#internals-and-peps | Python | Page 45

grave jolt Jun 1, 2020, 6:02 PM

#

A.__str__ is object.__str__?

whole glacier Jun 1, 2020, 6:02 PM

#

I see, that makes sense, so first it checks the current class __mro__

#

It doesnt have one, so __repr__ is called

spice pecan Jun 1, 2020, 6:02 PM

#

yeah, exactly

flat gazelle Jun 1, 2020, 6:02 PM

#

the __str__ of object checks for __repr__ afaik

whole glacier Jun 1, 2020, 6:02 PM

#

Well

grave jolt Jun 1, 2020, 6:02 PM

#

that would make sense

whole glacier Jun 1, 2020, 6:02 PM

#

!e ```py
def str(obj):
return getattr(obj, 'str', obj.repr)()

class A:
def str(self):
return 'hi'

print(str(A()))

fallen slateBOT Jun 1, 2020, 6:02 PM

#

@whole glacier :white_check_mark: Your eval job has completed with return code 0.

hi

whole glacier Jun 1, 2020, 6:02 PM

#

In this case it does pick up the __str__

#

So I'm leaning toward the mro

grave jolt Jun 1, 2020, 6:03 PM

#

We have to find the default __str__ implementation in the C source somehow

flat gazelle Jun 1, 2020, 6:03 PM

#

In [43]: class K:
    ...:     def __repr__(self):
    ...:         return 'AAAA'
    ...:

In [44]: object.__str__(K())
Out[44]: 'AAAA'

whole glacier Jun 1, 2020, 6:03 PM

#

Interesting lol

grave jolt Jun 1, 2020, 6:04 PM

#

so convenient to have cpython cloned, so I don't have to use github's search tool

#

.

#

ooh sorry

#

https://paste.pythondiscord.com/avovabudih.py

#

Yes, hastebin. That's totally python source...

flat gazelle Jun 1, 2020, 6:06 PM

#

pydis paste always uses .py

#

hastebin sometimes mistakes py for ruby

north root Jun 1, 2020, 6:06 PM

#

change the extension to .c and it'll highlight correctly: https://paste.pythondiscord.com/avovabudih.c

grave jolt Jun 1, 2020, 6:06 PM

#

    if (Py_TYPE(v)->tp_str == NULL)
        return PyObject_Repr(v);

#

actually, it highlights it correctly

#

oh, it doesn't

#

it stopped highlighting it correctly when I reloaded it lol

#

so, if it fails to find the object's __str__, it just calls __repr__.

#

And in the __repr__ source:

    if (Py_TYPE(v)->tp_repr == NULL)
        return PyUnicode_FromFormat("<%s object at %p>",
                                    Py_TYPE(v)->tp_name, v);

#

"hey, let's just print the pointer, it's so convenient"

spice pecan Jun 1, 2020, 6:10 PM

#

I never thought much about the fallback to __repr__, or rather how it's treated in derived classes

grave jolt Jun 1, 2020, 6:12 PM

#

I'd never though I would understand any of the cpython source

#

But it's written so well

undone hare Jun 1, 2020, 6:12 PM

#

It is actually pretty decent yeah

grave jolt Jun 1, 2020, 6:13 PM

#

👍

📎 unknown.png

#

who needs version control, you can just comment out some code

#

Oh, it's just an explanation

#

I didn't see that it was in Python, lol

spice pecan Jun 1, 2020, 6:14 PM

#

I really like that actually

#

Explaining Python's C sources with Python code

grave jolt Jun 1, 2020, 6:17 PM

#

That's an example of a big project made without having a 10-layer class hierarchy.

#

I never understood stuff like: https://github.com/fscdev/vkwave/blob/master/vkwave/bots/core/dispatching/events/base.py

#

creating a BaseX class for the only purpose of X inheriting from it, which only exists so that AX and BX inherit from it, while only differing in one class attribute

flat gazelle Jun 1, 2020, 6:21 PM

#

that is java style, and there is a reason java code is written this way. In python it is a lot more questionable though.

grave jolt Jun 1, 2020, 6:21 PM

#

yes, I felt I was reading Java

spice pecan Jun 1, 2020, 6:21 PM

#

It would be less questionable if it at least implemented the basic logic that was repeated in every child

snow olive Jun 1, 2020, 6:21 PM

#

hey guys is there anyone familiar with kivy? I'm in need of some help

flat gazelle Jun 1, 2020, 6:22 PM

#

#❓｜how-to-get-help or #user-interfaces

grave jolt Jun 1, 2020, 6:22 PM

#

You should ask in #user-interfaces of in #❓｜how-to-get-help. We have a few people who know kivy really well, hopefully they can help

snow olive Jun 1, 2020, 6:22 PM

#

got it! thank you!

atomic scarab Jun 1, 2020, 6:23 PM

#

Boys, is it possible to write an programm that will execute a command (pressing left click) when it hears a specific sound? I'm new into phyton

flat gazelle Jun 1, 2020, 6:23 PM

#

recognising sounds is quite difficult, but it possible in theory

grave jolt Jun 1, 2020, 6:24 PM

#

You too should probably ask in #❓｜how-to-get-help

#

but it's probably one of those problems that sounds really easy, but is really hard to implement

#

The code I linked is a part of an API wrapper for VK (the biggest social network in ex-USSR countries). It is not nearly as rich as the Discord API wrapper, and it is already ~22000 SLOC, while discord.py is 12000 SLOC or something.

#

So that java style inflates the code size as well.

#

At least in Python

somber halo Jun 1, 2020, 7:25 PM

#

that's an audio processing topic @atomic scarab, a more adequate channel for you problem might be #data-science-and-ml

#

I'm back with more quirky stuff from that string interning issue in Python's REPL

#

>>> a = "a b"
>>> b = "a b"
>>> c = "a b"
>>> a is b
False
>>> a is c
False
>>> b is c
False
>>> a = "ab" 
>>> b = "ab"
>>> c = "ab"
>>> a is b
True
>>> a is c
True
>>> b is c
True
>>> a = "a?b"
>>> b = "a?b"
>>> c = "a?b"
>>> a is b
False
>>> a is c
False
>>> c is b
False
>>> a = "a-b"
>>> b = "a-b"
>>> c = "a-b"
>>> a is b
False
>>> a is c
False
>>> c is b
False
>>> a = "a\nb"
>>> b = "a\nb"
>>> c = "a\nb"
>>> a is b
False
>>> a is c
False
>>> b is c
False

#

(sorry for the wall of text)

#

it seems to me that the decision rule is the exclusive presence of alphanumeric characters in string in order for it to be interned.

peak spoke Jun 1, 2020, 7:37 PM

#

I'm going to guess it's (ascii) valid identifiers below a certain len

somber halo Jun 1, 2020, 7:37 PM

#

maybe this should be taken to #esoteric-python no?

worldly hedge Jun 1, 2020, 7:46 PM

#

Can anyone give me a good explanation for the differences of a Semaphore and a Mutex

radiant fulcrum Jun 1, 2020, 7:48 PM

#

https://www.geeksforgeeks.org/mutex-vs-semaphore/ geeks for geeks do a pretty decent run down of it

GeeksforGeeks

Mutex vs Semaphore - GeeksforGeeks

A Computer Science portal for geeks. It contains well written, well thought and well explained computer science and programming articles, quizzes and practice/competitive programming/company interview Questions.

worldly hedge Jun 1, 2020, 7:52 PM

#

@radiant fulcrum oof thanks

#

I was wondering on what the real differences are 😛

oblique crystal Jun 1, 2020, 8:27 PM

#

ooph.... I don't know if this really fits the topic but here it goes: if, say I want to migrate code from CPython 3.5 to 3.7 (or better yet 3.8) is there is a chance that something will break? (assuming that I can install all same versions of the pypi packages that were used in the app)

#

as far as I saw till know usually code from newer Cpython might not run (or work as expected) on prior versions, but it seemed to me that the code from older (like 3.5) would run exactly the same way on new Cpython like 3.7/8

flat gazelle Jun 1, 2020, 8:29 PM

#

look at the changelogs and see if something got removed. As long as the code does mostly lawful python, it should be fine

oblique crystal Jun 1, 2020, 8:29 PM

#

well thing is the big parts of code weren't written by me

mellow widget Jun 1, 2020, 8:29 PM

#

can you update the version and build your app and run all tests against it?

oblique crystal Jun 1, 2020, 8:30 PM

#

and some part of app I've never even touched/seen 🙂

flat gazelle Jun 1, 2020, 8:30 PM

#

I would suggest looking at changelogs and seeing what behaviour changed

oblique crystal Jun 1, 2020, 8:30 PM

#

run all tests
yeah, tests.... not yet there. But the part I've been working with runs on py37

peak spoke Jun 1, 2020, 8:31 PM

#

there were very few breaking changes to the base language itself

#

for modules etc. you need to check yourself what has changed

wide shuttle Jun 1, 2020, 8:31 PM

#

There are some major changes to asyncio and some deprecations in that area

#

Although 3.5 already introduced async def/await

#

Still, that part of the library is very much being actively developed

#

Subtle things, like CancelledError no longer inheriting from Exception, but rather BaseException now

oblique crystal Jun 1, 2020, 8:33 PM

#

yeah quick glances told me that it seems like I would likely to be in trouble if it was 3.4 or older.
Asyncio is one reasons I would like an update: i.e. there is a small part of app I am writing where I want to use aiohttp and local asyncio.run. But the code itself did not use async in past.

