puzzle.txt"))
grid = []
operation
return grid
is really messy
can I somehow shove this all into one argument?
@stark bridge ^
umbral shuttle
map(lambda x:grid.append(list(map(int, x.split()))), file)
is not working
I am trying to use a map
instead of grid = []
then appending
then putting that in a list
YEAH!
I AM FEELING IT!
PYTHON POWER!
import functools
def readGrid(directory):
with open(directory) as file:
return list(map(lambda x:list(map(int, x.split())), file))
def printGrid(grid):
for row in grid:
print(functools.reduce(lambda x,y:x+" "+y, map(str, row)))
printGrid(readGrid("puzzle.txt"))
PYTHON POWER!
stark bridge
def readGrid(filename):
with open(filename) as file:
for line in file:
yield map(int, line.split())
def printGrid(grid):
for row in grid:
print(" ".join(map(str, row)))
printGrid(readGrid("puzzle.txt"))
don't need functools in mine
stark bridge
oy
it ... kinda ... returns, but saves its place, so that the next time you call the function, it takes up where it left off.
it's weird.
funny how the two functions are pretty similar 🙂
umbral shuttle
oh wow lol
.join make alot more sense here
instead of doing reduce + a lambda
def readGrid(filename):
with open(filename) as file:
for line in file:
yield map(int, line.split())
I don't understand yield here
you don't have a return statement?
this whole thing confuses me
return list(map(lambda x:list(map(int, x.split())), file))
isn't very good either though
because its very hard for me to understand
stark bridge
you gotta read about yield; it's as weird as you think it is
fiery cosmos
Hello
umbral shuttle
Hi
umbral shuttle
how can I have a function with attached storage?
such that the function contains a set and I can pass a flag to it which dumps the contents of the set or and also takes another argument which is stored regardless
I have heard this refer to as a functor
Also I am aware this could be done with a global variable
BUT I to encapsulate the contents of the function
And return a bool which is if the contents are unique or not
tiny jay
maybe a class, with a callable instance
feral gulch
last blaze
That'd probably be fine
thick eagle
guys whys it so hard to read x86 assembly
warm marsh
I don't think this is a question anybody can give a satisfying answer to 🤔
north zealot
Because ASM?
minor chasm
so i was trying to undersand the rasterisation rule(for without multi sampling) but i couldnt understand the top edge and left edge rule. can some pls explain it to me. ty 😄
actually i got it waat it means :() was just me being dumb lol
so from what i understoo it sud fill that top part based on top edge rule isnt it ???
daring plinth
Hello 🙂
I hope this is the right channel for a question like this.
I'm currently looking for an efficient way to read a random chunk from an audio file in order to train a neural network.
The most efficient way I could come up with was using
from scipy.io import wavfile
wavfile.read(file, mmap=True)
I'm not too familiar with mmap but it seems to be fast because it's not loading it in memory but using virtual memory to map to the storage drive.
I was interested if anyone knows a better way to do, like only reading the header of the file to get meta-data and then seek to the random position to read the sample.
If possible I'd like to keep files as mp3 or mp4 (audio only) in order to not take too much storage, but I didn't find any library with ~ same efficiency as scipy.wavfile.
Thanks for any insights 🙂
lucid nest
Write a function that takes the name of the shape as input and then automatically calculates the output for the respective shape. For simplicity, you may write the function for triangle and rectangle because both the functions take base and height as input.
Can anyone guide me ?!
candid osprey
Can you do this task if I told you to write it for just a triangle?
loud prism
Hi
silk stone
Hi,
While I'm learn the fundamentals of programming, through solving problems and coding challenges, but I have an important question:
Should I learn algorithms and data structures first or after spending some time with coding challenges ?
because I got stuck very while I solving problems now
thanks 😃
runic tinsel
don't focus on a thing maybe?
i bet you didn't choose between learning extensor control and playing with gravity prediction when you were learning to walk
fiery cosmos
name chacks out
fiery cosmos
Has anyone here used the Runestone course to learn data structures and algorithms? How was it?
spiral swallow
@thick eagle because it's very low level, and it has (afaik) a lot of very optimized instructions for particular situations, it's not done for humans, it's done for compilers, some cpu architectures have less instructions set, and so simpler circuits, and they can certainly be easier to read, but that's not what x86 (and x86_64 in particular) is designed for. Also, as it's low level, the code has to care about a lot of details, which makes it harder to understand the high level objective of the code.
