python-discussion

[granite wyvern]

A mix. Until something works may be a while

[frigid coral]

yea built some projects

[raw bramble]

You can do that with Flask too then! I’m sure there’s plenty of books, maybe someone else here can recommend some (I am not very intellect)

[opal gull]

very different answers here, first is architecture, second is freestyle, and third is both lol

i guess it depends on the person

[frigid coral]

ok thnx for your suggestion!

[crisp jay]

im trying to be more mature (in cs and software engineering) but there is a feeling that i hit a ceiling and cant go further

[raw bramble]

Who uses an architecture, if you know what you want to make then you jut make it

[vale wasp]

And the level of experience. And the complexity of the task.

[granite wyvern]

There’s a tutorial on their website

[raw bramble]

And if it doesn’t work the. You think about how it should work

[raw bramble]

And if that does work, you go back to part 1

[opal gull]

but i think ideally, you should get something working first to know what abstractions you need and don't need

[frigid coral]

ok i'll check it out

[raw bramble]

And you cycle between the 2 types of thinking till it all works

[opal gull]

basically the way the code is structured

every software has an architecture

[raw bramble]

I bet my software doesn’t

[opal gull]

yeah like for some tasks the architecture is already well known

like rest apis

[opal gull]

it definitely does

just that maybe it isn't good

spaghetti code is still, in a way, an architecture

[raw bramble]

Oh I hate spaghetti code

All classes at the start, unless they’re not relevant till way later on

Functions should try to be in the order in which they will be used

[vale wasp]

That's organization, not architecture.

[opal gull]

i usually follow the "newspaper article" pattern, where the important ideas are at the top, and the more concrete details go at the bottom

but yeah that is just organization, not architecture

[raw bramble]

Idk what you mean by architecture then

[vale wasp]

Designing APIs and interactions.

[granite wyvern]

Kind of the hierachy of classes or structures or layers

Not the lexical layout of things in your code as such

[raw bramble]

I avoid subclasses because I don't like them

I guess that's it

[granite wyvern]

Easily done. Provides you ignore that fact the every one of your classes subclasses object.

[opal gull]

composition over inheritance!

[vale wasp]

Inheritance can be useful.

[opal gull]

most of the time, i only want inheritance because i want shared helper stuff between subclasses
so i usually replace it with the context pattern

[granite wyvern]

I do a fair bit of inheritance, but deep stuff gets harder. Composition using mixins scales sideways fairly well. And composition using embedded things inside the main instance also works fairly well.

[raw bramble]

How do you do composition in Python?

[vale wasp]

There are times when inheritance makes sense. When it does I tend to use it. I've seen people go nuts with it though.

[vale wasp]

First you pick the key and tempo...

[raw bramble]

nice joke, but don't even start 😭 Honestly horrible advice there I have been given too many times

[coarse salmon]

When?

[vale wasp]

More seriously though, you try to figure out what each class needs and whether it should be a method or an external function.

[granite wyvern]

Can you elaborate a little on what "context pattern" means here?

[raw bramble]

Find what feeling you are trying to evoke from the listener, or what concept you are trying to embody, and work from there

[coarse salmon]

Dependency injection

[vale wasp]

I did a parser for cooklang. I created a base class for the Ingredient, Cookware, and Timing objects since they had enough in common that it made sense to have a base class for them.

[opal gull]

passing context to the objects basically

[coarse salmon]

Those seem completely different off the names lol

[granite wyvern]

I can imagine sharing counters, timing, units.

[raw bramble]

I need to start using python again and learn a whole lot more, but I don't have PyCharm or VSC rn 😭

[pallid garden]

just do it

[coarse salmon]

Yea

[granite wyvern]

Use whatever works. I just have an editor in one window and a shell in another to run stuff.

[vale wasp]

They are and they're not. Take a look.
https://github.com/brass75/cooklang-py/blob/main/src/cooklang_py/base_objects.py

[raw bramble]

I got really annoyed at Windows last night and just started uninstalling every single thing I don't frequently use, and today I've just been sorting out my documents to back up because I'm done with Windows

[raw bramble]

Today I used IDLE 😭

[vale wasp]

vim is your friend.

[raw bramble]

I use nvim but I don't have any plugins or anything at all, so it's just a fancy text editor for me rn 💔

[granite wyvern]

Mostly the same.

[raw bramble]

I've installed Lazyvim in the past but it's just so maximalist

And I don't really know how to install things for Neovim

[pallid garden]

so vim, but someone is playing roblox in the other room

[vale wasp]

There are LSPs for Python you can install. It's a reasonable code editor. I generally use it for smaller things, single file jobs.

[granite wyvern]

Yah ,trying ty as an LSP was what led me to try nvim.

[raw bramble]

roblox 💔

Roblox in the big 26 😔

To think prime Roblox was 10 years ago

[pallid garden]

i know it uses some fork of lua

[raw bramble]

It uses Luau

[crisp jay]

neovim

[opal gull]

luau is genuinely peak

[vale wasp]

Take a look at https://github.com/nvim-lua/kickstart.nvim @raw bramble @granite wyvern - it's a good starting point for configuring neovim.

[raw bramble]

scroll up

[vale wasp]

lua is a programming language. Luau is a polynesian party.

[raw bramble]

https://luau.org/

Ooh they updated their website

[vale wasp]

TIL.

[crisp jay]

you should use the package manager that Lazyvim uses (lazy.nvim),and start configuring in ~/.config/nvim

[raw bramble]

I really don't want to just download tools and start using them without understanding them

[crisp jay]

wdym

[raw bramble]

I don't really understand how Neovim works, I downloaded it and use it just for the modes. I barely know any commands other than the basic editing ones I need. I want to avoid installing a package manager or anything because all it'll do is introduce me to way too many plugins way too fast which is not what I want

[vale wasp]

neovim is a more modern version of vim. That's it.

.topic

[verbal wedge]

**What's your favorite PEP?**

Suggest more topics here!

[vale wasp]

!pep 810 is one I'm so looking forward to!

[edgy kraken]

**PEP 810 - Explicit lazy imports**

Status

Accepted

Python Version

3.15

Created

02-Oct-2025

Type

Standards Track

[jade robin]

how does that help you

[fiery yarrow]

505 695 and 701 gang represent

[crisp jay]

i think you should try out vim first then

[vale wasp]

I do a lot of CLI tools. It means that I don't need to import everything that all the commands need, I just need to import what the command run needs.

[fiery yarrow]

wonder how many years multi-pep support for this command has been suggested

[vale wasp]

For !pep?

[fiery yarrow]

yes

[crisp jay]

!pep 1

[edgy kraken]

**PEP 1 - PEP Purpose and Guidelines**

Status

Active

Created

13-Jun-2000

Type

Process

[fiery yarrow]

especially as this topic tends to flood the channel fairly quickly with !pep invocations

[crisp jay]

pep is a kinda cool thing,better than the C++ ISO

[jade robin]

yeah but how do cutting down on imports help you here, if the modules do take long to load, why not just import inside the function and do a check if it already exists before it?

[crisp jay]

that would be way too much boilerplate

[vale wasp]

I can do that. And there are places I might consider doing that. But this is letting the system handle it properly.

[bleak moth]

Yeah, I import everything I need at the beginning of the file.

[vale wasp]

Lazy imports should also help with circular import issues.

[jade robin]

yeah that was the only point it'll help me i think

the literal only library i've loading time issues with is pytorch

[pallid garden]

tfw when #ifndef INCLUDED_FOO_HPP

[jade robin]

the feeling when when

[vale wasp]

For some things it will probably help with less than half a second of improvement, but that's still faster.

[crisp jay]

is there a way to make 2 different modules that import one thing in common be imported only once?

[vale wasp]

I don't think so. I don't think it would be desirable either.

[crisp jay]

yea that would be confusing to work on

[subtle coyote]

Why is 2 different modules having a thing in common anyway 😭

[vale wasp]

I think it doesn't parse more than once but it creates objects for each module that imports it, but I'm not 100 sure.

[subtle coyote]

Just make a separate module

[pseudo coyote]

What does parse mean

[vale wasp]

This is the case of 2 files having import foo

[vale wasp]

In this case it means processing the file and converting it to Python objects.

[subtle coyote]

Oh so just write import foo in those two files no? Create different objects?

Or do you mean to make a singleton?

[vale wasp]

TBH I'm not sure how it handles it. I'd have to look at the relevant PEPs and the source code.

[subtle coyote]

Ping me pls my phone's bout to die

[whole snow]

hey can anyone guide through to learn Python . where do i start and how can i speed up BY using AI.

[vale wasp]

!res

[edgy kraken]

Resources

The Resources page on our website contains a list of hand-selected learning resources that we regularly recommend to both beginners and experts.