mellow widget Jun 1, 2020, 8:34 PM

#

Best to read through the changelog, and check for any obvious changes which could cause errors, should be able to search your project for changes like Ves said above

oblique crystal Jun 1, 2020, 8:43 PM

#

also as far as I see in many cases when things are meant to be deprecated they first introduce only a deprecation warning

#

which would persist over several versions

#

before it is fully depcreated

#

I think even if there are some incompatible things they are likely easily fixe-able so I guess I will make a case for upgrading from 3.5 -> 3.8 to my collegues

spiral willow Jun 1, 2020, 8:49 PM

#

LossBerg, yall containerize?

mellow widget Jun 1, 2020, 8:50 PM

#

yeah if you're using docker to build your app and run your tests against you could just upgrade your python image version

#

and test it locally

spiral willow Jun 1, 2020, 8:51 PM

#

yep

#

simple edit

#

FROM python:3.5-slim to FROM python:3.8-slim

oblique crystal Jun 1, 2020, 8:51 PM

#

I don't think that app has any automated tests as of now

#

about containerization, I should ask

spiral willow Jun 1, 2020, 8:52 PM

#

great way to test upgrades

oblique crystal Jun 1, 2020, 8:52 PM

#

locally I only used venv but in production I believe it is docerized

mellow widget Jun 1, 2020, 8:53 PM

#

if you have a dockerfile or docker-compose.yml then you can build locally

spiral willow Jun 1, 2020, 8:53 PM

#

python 3.5 container is still available

#

https://hub.docker.com/layers/python/library/python/3.5.9-slim/images/sha256-f8cc6f7262bf9950ef040c4c8e90192b12b804fd49e4705104fb16258291750e?context=explore

snow kindle Jun 1, 2020, 8:58 PM

#

I struggle with decorator:

from libs.JsonManager import JsonManager


class ServerData(JsonManager):

    def __int__(self, ServerId):
        super().__init__('./storage/ServerData.json')
        self._Server_Id = str(ServerId)

    '''
    
    decorator
    
    '''

    def conditional_decorator(cls, condition):
        def decorator(func):
            if not condition:
                # Return the function unchanged, not decorated.
                return func
            return cls(func)

        return decorator

    '''
    
    Server Data
    
    '''

    def ServerInitialization(self):
        return True

    '''
    
    User Data
    
    '''

    @conditional_decorator(condition=True)
    def UserMaster(self, user_id):
        try:
            user_id = int(user_id)
        except ValueError:
            raise Exception('Please chose a valid int')
        if str(user_id) in [k for k in self.json_manager["user"]]:
            return self.json_manager["user"][user_id]
        else:
            new = self.json_manager
            new["user"][user_id] = {
                "is_kick": False,
                "is_ban": False,
                "is_mute": False,
                "is_admin": False,
                "reason": None,
                "time": None
            }
            self.json_manager = new
            return self.json_manager

How I can get is right ? (if a great people can correct it fast it will be the best)

mellow widget Jun 1, 2020, 8:59 PM

#

you don't understand decorators?

snow kindle Jun 1, 2020, 8:59 PM

#

I want to learns them

mellow widget Jun 1, 2020, 8:59 PM

#

think of your decorator like this

snow kindle Jun 1, 2020, 9:00 PM

#

I just want this right for get a correction and see where is my error

mellow widget Jun 1, 2020, 9:00 PM

#

conditional_decorator(condition=True, UserMaster(self, user_id))

#

UserMaster will execute first and return a value(hopefully), then your decorator will pass the returned value into it

snow kindle Jun 1, 2020, 9:01 PM

#

No it still wrong

#

can I dm you for send you all the code (in edition) in link ?

mellow widget Jun 1, 2020, 9:02 PM

#

https://realpython.com/primer-on-python-decorators/

Primer on Python Decorators – Real Python

In this introductory tutorial, we'll look at what Python decorators are and how to create and use them.

#

I'm just about to go offline, but take a read of that. Hopefully that will help u understand decorators and how to use them. If you're still confused ask someone in the chat 👍 good luck

snow kindle Jun 1, 2020, 9:06 PM

#

some one else ?

hollow crane Jun 1, 2020, 9:07 PM

#

what's the issue

snow kindle Jun 1, 2020, 9:07 PM

#

I don't undestand decorator I want some one make my code working for see my error

#

https://go.codetogether.com/#/5f16e44c-a823-488c-8288-a648c4c6e570/rQub4HJAV1pOOIHCQ0dPW8

If you want to see and edit @hollow crane You will be the best

hollow crane Jun 1, 2020, 9:26 PM

#

session expired

snow kindle Jun 1, 2020, 9:26 PM

#

https://go.codetogether.com/#/5f16e44c-a823-488c-8288-a648c4c6e570/vs8dJZeDArf6YzxotBpPFv

#

@hollow crane again ? (you can abuse mentioning me because I do it mechanically when I begin to be angry)

hollow crane Jun 1, 2020, 9:28 PM

#

what's the error then

shy vine Jun 1, 2020, 9:29 PM

#

Please see: #❓｜how-to-get-help , this is not a help channel

hollow crane Jun 1, 2020, 9:30 PM

#

oh yeah

snow kindle Jun 1, 2020, 9:30 PM

#

I know this

#

I want knowleadge

shy vine Jun 1, 2020, 9:32 PM

#

"I want someone to make my code work" is seeking help.

lost nexus Jun 1, 2020, 9:51 PM

#

will IOR overwrite keys or append a duplicate?

for example

d = {'one': '1', 'two': '2'}

d |= {'one': '700'}

# desired result is {'one': '700', 'two': '2'}

shy vine Jun 1, 2020, 9:53 PM

#

Overwrite

#

https://www.python.org/dev/peps/pep-0584/#specification

Python.org

PEP 584 -- Add Union Operators To dict

The official home of the Python Programming Language

lost nexus Jun 1, 2020, 9:54 PM

#

alright

golden imp Jun 2, 2020, 3:05 AM

#

oh sweet got renamed

rich wharf Jun 2, 2020, 4:25 AM

#

python used to have argument unpacking

#

def f(x, (y, z)): pass```

#

but they removed it

languid dagger Jun 2, 2020, 4:28 AM

#

How long ago was that?

rich wharf Jun 2, 2020, 4:29 AM

#

https://www.python.org/dev/peps/pep-3113/

Python.org

PEP 3113 -- Removal of Tuple Parameter Unpacking

The official home of the Python Programming Language

#

The pep was March 2007

languid dagger Jun 2, 2020, 4:32 AM

#

Huh 3.0 so you can still do it in Python 2

tawdry gulch Jun 2, 2020, 4:47 AM

#

Whattttt, thats so cool

#

It's not useful, but its cool

#

well, imo not very useful

grave jolt Jun 2, 2020, 4:54 AM

#

It's kinda useful in lark-parser because it may pass a nested list/tuple tree transformer sometimes

rich wharf Jun 2, 2020, 6:16 AM

#

I'm working on an iterable class

#

It can map, slice, filter, reduce, sum, and exhaust iterables.

#

It has four states

snow olive Jun 2, 2020, 6:18 AM

#

if there is anyone familiar with kivy could someone help me out in #help-lollipop ;-; thank you!

rich wharf Jun 2, 2020, 6:18 AM

#

@snow olive This isn't the right channel for that

snow olive Jun 2, 2020, 6:18 AM

#

got it ty

rich wharf Jun 2, 2020, 6:23 AM

#

Iterable Wrapper
States
lazy Lazily generates values from mapping/filtering/slicing
exhaust Generally lazy, but iterates over the iterable to find a match when needed
generate Generally lazy, but iterates over the iterable to find a match when needed, and stores previously exhausted values
sequence Converts the iterable to a list and stores it internally to manage operations
Methods
map Returns a mapped iterable from a function returning the replacement for the previous value
filter Returns a filtered iterable from a function returning a boolean for keeping or removing the value
reduce Takes two values of the iterable, applies a function to them, and repeats with the new value and the next value of the iterable
sum Gets the sum of the iterable
product Gets the product of the iterable
reverse Returns a reversed version of the iterable
list Converts the iterable to a list
len(iterable) Gets the length of the iterable
value in iterable Checks if the iterable contains a value
iterable[index] Gets the value at the index of the iterable
iterable[start:stop:step] Returns a sliced version of the iterable

def gen():
  c = 0
  while True:
    c += 1
seq = iterable(gen())
for i in seq[3::5]:
  print(i)

rich wharf Jun 2, 2020, 6:57 AM

#

I added a couple new functions

#

tuple, dict, and set return the iterable as a tuple, dict, or set

#

apply takes in a lambda of (keep, value) with keep being a condition of if to keep the value or not and value being the new value, returning a new iterable. It's essentially a merger of filter and map

#

chain allows you to chain one or more iterables to this one

#

+ allows you to add another iterable to this one

#

repeat and * allow you to repeat this iterable a certain amount of times

#

chunks allows you to split your iterable into chunks depending on the size, which works with or without the iterable's length, and optionally allows you to get the remainder piece

#

you can also use / on an iterable to divide it into chunks

cloud crypt Jun 2, 2020, 7:13 AM

#

I wrote kind of the same thing but it was an Iterator

#

https://github.com/NeKitDS/iter.py/blob/master/iter/iter.py here

rich wharf Jun 2, 2020, 7:25 AM

#

interesting

#

I do see a difference though

#

Mine takes in a state on what type of iterator while yours doesn't

#

But they are quite similar

#

Oh, the difference between yours and mine is that yours clones itself

#

While mine doesn't use cloning and uses other ways to maintain state

#

Anyways I'm going to sleep now, gn

dire hull Jun 2, 2020, 11:33 AM

#

how advanced do i need to be to use this channel

#

im scared my pleb questions gonna get me banned

#

adapride

gray mirage Jun 2, 2020, 11:59 AM

#

It's not a help channel, more a discussion channel

wide shuttle Jun 2, 2020, 12:00 PM

#

Yeah, it's not a help channel; it's in the discussion category

#

It's meant for discussion of the topics mentioned in the channel topic:

#

Discussion on the use cases, implementation and future of the Python programming language including PEPs, advanced language concepts, new releases, the standard library, and the overall design of the language.

flat gazelle Jun 2, 2020, 12:01 PM

#

well, some questions do fit, but more open ended ones, like What decides whether a string gets interned

dire hull Jun 2, 2020, 12:09 PM

#

wait this is the old python language lel i lost my brain

#

intimidated by the new name lol i see

blazing lotus Jun 2, 2020, 12:11 PM

#

is there an alternative to cx_Oracle package?

deft pagoda Jun 2, 2020, 12:13 PM

#

the idea i had to reduce decorator boiler plate has already been born out by several modules after i searched --- the wrapt module in particular seems pretty sweet here as it preserves function signatures

#

i'm surprised i hadn't heard of it before --- functools.wraps is all i've ever seen

undone hare Jun 2, 2020, 1:24 PM

#

So that's unh.. Hacky, let's say https://rushter.com/blog/python-virtualenv/

Artem Golubin

How virtual environment libraries work in Python

An introduction to virtual environment internals

wide shuttle Jun 2, 2020, 1:32 PM

#

I've never looked at the internals of virtual environments before

unkempt rock Jun 2, 2020, 1:32 PM

#

pls don't ban for me this https://www.programcreek.com/python/example/107431/discord.PermissionOverwrite

discord.PermissionOverwrite Python Example

This page provides Python code examples for discord.PermissionOverwrite.

#

lol

wide shuttle Jun 2, 2020, 1:33 PM

#

I don't think that's worthy of a ban but it also doesn't quite belong in this channel

unkempt rock Jun 2, 2020, 1:33 PM

#

oh ok

#

where does it belong

wide shuttle Jun 2, 2020, 1:33 PM

#

It's a #discord-bots-related example; do you have questions about it?

glass robin Jun 2, 2020, 1:35 PM

#

has the purpose of the channel changed?

wide shuttle Jun 2, 2020, 1:36 PM

#

No, the topic is still the same: advanced discussion of the Python programming language from a higher-level perspective

#

We're just experimenting with the name to see if that reduces the number of help questions that get posted here

undone hare Jun 2, 2020, 1:38 PM

#

I've never looked at the internals of virtual environments before
@wide shuttle basically, it is copying the python core interpreter and symlinking all the other files it has to use to the system interpreter, just so when you enter the interpreter it is in a different folder than the system one, and it will look for site packages in another folders

#

But it is a really dirty hack, they are symlinks and actual files here and there, and it is messing with your path

peak spoke Jun 2, 2020, 1:40 PM

#

How else would it work without messing with the path

undone hare Jun 2, 2020, 1:41 PM

#

I mean, the whole venv system is hacky, I'm surprised we don't have any pipenv-like system built into the interpreter yet

full jay Jun 2, 2020, 1:41 PM

#

Not to mention keeping it functional while still keeping the venv portable

undone hare Jun 2, 2020, 1:41 PM

#

Like, activating nested venv will break your PATH for sure

full jay Jun 2, 2020, 1:41 PM

#

Wall actually, it's funny you mention that, Ak

#

I think there is a spec for .pipfiles in one of the PEPs

#

So the structure is there, but it's just that there's been nothing built in for it. And I kind of get why.

wide shuttle Jun 2, 2020, 1:42 PM

#

Pipfile?

#

I know there's for pyproject.toml

full jay Jun 2, 2020, 1:42 PM

#

Hold on, now I'm second guessing myself

#

Ah no you're right

#

I knew it was one or the other

undone hare Jun 2, 2020, 1:43 PM

#

Yeah pyproject is a thing, but I looked into it, the support isn't great, and the functionalities aren't too

wide shuttle Jun 2, 2020, 1:43 PM

#

@undone hare Yeah, I know a bit how it works, but I've never looked at it in detail.

full jay Jun 2, 2020, 1:44 PM

#

It's what poetry uses for its purposes

wide shuttle Jun 2, 2020, 1:44 PM

#

pyproject.toml is pretty nice for packaging and publishing to PyPI

full jay Jun 2, 2020, 1:45 PM

#

The use cases for poetry and pipenv are a bit different, and I find it hard to justify one being in the stdlib over the other, and even harder to justify having both in there. It makes more sense to have the core still be there (venv) and have other 3rd party libraries extrapolate from that

#

Because at the end of the day, all the solutions are still using venv

#

From my loose understanding, pipenv is better for what I consider "living" projects. Ones in constant active development like the site and the bot here. poetry seems to be better for packaged and distributed projects, like what you would put in PyPi

#

I think

undone hare Jun 2, 2020, 1:47 PM

#

Yeah, making them into the stdlib would be silly, they don't belong here

#

But something that would automatically sync the requirements file before running code would be enough IMO, I just want to stop messing around with venvs

#

And some people never managed to make poetry working, and pipenv isn't getting any new update, so..

full jay Jun 2, 2020, 1:49 PM

#

It is

#

Just slowly

peak spoke Jun 2, 2020, 1:49 PM

#

pipenv is being worked on and there are very recent releases

full jay Jun 2, 2020, 1:49 PM

#

Just not uploaded to pypi quite yet

#

But the last merged commit was 8 hours ago on pipenv's github

#

So there's still life in it

peak spoke Jun 2, 2020, 1:50 PM

#

pypi has that release

undone hare Jun 2, 2020, 1:50 PM

#

Fair enough

full jay Jun 2, 2020, 1:50 PM

#

Oh does it?

#

Wait is that today?

peak spoke Jun 2, 2020, 1:50 PM

#

https://i.imgur.com/ehqlLwT.png

Imgur

undone hare Jun 2, 2020, 1:50 PM

#

I'm still surprised that there is still no form of concurrent locking yet

full jay Jun 2, 2020, 1:50 PM

#

Huh, I'll be damned

undone hare Jun 2, 2020, 1:51 PM

#

Oh nice

full jay Jun 2, 2020, 1:51 PM

#

Literally today

undone hare Jun 2, 2020, 1:51 PM

#

Can you link the changelog please? :)

full jay Jun 2, 2020, 1:51 PM

#

https://github.com/pypa/pipenv/blob/master/CHANGELOG.rst

GitHub

pypa/pipenv

Python Development Workflow for Humans. Contribute to pypa/pipenv development by creating an account on GitHub.

undone hare Jun 2, 2020, 1:53 PM

#

That's pretty cool, thanks

#

pipenv install and pipenv sync will no longer attempt to install satisfied dependencies during installation.
Finally

full jay Jun 2, 2020, 1:55 PM

#

Still looks like it's stuck on 3.7, though

#

But I guess that's not a huge issue

undone hare Jun 2, 2020, 1:56 PM

#

Yeah, I don't think it is

#

It handles 3.9b1 just fine haha

full jay Jun 2, 2020, 1:57 PM

#

I might fiddle with that then

unkempt rock Jun 2, 2020, 2:35 PM

#

hello

#

has someone tried using pip with python compiled from source?

wide shuttle Jun 2, 2020, 2:37 PM

#

Yes, I'm mainly using pip with versions of Python I've compiled myself

undone hare Jun 2, 2020, 2:38 PM

#

I think pyenv is source compiling them, so me too

regal frigate Jun 2, 2020, 3:09 PM

#

i am trying to understand cpython. do anyone can share links, tutorials, books so it can helpful for me.
thanks

full jay Jun 2, 2020, 3:19 PM

#

@regal frigate What all are you trying to understand about it? And do you mean CPython the interpreter you get from the python.org site or do you mean Cython the compiler that uses Python as a base?

regal frigate Jun 2, 2020, 3:21 PM

#

i already studied how interpreter works.. ie. lexer parser thing.
but when i look at cpython source i am not able to understand it.

#

i am interested in interpreter

#

i know that its not easy to understand. i am trying my best.

undone hare Jun 2, 2020, 3:26 PM

#

They are some great references here on the devguide https://devguide.python.org/exploring/#additional-references

snow kindle Jun 2, 2020, 3:28 PM

#

I ask my self do I can with out publish put my own lib in pip libs ?

boreal umbra Jun 2, 2020, 3:29 PM

#

For pathlib.Path, I feel like __bool__ should be the same as exists

flat gazelle Jun 2, 2020, 3:30 PM

#

I can imagine some issues with PurePath, but it does not seem like a bad idea

boreal umbra Jun 2, 2020, 3:30 PM

#

what's PurePath?

void sonnet Jun 2, 2020, 3:30 PM

#

hmm. i don't initially disagree. but, yeah, i imagine there are some pitfalls.

peak spoke Jun 2, 2020, 3:30 PM

#

Don't really like it as it involves an os call

flat gazelle Jun 2, 2020, 3:31 PM

#

A class in pathlib that allows path manipulation without actually being able to interact with real files

sacred tinsel Jun 2, 2020, 3:32 PM

#

eeh I'd probably end up doing if my_path: # If file exists lol

boreal umbra Jun 2, 2020, 3:32 PM

#

Right now I have txt_path = potential_txt_path if potential_txt_path.exists() else None

#

but it could just be txt_path = potential_txt_path or None

void sonnet Jun 2, 2020, 3:32 PM

#

my first question might be this: what is truthiness of a Path object?

boreal umbra Jun 2, 2020, 3:33 PM

#

I assume they're always true. I'll check.