gusty grove
most of the time you use some kind of pathfinding (something like A* or less advanced)
bronze gazelle
You might be able to glean some insight from some of the modding tools folks have created. At a glance it looks like they're using splines but I am totally unfamiliar with pathfinding algorithms
stable pecan
you can also hard code millions of paths pretty cheaply and just use one at random
i have no idea if that's what they did
gusty grove
normally they just mark the vertices
this is how unity's navmesh agents handle it. id assume others do it similarly
quasi oracle
Considering the police can track you down I doubt the paths are hard coded
bronze gazelle
This gives how the nodes are defined for GTA V at least https://gta.fandom.com/wiki/Paths_(GTA_V)
spiral swallow
I'd just make the vehicles take random directions at intersections, its not like the player is going to check if it makes sense, and yeah, maybe have a small amount of them and move them around when the player doesn't look
north zealot
The fancy way : GTA V actually destroy any vehicules or NPC out of range, so you could just spawn them using a circle around the player, and target an intersection node at other intersection on the circle and recalculate them as the player moves
fiery cosmos
Just make sure you don’t duplicate The GTA methods disappearing vehicles right next to you because you panned the camera. Us camera whippers notice that lazy coding real quick.
sour wigeon
round plume
Everytime i download python it DELETES ALL MY FILES
gusty grove
Yeah probably they just pathfind to the player. But im pretty sure gta wil spawn cops in front of the player
spare tree
is there any name for a diagram that just shows the relationship between the methods in a class?
spiral swallow
@spare tree what kind of relationships? method1 calling method2 ? i guess i flowchart could do, or maybe boxes nested in boxes
but i don't know how that last one would be named
spare tree
yeah basically that's my intent
and is it good to have more classes?
in general, for places where you dotnt neded as many classes
spiral swallow
no, don't do classes you don't have a need for
you can make one that you know are going to be handy, but only if you know they are going to be handy really soon, not in some distance future
because we always get that wrong and it makes things unnecessarily complex, and makes the design harder to evolve
stark bridge
man I hope you're joking
last blaze
In the UK w/ a levels, you'd get more marks for a massively overcomplicated oop structure than for actual good code. It's ridiculous, but exam boards are dumb like that
stark bridge
sigh
gusty grove
how would a 3d scanline boundary/flood fill work?
fiery cosmos
@last blaze i can vouch for this
graceful forum
oh dear im planning to take a levels
i can code decently already but something about it scares me
boreal verge
@graceful forum im doing a levels rn
Im doing the international Cambridge 9608 syllabus
If u have any questions, Id be more than happy to help
graceful forum
aye thanks
fiery cosmos
is anyone here good at scheme? @/dm if you are pls
stark bridge
I used to be but a) I'm only here sporadically and b) I don't do direct messages
fiery cosmos
y is this giving an error
plt.xlabel("correct" + class_names[test_images[i]])
TypeError: only integer scalar arrays can be converted to a scalar indexerror
hot latch
@fiery cosmos You'd probably be better asking that in one of the #help channels, from 0 to 7, that's more where you're going to get help for things like this. #help-kiwi looks like it's been free for some time now. You can check free channels in #bot-commands by doing !f.
Looking at this, it looks like the test_images array you're using is meant to hold a different data type than what can be accepted as an array index.
I don't know what libraries you're using, and am only* assuming this is python, so I'd recommend adding more information to the question and posting it to the other channel ( #help-kiwi ).
hot latch
Something I wanted to discuss/ask about:
Can you have a class which as a part of it's constructor, creates a new instance of that same class, except encapsulated within the first instance?
Is this recursion?
How does this relate to first-class objects/functions/etc?
stark bridge
I guess you could do that, but eww
how would the newer one know not to itself create third sub-instance?
hot latch
MAybe have the constructor function pass some number to the next instance, so that it only goes down a certain amount of times?