[subtle coyote]

You don't speed up by using ai when learning

[fiery yarrow]

using llms or any other type of ai is considered very dangerous for newbies like yourself. they're prone to lying, and you're not knowledgeable enough yet to be able to tell when that happens

[subtle coyote]

You learn by struggling

[crisp jay]

so my main.py have a object that is passed by to other modules (files),they need the same tools but for different jobs(ex: import an exception of the object)

[heady basin]

why are keyword only paramteres not allowed after * parameters?

def foo(*bar, *, baz):
    pass

foo(1, 2, 3, baz=4)``` this wouldnt break anything, no?

[ionic bluff]

lol im gonna be real strong then

[vale wasp]

Module 1 has an object that you want modules 2 and 3 to both access and modify?

[vale wasp]

In the function header or in the call?

[subtle coyote]

Isn't that invalid python

Does it work?

[heady basin]

yes its gonna error at function definition

but I dont understand why that rule is in place, it can just work, no?

[rugged viper]

if your question is about *, that does work

[vale wasp]

If you remove the , *, does that fix it?

[subtle coyote]

You can't have a bare *, after ```py
*args

[heady basin]

oh youre right, that would be the exact same

[rugged viper]

yeah

[crisp jay]

yea,but module 2 and 3 does different things that making them one wouldnt be very cool to work on

[subtle coyote]

So you're wrong it doesn't work lol

[rugged viper]

also just use backslashes:

"You can't have a bare *, after *args"

[rugged viper]

??? did you even read what i said

i meant *, is valid syntax, just not in that context.

[subtle coyote]

U said that does work no?

Oh

[rugged viper]

*, <-- this works

[vale wasp]

I think if you have a global and then a function that yields that global in an infinite loop it might do what you want? But I'd have to play around with it.

[subtle coyote]

Not after

[rugged viper]

*bar, *, baz doesnt work

[subtle coyote]

*args tho

[crisp jay]

nah

[vale wasp]

#bot-commands message it does work as def foo(*bar, baz=None)

.topic

[verbal wedge]

**In what ways has Python Discord helped you with Python?**

Suggest more topics here!

[vale wasp]

It has shown me new perspectives on things. I have learned some nuances (and things) that I didn't know before.

[granite wyvern]

The Mac pyobjc libraries seem to take forever.

[granite wyvern]

The one thing in common is only loaded once.

[crisp jay]

i might make a dumb question but is it possible to optimize python bytecode when compiling to it?

like what c/c++ compilers does

[vale wasp]

There has been a lot of work that is going into that. Including more specialized bytecode and the JIT.

[silver plover]

That's behind the bytecode, not optimizing the bytecode itself

Like, as a user, rewriting the bytecode won't get your far

[vale wasp]

It is both... The specialized bytecode helps simplify and make less things.

Take a look at this talk from PyCon 2025 which goes into it somewhat.
https://youtu.be/NE-Oq8I3X_w?si=yQ_ve3tGPF9CahVI

[silver plover]

The point I'm making is that: as a user, you can't 'hand write' optimized code

[vale wasp]

Definitely.

[silver plover]

And that specialization is not at compile time, there's no way to optimize early... afaik with how specialization is instrumented

[vale wasp]

I know. I'm just saying that the core devs have been working hard on performance improvements for everything.

[silver plover]

Yup, I'm just being literal about their question

[vale wasp]

If you understand how certain things are interpreted into the bytecode you might be able to do some things that will result in more optimal bytecode, but I don't think there is a flag you can set (other than to turn the JIT on) that will do optimizations.

[silver plover]

That was one of the things I had wondered about, how to analyze / profile specialization

[vale wasp]

Damned if I know. I just know it should be possible.

[silver plover]

Last time I asked, this is what kenjin said: #python-discussion message

[vale wasp]

Nice!

[pallid tiger]

Ah this logo is better then the previous one

[dry pike]

this is the default one

[pallid garden]

not python: is it possible to allow connection to a postgres for only a specific docker container?

[ashen cipher]

mandatory usecase?

[pallid garden]

i dont understand

[ashen cipher]

what's your usecase

[pallid garden]

to use psycopg2 to connect to it from the other container

[ashen cipher]

can you not use them in the same container?

[pallid garden]

well, it's a proof of concept

im trying to mimic a network of servers

[pallid tiger]

There was a blue and yellow one also

[dry pike]

pretty sure this is the default one for the discord server

[ashen cipher]

is there anything wrong with just making a user/password for the container

[pallid garden]

and expose to 0.0.0.0?

[ashen cipher]

yeah

[pallid garden]

seems like a bad network security idea

[ashen cipher]

postgres has oauth

erm

[pallid garden]

im just going to do a subnet

[peak relic]

And expose port 5432 if you haven't changed the port yet

[pallid garden]

docker ps
9273fca604f5 toy-darts-omop-db "postgres -D /etc/po…" 3 minutes ago Up 3 minutes 0.0.0.0:5432->5432/tcp, [::]:5432->5432/tcp toy-darts-omop-db-1

i also added host omop postgres 172.18.0.0/16 md5 to pg_hba

but it's not working

[peak relic]

Where are you accessing the postgres container from? The same machine or over the network?

[pallid garden]

another docker container

on the same subnet

[peak relic]

Not in a docker-compose.yml setup?

[pallid garden]

the docker bridge is working, im getting (172.18.0.3), port 5432: Connection refused

[pallid garden]

it is a compose setup

[peak relic]

If its in a compose setup, then they should already be in the same default docker network. Probably don't need to edit the pg_hba.conf. Just pass the usual postgres credentials

example docker-compose.yml

version: '3.8'
services:
  postgres:
    image: postgres:16
    environment:
      POSTGRES_DB: mydb
      POSTGRES_USER: myuser
      POSTGRES_PASSWORD: mypass
    ports:
      - "5432:5432"  # Optional: for host access

  app:
    image: myapp:latest  # Or build: .
    depends_on:
      - postgres
    environment:
      DB_HOST: postgres  # Use service name as hostname
      DB_PORT: 5432
      DB_NAME: mydb
      DB_USER: myuser
      DB_PASS: mypass

[pallid garden]