#

>>> from pathlib import Path
>>> scratch = Path("./scratch")
>>> bool(scratch)
True
>>> fake = Path("./asdf")
>>> bool(fake)
True

#

I don't have an asdf directory.

peak spoke Jun 2, 2020, 3:36 PM

#

Not a fan of it because of the expensive call it does and not being really clear

full jay Jun 2, 2020, 3:38 PM

#

The path only creates the appropriate file path, it doesn't check for its existence

#

And there's a logical reason for that

#

If you wanted to create a file at a certain location, you don't want it to fail because that location doesn't exist. You're going to be creating it, after all

boreal umbra Jun 2, 2020, 3:39 PM

#

Also, isn't there a way to create a function that applies a certain operator between every object in a sequence?

#

I forgot the formula, except importing that operator's object from operator

flat gazelle Jun 2, 2020, 3:40 PM

#

functools.reduce?

void sonnet Jun 2, 2020, 3:40 PM

#

right. but having __bool__ wouldn't necessarily block that, would it Hem?

primal girder Jun 2, 2020, 3:41 PM

#

It does have a .exists method

void sonnet Jun 2, 2020, 3:42 PM

#

For pathlib.Path, I feel like __bool__ should be the same as exists

full jay Jun 2, 2020, 3:43 PM

#

Fair

#

I'm not sure then

primal girder Jun 2, 2020, 3:43 PM

#

So you can do

scratch = Path("./scratch")
print(scratch.exists())

full jay Jun 2, 2020, 3:43 PM

#

^

#

I'm just not sure what the bool value is meant to imply

#

That is a fair point

#

I do think the .exists() method is more explicit, so no complaint on that front

peak spoke Jun 2, 2020, 3:44 PM

#

The bool can be interpreted in multiple ways that make the same sense to the programmer without knowing more about Path

#

if path is not exactly clear imo unlike the exists method

flat gazelle Jun 2, 2020, 3:45 PM

#

It is a bit different from all the other __bool__ in the standard library

void sonnet Jun 2, 2020, 3:45 PM

#

i don't dislike the idea, but i can see the system call, as Numerlor points out, being an obstacle. my previous question was not worded well, and should've been more along the lines of what the expected truthiness of a Path object should be.

its clear (i think...) that it currently operates solely on the len(Path.__str__), which will always be true since an empty path is always eval'd as '.'.

primal girder Jun 2, 2020, 3:45 PM

#

My only gripe with pathlib so far is needing to use str(scratch) to get at the path if you're needing to pass the path string to something.

flat gazelle Jun 2, 2020, 3:45 PM

#

most python builtins take both Path and strings

paper echo Jun 2, 2020, 3:48 PM

#

yep, imo any function that doesn't accept an os.PathLike instead of a string is in error imo

#

and it's silly that e.g. sqlite3.connect doesn't support it

#

(seems like an easy PR)

full jay Jun 2, 2020, 3:51 PM

#

Wait wait wait, it can't? (also hey salt, I've missed you <3)

#

I'm assuming you can at least do str(os.Pathlike) and it'll take it, right?

molten onyx Jun 2, 2020, 4:03 PM

#

Today I discovered a "weird" inconsistency in Python that I can only assume is due to special method lookup semantics:

>>> def _getitem(_, item):
...     if item <= 5:
...         return item
...     raise StopIteration
...
>>> class SillySequenceSemantics:
...     __getitem__ = _getitem
...
>>> for item in SillySequenceSemantics(): print(item)
...
0
1
2
3
4
5
>>> class Meta(type):
...     __getitem__ = _getitem
...
>>> class ClassLevel(metaclass=Meta):
...     ...
...
>>> for item in ClassLevel: print(item)
...
0
1
2
3
4
5
>>> class Inconsistency:
...     __class_getitem__ = _getitem
...
>>> Inconsistency[0]
0
>>> for item in Inconsistency: print(item)
...
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: 'type' object is not iterable```

boreal umbra Jun 2, 2020, 4:04 PM

#

is this #esoteric-python ?

molten onyx Jun 2, 2020, 4:04 PM

#

I don't think so no since this isn't CPython abuse

#

anyway I don't really consider it too noteworthy

#

if someone thinks it's BPO-worthy they can go ahead and open a ticket

primal girder Jun 2, 2020, 4:06 PM

#

Your object needs to define __iter__() and __next__() to be iterable

peak spoke Jun 2, 2020, 4:08 PM

#

it only needs iter to be iterable, and there is a fallback to a method that uses getitem as shown above

boreal umbra Jun 2, 2020, 4:10 PM

#

    def count_tags(self) -> Counter[str, int]:
TypeError: 'type' object is not subscriptable

#

Did they not implement __class_getitem__ until 3.8?

primal girder Jun 2, 2020, 4:11 PM

#

3.7

peak spoke Jun 2, 2020, 4:11 PM

#

3.9 brings it to standard collections

#

above doesn't work for the same reason list[str] doesn't

boreal umbra Jun 2, 2020, 4:14 PM

#

I thought the typing types in typing are just for builtins rather than stdlibs

peak spoke Jun 2, 2020, 4:16 PM

#

there is a typing.Counter that you can use for the above

#

and the subscription behaviour will be brought to collections.Counter in 3.9

#

!pep 585

fallen slateBOT Jun 2, 2020, 4:16 PM

#

**PEP 585 - Type Hinting Generics In Standard Collections**

Link

Status

Accepted

Python-Version

3.9

Created

03-Mar-2019

Type

Standards Track

glass robin Jun 2, 2020, 4:18 PM

#

typing module needs to be easier to read for function IMHO

peak spoke Jun 2, 2020, 4:19 PM

#

What do you mean?

molten onyx Jun 2, 2020, 4:21 PM

#

@primal girder the whole point of this demonstration was that you do not in fact need __iter__

#

I very much do not like the __getitem__ fallback system in Python

boreal umbra Jun 2, 2020, 4:25 PM

#

what do you mean, the getitem fallback?

molten onyx Jun 2, 2020, 4:27 PM

#

not needing an explicit __iter__ method to be iterable

spice pecan Jun 2, 2020, 4:27 PM

#

I thought the fallback required __len__ as well

molten onyx Jun 2, 2020, 4:27 PM

#

nope

#

if the class follows sequence semantics with __getitem__ alone (0, 1, 2, ... as indices) it is iterable

spice pecan Jun 2, 2020, 4:28 PM

#

That's nice

boreal umbra Jun 2, 2020, 4:32 PM

#

suppose you have a class that contains a subscriptable

#

could you just do

#

__getitem__ = self.some_thing.__getitem__?

peak spoke Jun 2, 2020, 4:32 PM

#

you won't have access to self where that would be relevant

boreal umbra Jun 2, 2020, 4:33 PM

#

oh good point.

boreal umbra Jun 2, 2020, 5:38 PM

#

Another thing I noticed, there doesn't appear to be a simple way to import a module from its path as a string/os.PathLike

#

so I don't think you can do

#

>>> my_module = some_func('/path/to/my/file.py')
>>> my_module.another_func()

#

for any some_func that's defined in the stdlib.

true ridge Jun 2, 2020, 5:42 PM

#

for any some_func that's defined in the stdlib.
@boreal umbra I'm pretty sure there are other ways, but I use this all the time https://github.com/isidentical/BRM/blob/3627676f55851d7ce57f03221b8da43f0d1ce680/brm.py#L458-L461

GitHub

isidentical/BRM

Bicycle Repair Man - Rewrite Python Sources with Your Additions - isidentical/BRM

boreal umbra Jun 2, 2020, 5:43 PM

#

I saw a similar solution elsewhere. I thought it looked a bit clunky, but I guess I just don't appreciate what the different steps are for.

peak spoke Jun 2, 2020, 5:43 PM

#

for simple things this should work nicely

module_dict = runpy.run_path("file.py")
module=SimpleNamespace(**module_dict)

boreal umbra Jun 2, 2020, 5:43 PM

#

is runpy a stdlib module?

peak spoke Jun 2, 2020, 5:43 PM

#

yes

boreal umbra Jun 2, 2020, 5:44 PM

#

😄

true ridge Jun 2, 2020, 5:44 PM

#

It does a bit of abstraction, but yea runpy is nice.

peak spoke Jun 2, 2020, 5:45 PM

#

while not that nice I think the snippet linked above is the "proper" way

true ridge Jun 2, 2020, 5:46 PM

#

AFAIK they both run in the same way for normal source files (without silly details)

boreal umbra Jun 2, 2020, 5:46 PM

#

I feel like there should be a one-function-call solution in importlib

rich wharf Jun 2, 2020, 5:58 PM

#

I added more features to my iterable

#

The iterable class is very powerful

#

Nekit made a similar iterable class, the main difference is that their iterable class constantly depends on itertools.tee to clone the iterable

#

And tee isn't threadsafe

#

While mine depends on the user-provided state for how to act on the iterable

wide shuttle Jun 2, 2020, 7:03 PM

#

Sure

#

Still, I'm trying to think of a good reason to do this

bold valve Jun 2, 2020, 7:04 PM

#

how can i check for the message author role on discord server

wide shuttle Jun 2, 2020, 7:05 PM

#

That isn't really a question for this channel, although the previous question wasn't probably either