Also, this is recursion then?
stark bridge
heh
it is but a strange kind
class Turtle:
def __init__(self, size):
if size > 0:
self.beneath = Turtle(size - 1)
# >>> t = Turtle(10)
# >>> t
# <__main__.Turtle object at 0x106f3c430>
# >>> t.beneath
# <__main__.Turtle object at 0x107163fa0>
# >>> t.beneath.beneath
# <__main__.Turtle object at 0x107146970>
# >>> t.beneath.beneath.beneath.beneath.beneath.beneath.beneath.beneath
# <__main__.Turtle object at 0x1071b8340>
# >>> t.beneath.beneath.beneath.beneath.beneath.beneath.beneath.beneath.beneath
# <__main__.Turtle object at 0x1071b8820>
# >>> t.beneath.beneath.beneath.beneath.beneath.beneath.beneath.beneath.beneath.beneath
# <__main__.Turtle object at 0x1071b8580>
# >>> t.beneath.beneath.beneath.beneath.beneath.beneath.beneath.beneath.beneath.beneath.beneath
# Traceback (most recent call last):
# File "<stdin>", line 1, in <module>
# AttributeError: 'Turtle' object has no attribute 'beneath'
# >>>
hahahaha
stark bridge
I guess there might be a legit use for this, but I've never run across one
hot latch
So is there a way to then do something where you'd build/draw a sort of tree like this, in turtle?
Set the angles a bit off, draw a line forward a bit, repeat?
stark bridge
I guess, but it seems inconvenient
hot latch
I just don't know where to put the self.forward()call
stark bridge
I would rather have a simple list of things, than something that you have to traverse using attribute access
hot latch
What is allowed in the __init__ of objects like that?
Can methods of an object be called in it's init?
stark bridge
I guess, seems like a bad idea though
I keep __init__ about as bare-bones and dumb as I can
hot latch
as little work as possible done there
yeah makes sense to do that
What if you were to kill the instances off as they completed some task?
Could this be used for some kind of recursive search around a filesystem?
or is that better done with traditional recursion
stark bridge
if I were searching a filesystem in python, I'd use os.walk, like my granddaddy did
errant ferry
Learning about ADTs
Wouldn't all the data types in Python be considered abstract data types?
last blaze
They would be implementations of those adts
spark night
Is anyone here good at decoding
RID THAT DUES TOY HEM FONT THAT WON TITHE FARE NO NO
How do I decode this
stark bridge
I think "abstract" means: certain operations are deliberately undefined, with the expectation that someone will subclass the type, and provide those operations
magic crypt
Hello, I'd like to learn about Natural Language Processing(NLP).
so far I got lost because it's too vast(for me). The more I read or watch some tutorial the more confused I get, so I don't know where to start anymore.
I'm thinking about trying with simple data like CSS files to understanding how NLP does. since it's also easy to get CSS files anywhere to train with.
and so which module I need to use and what tutorial I need to read about this case?
dry agate
@spark night you want to program a decoder or decode this specific sentence?
spark night
No, I just want to decode it
spark night
Like, it’s an anagram that says a day and time in it, and I can’t even find that itself
stark bridge
I'd pull out "Monday", "Tuesday", etc and see what's left
unfortunately that's long enough that it'd take ages to compute all the anagrams (at least, with my not-terribly-fast anagram checking thingy)
might's well try months, too, since those are pretty big words; if you find a month in there, removing those letters might let you brute-force what remains
jolly seal
magic crypt
why not 'else' for the last 'if'?
tiny jay
because else can be related only to a single if
tiny jay
run these commands one by one and tell me the output:
py -V
python -V
python3 -V
py -0 too [that is a zero]
open flume
@tiny jay r u talking to me
tiny jay
yes
tiny jay
@open flume ok so py
@open flume next question - are you using some sort of editor to write your code? A programming editor perhaps?
open flume
yes
the python idle
wait i figured out y python wasnt working
i didnt add it to system path
tiny jay
I strongly suggest using a better editor than IDLE - it is really bad
open flume
like
tiny jay
I suggest https://thonny.org - it is based on IDLE so you can consider it a upgrade - and it has very nice features for beginners - including a graphical package manager frontend to pip
tiny jay
if you did the path thing correctly, yes
otherwise py -m pip install django
to use pip from py
charred reef
Hey are there any good math geniuses here? Stuck on something!? Would really appreciate some help! cheers
worthy star
@charred reef How complex?
charred reef
Well im trying to make an algorithm that can cycle through each possibility
In other words, I need the system to cycle through each buy and sell possibility that I have in my txt file. (Its a trading system) i want to force it to buy and sell everything so i can understand what factors makes a profit of loss
If that makes sense @worthy star
worthy star
Sort of, but what maths do you envisage it'll require?
charred reef
are you willing to jump into a quick voice chat? Its almost impossible to explain with words xD
worthy star
Don't have a mic, so won't be able to speak.