i have

    networks:
      - testing-subnet
```on both

docker network inspect also confirms that both are on the same subnet

[peak relic]

hmm

[peak relic]

Can you share your docker compose file?

[pallid garden]

hold on imma push it

https://github.com/zehata/Toy-DARts/blob/main/compose.yml#L35-L75

i can confirm that i have EXPOSE 5432 on the db dockerfile

oh wait

let me try to add the -i and -h

[peak relic]

How do you currently connect from your ingress to db?

[pallid garden]

give me a moment im just quickly trying this

perfect

turns out just having it inside the pg_hba.conf is not sufficient

you need the -i as well

sorry about that friends

[charred tusk]

That doesn’t do anything

EXPOSE “advertises” what port is being sued
But if you don’t add the correct --port flag to docker run then it doesn’t actually expose anything.

[pallid garden]

(it worked though?)

[charred tusk]

[pallid garden]

strange, right?

[charred tusk]

It’s not supposed to

[pallid garden]

docker is just a jester like that

[charred tusk]

Did you define the port in your args or your compose.yaml?

[pallid garden]

this is the CMD: CMD ["postgres", "-i", "-D", "/etc/postgresql/16/main/"]

and compose.yaml is unchanged from the current committed file on github

to up it im just running sudo docker compose up <db-name>

[charred tusk]

https://github.com/zehata/Toy-DARts/blob/main/compose.yml#L72-L73

[edgy kraken]

compose.yml lines 72 to 73

ports:
  - "5432:5432"```

[charred tusk]

This does the actual binding

[pallid garden]

i see, so there is no need to

expose:
  - 5432

TIL

[charred tusk]

Not sure
I didn’t even know that could go there

[charred tusk]

So imagine you have 8 different Dockerfile(s) that all run their app on port 8080.

In the first you could put ports: 9001:8080
Then in the second put ports: 9002:8080
And so on and so forth

The ports mapping is where you tell Docker what port to get from the host network.

[pallid garden]

so i would go to localhost:9001 to access the first app?

[charred tusk]

Yea

This is just metadata for documentation
https://docs.docker.com/reference/dockerfile/#expose

This is woefully vague and I still don’t understand what it actually does
https://docs.docker.com/reference/compose-file/services/#expose

[stray field]

So EXPOSE is like MAINTAINER

Completely useless

[modest obsidian]

How do I make a crypto mixer in python

[ionic bluff]

you write the code to perform the tasks that a "crypto mixer" would do.

[fallen locust]

Hi, I’m learning Python with 100 Days of Code. Happy to join 👋

[ionic bluff]

Hi! welcome aboard 🙂

[modest obsidian]

This doesn’t help

?

You can’t speak proper english

Refrain from speaking to me mate

[ionic bluff]

why do i get the feeling this is the same guy talking to himself?

[modest obsidian]

Mixers aren’t illegal

do your research before talking

Who are you again

Ur an attention seeker move

Nobody cares about you this enough

<@&831776746206265384>

[visual juniper]

what even is a crypto mixer

[modest obsidian]

Yes boss, we can talk once you learn proper English

Now stop wasting my time

[ionic bluff]

I think he meant "miner"

[modest obsidian]

Google exists

[modest obsidian]

Mixer

Lol this is not what a mixer is but okay

[fallen locust]

Anyone here also doing 100 Days of Code? Would love an accountability buddy.

[modest obsidian]

Mixers are used to obscure the transaction origins, it’s used for privacy matters

You obviously don’t know what you’re talking about

[visual juniper]

i wanted a TLDR and tell u if its possible or not
seeing that u want to dont give a TLDR , good luck

<@&831776746206265384>

[modest obsidian]

Mixers aren’t illegal in neither the US or EU if no illicit funds are passed through it

[modest obsidian]

He’s just a troll don’t mind him

[visual juniper]

i can tell

we need more mods with eastern timezones

[ionic bluff]

yeah lol

[modest obsidian]

7 moderators online, not one in the chat rn

[fallen locust]

Hi 👋
I’m currently on Day 3 of 100 Days of Code.
I’m serious about building strong Python foundations and moving into AI long-term.
Looking for an accountability partner at a similar level to grow consistently.
Daily check-ins + small weekly projects.
Anyone interested?

[modest obsidian]

Can you please stop coping with the fact that you’re poor @silent spindle

We don’t want to hear it

you’ve never touched 4f in your life please stop

lmfao

[visual juniper]

its hard to find partners like that , 99.99% of the time it just doesnt work out
its better to just start on your own , you can always come here to hang out and chat
dont wait on somenoe

[kind thicket]

<@&831776746206265384> ableist troll

[modest obsidian]

Can the moderators get on already?

[jagged belfry]

!mute 1462338246775148691

[edgy kraken]

:incoming_envelope: :ok_hand: applied timeout to @silent spindle until <t:1772348203:f> (1 hour).

[modest obsidian]

👍

[modest obsidian]

Also I’m a little confused what is this supposed to mean

What does TLDR stand for

[jagged belfry]

!cleanban 1462338246775148691 troll

[visual juniper]

#python-discussion message

[edgy kraken]

:incoming_envelope: :ok_hand: applied ban to @silent spindle permanently.

[modest obsidian]

Oh I explained it

[modest obsidian]

^^

[fallen locust]

Thanks, I appreciate your advice. 🙏
I’ll keep building consistently.

[velvet trout]

class Hub:
    @staticmethod
    def create(
        kind: type[EventA] | type[EventB],
    ) -> Ticket[EventA] | Ticket[Bundle]:
        return Ticket[?????]()

        #if kind is EventA:
        #    return Ticket[EventA]()
        #else:
        #    return Ticket[Bundle]()

Any idea what to put in it? The comment version works but i want not to repeat.

Ticket is Generic over T, where T is EventA | Bundle.

[marsh obsidian]

Guys i won my first high-school microsoft hackathon today.

I am 13 years old

I got 300 dolllars and a trip to seattle for the national competition

[visual juniper]

thats very good
congratulations 👏

[quartz fulcrum]

well done

[kind thicket]

congrats!
What did you do to win?

and why was it a python program?

[marsh obsidian]

THere were two categores: Health, and Finance. I built a ai powered public speaking app that helps you get better.

[kind thicket]

nice!

[pallid garden]

wait, which category does that fall into?

[marsh obsidian]

and no, it wasn't python. It was SwiftUI and CoreML, and ARKit

[marsh obsidian]

fell into the subisdary of health (solving a problem faced by many)

[kind thicket]

not python discussions some would say

[marsh obsidian]

yea sorry

[marsh obsidian]

just wanted to share the achievement

[kind thicket]

np. You are happy to have won. It's fair

[quartz fulcrum]

np...its not an easy thing to achieve anything

[velvet trout]

A 13 year old won a 300 dollars price and a trip from a Microsoft hackathon, meanwhile my 45 years old brain getting tortured by some generics/typing in Python.

😭 👍

I am happy for you.

[marsh obsidian]

Thank you all!

[ashen cipher]

okay my brain is confused i need someone sane

how many elements are in my_list[1:2]

[quartz fulcrum]

its ok Aji,
everyone has different path with different achievements

[quartz fulcrum]

1?

[ashen cipher]

alr ty

[velvet trout]

Exactly 1 item from index 1

end is exclusive, meaning it won't include the item of end index in the slice

[ashen cipher]

look its midnight my brain doesnt process past first order thinking

[quartz fulcrum]

remember, just don't include the last one

[velvet trout]

relatable i am running wild over typing in python.

[ashen cipher]

i could probably work that out but like i dont have the mental capacity at 12:08

[pallid garden]

set your system clock to 10am, spoof your biological clock

[ashen cipher]

lmao

[quartz fulcrum]

are you following 'If you need a good sleep, then stay awake'?

[ashen cipher]

i am following "im going to sleep later"

[velvet trout]

Anyone?
Wait should i post this in #1035199133436354600 ?

[ashen cipher]

i forget python has generics now

it is way better than typevar or whatever used to happen but i just dont write much python now

[cyan scaffold]

The comment version is perfect. Here's a shorter version if you're still set about sacrificing readability:

return (Ticket[EventA] if kind is EventA else Ticket[Bundle])()

[velvet trout]

mypy didn't say there was an issue, when i put x = Ticket() reveal_type(x) it says need type annotation... What was bro doing before ?

[velvet trout]

Thanks. I also used overloads on create func. It seems like it just thinks return Ticket() is Ticket[Any] ? Is that healthy?

[tacit tapir]

anyone else love walrus

[velvet trout]

Overload helps to clarify that passed EventA/B returns Ticket[EventA/Bundle]

[velvet trout]

Me.

I love [] type parameterization syntax too

feels like am a hacker or something lmao

[white knot]

hi

[velvet trout]

import hello

[white knot]

print("hello")

input("how are u")

[tacit tapir]

you gotta speak more pythonicly i have no idea what this means

[velvet trout]

Just got fine, thanks to my previous replies. How about you?

[fiery yarrow]

!pep 695

[edgy kraken]

**PEP 695 - Type Parameter Syntax**

Status

Final

Python Version

3.12

Created

15-Jun-2022

Type

Standards Track

[fiery yarrow]

way to express generics

[velvet trout]

Previous repliers*

[cyan scaffold]

That doesn't sound right. I feel like it should infer the usage and the parameter types and return type as you pass in arguments

[velvet trout]

My bad, i vaguely remember the term for it

[tacit tapir]

oh i see i just didnt know thats what it was called

[white knot]

MemoryError

[cyan scaffold]

What is "it" that thinks it's Ticket[Any]?

[tacit tapir]

nah walrus is goated tho that was like the first new feature that got added when i started using python and i use it all the time

[white knot]

;=

why is it the goat

[tacit tapir]

cuz its useful frequently

[cyan scaffold]

Oh never mind Ticket() should just be Ticket[Any]

[white knot]

i see why cant we use a extra line

which is more pythonic

[velvet trout]

Opps. Wait, i actually ran this:

class Hub:
    @overload
    @staticmethod
    def create(kind: type[EventA]) -> Ticket[EventA]: ...
    @overload
    @staticmethod
    def create(kind: type[EventB]) -> Ticket[Bundle]: ...

    @staticmethod
    def create(
        kind: type[EventA] | type[EventB],
    ) -> Ticket[EventA] | Ticket[Bundle]:
        t = Ticket()
        return t
        #if kind is EventA:
        #    return Ticket[EventA]()
        #else:
        #    return Ticket[Bundle]()

t1  = Hub.create(EventA)
t2 = Hub.create(EventB)
reveal_type(t1)
reveal_type(t2)

x.py:46: error: Need type annotation for "t"  [var-annotated]
x.py:47: note: Revealed type is "x.Ticket[Any]"
x.py:56: note: Revealed type is "x.Ticket[x.EventA]"
x.py:57: note: Revealed type is "x.Ticket[x.Bundle]"

Is that healthy?

[white knot]

what does overload do. python has method overloading?

[golden mortar]

I wouldn't say that adding extra lines is Pythonic necessarily

[tacit tapir]

i dont rly think using the extra line is more pythonic

[velvet trout]

Python 3.8 — late 2019

[golden mortar]

It's just for the type checker

[cyan scaffold]

This is expected... You should use the commented version

[velvet trout]

It's for functions whose type of parameter changes the return type.

[cyan scaffold]

Else why did you care about kind

[velvet trout]

I actually don't care about kind, it only affects generic T of Ticket and i want to statically type that Ticket[T] will have different T type depending upon the kind's type, the comment version was used due to earlier type errors and was working.

Is the current code healthy and should be like that given I don't care about kind but for type checking only?

[icy quest]

hi guys

has anyone hosted a website with premium features?

[white knot]

why is my algorithm analysis 50hr course finishing in like 30 pages of notes

all the topics feel very small

like this is one module for example The Control Abstraction- The Optimality Principle- Matrix Chain Multiplication-Analysis, All Pairs Shortest Path Algorithm - Floyd-Warshall Algorithm-Analysis. The Control Abstraction of Back Tracking – The N Queen’s Problem. Branch and Bound Algorithm for Travelling Salesman Problem.

thats 4 algorithms and chapter is over

am i not learning properly

[burnt cargo]

I can't speak L

[white knot]

can you explain? you mean a SAAS product?

[icy quest]

yess

i wanna host my project with free and premium mode

[white knot]

What does your project do for premium

You already have a website made?

[icy quest]

real-time collaborative coding

[velvet trout]

class Hub:
    @overload
    @staticmethod
    def create(kind: type[EventA]) -> Ticket[EventA]: ...
    @overload
    @staticmethod
    def create(kind: type[EventB]) -> Ticket[Bundle]: ...

    @staticmethod
    def create(
        kind: type[EventA] | type[EventB],
        #x: int
    ) -> Ticket[EventA] | Ticket[Bundle]:
        return Ticket(x)

Well, i actually want to introduce more non related params like x
But mypy yells:

error: Overloaded function implementation does not accept all possible arguments of signature 1  [misc]
error: Overloaded function implementation does not accept all possible arguments of signature 2  [misc]

[icy quest]

w interview, etc features

[icy quest]

yep

[white knot]

Nice so you are searching for features that you can make premium?

[velvet trout]

I don't want to spam x every func overload when its not related to typing overloads

[cyan scaffold]

It's healthy as far as telling you that your types don't match up, because Ticket() is of type Ticket[Any], which is not the expected return type Ticket[EventA] | Ticket[Bundle]

It wouldn't tell you this if the return type and t's type were switched, but why not just match them to the narrow Ticket[EventA] | Ticket[Bundle] from the start?

[icy quest]

no ik im just asking is it worth? cuz will people even use it or not? if someone has done it what were their experience when they got users?

[velvet trout]

Makes sense but it repeats the code in if-else branches

[cyan scaffold]

Well you need one overload with x: int in its signature otherwise the implementation accepts x: int for no reason

[velvet trout]

Should i just go union over Generic ?

[velvet trout]

I see

[velvet trout]

Like Ticket.value: list[EventA | Bundle] = []?

I am kind of unsure when to use Generics and Unions

[cyan scaffold]

You could do

return Ticket[EventA if kind is EventA else Bundle]()
``` but idk if the type checker gets this right