#

We do have a #discord-bots channel

narrow kettle Jun 2, 2020, 7:05 PM

#

Current discussion is in general

bold valve Jun 2, 2020, 7:05 PM

#

oh i wasnt sure ^^

wide shuttle Jun 2, 2020, 7:06 PM

#

This channel is meant to discuss the Python programming language itself, from a more advanced/higher level perspective

#

Think of things like the implementation, advanced language features, PEPs, the future of the language

cloud crypt Jun 2, 2020, 7:07 PM

#

interested to see python without GIL and with JIT haha

#

in the standard implementation, that is

primal girder Jun 2, 2020, 7:08 PM

#

Probably gonna be a while. Removing the GIL adds a lot of overhead

cloud crypt Jun 2, 2020, 7:08 PM

#

indeed it does, generally speeding python up will be a nice addition

gray mirage Jun 2, 2020, 7:09 PM

#

The GIL is a speed optimisation though

#

That's why it's hard

spiral willow Jun 2, 2020, 7:09 PM

#

GIL makes single thread faster

tacit hawk Jun 2, 2020, 7:09 PM

#

what about subinterpreters, I saw one of the goals is to have one GIL per subinterpreter

peak spoke Jun 2, 2020, 7:09 PM

#

python's speed benefits from the gil in probably more than 90% of its uses

#

The per interpreter gil seemed to be actively worked on last time I checked

spiral willow Jun 2, 2020, 7:11 PM

#

in some cases, you can easily scale by spinning up multiple instances

#

like web hosting

cloud crypt Jun 2, 2020, 7:13 PM

#

switching from python to rust and back feels so weird in terms of speed

rich wharf Jun 2, 2020, 7:13 PM

#

nekit, your iterator uses tee a lot

#

so that's the main difference I saw with yours from mine

spiral willow Jun 2, 2020, 7:13 PM

#

is rust significantly faster for your needs?

cloud crypt Jun 2, 2020, 7:13 PM

#

well yeah I needed to copy an iterator

rich wharf Jun 2, 2020, 7:14 PM

#

yeah, mine looks like it works differently in that case

cloud crypt Jun 2, 2020, 7:14 PM

#

hm, and what do you do?

rich wharf Jun 2, 2020, 7:14 PM

#

it either exhausts the original iterator

#

or it stores previously exhausted items

peak spoke Jun 2, 2020, 7:14 PM

#

Last time I needed speed for something was a brute force algo fo some math that was easily solved by one numba decorator (or could've been solved by spending a bit more time on memoization)

rich wharf Jun 2, 2020, 7:14 PM

#

or it converts the entire iterable to a list

cloud crypt Jun 2, 2020, 7:14 PM

#

@spiral willow rust is just overall much faster, yet still provides some neat abstractions

rich wharf Jun 2, 2020, 7:14 PM

#

depending on which state you set the iterator to

cloud crypt Jun 2, 2020, 7:15 PM

#

ah okay

rich wharf Jun 2, 2020, 7:15 PM

#

so it doesn't depend on tee

peak spoke Jun 2, 2020, 7:15 PM

#

I do wonder why the speed keeps being brought up so much

cloud crypt Jun 2, 2020, 7:15 PM

#

I don't remember how tee exactly works

rich wharf Jun 2, 2020, 7:15 PM

#

itertools.tee(iterable) returns two copies of the iterable

#

but you can't use the original one

cloud crypt Jun 2, 2020, 7:15 PM

#

yeah, yeah

#

but what does it do under the hood

rich wharf Jun 2, 2020, 7:15 PM

#

def tee(iterable, n=2):
    it = iter(iterable)
    deques = [collections.deque() for i in range(n)]
    def gen(mydeque):
        while True:
            if not mydeque:             # when the local deque is empty
                try:
                    newval = next(it)   # fetch a new value and
                except StopIteration:
                    return
                for d in deques:        # load it to all the deques
                    d.append(newval)
            yield mydeque.popleft()
    return tuple(gen(d) for d in deques)

roughly this

#

also, tee iterators are not thread safe

cloud crypt Jun 2, 2020, 7:16 PM

#

yeah I saw

rich wharf Jun 2, 2020, 7:16 PM

#

so that's the difference between yours and mine

#

yours is easier to use with any iterator (you don't have to manually set a state), but mine is better for thread-safe cases

spiral willow Jun 2, 2020, 7:17 PM

#

Sure

cloud crypt Jun 2, 2020, 7:17 PM

#

I could also implement this at some point

#

I made that iter project for fun in like, several hours, without putting much thought :p

rich wharf Jun 2, 2020, 7:17 PM

#

also I took some of your function ideas and put it in my iterator

cloud crypt Jun 2, 2020, 7:17 PM

#

sure

rich wharf Jun 2, 2020, 7:17 PM

#

like that funcmap method

spiral willow Jun 2, 2020, 7:17 PM

#

I'd love to see CPYthon move to JIT

rich wharf Jun 2, 2020, 7:17 PM

#

but I changed it a bit so that it's not a lambda

#

and seq.fmap('x, y', body) unpacks a single argument as x, y

#

instead of having two args

cloud crypt Jun 2, 2020, 7:19 PM

#

I wrote func_map because someone told me they don't like writing lambda haha

rich wharf Jun 2, 2020, 7:20 PM

#

mine is basically an equivalent version of yours

#

except instead of using 'x, y' as 2 args

#

it uses 'x, y' as unpacking 1 arg

#

and I think it's a neat idea

undone hare Jun 2, 2020, 7:20 PM

#

What on earth is that

rich wharf Jun 2, 2020, 7:20 PM

#

so you can do

#

seq = iterable([(1, 2), (3, 4), (5, 6)])
for i in seq.fmap('x, y', 'x + y'):
  print(i) #3 then 7 then 11

cloud crypt Jun 2, 2020, 7:21 PM

#

from iter import iter
print(iter(range(10)).func_map("x", "x * x").chain_with("nekit").zip("lemon_ves_zappa").collect(dict))``` haha

rich wharf Jun 2, 2020, 7:22 PM

#

what is chain_with?

#

oh wait

#

I see

cloud crypt Jun 2, 2020, 7:22 PM

#

hm I don't actually remember if that is correct way haha

rich wharf Jun 2, 2020, 7:23 PM

#

from iterable import iterable
print(
  iterable(range(10)).fmap('x', 'x ** 2').chain('nekit').zip('lemon_ves_lappa').collect(dict)
)

#

but I haven't added zip method yet

cloud crypt Jun 2, 2020, 7:23 PM

#

aha

rich wharf Jun 2, 2020, 7:23 PM

#

alright implemented

#

{0: 'l', 1: 'e', 4: 'm', 9: 'o', 16: 'n', 25: '_', 36: 'v', 49: 'e', 64: 's', 81: '_', 'n': 'l', 'e': 'a', 'k': 'p', 'i': 'p', 't': 'a'} I got this as my output

#

also with mine

#

you can replace .collect(dict) with .dict()

#

from iterable import iterable
print(
  iterable(range(10)).fmap('x', 'x ** 2').chain('nekit').zip('lemon_ves_lappa').dict()
)

cloud crypt Jun 2, 2020, 7:25 PM

#

indeed

#

Iterator.collect(func) is just func that takes the iterator and does the job

rich wharf Jun 2, 2020, 7:25 PM

#

yeah

cloud crypt Jun 2, 2020, 7:26 PM

#

at some point I want to write a library with nifty extensions to standard classes and also adding nifty functions but idk how to call it

rich wharf Jun 2, 2020, 7:26 PM

#

call it nifty

cloud crypt Jun 2, 2020, 7:27 PM

#

haha

#

@extends(Iterator)
def zip(self, iterable):
    return self.__class__(zip(self, iterable))``` something like this, for example

rich wharf Jun 2, 2020, 7:27 PM

#

I'm also going to work on a library I call ease

undone hare Jun 2, 2020, 7:27 PM

#

That looks fun, even if I didn't understood everything haha

rich wharf Jun 2, 2020, 7:27 PM

#

@undone hare iterable(range(10)) is a special class that handles iterators

#

.fmap('x', 'x ** 2') is equal to .map(lambda x: x ** 2) which is essentially equal to map(lambda x: x ** 2, the_iterable)

#

.chain('nekit') is the same as itertools.chain(the_iterable, 'nekit')

#

.zip('lemon_ves_lappa') is the same as zip(the_iterable, 'lemon_vas_lappa')

undone hare Jun 2, 2020, 7:28 PM

#

Hmmmmmmmm

rich wharf Jun 2, 2020, 7:28 PM

#

and .dict() is the same as dict(the_iterable)

cloud crypt Jun 2, 2020, 7:28 PM

#

I think itertools.chain(the_iterable, *"nekit") though

rich wharf Jun 2, 2020, 7:28 PM

#

not in my case

undone hare Jun 2, 2020, 7:29 PM

#

So that's just a big boilerplate library to do funny functional things?

rich wharf Jun 2, 2020, 7:29 PM

#

  def chain(self, *iters):
    return iterable(itertools.chain(self, *iters), state = self.state)

#

It's a convenient way to handle iterables

cloud crypt Jun 2, 2020, 7:29 PM

#

call it as you want

#

lol

rich wharf Jun 2, 2020, 7:29 PM

#

You can even copy iterables

#

iterable(range(10)).copy()

undone hare Jun 2, 2020, 7:29 PM

#

I mean, the description can sound boring, but that's is pretty fun

cloud crypt Jun 2, 2020, 7:29 PM

#

iterators, not iterables

rich wharf Jun 2, 2020, 7:30 PM

#

what's the difference

cloud crypt Jun 2, 2020, 7:30 PM

#

iterable is anything you can iterate over

#

including sequences, for example

undone hare Jun 2, 2020, 7:30 PM

#

You should try to implement all the kotlin iterator functions haha

cloud crypt Jun 2, 2020, 7:30 PM

#

iterator is something that can be generally exhausted

rich wharf Jun 2, 2020, 7:30 PM

#

oh

cloud crypt Jun 2, 2020, 7:31 PM

#

@undone hare my Iterator is inspired by std::iter::Iterator from guess the language

undone hare Jun 2, 2020, 7:31 PM

#

C++?