I'm not a maths genius either, but your problem doesn't seem to require complex maths.
Calc is my limit (pun intended).
charred reef
So as a quick example @worthy star I have a time period from 1 - 10 (lets say minutes) and at each time frame I have price data values so: (open, close, low, high etc.) What i need the program to do is place a trade, so buy and sell, at each time frame. So: BUY @ 1 - SELL @ 2, then : BUY @ 1 - SELL @ 3, then: BUY @ 1 - SELL @ 4 ---- i need it to keep doing this until it runs out of buy and sell combinations so it should end at BUY @ 10 - SELL @ 10
Is this a maths problem or just a simple loop code?
charred reef
is that legit it?
halcyon plankBOT
:incoming_envelope: :ok_hand: applied mute to @worthy star until 2020-03-26 15:07 (9 minutes and 59 seconds) (reason: newlines rule: sent 106 newlines in 10s).
charred reef
Lol wtf?
cobalt crater
!unmute 566591369309847567
halcyon plankBOT
:incoming_envelope: :ok_hand: pardoned infraction mute for @worthy star.
charred reef
Well thanks @worthy star anyways!
mossy gorge
!paste
halcyon plankBOT
worthy star
Haha.
charred reef
lmao
mossy gorge
Might want to use the hastebin we host for long outputs like that
worthy star
Thanks @cobalt crater .
mossy gorge
The bot gets a bit touchy when it sees that many newlines
worthy star
I see, thanks.
@charred reef The code is:
for buy in range(1, 11):
for sell in range(1, 11):
print('Buy:', buy, 'Sell:', sell)
That should work.
worthy star
You're welcome!
wild bronze
hey I have a question about RSA keys. I have a file that contains rsa public key. I looked it up and got the idea how they are generated by two prime numbers. So my file (I think) contains a number which is base 64 encoded but when I decode it back with base64 I get unprintable characters not numbers(I expected numbers). Does that mean my RSA key numbers are encoded with a different encoding rather than utf and then base64 encoded??
north zealot
No they're just raw bytes representing your key
wild bronze
but when I look at the wikipedia page
I see statements like
c(m) = m^17 mod 3233``` so do raw bytes not represent 3233 and 17?@north zealot
north zealot
I think they do, but I'm not sure
stark bridge
I'd be surprised if the numbers were as small as 3233 and 17, but I think those raw bytes do indeed encode integers
wild bronze
could this be the case like the difference between binary 1 and printable character 1?
lets say my key is (n = 255 e=5)
11111111 00000101 binary version
stark bridge
something like that
there are lots of ways to encode numbers into bytes, and I don't know which one the RSA key uses -- but I bet it's a fairly common, and well-documented one
wild bronze
How can you tell looking at the base64 encoded 'G+glEae6W/1XjA7vRm21nNyEco/c+J2TdR0Qp8dcjP' string with a lenght of 42 characters is 32 bytes long?
north zealot
Yes you can
Base64 have an overhead of exactly (100/3)%, rounded up
So you can exactly calculate the length of the decoded string
charred reef
Hi guys, can someone tell me why this code is in-consistent? :
''' for buy in dataPrice:
for sell in dataPrice:
if sell >= buy:
outcome = ('\x1b[0;31;40m' + "Loss")
else:
outcome = ('\x1b[0;34;40m' + "Profit")
print("Buy:", buy[3], "Sell:", sell[3], outcome) '''
and how the hell do I paste code in discord correctly lol
worthy star
Use three back-ticks before the code and three after.
@charred reef Back-tick is the (`) symbol next to the 1 key (on a QWERTY keyboard)
charred reef
for sell in dataPrice:
if sell >= buy:
outcome = ('\x1b[0;31;40m' + "Loss")
else:
outcome = ('\x1b[0;34;40m' + "Profit")
print("Buy:", buy[3], "Sell:", sell[3], outcome) ```Ah lovely
cheers
worthy star
What's the problem?
charred reef
Heres a little snippet of the output code:
Buy: 1.5016 Sell: 1.5017 Profit
Buy: 1.5016 Sell: 1.5015 Profit
Buy: 1.5016 Sell: 1.5017 Profit
Buy: 1.5016 Sell: 1.5016 Loss
now the 2nd line, the Sell value is lower than the buy value, but it's still registering as a profit where it should be a loss
cant figure out why
charred reef
3
3.7
even though in my code it states that if sell >= buy:
But in this case sell is less than buy
but output still seems to be a profit
worthy star
You are checking sell >= buy yet you are printing buy[3] and sell[3]. Print buy and sell instead and see the result.