[solemn rain]

Hi

I have one question

[visual juniper]

nice

[solemn rain]

Can I ask

[visual juniper]

you already know that you shouldnt ask to ask

you have been here long enough to know to just ask the question

[solemn rain]

i have a zip file now if i share that with my friend who uses windows
The packages inside site package or node modules won't work properly ?
Because i am using Ubuntu

[visual juniper]

send them the source code and have them install depedencies

[solemn rain]

I did try to make them clone the repository
But they don't have a git account and are very new to software development

They just wanted to edit some frontend html css

of the project

[visual juniper]

you can share your folder with google drive or any other method
just dont send your site pacakges etc in it

you can also share just the files they want to edit

[solemn rain]

And they send me the edited file back?

[visual juniper]

mhm

[velvet trout]

Expressions are not allowed inside []

I am thinking of shifting to not using Generic

I am spending too much time on fixing this generic

[visual juniper]

also why did u care if the site packages etc work or not if they just wanted to edit some HTML CSS ?

[velvet trout]

And i am not sure what's the correct path

[solemn rain]

I am using react and they are going to edit tailwind I think

They said they want to edit styles

[visual juniper]

so they want to run your app ?

[solemn rain]

They wanted to contribute to the project

[visual juniper]

setup a github repo , have them make a github account , tell them to learn how to use git

i thought this was like , you were selling ur client something

[solemn rain]

No its a group project

[visual juniper]

well , you have multiple options
base line is that , you send just the source code , and the other user sets it up on their end and developes

[short solstice]

Check out this extension https://marketplace.visualstudio.com/items?itemName=MS-vsliveshare.vsliveshare

You can share your current project and terminal, and assign permissions so he can edit files and access terminals, etc.

[visual juniper]

you dont send node modules and python modules and stuff liek that

[icy quest]

who sends node modules

they are a lot

[velvet trout]

Playing Visual Studio Code

[pallid garden]

for 80 days

[quartz fulcrum]

why there is a red line on the whole function and still working but says 'Explicit return statement expected'

[golden mortar]

Python doesn't require explicit return statements, a function without a return statement will return None, but it's a good practice to include one for clarity.

[lament bobcat]

Is there a self promo section on this server?

[gleaming knoll]

there is #1468524576479641744 but its not for literal promotion

[craggy willow]

it's for promotion literals

[gleaming knoll]

p-string

makes your string bold and annoying

[craggy willow]

p-string
https://www.youtube.com/watch?v=kF1W26OUAy0

ppppppppp

[quartz fulcrum]

lol

i was learning about functions calls(call stack) and i think i got the basic i.e. python doesn't forget its top function call.
if there is a nested function inside the function call then python will call that function, does whatever is written in it and comes back the main function..

but they also talked about reuturns, which is not clicking in my mind..can anyone help me with this

[craggy willow]

return leaves a function immediately, back to the surrounding scope

with the value after the return

[gleaming knoll]

(to the place where the call was, which is remembered by putting its address on the stack)

[craggy willow]

!e

def main():
    def foo():
        if True:
            return 1
        return 2
    a = foo()
    print(a)
main()

[edgy kraken]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

1

[craggy willow]

return 2 is never reached in above

[quartz fulcrum]

and what about def foo().
it has no data

[gleaming knoll]

no parameters, you mean? how does that change anything?

[quartz fulcrum]

no code

yes

[gleaming knoll]

def foo():
    if True:
        return 1
    return 2

is this "no code"?

[raw bramble]

pass, continue and return all have that same kind of functionality, right?

[gleaming knoll]

pass doesn't do anything to control flow, continue goes to the loop header, return goes to the call address on the stack

[craggy willow]

gotta have the indentity control flow

[quartz fulcrum]

oh my bad

[bronze dragon]

they all do different things

[quartz fulcrum]

where the return 2 stands

[raw bramble]

Yeah, but I mean the sort of jumping out of their scope

I don’t know proper terminology

[gleaming knoll]

continue stays in the same scope, the loop scope
well, python loops dont have name scopes, but you get the point, the loop.. block

[raw bramble]

Mhm

[gleaming knoll]

return is specific in that regard as the place it goes to depends on who called the function, not the return itself
that is why the address it goes to is passed in the program as data on the stack

[raw bramble]

How can you use pass outside of just making sure your code doesn’t break when you’ve not implemented something yet like a function?

[gleaming knoll]

you cant, it does not do anything, it is simply a blank statement to put where a statement is expected when you dont have anything better

[raw bramble]

so it’s only purpose is just to stop my code from breaking till I have the energy to finish it

🙂

[craggy willow]

there's some interesting uses of pass in if-elif-branches

if a:
    pass
elif b:
    do_something()
elif c:
    ...

something like

[granite wyvern]

It's a placeholder which does nothing where syntacticly you need a statement.

[gleaming knoll]

i mean, pass is not doing anything specific here, it could be any statement, including like.. 0

[raw bramble]

What does this do? Why not just have if b elif c?

[granite wyvern]

To only test that if a is false.

[gleaming knoll]

if/elif are mutually exclusive, so if if a ran other stuff wont

[craggy willow]

sure, but it's also a more interesting use case of "do nothing" here

[raw bramble]

oh yeah I know that 🧠🧠

[gleaming knoll]

this is indeed a neat trick to reduce 1 indentation from

if not a:
  if b:
    do_something()
  elif c:
    ...

[granite wyvern]

Without it the following conditions with if if not a and b and if not a and c. And so on. CUmbersome. I might be belabouring the obvious.

[craggy willow]

i used to use something like this for parsing ansi, but not anymore

[raw bramble]

I do this sort of thing everywhere because I hate having long if statements

if (a and not b) or (c and not d)
Just looks horrible

[fossil steeple]

never nester ah code

i just use dicts to avoid lots of elifs

[gleaming knoll]

how would you use dicts in this case?

[quiet shard]

is it possible to make a python interpreter using python

[visual juniper]

yes

[cobalt flax]

Isnt pass just for not doing anything?

[visual juniper]

yes

[gleaming knoll]

kind-of, but the more things you actually reimplement rather than just reuse, the slower it will be
(specially the object model)

and well you'll still need a python interpreter to run your python interpreter written in python, so it wont be self-sufficient, so if thats a goal then it doesnt really work out

[velvet trout]

Eat 5star, do nothing.

[pseudo coyote]

What are classes used for in short

[quiet shard]

to create objects

[slow rivet]

Make grouping code and data together easier

[pseudo coyote]

What init

[quiet shard]

constructor i think (i don't use python)

[nimble isle]

a = []
b = a.copy()

Is this equivalent to ```py
a = []
b = []
b[:] = a

[fossil steeple]

what's the case now

[fossil steeple]

it's like called when you make the class

[pseudo coyote]

Thanks

[fossil steeple]

it like defines how you should make the instance and like what vars it should have and stuffs

[fossil steeple]

class Foo:
  def __init__(self, name, age):
    self.name = name
    self.age = age

john = Foo("John Doe", 22)

like this

[gleaming knoll]

the result is the same, yes

[nimble isle]

do they run in equivalent time?