cloud crypt Jun 2, 2020, 7:31 PM

#

rust

undone hare Jun 2, 2020, 7:31 PM

#

Ah I was going to say that haha

cloud crypt Jun 2, 2020, 7:31 PM

#

@rich wharf I think I used chain_with wrong haah

undone hare Jun 2, 2020, 7:31 PM

#

Forgot that you are the rust guy :>

rich wharf Jun 2, 2020, 7:31 PM

#

I just made mine randomly inspired by JS

#

[1, 2, 3].map(x => x + 1)

cloud crypt Jun 2, 2020, 7:32 PM

#

yeah I like => syntax

rich wharf Jun 2, 2020, 7:32 PM

#

iterable([1, 2, 3]).map(x => x + 1)

#

yep, I want to get it added to Python

cloud crypt Jun 2, 2020, 7:32 PM

#

though more a fan of ->

rich wharf Jun 2, 2020, 7:32 PM

#

I like => slightly more

#

but either is good

cloud crypt Jun 2, 2020, 7:33 PM

#

well I mean

#

if you think about it, you can implement => with tokens and encoding for now

#

haha

undone hare Jun 2, 2020, 7:33 PM

#

One thing that would be pretty handy is to be able to convert dicts to lambda haha

{"var" : "print(var)"} ```->
```py
lambda var: print(var)```

rich wharf Jun 2, 2020, 7:33 PM

#

really?

cloud crypt Jun 2, 2020, 7:33 PM

#

yeah, why not

rich wharf Jun 2, 2020, 7:33 PM

#

@undone hare you'd need a variable called var

undone hare Jun 2, 2020, 7:33 PM

#

Well, you can do everything with custom tokens haha

#

True that

#

Like this

rich wharf Jun 2, 2020, 7:34 PM

#

f = x => 5 + x
print(f(3)) #8
f = x => (
  if True:
    print(x + 3)
)
f(3) #prints 6

cloud crypt Jun 2, 2020, 7:34 PM

#

# coding: arrow_lambda
func = x => x + 1``` can be translated to ```py
func = lambda x: x + 1```

rich wharf Jun 2, 2020, 7:34 PM

#

yeah

#

would that work for my complex lambdas

flat gazelle Jun 2, 2020, 7:34 PM

#

dict(var='print(var)') quoting dict keys is for suckers

undone hare Jun 2, 2020, 7:34 PM

#

It would be very nice to not need one, but I don't think that would be any possible

rich wharf Jun 2, 2020, 7:34 PM

#

f = x => (
  if True:
    print(x + 3)
)
f(3) #prints 6

cloud crypt Jun 2, 2020, 7:34 PM

#

it has an if

#

so well, no I guess

rich wharf Jun 2, 2020, 7:35 PM

#

oh well

cloud crypt Jun 2, 2020, 7:35 PM

#

@flat gazelle this post was made by js gang

undone hare Jun 2, 2020, 7:35 PM

#

It would be a use case for braces.py blobthinkingdown

spiral willow Jun 2, 2020, 7:35 PM

#

nekit, well there goes all my RAM then

rich wharf Jun 2, 2020, 7:35 PM

#

question

worldly hedge Jun 2, 2020, 7:35 PM

#

f = x => (
  if True:
    print(x + 3)
)
f(3) #prints 6

@rich wharf seems like an arrow function Thonking

rich wharf Jun 2, 2020, 7:35 PM

#

that's the point

#

how would you convert

'''
  if True:
    print(3)
'''
``` to
```python
'''
if True:
  print(3)
'''

#

like remove the lowest indent

undone hare Jun 2, 2020, 7:36 PM

#

Dedent it?

#

There's a function for that

cloud crypt Jun 2, 2020, 7:36 PM

#

why would you even do it

undone hare Jun 2, 2020, 7:36 PM

#

!d textwrap.dedent

fallen slateBOT Jun 2, 2020, 7:36 PM

#

`textwrap.dedent`

textwrap.dedent(text)```
Remove any common leading whitespace from every line in *text*.

This can be used to make triple-quoted strings line up with the left edge of the display, while still presenting them in the source code in indented form.

Note that tabs and spaces are both treated as whitespace, but they are not equal: the lines `" hello"` and `"thello"` are considered to have no common leading whitespace.

Lines containing only whitespace are ignored in the input and normalized to a single newline character in the output.

For example:

```py
def test():
    # end first line with \ to avoid the empty line!
    s = '''\
    hello
      world
    '''
    print(repr(s))          # prints '    hello\n      world\n    '
    print(repr(dedent(s)))  # prints 'hello\n  world\n'

undone hare Jun 2, 2020, 7:36 PM

#

Yeaaaahh

rich wharf Jun 2, 2020, 7:36 PM

#

does that work for an arbitrary amount of them

undone hare Jun 2, 2020, 7:36 PM

#

Yup

rich wharf Jun 2, 2020, 7:38 PM

#

f = single_function(
    'x',
    '5 + x'
)

print(f(3))

f = single_function(
    'x',
    '''
if True:
    print(45)
    '''
)

f(6)

cloud crypt Jun 2, 2020, 7:38 PM

#

sign = x => {
    if x < 0 {
        return -1;
    } elif x > 0 {
        return 1;
    } else {
        return 0;
    }
}``` now that I think about it...

rich wharf Jun 2, 2020, 7:38 PM

#

I just got this to work

#

ahhh braces in Python

#

I like this more

#

sign = x => (
  if x < 0:
    return -1
  elif x > 0:
    return 1
  else:
    return 0
)

cloud crypt Jun 2, 2020, 7:39 PM

#

# coding: braces
def square_int(x: int) { print("braces gang bois"); return x * x }```

undone hare Jun 2, 2020, 7:39 PM

#

Parsers would not be happy about it afaik procez

rich wharf Jun 2, 2020, 7:39 PM

#

I think it's possible to parse

#

sign = x => (
  if x < 0:
    return -1
  elif x > 0:
    return 1
  else:
    return 0
)

#

could be converted to

#

sign = x => (
if x < 0:
  return -1
elif x > 0:
  return 1
else:
  return 0
)

#

and it could parse that dedented inner-body just like any other python body

undone hare Jun 2, 2020, 7:40 PM

#

I don't remember if the parenthesis tokens are discarded or not hmm

cloud crypt Jun 2, 2020, 7:40 PM

#

how would you put it into a function

undone hare Jun 2, 2020, 7:40 PM

#

No they are not, everything's fine

rich wharf Jun 2, 2020, 7:40 PM

#

print(x => (
  if True:
    print(x)
))

#

or

undone hare Jun 2, 2020, 7:40 PM

#

x => should be in the parenthesis IMO

rich wharf Jun 2, 2020, 7:40 PM

#

print(
  x => (
    if True:
      print(x)
  )
)

undone hare Jun 2, 2020, 7:41 PM

#

And future you will thank you for that

rich wharf Jun 2, 2020, 7:41 PM

#

x => is ugly

#

x => (\n...\n) is better

undone hare Jun 2, 2020, 7:41 PM

#

-> maybe?

rich wharf Jun 2, 2020, 7:41 PM

#

x -> is ugly

#

x -> (\n...\n) is better

#

print(
  x ->
    if True:
      print(x)
)

undone hare Jun 2, 2020, 7:41 PM

#

I mean

print(
  (x => 
    if True:
      print(x)
  )
)```

#

This

rich wharf Jun 2, 2020, 7:41 PM

#

I prefer

#

print(
  x => (
    if True:
      print(x)
  )
)

#

I feel like surrounding the body is better

#

It just looks better

#

Your syntax looks like coffeescript

undone hare Jun 2, 2020, 7:42 PM

#

F for backtracking

cloud crypt Jun 2, 2020, 7:42 PM

#

how
do you
plan
to put
this
into the function body

rich wharf Jun 2, 2020, 7:42 PM

#

wdym put this into the function body

#

@cloud crypt can you elaborate

#

i'm confused

cloud crypt Jun 2, 2020, 7:44 PM

#

lambdas can not contain statements

rich wharf Jun 2, 2020, 7:44 PM

#

oh, that's what you mean

#

it'll try to evaluate it as an expression first

#

but if it fails and there are (), it'll evaluate it as statements

#

x => (5 + x) #treated as expression
x => (return 5 + x) #treated as statement

cloud crypt Jun 2, 2020, 7:45 PM

#

how would you make python understand it is a function

undone hare Jun 2, 2020, 7:45 PM

#

I'm pretty sure making multi-line lambdas is just a matter of changing 3 lines in the grammar file haha, like the relaxed decorator syntax was

rich wharf Jun 2, 2020, 7:45 PM

#

it would be built into the syntax?

undone hare Jun 2, 2020, 7:45 PM

#

With some internal calls I'm pretty sure that would be possible

rich wharf Jun 2, 2020, 7:45 PM

#

not the current python implementation

#

but if you changed python's grammar and parsing you could

cloud crypt Jun 2, 2020, 7:46 PM

#

well who is against you forking cpython

#

except the size LOL

undone hare Jun 2, 2020, 7:46 PM

#

It is just the grammar here

#

It is really small actually haha

#

The whole work dir is 30Mb

cloud crypt Jun 2, 2020, 7:46 PM

#

of lines

undone hare Jun 2, 2020, 7:46 PM

#

I don't know for the repo itself though

#

Do you count your disk usage with lines? blobthinkingdown

rich wharf Jun 2, 2020, 7:47 PM

#

alright I added a new feature to my iter class

#

(renamed it to iter)

undone hare Jun 2, 2020, 7:47 PM

#

F for the built in iter?

rich wharf Jun 2, 2020, 7:47 PM

#

f

cloud crypt Jun 2, 2020, 7:48 PM

#

new iter is just iter but with wrapping

#

lol

rich wharf Jun 2, 2020, 7:48 PM

#

but why would need it when you have mine

cloud crypt Jun 2, 2020, 7:48 PM

#

from iter import iter
print(iter([1, 2, 3, 4, 5]).join(", "))``` js gang

rich wharf Jun 2, 2020, 7:48 PM

#

you added join?