charred reef
Nope same issue, I just get the entire list printed rather than that column [3]
worthy star
Can you show the output?
charred reef
Buy: ['GBPUSD', 20010102.0, 234900.0, 1.5011, 1.5012, 1.5009, 1.5009, 4.0] Sell: ['GBPUSD', 20010102.0, 230100.0, 1.5021, 1.5021, 1.5018, 1.5018, 4.0] Profit
Buy: ['GBPUSD', 20010102.0, 234900.0, 1.5011, 1.5012, 1.5009, 1.5009, 4.0] Sell: ['GBPUSD', 20010102.0, 230200.0, 1.5019, 1.5019, 1.5019, 1.5019, 4.0] Profit
thats just 2 rows, i have thousands
hold on
actually it technically worked
but i only wanted it to print column [3] instead of all of them
charred reef
for sell in dataPrice:
if sell[3] >= buy[3]:
outcome = ('\x1b[0;31;40m' + "Loss")
else:
outcome = ('\x1b[0;34;40m' + "Profit")
print("Buy:", buy, "Sell:", sell, outcome)```Now just change print("Buy:", buy, "Sell:", sell, outcome) to print("Buy:", buy[3], "Sell:", sell[3], outcome)?
charred reef
ah
let me try
its worked
Buy: 1.5011 Sell: 1.5009 Profit
Buy: 1.5011 Sell: 1.5011 Loss
Buy: 1.5009 Sell: 1.5021 Loss
but the values are opposite let me cvhange that
yeah perfect its working
One more thing @worthy star Any idea why the colours are abit dodgy?
worthy star
Yeah, because it's not being reset. Try using a terminal colour library such as colourama (https://github.com/tartley/colorama).
charred reef
lovely thanks alot guys
sharp atlas
can someone explain how pyramid image blending works
like i get the whole gaussian and laplacian pyramid part but im having trouble visualizing how it would make a nicely blended image
like how does blending the resulting laplacian pyramid mashed together make a normal image
it seems that most of the colors in the laplacian pyramid are just black
sharp atlas
this is probably easier
my confusion starts past line 35, because i dont understand why apple_orange_pyramid[0] is a small color blended image when its generated from a laplacian pyramid
it looks like the laplacian images are mostly black with outlines of the objects which makes sense
or is the first image so small that those outlines are enough to make up the entire image?
lean pawn
Hi everyone. I just used python for computer vision. I need lots of references for pixel interpolation in images using python.
I have an RGB image like the following. In that picture, I have a white unknown pixel. I want to interpolate the unknown pixel starting from the point (x0, y0) using the colour intensity information along the blue line. Can someone provide code samples for interpolating at point (x0, y0) with any interpolation technique? The input is a collection of coordinate colour references x,y = np.where(np.all(img == [0,0,255],axis=2)) and pixel coordinates (x0, y0). The output is pixels (x0, y0) which have been interpolated.
How to Interpolate Unknown Pixel given Coordinates references?
fiery cosmos
I want to make sure I am on point here. A word is a simple piece of data that is used within an instruction. An instruction is stored in a register?
I am reading a book on computer architecture. I get confused on some things.
stark bridge
a "word" is a unit of storage. How big it is depends on the computer
I didn't think anyone talks about "words" anymore
some ancient computers had 36-bit words
instructions can be stored in registers; in fact they have to be put into a register before being used, iirc
most of your instructions, though, are in memory, waiting to be run
warm marsh
I think the unit "word" still used in context of assembler or lower level C code.
bright cedar
hi could someone help me im trying to write a function to create a lightened grayscale image. image by adding 75 to the red, green, and blue components of every pixel. Since higher numbers are closer to white, this make the pixel lighter. Now, grayscale for the new image
drifting drum
@bright cedar iterate through all the pixels and add 75(Making sure to actually take max(value + 75, 255) in case you overshoot) ?