[pseudo coyote]

What is self used for then?

[solemn rain]

Run and find out

[nimble isle]

....how

[fossil steeple]

the way i see self is you're referring to the object, so like you're basically doing john.age in that example

[quiet shard]

what does : do

[nimble isle]

basically the same as [0:-1]

it's a slice

but it slices the whole list

[solemn rain]

Run the file

[nimble isle]

but how do i tell the time complexity from that

[fossil steeple]

@pseudo coyote check bot commands

[pseudo coyote]

Thanks

[solemn rain]

You can see the execution speed

[gleaming knoll]

the time complexity is obviously the same, but the exact time it takes will be slightly different (just because they go through slightly different code paths to end up doing the same copying loop), but it doesnt matter
imo copy is more direct to understand so use that
or atleast b = a[:] rather than empty initialize and then slice assign

[nimble isle]

...?

[quiet shard]

forget what i said because its not python

[fossil steeple]

slices the yoke

<@&831776746206265384> @wise yarrow be messing

[wise yarrow]

!e not that, you can't really describe None for the stop argument with an integer in a way that works for any list, tho for fixed sizes the length works

a = [1, 2, 3]
print(a[:], a[0:-1], a[0:3])

[edgy kraken]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

[1, 2, 3] [1, 2] [1, 2, 3]

[wise yarrow]

?

I haven't even typed anything yet...

[gaunt badge]

What's going on, what do you mean?

[fossil steeple]

wait maybe it wasn't u then who did htat

[fossil steeple]

someone sent like a free sex discord link and deleted it

[gleaming knoll]

the bot deleted it

[gaunt badge]

Ah, that was someone else. Our bot deleted it and banned them

[fossil steeple]

mb maybe they had like a similar username

[gleaming knoll]

why would you think stickie do that 😭

[fossil steeple]

bc stickie was the only one typing lol

[solemn rain]

Poor stickie

[nimble isle]

can you use complex numbers as keys for a dictionary?

[gleaming knoll]

yes, why not? they are immutable, and consist of hashable parts (floats), so obviously they have a stable hash

[harsh swallow]

... Paste+send doesn't show as "typing"

[wise yarrow]

sure, i do it all the time for AoC problems

[fossil steeple]

that's prbly what happened then

[winged shale]

Python as first programming language and what second? Should i learn C (i will need to learn it anyway in school)

[nimble isle]

I recently discovered tuples can be used as dict keys

[wise yarrow]

anything hashable can be used as a dict key

[fossil steeple]

do you just have to type ur message as quickly as possible and then edit it like 5 times lol

[timid ember]

Sure

[gleaming knoll]

yes

[fossil steeple]

i looked out the windows for 3 seconds...

[wise yarrow]

and for stdlib types, if something is immutable it's probably hashable? I can't think of any counterexamples immediately at least

[pseudo coyote]

What does .append do

[nimble isle]

adds a value to the end of the list

[solemn rain]

appends to the list

[pseudo coyote]

Cool

How to round to 2 decimals

[fossil steeple]

round(num,2)

[solemn rain]

!d round

[edgy kraken]

round

round(number, ndigits=None)```
Return *number* rounded to *ndigits* precision after the decimal point. If *ndigits* is omitted or is `None`, it returns the nearest integer to its input.

For the built-in types supporting [`round()`](https://docs.python.org/3/library/functions.html#round), values are rounded to the closest multiple of 10 to the power minus *ndigits*; if two multiples are equally close, rounding is done toward the even choice (so, for example, both `round(0.5)` and `round(-0.5)` are `0`, and `round(1.5)` is `2`). Any integer value is valid for *ndigits* (positive, zero, or negative). The return value is an integer if *ndigits* is omitted or `None`. Otherwise, the return value has the same type as *number*.

For a general Python object `number`, `round` delegates to `number.__round__`.

[pseudo coyote]

Huh

I tried that earlier and it didn’t work

[winged shale]

Kinda off topic but is c, c++, c# almost the same? Like if i learn c will i be easily learn those?

[fossil steeple]

prbly did it wrong

[gleaming knoll]

no, they are not even remotely "almost the same", but they do have some overlap

[winged shale]

I see

[fossil steeple]

i believe c++ is a superset of c but i think c# is completley different

[winged shale]

Thanks for the info

[gleaming knoll]

(a non strict superset, there are loads of c code that wont get compiled by a c++ compiler and you dont even have to try hard to do that, just dont cast malloc for example, in C void* <-> T* is implicit unlike C++)

[fossil steeple]

yeah okay

[fossil steeple]

i was using arduino the other day and i couldn't tell if i was learning c or c++

[gleaming knoll]

arduino is C++ but you specifically might only be using C stuff

[fossil steeple]

but maybe i'm just mad

[wise yarrow]

there's some syntax similarities, but imo they write completely differently

[fossil steeple]

.topic

[verbal wedge]

**Have you ever tried making your own programming language?**

Suggest more topics here!

[harsh swallow]

You still have the float shenanigans happening in the background.
round(0.1, 2) + round(0.2, 2) will still be 0.30000000000000004 because even with rounding, those numbers can't be stored as their exact value

[pseudo coyote]

What’s the coolest sht you guys have made

[pseudo coyote]

What’s it like being neither genders

[solemn rain]

I made a advanced calculator that can do arithmetic operations

[pseudo coyote]

No hate

[harsh swallow]

That is offtopic for this channel. If you genuinely wish to talk about queer stuff, tag me in offtopic

[fossil steeple]

i made gloop which is a simple interpreter with 5 "gloops", which are:
print the current value
add a gloop to current value
add a gleep to current value
2 more which are same as above ubt instead remove

[pseudo coyote]

Sorry, but how’s that different to a normal calculator

[fossil steeple]

it's more advanced

[solemn rain]

Because its advanced

[pseudo coyote]

Cool

What’s an easy game to automate

[fossil steeple]

how does one automate a game

do u mean code?

[pseudo coyote]

Yh mb

[fossil steeple]

tic tac toe

[pseudo coyote]

Ooh

That’s a good idea

And I could use ai aswell to make him smart

It*

[fossil steeple]

good luck with

that

[pseudo coyote]

No I sleep now it’s 5am

[fossil steeple]

concerning

[solemn rain]

I am going to study travelling salesman problem

I saw a small video and i don't understand anything 😞

[graceful oracle]

Fellas what IDE/Code editor do yall use?

[craggy willow]

[junior wraith]

Hello. How do I use a third party library? I am new to python specifically.

[fossil steeple]

peak

[graceful oracle]

Using a package manager like pip

[gleaming knoll]

i was using vscode with the vscode-neovim extension, but the extension downloads became unreachable from russia so i just switched to using neovim

[opal gull]

vscode is genuinely so peak

[graceful oracle]

Yeaah I feel that CDNs are broken in russia. Same for me I use the Lunar Vim distro of NeoVim

[autumn forge]

Zed https://zed.dev

[graceful oracle]

All valid honestly

[young flare]

vscodium

[graceful oracle]

:0

[gleaming knoll]

i got most of what i'd want from a "neovim distro" from https://github.com/folke/snacks.nvim without actually giving up my config fully to it

[graceful oracle]

Ooooh thats nice!

Using Linux and Neovim taught me some best practices for python

[gaunt badge]

[graceful oracle]

Oh hell yeah

[opal gull]

not even neo?

[gaunt badge]

Nope, I don't see why I would switch

[craggy willow]

inb4 ed

[gleaming knoll]

who is that one Ed who edits everyone's text
how is he not tired

[junior wraith]

I tried ed with rust once and suffice to say it was not fun

[quiet shard]

aseprite text tool

[wise yarrow]

I used to use pycharm, but ever since I've setup neovim for python dev I've been way too used to neovim to switch back

[simple willow]

bit late bit I use IDLE

[junior wraith]

Anyways, I can't seem to escape vim (on termux)

[quiet shard]

:q

[wise yarrow]

my neovim navigation skills are lacking tho

[simple willow]

I did try neovim on my laptop and I broke it

[graceful oracle]

Interesting

[simple willow]

so I went back to IDLE

[graceful oracle]

Primeagen has a vim game I think

[graceful oracle]

Oooh

[wise yarrow]

I used IDLE for my first python project in school

I'm so glad I never have to use IDLE again

[simple willow]

I do really like IDLE and it does everything I need to do

[junior wraith]

The reason I'm considering hopping on python in the first place is because of the brevity

Compared to... c

[kind helm]

I am facing issues in running my python code in visual studio
Which environment to choose for

[golden mortar]

Have you tried Thonny?