#

I need to implement that ahh

#

also

cloud crypt Jun 2, 2020, 7:48 PM

#

a while ago

#

ahah

rich wharf Jun 2, 2020, 7:48 PM

#

I want to add this

#

iter[1, 2, 3, 4, 5].join(', ')

cloud crypt Jun 2, 2020, 7:48 PM

#

ah that's simple

rich wharf Jun 2, 2020, 7:49 PM

#

yeah, hooking into getitem

#

and this

#

iter(1, 2, 3, 4, 5) being equal to iter((1, 2, 3, 4, 5))

cloud crypt Jun 2, 2020, 7:49 PM

#

metaclass go weeeeeeeeeeee

rich wharf Jun 2, 2020, 7:49 PM

#

yep

undone hare Jun 2, 2020, 7:49 PM

#

All the iter functions are still lazy, right?

cloud crypt Jun 2, 2020, 7:50 PM

#

yeah

rich wharf Jun 2, 2020, 7:50 PM

#

yeah

undone hare Jun 2, 2020, 7:50 PM

#

That's pretty cool

cloud crypt Jun 2, 2020, 7:50 PM

#

I am not familiar with thread-safety, but why is tee unsafe

undone hare Jun 2, 2020, 7:50 PM

#

You have things like filter, sort, group, and funny functions like that?

rich wharf Jun 2, 2020, 7:50 PM

#

I have filter

cloud crypt Jun 2, 2020, 7:50 PM

#

sort is uh

#

no

rich wharf Jun 2, 2020, 7:50 PM

#

not sure about sort and group

cloud crypt Jun 2, 2020, 7:50 PM

#

group is what

rich wharf Jun 2, 2020, 7:51 PM

#

I also have an apply function

#

which takes in the item

#

and returns a tuple (keep, new_value)

undone hare Jun 2, 2020, 7:51 PM

#

Make a dict based on the return value of a lambda

rich wharf Jun 2, 2020, 7:51 PM

#

I have .dict()

#

which is like dict(iterable)

undone hare Jun 2, 2020, 7:51 PM

#

A dict -> list -> values with the same output

cloud crypt Jun 2, 2020, 7:52 PM

#

I thought .collect(list|dict|set) is fine

rich wharf Jun 2, 2020, 7:52 PM

#

I have both

cloud crypt Jun 2, 2020, 7:52 PM

#

@undone hare what did you mean under group function though

rich wharf Jun 2, 2020, 7:52 PM

#

.collect(list)

#

.list()

undone hare Jun 2, 2020, 7:52 PM

#

Sort wouldn't be lazy yeah, but that would be really useful

rich wharf Jun 2, 2020, 7:52 PM

#

I'll add it

cloud crypt Jun 2, 2020, 7:53 PM

#

oh I know, I know

undone hare Jun 2, 2020, 7:53 PM

#

You have a dict output -> list, where all the values in a list have returned the same output through a lambda

#

Time to go to bed, cya guys! pyllow

cloud crypt Jun 2, 2020, 7:53 PM

#

import gc
gc.get_referents(object.__dict__)[0].update(iter=lambda self: iter(self))```

#

esoteric but magical

#

print([1, 2, 3, 4, 5].iter().map(lambda x: x * x).step_by(2).join("-"))```

undone hare Jun 2, 2020, 7:55 PM

#

wut

rich wharf Jun 2, 2020, 7:55 PM

#

wait does that work

cloud crypt Jun 2, 2020, 7:55 PM

#

mhm

rich wharf Jun 2, 2020, 7:55 PM

#

!e

import gc
gc.get_referents(object.__dict__)[0].update(into_iter=iter)
print([1, 2, 3, 4, 5].into_iter().map(lambda x: x * x).step_by(2).join("-"))

fallen slateBOT Jun 2, 2020, 7:55 PM

#

You are not allowed to use that command here. Please use the #bot-commands channel instead.

cloud crypt Jun 2, 2020, 7:57 PM

#

please don't do that in production tutu

rich wharf Jun 2, 2020, 7:59 PM

#

alright

cloud crypt Jun 2, 2020, 7:59 PM

#

I am still wondering what Akarys meant under group

rich wharf Jun 2, 2020, 7:59 PM

#

same

#

wait question

#

if we can add .iter

#

why not just add all of the methods to existing objects

#

[1, 2, 3].map(lambda i: 1 + i)

cloud crypt Jun 2, 2020, 8:00 PM

#

I mean you can

#

but you would need to convert it to iter

rich wharf Jun 2, 2020, 8:01 PM

#

so what

#

[1, 2, 3].map(lambda i: 1 + i) could return an iter object

cloud crypt Jun 2, 2020, 8:01 PM

#

assert iter.new(1, 2, 3, 4).chunk(2).collect(dict) == {1: 2, 3: 4}``` I think he meant this under groups

rich wharf Jun 2, 2020, 8:01 PM

#

wait

#

does .chunk(2) split it into two chunks

#

or chunks of size 2

cloud crypt Jun 2, 2020, 8:02 PM

#

two chunks

rich wharf Jun 2, 2020, 8:02 PM

#

I have a dif method name for that

#

.partition(2) splits it into chunks of size 2

#

.split(2) splits it into 2 chunks

#

both have optional remainder methods

cloud crypt Jun 2, 2020, 8:02 PM

#

fun

rich wharf Jun 2, 2020, 8:03 PM

#

I'm going to add a likely_max() method

cloud crypt Jun 2, 2020, 8:03 PM

#

so we're essentially copying each other's work

rich wharf Jun 2, 2020, 8:03 PM

#

yeah haha

cloud crypt Jun 2, 2020, 8:03 PM

#

except that I don't have your source LOL

rich wharf Jun 2, 2020, 8:03 PM

#

I can put it on git rn

#

but likely_max() will work like this

#

it looks for round(len(iterator) / math.e + .5) items

#

and then gets the biggest item after that

#

so it gets the likely max item without going through every item

cloud crypt Jun 2, 2020, 8:04 PM

#

how would you get length of an iterator

rich wharf Jun 2, 2020, 8:04 PM

#

  def __len__(self):
    return len(self.array)

#

self.array is a property that only works for iterators of state sequence

cloud crypt Jun 2, 2020, 8:05 PM

#

ah well

rich wharf Jun 2, 2020, 8:05 PM

#

so only iterators that allow you to convert it to a list before starting

#

  @property
  def array(self):
    self.block(state = 'sequence')
    if self._array is None:
      self._array = list(self.obj)
    return self._array

cloud crypt Jun 2, 2020, 8:05 PM

#

just put it on GitHub lemon_pleased

rich wharf Jun 2, 2020, 8:06 PM

#

https://github.com/Procez/ease/blob/master/iter.py

GitHub

Procez/ease

This is just some modules that I created to mess with Python. - Procez/ease

cloud crypt Jun 2, 2020, 8:06 PM

#

ah so ease

rich wharf Jun 2, 2020, 8:06 PM

#

yeah, basically a collection of random python modules I made

cloud crypt Jun 2, 2020, 8:07 PM

#

type(self) < self.__class__

rich wharf Jun 2, 2020, 8:07 PM

#

why?

#

@cloud crypt why is self.__class__ better

cloud crypt Jun 2, 2020, 8:10 PM

#

it is faster

rich wharf Jun 2, 2020, 8:11 PM

#

alright

cloud crypt Jun 2, 2020, 8:11 PM

#

and in almost all cases the same as type(self)

rich wharf Jun 2, 2020, 8:11 PM

#

ok, I'll switch

cloud crypt Jun 2, 2020, 8:11 PM

#

  def all(self):
    return sum(self)``` when Procez codes too fast

rich wharf Jun 2, 2020, 8:11 PM

#

haha

cloud crypt Jun 2, 2020, 8:11 PM

#

I feel like we could go to DM’s since we are the only participants here KarenWhat

rich wharf Jun 2, 2020, 8:12 PM

#

sure

rich wharf Jun 3, 2020, 12:28 AM

#

I made some code that gets this to work on CPython (maybe PyPy too)

#

SOURCE

for (current, after) in range(10).context(count = 1):
  print(current, after)

OUTPUT

twilit garnet Jun 3, 2020, 12:30 AM

#

did you add an entry to range's __dict__ with ctypes?