gusty grove
if its already a np array you can just add it directly
gray = np.clip(gray + (75, 75, 75), a_min=none, a_max=255)
@bright cedar that should be a lot faster than looping over it
and more pythonic
with numpy you will always want to use numpy operations if you can
stable pecan
you can also use cv2 to do it for you
In [51]: arr = np.random.randint(0, 256, size=(10, 10, 3), dtype=np.uint8)
In [52]: cv2.cvtColor(arr, cv2.COLOR_RGB2GRAY)
Out[52]:
array([[ 64, 120, 128, 123, 69, 169, 100, 165, 111, 101],
[179, 72, 93, 200, 87, 66, 207, 198, 42, 93],
[ 58, 141, 177, 17, 34, 177, 156, 126, 107, 56],
[ 85, 99, 75, 106, 143, 74, 114, 177, 189, 92],
[206, 189, 72, 111, 85, 32, 230, 45, 208, 42],
[ 53, 153, 166, 117, 81, 98, 136, 141, 192, 93],
[133, 121, 84, 159, 105, 117, 137, 159, 93, 134],
[ 91, 175, 68, 206, 109, 108, 58, 125, 35, 88],
[ 80, 200, 86, 202, 187, 63, 135, 198, 212, 39],
[ 63, 117, 72, 114, 238, 81, 52, 173, 140, 186]], dtype=uint8)
um, this is kind of a stupid question but how do i run my python script in vs 2019?
pls ping me
night ruin
VS2019 iirc has a Python runner
night ruin
that sounds like too much for Python
fiery cosmos
its just a python script i made long ago, i want to test it out with vs
night ruin
you need to run the installer, select Modify, then select the Python Development workload
then continue
night ruin
np
fiery cosmos
so..?
night ruin
you have python installed on yoru pc right
night ruin
then you need to create a python environment so vs knows you want to run python
fiery cosmos
how?
already have a couple of envs installed
python 3.7.5 32-bit and 64-bit
i just want it to run on cmd, like a normal script lol
i guess i could just code on vs and then run the script from the .py file itself
azure inlet
@fiery cosmos @night ruin I don't think this is related to computer science.
Feel free to use a free help channel.
fiery cosmos
oh, apologies
azure inlet
No problem.
errant ferry
Learning about Big-O etc
def convert_dec(n):
binary = 0
for i, _ in enumerate(str(n)[::-1]):
binary += i * 2
return binary
Would this be O(n)?
last blaze
Yes, it would be
errant ferry
@last blaze
If I added a random x = 2 at the bottom, that does nothing
It would be 1n + 2
Still O(n) right?
north zealot
Yes
errant ferry
Thanks
north zealot
O(n) just says that the computation time is proportional to the input size
bright cedar
@gusty grove @stable pecan @fiery cosmos sorry kinda fell asleep im still a beginner in python and my professor required we use the latest verison of pycharm to grayscale from rgb
so the question that was ask was Write a function to create a lightened grayscale image. First, lighten the image by adding 75 to the red, green, and blue components of every pixel. Since higher numbers are closer to white, this should make the pixel lighter. Now, grayscale the new image.
fiery cosmos
@bright cedar lol i think you tagged the wrong ppl haha
bright cedar
ah shoot my bad
drifting drum
you can also use cv2 to do it for you
@stable pecan the code after that, you converted rgb to gray, where's it adding the 75s though?
stable pecan
it's not, just do arr + 75 before hand
clear rose
is there any advantage of using apache spark over pandas dataframes? How would you integrate both?
atomic olive
I think this is appropriate channel. The ACM digital library is free till June due to covid
It’s full of research papers on various topics.
spiral swallow
nice
feral spindle
1585674071
what kind of time is this?
how can i turn it into hours days or whatever
stark bridge
datetime.datetime.fromtimestamp iirc
>>> import datetime
>>> import pytz
>>> datetime.datetime.fromtimestamp(1585674071, tz=pytz.utc).isoformat()
'2020-03-31T17:01:11+00:00'
>>>
valid mantle
Looking for some help on getting started with ML classification.
I have product listings, categories, and descriptions from a number of websites.
I have my own categories that I want to group these listings.
I have manually categorized a large number of the listings.
What's next? Does anyone have a good example project or guide that goes through something similar.
grand barn
Tensor Flows and Pytorchs websites are quite good. Most documentation in this field is quite good as well
Although be warned, they don't hold back on the maths so you probably need to know calc and matrix operations at a minimum
valid mantle
I'm looking into pytorch and text classification now and finding useful resources
thank you!
grand barn
No worries
split geyser
Guys, Does anyone know any good way to train a chatterbot?
stark bridge
I think I seeded mine with a bunch of hand-selected quotes, but I also let it just accumulate stuff from the channel
fiery cosmos
@split geyser This might help
Using a corpus file instead of seeding it with bunch of quotes!
split geyser
Tyy
woven wind
Hello, does anyone have an idea how I can make a minimax function given the state of a tic tac toe board?
fiery cosmos
does anyone know why in boolean logic a inverted gate requires two transistors
while a non inverted boolean gate requires 3 transistors?