[kind helm]

The program won't run
The terminal opens, shows the path stays there freezed

[golden mortar]

That's a good reason

[junior wraith]

I kinda want to have fun first before I go fast

[nimble isle]

I wish for loops allowed expression in their assignments: ```py
for x in range(5):
print(x)

0
1
2
3
4

But you could also do ```py
for x+1 in range(5):
  print(x)
>>> 1
2
3
4
5

[junior wraith]

The joy of actually having a finished piece of software

Does this server have a bot that runs python code?

[cerulean ravine]

that seems really odd. You also can't do x + 1 = 0

[nimble isle]

?

x + 1 = 0 -> x = -1

[golden mortar]

You don't necessarily have to choose. Python can be as fast as you need it to be for many real-world applications.

[cerulean ravine]

for loop is assignment on repeat, so for SOMETHING in .. works if SOMETHING = .. works.

[gleaming knoll]

this is not what your example does, by the way, its not solving for x in x+1, its adding 1 to the value in the iterable and that is what it assigns to x

[nimble isle]

well then for loops should solve algebraic equations

[golden mortar]

Yes

[cerulean ravine]

we can debate what the word "should" means there...

[golden mortar]

!e ```py
print("hello world")

[edgy kraken]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

hello world

[cyan scaffold]

Assignment isn't commutative

[golden mortar]

But only staff will see the output in this channel

Other people using it will have their output redirected to #bot-commands

[nimble isle]

for loops need to solve algebraic equations

[cyan scaffold]

No they don't need to solve algebraic equations

They need to loop

[nimble isle]

they need to solve algebraic equations such that no human has ever solved before

[cerulean ravine]

it alarms people, but you can do this:

nums = [1, 2, 3]
for nums[1] in range(3):
    print(nums)

[cyan scaffold]

Iterate over items specifically

[gleaming knoll]

this really just doesnt work out with more complicated expressions

[nimble isle]

they need to determine if every statement in this universe is true or false

[cerulean ravine]

you need to come back down to earth 🙂

[nimble isle]

and it just assigns the 2 to each of 0, 1 and 2?

[cyan scaffold]

Wait are walrus expressions valid on the left hand side of assignments?

[gleaming knoll]

nums[1] is the assignment target (the thing that will be assigned to), not the value that is assigned

[nimble isle]

every for loop must be able to determine if it will terminate or not

[nimble isle]

yeah I know

[cerulean ravine]

this does nums[1] = 0; nums[1] = 1; nums[1] = 2

[nimble isle]

yeah that's what I meant to say

[gleaming knoll]

no

[cyan scaffold]

Yeah I just checked

[cerulean ravine]

(x := 2) is 2, so (x := 2) = 3 is like 2 = 3, which doesn't make sense.

[craggy willow]

i've done something like:

for texture[:] in colors:
    display_texture(texture)

[cyan scaffold]

Right

[gleaming knoll]

if you apply the same logic to your previous example it doesnt work out
i mean, you get:
nums[1] is 2 so for nums[1] in range(3) is for 2 in range(3)

[cerulean ravine]

i've used for d["page"] in range(10):

[cerulean ravine]

true, it's difficult to explain briefly

[craggy willow]

using a dict is fun

[nimble isle]

would it be confusing for python to bring back the goto from basic?

[cyan scaffold]

No I think it uses the second item slot to keep all the iteration objects instead of in a new variable

[craggy willow]

it would be if it used line numbers like basic

[nimble isle]

if it used labels though?

[craggy willow]

!pypi goto

[edgy kraken]

goto v0.1.3

Urllib object wrapper

Released on <t:1320774534:D>.

[gleaming knoll]

yes, because this explanation is wrong, the assignment target is not evaluated, its a different kind of code

[craggy willow]

dammit

!pypi goto-statement

[edgy kraken]

goto-statement v1.2

A function decorator, that rewrites the bytecode, to enable goto in Python

Released on <t:1546111106:D>.

[cyan scaffold]

Oh okay yes I see

[craggy willow]

there we go

cursed python

[gleaming knoll]

goto that would go between scopes is cursed, goto in 1 scope is fine
problem is each assignment to a "new" local variable is like a new scope but not really

[nimble isle]

between scopes is confusing more than anything

[craggy willow]

there's been proposals for labeled breaks i think

!pep 3136

[edgy kraken]

**PEP 3136 - Labeled break and continue**

Status

Rejected

Python Version

3.1

Created

30-Jun-2007

Type

Standards Track

[gleaming knoll]

damn that pep is almost as old as i am. its like, 3 weeks short

[nimble isle]

I've occasionally wanted to break out of multiple loops at once.
But in those cases my code's been in a big mess anyway

[nimble isle]

it's older than I am

[cyan scaffold]

You're way younger than I thought you were

[craggy willow]

i was older than yall when this pep was written

[gleaming knoll]

was? are you no longer older than us? did bro stop aging?

[craggy willow]

😢

i was, but am still

[gleaming knoll]

salt die are your parents older than you

[craggy willow]

not yet

[cyan scaffold]

[nimble isle]

I know someone who's younger than their niece

[slow rivet]

Dice do not age like humans

[cyan scaffold]

Oh

Age preservation? Username does check out

[gleaming knoll]

"way younger"? how old did you think i am, like 30? i thought its quite obvious that im basically a kid

[quartz fulcrum]

can i ask a non python question here..

what i am doing is very cumbersome

[cyan scaffold]

You're so much wiser than me for basically a kid

Look at you FP master

[cyan scaffold]

If it has absolutely nothing to do with Python maybe one of the off topic channels will do

[slow rivet]

!ot not in this channel, but you can in the off topic ones

[edgy kraken]

Off-topic channel

#ot2-never-nester’s-nightmare

Please read our off-topic etiquette before participating in conversations.

[gleaming knoll]

i actually consider my wisdom stat to be insanely low im making stupid life decisions
i just read' a couple things about programming thats it

[cyan scaffold]

You're more impressive than you realize

[craggy willow]

so not a cleric or druid

[amber vine]

Do you know about antigravity?

[gleaming knoll]

can one even be called an fp master if they are not actually making any projects

[gleaming knoll]

sorry i dont know the basements and lizards class chart 😭

[craggy willow]

prbably not a monk either

[mossy sigil]

[craggy willow]

but wis is a dump stat everywhere else, so could be anything else

[gleaming knoll]

what class would diogenes be
i want to be that

[simple willow]

No

[cyan scaffold]

Do projects if you feel like it but I wouldn't say it's a prerequisite because you already prove yourself when talking about FP in this server

[craggy willow]

i feel like monk, but maybe, like, homeless rogue

[gleaming knoll]

FUCK
of course this joke has already been done
nothing ever happens
nothing new under the sun

[cyan scaffold]

The line is blurry 💀

[verbal citrus]

what's an FP master

[craggy willow]

well diogenes had like 22 wisdom

[opal gull]

i feel like KISS isn't emphasized enough

[cyan scaffold]

Formidable Python

[opal gull]

some people just like to spam design patterns

[verbal citrus]

[gleaming knoll]

its a title a person who spent time on bullshit makes up to feel better about themselves

[cyan scaffold]

I'm kidding it stands for functional programming master

[jade robin]

you haven't made a sentence factory for this sentence to exist

[jade robin]

nvm

[verbal citrus]

ah so kinda like those people who spend time speedkubing

[split onyx]

I think none of the software engineering acronyms make seense (e.g., KISS, YAGNI, ...) no one decides things in a vacuum, everything is situational 🙂

[opal gull]

clearly i have to inject an abstract ISentenceFactory

[opal gull]

the other end of the spectrum would be SOLID

most of the time, where do you want to be

[gleaming knoll]

at home?

oh sorry, wrong interpretation

[verbal citrus]

lol home is good

[opal gull]

me, personally, i prefer to be outside

but unfortunately coding outside is ehhh

[jade robin]

clearly a SOLID enthusiast with 99 years of experience

[gleaming knoll]

#ot0-psvm’s-eternal-disapproval message

[opal gull]

that's just the average java developer

[split onyx]

Personally my style (at work) is that I don't like making "big bets" on features or architecture. I push things out ASAP, get feedback and iterate. Eventually your abstractions will no longer fit the direction of your project and then you rewrite (some of) it

[opal gull]

that thing runs discord btw

[gleaming knoll]

a custom discord client written in grasst, the good brother of rust

[opal gull]

so, KISS then?

[solemn rain]

what is P NP NP HARD NP COMPLETE

[opal gull]

pretty much dont abstract until you need it

[gleaming knoll]

words and acronyms written in capital latin letters
what are*

[cyan scaffold]

Idk why but it reminded me of PPAP

[jade robin]

PPAD is a class too

[solemn rain]

i mean why should i categorize problems i don't even understand them

[split onyx]

Yes and no - sometimes you already know that a feature needs to be more complex than your baseline because of certain functional/non-functional requirements, and then you break "KISS" and "YAGNI"

[craggy willow]

tf is yagni

[golden mortar]

If you like the simplicity of IDLE, Thonny is also a Python IDE designed for simplicity, but in my opinion superior to IDLE.

[split onyx]

you aint gonna need it 🫠

[gleaming knoll]

i mean i think the https://en.wikipedia.org/wiki/P_versus_NP_problem page does a good job at explaining
but if you think this is just useless theoretical stuff then you are half right, it wont help you solve any of your problems, it will just give you insights on what kind of problem you're dealing with computationally

[opal gull]

but couldn't you argue that you do "need it" then?

[split onyx]

In a way yes

[cerulean ravine]

where did you read about P vs NP?

[split onyx]

But yeah, I think a lot of our profession is about predicting the future (with the abstractions we use). I like predicting nothing and then doing rewrites when needed

but your organisational culture needs to allow for this ofc

[golden mortar]

You keep it as simple and stupid as is possible given the available information, basically.

[opal gull]

that's true

[split onyx]

In some places rewriting is an absolute no-go, then I'd do things differently

[junior wraith]

I want to take an input image (.png), split it into a grid, output the images in the grids. Not sure where to start

[solemn rain]

my university syllabus

[cyan scaffold]

So you want sub-pictures?

[slow rivet]

I generally like to do a mix where the architecture is modular and flexible enough to avoid big rewrites, but the actual implementations are as simple as they can be

[golden mortar]

These concepts usually come up as part of an introductory course on the theory of computation at uni, and it's a lot easier to explain (P, NP, NP-hard, NP-complete) if you have the foundations from that course.

[solemn rain]

alright

An answer to the P versus NP question would determine whether problems that can be verified in polynomial time can also be solved in polynomial time. If P ≠ NP, which is widely believed, it would mean that there are problems in NP that are harder to compute than to verify: they could not be solved in polynomial time, but the answer could be verified in polynomial time.

whats the point of verifying output and categorizing them as NP.

The problem has been called the most important open problem in computer science

what problem NP == P ? i cant get that

and what is this NP hard

[split onyx]

In "corporate programming" this is unavoidable because eventually someone will bring features that go against the assumptions that your abstractions make

[solemn rain]

we already had TOC this is not TOC

[golden mortar]

Yes, it is?

[solemn rain]

toc talks about Formal languages

[tame vapor]

guys what is the best GUI framework for python that looks good ? i want to make a simple GUI

[gleaming knoll]

i'd just like, think how to split the rectangle into a grid of other rectangles, what the bounds of those rectangles would be (which would end up just being dividing into equal pieces with the usual division)
then find a library to load an image into an array and then do this array splitting and then save it back to different pngs
https://stackoverflow.com/questions/5953373/how-to-split-image-into-multiple-pieces-in-python

also, no, the c++ mascot is not a diseased rat named Keith

[solemn rain]

not problems

[split onyx]

But yes, I also try designing things to be modular with simple implementations. Rule of thumb is an interface with an implementation and having none of the modules depend on other ones

[junior wraith]

Yes.

Unrelated, but part of why I'm considering python is so that I don't have to spend a lot of time dealing with project setup and linker errors

[golden mortar]

ToC is not only about formal languages.

[jade robin]

this is a part, do you know about non deterministic turing machines?

[slow rivet]

Yeah like in my current project if I want to switch from a pull based scheduler to one that does up front graph analysis I just need to rewrite the scheduler and node implementations, the API wrappers and compiler can stay untouched. But for now pull based is the simplest option and it works

[golden mortar]

Did you not cover anything about computability?

The halting problem?

[tame vapor]

.

i want it to look modern

[solemn rain]

ah i think what we learnt in uni just taught us Formal languages as TOC. since it deals with automata, computability and stuff

this does not feel like that

[junior wraith]

Okay, what is this library in question?

As for the other statement, I know. I put it up because it is funny.

[solemn rain]

yea i remember that in tm but i actually forgot that

[golden mortar]

What about complexity theory?

[pallid garden]

it's a diseased rat named Keith with his leg blown off

[solemn rain]

we learn that in another subject called algorithm analysis

[split onyx]

I look at it like the ship of theseus 😄 you change 1 small part at a time (no big rewrites) but after doing enough of these you've probably changed everything

[jade robin]

looking at algorithm classes from the perspective of turing machines often make way more sense

[tame vapor]

guys

[golden mortar]

Alright, well, do you know what polynomial time means in the context of complexity theory?

[gleaming knoll]

i think people usually use PIL to deal with images in python

[craggy willow]

split it with any other restriction? just some regular division?

[solemn rain]

yea i think it just means n^x like n n^2 etc

[cyan scaffold]

What lambda said.
I think you can use just Pillow for all of this. Personally I would parameterize the number of rows and columns you want to slice and slice relatively to the size of the source image, but I don't know if you need each sub-picture to be in certain dimensions

[golden mortar]

Right. And do you know the difference between solving a problem and verifying a solution?

[solemn rain]

no thats something i came up with today. what does verifying mean just looking whether the output is correct?

[tame vapor]

how can not a single person know

[solemn rain]

why do we learn that

[gleaming knoll]

does verifying mean just looking whether the output is correct?
yes

[golden mortar]

It's verifying whether a given solution solves the given problem or not. I. e. you get handed a solution and you check if it's a correct solution.

[cyan scaffold]

PySide6. You can make simple GUIs with it and you can make look very modern. So can tkinter and maybe custom tkinter if you prefer that, but I guess you're staying far away from tkinter

[golden mortar]

It's a different task from coming up with the solution from scratch, and it can have a different complexity.

[solemn rain]

using a automata?

[golden mortar]

Using anything.

[lunar thorn]

i'm curious. imaging that i have a old keyboard with no customization app. Is it possible to make a app that customize the keyboard keys

[gleaming knoll]

why do we learn anything?

[tame vapor]

do you have an example of something made with it ? or how the default widgets look ?
i tried tkinter and it looks so old , i want something better

[cyan scaffold]

I'm sure there are examples on the internet

[solemn rain]

i mean we learn about TC because it helps us to study algorithms but this does not makes sense. we are making sure the answer we got is correct?
we can just look at the output

[cyan scaffold]

You can style tkinter apps with... Styles in themed tkinter. And custom tkinter doesn't look that bad but I figure it's just as rigorous

[solemn rain]

and why do we prove this

[golden mortar]

So some problems have a known non-polynomial complexity for solving the problem, but a polynomial complexity for verifying a solution, like 3-SAT.

[gleaming knoll]

if you care about the time complexity of an algorithm in general, why would you not care about the time complexity of an algorithm which verifies a solution? is verification just.. not interesting?

[tame vapor]

pyside6 looks ggood , thanks
do you have any good tutorial for it ? or should i just google and use whatever ?

[jade robin]

if I hand you an answer and hide my algorithm for coming up with the answer, you should be able to tell me if my answer is correct in polynomial time (sudoku is an easy example if you don't know about sat solving, there are problems which don't fit this ofc)

[craggy willow]

what's the time complexity of explaining this problem

[golden mortar]

Not always. With Traveling Salesman, you can't just look at the solution and know if it's correct.

[cyan scaffold]

There's something hold on

[golden mortar]

The output is a path, but you don't know if it's the shortest path or not.

[tame vapor]

O(n^n)

[solemn rain]

NP are the ones which can be verified in polynomial time right.
then i see things like hard NP NP complete

[gleaming knoll]

O(h, i dont know)

[solemn rain]

what do they mean when they say when a problem is hard

[craggy willow]

the version they use is to check if some path is less than n

checking is fast, finding such a path not so

[golden mortar]

Yeah, that's a variant with P verification. I was mainly just explaining how verifying isn't necessarily easy.

[simple willow]

I might try it thank you

[cyan scaffold]

This is no official documentation but it's a good start: https://www.pythonguis.com/pyside6-tutorial/
The official Qt docs are good, just they're for C++. When you need to use those, keep in mind the naming is the same but the syntax is different.
Oh, there are official PySide6 docs: https://doc.qt.io/qtforpython-6/index.html

[golden mortar]

They don't just say it's "hard", "NP-hard" is a specific term with a specific meaning.

[jade robin]

i think general computational hardness just means "more than polynomial"

[solemn rain]

yes i got it i think, we check our answer is correct or not.

so sudoku is NP ?

[jade robin]

yeah essentially

[solemn rain]

yea what i see is that "If every problem in NP can be polynomially reduced to NP hard then its NP hard for example travelling salesman"
but i don't understand this sentence

[golden mortar]

It basically means, if you can solve that NP-hard problem in polynomial time, then you can solve all NP problems in polynomial time, because you can transform all NP problems into your NP-hard problem in polynomial time and then solve it.

[tame vapor]

thank you very much bro you are the goat

[cyan scaffold]

You're welcome, thanks

[golden mortar]

So solving an NP-hard problem in polynomial time is a shortcut to proving P = NP

[hybrid nebula]

technically P = NP is always false

[tame vapor]

guys are you talking about semiconductors ? i am just hearing P and NP

[hybrid nebula]

what we want to prove is NP = {}

[jade robin]

huh

[golden mortar]

No, I don't think that's true.

[tame vapor]

what is P and NP

[grim axle]

NP - non deterministic polynomial (time)

[tame vapor]

huh

[golden mortar]

NP is the class of problems with proofs verifiable in polynomial time.