I am reading a computer architecture book and it is not really explained well
half forum
@woven wind tic tac toe is either a win/loss or tie
so the psuedocode would look something like this
self.game = Game() # surrounding scope of the MinMax function
def MinMax(self):
if self.game.Winner() == self.game.CurrentPlayer(): # current is X or O
return WIN
elif self.game.Winner() == opposite(self.game.CurrentPlayer():
return LOSS
elif self.game.Winner() == TIE:
return TIE
result = LOSS
for move in self.game.LegalMoves():
self.game.Play(move)
tmp = oppositeResult(MinMax()) # opposite turns WIN into LOSS, LOSS into WIN and TIE into TIE
self.game.Undo()
result = max(tmp, result) # assumes WIN > TIE > LOSS
return result
important functions here are checking terminal state, playing and undoing moves, and being able to find all legal moves for a given state (empty points in tic tac toe)
lean sleet
why do yall like python so much?
soft patrol
development time
it's very easy to get something pumped out and functioning with python, much faster than in other languages
there's a performance cost, but there are ways you can take advantage of multithreading and async and call C code to speed up your program
it's also not as broken as JavaScript
woven wind
@half forum so opposite result is the recursive function?
I have a terminal function, a result function and I'm gonna create an undo function
woven wind
it's supposed to return the optimal action for the AI tho, not the result
elfin sundial
@half forum thats litrally python not pseudo code
there is no defMinMax(self): in pseudo
woven wind
Python is very similar to pseudocode you could just call it pseudocode lol
wild bronze
I want to ask something about Kernel and CPU.
What exactly is a syscall?
Assume I have an application which just prints 'hello world' by printf().
I think what happens in the background is that Printf() makes a syscall to write(). In lower level, does it mean my application running in the user space has an INTx80 instruction which is basically a jmp instruction in assembly which jumps to kernel application in memory to execute codes
half forum
if its on a linux system then yes
and its not exactly a jmp instruction
if you look up what an interrupt is, it saves the state of the currently running process, then handles the interrupt in kernel space
wild bronze
I always think of kernel as a process that I can see in gdb
when you say it handles the interrupt in kernel space I think of code execution in other process which is not accessable by gdb
@half forum on 32 bit systems, kernel space would be 1 gb, user space would be 3 gb in ram. As far as I know cpu won't execute some instructions in user space because this area is protected. Kernel mode however I don't know what it means in this context
half forum
yeah kernel mode
fiery cosmos
Hi
cursive owl
I'd like to build a simple word-order-based LSTM RNN to produce imitations of people in a chat room. I've found the keras library and wow is there a lot going on there. Could anyone recommend any resources or terms to research to send me in the right direction?
cursive owl
I tried my hand at figuring it out on my own and managed to produce the word "wererererere" from a few paragraphs of classic literature. 🤔
soft patrol
why is it that "werererere" is easier to understand than shakespeare?
seems like your RNN is learning well
rancid birch
Hi everyone. I'm trying to write a ast.NodeTransformer subclass that does something similar to constexpr in C++. My attempt and the ideal desired API can be found here: https://hackmd.io/7Mr6f1HORxeg_RD5viZ_IA
Full disclaimer, I'm a newbie at AST manipulations, so be gentle.
hearty maple
Hello, I am sort of a beginner to Python. I am currently in grade 11 so don't know much about advanced python coding. I am trying to make a 2d game where upon screen update the enemy will go towards the player. My problem is it happens too fast and i would like to make it so it takes a while for the enemy to get there. If anyone can help me with that it would be awesome
spiral swallow
well, "toward" is a direction, you want to be able to define the distance, and also, "upon screen update", you want to compute the amount of time that was spent since the previous update, and use that to adjust the distance.
fixing framerate to a fixed value usually make things easier, if you can.
sturdy prairie
What is an algorithm to find the farthest vertex from a given vertex, X, in an undirected graph?
last blaze
two elements with the largest possible shortest path between them?