[hybrid nebula]

oh wait yeah my bad

got confused

[golden mortar]

We already know it's not empty.

[hybrid nebula]

we want to prove NP \ P = {}

[golden mortar]

Yeah, which is equivalent to P = NP

I think?

[jade robin]

everything P is already NP

[golden mortar]

yes

[hybrid nebula]

yes, since P is a subset of NP

[golden mortar]

P is a subset of NP

[jade robin]

we know NP is a super set of P, we want to prove if P is the only set NP contains

[verbal citrus]

aight i see i need to jump in, there is decidability and undecidability ...kay? (let's ignore semi-decidibaility for now, that's a special case)

so a problem is one of 2 ...decidable of undecidable..k?

all ez probs (ones that can be solved fast are called P solvable problems)

so these ez probs are all a subset of harder problems called NP problems (gg problems)

now these gg problems have legenedary tier (decidable) and god tier players in them (undedicable)

the undecidable ones are called NP-hard problems, the decidable ones are called np-complete ..

[cyan scaffold]

Deterministic polynomials are a subset of nondeterministic polynomials? Do I have this right?

[grim axle]

e.g. suduko.
although solving it is difficult (NP),
verifying it is rather easy, which usually takes "polynomial time" (P) or less.
like O(n¹), O(n³) or any O(n^k) for example.

[verbal citrus]

decidable just means can the algo give yes or no as an anser to a question

[jade robin]

huh

[worn moat]

Hey everyone 👋
I’m Avi, currently learning Python. I’ve completed the basics (syntax, loops, functions, basic OOP) and now I want to level up from beginner to strong intermediate.
My goal is to get solid in python eventually move towards data science + ML.

[jade robin]

decidability doesn't relate to NP hardness afaik

did i misinterpret it all

[hybrid nebula]

'does this C++ program compile' is undecidable, for example

[soft coral]

Hey! I liked the book "Fluent Python", maybe you'd like it too

[jade robin]

i think i'm misinterpreting all of it so it should definitely be phrased better if it's correct at all

[verbal citrus]

it does, because turing undecidable problems are np hard - eg halting problem or diophantine eqs

[gleaming knoll]

because of comptime (const..eval? constexpr? what do they call it) and templates?

[jade robin]

halting problem isn't np hard

[verbal citrus]

it's the quintessentional example of an np hard problem

[jade robin]

h u h

it's very literally undecidable by computation aka a logical paradox which can't be computed by any algorithm

NP hard problems can be computed even if inefficiently

[hybrid nebula]

templates specifically.
Do you know sheafification of G?

[gleaming knoll]

the youtuber? yeah

https://youtu.be/cFtymODJEjs ?
i didnt watch it

[hybrid nebula]

he made an interpreted language within the C++ compiler

using templates etc.

[jade robin]

atleast what i thought, turns out i'm wrong but i'm also interpreting all of it wrong

[hybrid nebula]

he also made world's best guide to Python

[verbal citrus]

that's false, cook levin says np complete can computed, if np hard is solved in p then all np complete problems are solved in p because you've solved an undecidable problem, and by proxy solved np complete since all np-hard problems are more expressive than np-complete

[gleaming knoll]

https://github.com/SheafificationOfG/based-cpp/blob/main/gil/gil.impl.hpp okay yeah i see

[hybrid nebula]

https://youtu.be/z_itQLvQxtk

[hybrid nebula]

lol yea

[gleaming knoll]

c++ templates are just a bad functional programming language

[hybrid nebula]

just like Rust types

[golden mortar]

NP-hard problems can't necessarily be computed.

Some NP-hard problems can, others can't.

[jade robin]

that's false, cook levin says np complete can computed
i do agree with this
if np hard is solved in p then all np complete problems are solved in p because you've solved an undecidable problem
not all np hard are undecidable. if you solve halting problem, it's np hard because it's undecidable, it's like solving a far harder problem than you need to show P = NP

[gleaming knoll]

rust traits are just a bad prolog
and haskell typeclasses

[pallid garden]

holy shit docker prune Total reclaimed space: 16.25GB

[golden mortar]

Yes, I don't think NP-hardness has anything to do with decidability.

[hybrid nebula]

what?

[verbal citrus]

i agree, by i guess tsp is np hard and isn't undecidble

[hybrid nebula]

NP-hard means

for every problem L which can be solved in non-deterministic polynomial-time, there is a polynomial-time reduction from L to H

[craggy willow]

16GB, the amount of space needed for a single picture of your mom

[split onyx]

my server has crashed twice because of disk space being eaten my docker logs 🥴

[golden mortar]

NP-hard problems are just problems that can be reduced to from an NP problem in polynomial time. It says nothing about how hard it is to solve.

[hybrid nebula]

every NP problem can be solved by a computer with sufficient memory in sufficient time

[golden mortar]

It could be uncomputable.

[gleaming knoll]

that would be a small human actually if you want an accurate picture
if you want each (surface) cell to be visibile

[golden mortar]

NP-hard != NP

[split onyx]

The first time it was me being naive and not knowing this was a problem, the second time around I have a grafana setup and alerting but my sendgrid API key had expired in the mean time lol

[craggy willow]

ur mom so small...

[pallid garden]

lol

[hybrid nebula]

NP-hard means, if you can solve a NP problem in time T, you can solve it in p(T) for polynomial p

[gleaming knoll]

pixels are a low quality approximation
scans ftw

[golden mortar]

Yes, it does mean that, but it doesn't only mean that.

In computational complexity theory, a computational problem H is called NP-hard if, for every problem L which can be solved in non-deterministic polynomial-time, there is a polynomial-time reduction from L to H.

[steel whale]

i'm making an image hosting service and i would like to know how i can add verification checks to uploads, to make sure that the received upload isnt a binary file, just an image

[golden mortar]

This doesn't specify if H can be solved at all.

[hybrid nebula]

what is the distinction?

[golden mortar]

Just that you can reduce an NP problem to it.

[hybrid nebula]

you can check for well-formed PNGs and remove all unnecessary chunks

[hybrid nebula]

every NP problem can be solved by a sufficiently powerful computer

[steel whale]

so PIL.Image.verify()?

[verbal citrus]

even the halting problem?

[golden mortar]

But NP is a subset of NP-hard.

[hybrid nebula]

the halting problem is not NP. it's undecidable

[golden mortar]

The halting problem is not in NP.

[verbal citrus]

i thought it was...hmmm

[hybrid nebula]

if you can solve every NP problem any amount of times, you can solve every NP-hard problem.

[solemn rain]

yea we classify them based on whether it can be solved and verified
the easy problems can be solved in polynomial time so that means its verification is also polynomial

the NP problems like sudoku can be verified easily but its hard to find the answer with some fixed algorithm (but can't we do something like checking all possbilities like bruteforce that solves it)
so that question asked in the wikipedia P vs NP means. "are all NP problems which can be verified P problems" ?

[golden mortar]

No? Why?

[pallid garden]

apparently you can have decidable problems that are not polynomial time, TIL

[craggy willow]

and then as a follow-up project, storing binaries in pngs

[jade robin]

halting problem is a paradox in the true sense of "this sentence is false, figure out the truth value of this statement"

[verbal citrus]

interesting...

[hybrid nebula]

every NP-hard problem can be solved by solving a NP problem a polynomial amount of times

[steel whale]

by a binary i mean an executable file

[golden mortar]

How is that?

[pallid garden]

it has to be solved by solving it a particular number of times, right?

[jade robin]

i don't think it's very helpful for finite systems either, determining if a turing machine with a finite number of states and memory cell halts is exp time complete iirc

[hybrid nebula]

oh wait I think I messed up the order

[jade robin]

and most of our real systems are finite

[hybrid nebula]

wait

[hybrid nebula]

all of them are

[jade robin]

idk i don't wanna trigger some quantum computing dude

[gleaming knoll]

the observable universe is finite and i dont think we have systems outside it

[hybrid nebula]

we def can't interact with them

[golden mortar]

Like, perhaps it's not very interesting, but technically you could reduce for example SAT into the halting problem.

So the halting problem should be NP-hard.

[jade robin]

that's stretching the word "reduce"

[solemn rain]

ok computer science is hard

i agree

[pallid garden]

but is it NP-hard?

[solemn rain]

NP HARD MAY NOT BE NP SINCE THEY ARE HARD AND CAN BE EXPONENTIAL TIME for verifying

[steel whale]

https://github.com/libvips/pyvips

seems like a more efficient alternative to PIL

[raw bramble]

Does the heat death of the universe not solve the halting problem? Every program will end, whether it’s designed to or not

[craggy willow]

maybe something will warm the universe back up

[peak relic]

Definitely solves cold booting fr

[craggy willow]

a space heater

[steel whale]

so one can die from being too hot

[pallid garden]

computer scientists live outside the bounds of reality

they can run their turing machine forever in violation of 2LOTD

[steel whale]

a hair dryer

[raw bramble]

Universe isn’t hair