#

i think i did similar things using a helper module that i didn't complete

#

https://discordapp.com/channels/267624335836053506/470884583684964352/611320328659927050

#

actually, it starts more like here: https://discordapp.com/channels/267624335836053506/470884583684964352/611239643878522908

rich wharf Jun 3, 2020, 12:33 AM

#

not just ranges

#

gc.get_referents(object.__dict__)[0].update(...)

twilit garnet Jun 3, 2020, 12:34 AM

#

that's an interesting way of doing it, but definitely more #esoteric-python than #internals-and-peps

#

:D

rich wharf Jun 3, 2020, 12:34 AM

#

huh

#

@twilit garnet it actually works by converting the object to a class and then calling the method from the class object

#

you can do range(10).context(count = 1) with the trick

#

or iter(range(10)).context(count = 1)

twilit garnet Jun 3, 2020, 12:35 AM

#

yeah, my method allows that too theoretically

#

since it does the same thing

#

modifies the same attribute, i mean

#

mine is just a bit hackier, though

rich wharf Jun 3, 2020, 12:36 AM

#

that's from a really long time ago

twilit garnet Jun 3, 2020, 12:36 AM

#

haha yeah

rich wharf Jun 3, 2020, 12:36 AM

#

my goal though is to make a convenient iterable class

twilit garnet Jun 3, 2020, 12:36 AM

#

i used to do a lot of stuff like that back then, haven't had much motivation since then

rich wharf Jun 3, 2020, 12:36 AM

#

I think this is readable and convenient

twilit garnet Jun 3, 2020, 12:36 AM

#

kinda like rust's iterables?

rich wharf Jun 3, 2020, 12:36 AM

#

I haven't seen them

twilit garnet Jun 3, 2020, 12:37 AM

#

oh they're fucking beautiful

#

i once made an entire math expression parser using just a single chain of iterator methods

rich wharf Jun 3, 2020, 12:37 AM

#

interesting

#

wait what

#

an entire MATH expression parser

twilit garnet Jun 3, 2020, 12:37 AM

#

it was a simple one

#

and there were some frills that weren't included in the iterator chain

#

but it was still mostly put to work with the iterator IIRC

rich wharf Jun 3, 2020, 12:38 AM

#

for i in iter([1, 2, 3, 4, 5]).context(count = 1).fmap(
  'item, after', '{"item": item, "after": after, "sum": item + after}'
):
  print(i)

Output

{'item': 1, 'after': 2, 'sum': 3}
{'item': 2, 'after': 3, 'sum': 5}
{'item': 3, 'after': 4, 'sum': 7}
{'item': 4, 'after': 5, 'sum': 9}

#

These are an example of my iterables

#

(item, after) is the item that we're iterating over, and then the item after it

#

.fmap is a good function when you need to unpack variables

#

.map doesn't let you unpack values

#

I'm going to check out rust iterables

twilit garnet Jun 3, 2020, 12:40 AM

#

looks a lot like iterators in rust to me haha

#

TokenKind::Identifier => string
    .char_indices()
    .take_while(|&(_, chr)| chr.is_ascii_alphabetic())
    .last()

#

(don't hurt me mods, i'm just using this as a comparison! :D)

rich wharf Jun 3, 2020, 12:42 AM

#

I'm going to add .peek to my iterables

twilit garnet Jun 3, 2020, 12:43 AM

#

but still, modifying builtins like this is definitely more on the lines of #esoteric-python, to be honest.

rich wharf Jun 3, 2020, 12:44 AM

#

yeah, but I'm not talking about modifying builtins

#

I'm talking about my general iterables now

#

which is on the topic of advanced python

twilit garnet Jun 3, 2020, 12:44 AM

#

oh sure, that's fair.

brave badger Jun 3, 2020, 12:44 AM

#

.fmap is a good function when you need to unpack variables
@rich wharf What does the f in fmap stand for?

rich wharf Jun 3, 2020, 12:44 AM

#

funcmap

#

since it generates a function

brave badger Jun 3, 2020, 12:45 AM

#

Ah I see, kinda confused it with Haskell's fmap for Functors

rich wharf Jun 3, 2020, 12:51 AM

#

alright

#

I'm going to take a bunch of rust's functions for iterators and add them to my class

twilit garnet Jun 3, 2020, 1:02 AM

#

i can approve of that

rich wharf Jun 3, 2020, 1:24 AM

#

I added a few new methods

#

peek

#

skip_while

#

take_while

#

try_map

#

try_filter

#

find

#

sort

#

cycle

deft pagoda Jun 3, 2020, 1:29 AM

#

we made a peekable a week or two ago

rich wharf Jun 3, 2020, 1:29 AM

#

we who is we?

north root Jun 3, 2020, 1:29 AM

#

peekable iterator? that was cool

deft pagoda Jun 3, 2020, 1:29 AM

#

i don't remember exacts, but i have the file

rich wharf Jun 3, 2020, 1:29 AM

#

ok

#

who else made it?

deft pagoda Jun 3, 2020, 1:30 AM

#

https://github.com/salt-die/Snippets/blob/master/peekable.py

rich wharf Jun 3, 2020, 1:30 AM

#

also

#

peekable can be done using tee

deft pagoda Jun 3, 2020, 1:30 AM

#

yep, but i used a deque

#

because you can pass in a default value when you peek if you want

rich wharf Jun 3, 2020, 1:31 AM

#

huh

deft pagoda Jun 3, 2020, 1:31 AM

#

if the iterable is exhausted

#

we also had the .__ magic

#

In [134]: a = peekable(range(100))

In [135]: a._
Out[135]: 0

In [136]: a.__
Out[136]: 1

In [137]: a.___
Out[137]: 2

#

this is kind of cute

rich wharf Jun 3, 2020, 1:32 AM

#

I don't like it ahhh

#

that's #esoteric-python

#

but it's interesting

#

I'm going to add a .some method

deft pagoda Jun 3, 2020, 1:33 AM

#

well, the most often used mode is .peek(1) so we just shortcut to ._

rich wharf Jun 3, 2020, 1:33 AM

#

for me, it's just .peek()

#

you can't go further

#

but I'll add that option

#

also I'm going to do .peek(0) for starting

#

not .peek(1)

deft pagoda Jun 3, 2020, 1:34 AM

#

peek(0) returns a () for what we ended up with

rich wharf Jun 3, 2020, 1:34 AM

#

alright

deft pagoda Jun 3, 2020, 1:34 AM

#

peek() returns the next item, and peek(n) returns a tuple of length n

rich wharf Jun 3, 2020, 1:35 AM

#

oh

#

I think I'll implement that too

#

since it's a good idea

#

anyways I'm going to add in .some

deft pagoda Jun 3, 2020, 1:36 AM

#

In [138]: a = peekable(range(3))

In [139]: a.peek(5, default='python')
Out[139]: (0, 1, 2, 'python', 'python')

unkempt rock Jun 3, 2020, 1:37 AM

#

?

#

i just learned printing today XD

rich wharf Jun 3, 2020, 1:37 AM

#

This is a more advanced channel @unkempt rock

#

So you might not get the examples

#

I suggest you check out #python-discussion for more beginner-friendly conversation

#

seq = iter(range(10))

for i in seq.some(lambda i: i ** 2 if i != 5 else None):
  print(i)

Output

#

Should I do that

#

Or

#

seq = iter(range(10))

for i in seq.some(lambda i: i == 5).map(lambda i: i ** 2):
  print(i)

Output

unkempt rock Jun 3, 2020, 1:40 AM

#

k

rich wharf Jun 3, 2020, 1:40 AM

#

I think I'm going to do the latter

deft pagoda Jun 3, 2020, 1:40 AM

#

should implement islice as a __getitem__

rich wharf Jun 3, 2020, 1:42 AM

#

alright

#

and I'm going to implement peek defaults too

#

for i in (
  iter(range(10))
  .context(count = 2)
  .fmap('a, b, c', '{"a": a, "b": b, "c": c}')
  .some(lambda i: i['a'] == 5)
):
  print(i)

Output

{'a': 0, 'b': 1, 'c': 2}
{'a': 1, 'b': 2, 'c': 3}
{'a': 2, 'b': 3, 'c': 4}
{'a': 3, 'b': 4, 'c': 5}
{'a': 4, 'b': 5, 'c': 6}
{'a': 5, 'b': 6, 'c': 7}

brave badger Jun 3, 2020, 1:45 AM

#

What could be a rational way of implementing Scala's placeholder syntax e.g. List(1, 2, 3).map(_ + 2)? I'm currently managing the arithmetic operations as a list and then processing it through there.

rich wharf Jun 3, 2020, 1:46 AM

#

@brave badger Maybe have _ be some sort of Placeholder object

#

With __add__ and __sub__ and etc methods

brave badger Jun 3, 2020, 1:46 AM

#

Yeah I have those, was thinking of a better way of managing the functions

rich wharf Jun 3, 2020, 1:46 AM

#

Wdym?

#

You mean func(_)?

#

func(placeholder) could be converted to placeholder.call(func)

brave badger Jun 3, 2020, 1:47 AM

#

The _Wildcard class I have has a func list attribute that holds all of the functions of which I compose by iterating through them

rich wharf Jun 3, 2020, 1:48 AM

#

I think that's fine

brave badger Jun 3, 2020, 1:52 AM

#

I have to somehow preserve operator precedence though, since having _ + _ * _ would mean a list of [add, mul]

rich wharf Jun 3, 2020, 1:53 AM

#

That might be tricky

#

@brave badger I think that's quite esoteric though, maybe that's more of an #esoteric-python project

brave badger Jun 3, 2020, 1:54 AM

#

I could use the easy route and write a parser that builds the function for me pithink

rich wharf Jun 3, 2020, 2:00 AM

#

for i in (
  range(15)
  .context(count = 4)
  .fmap('a, b, c, d, e', '{"a": a, "b": b, "c": c, "d": d, "e": e}')
  .some(lambda i: i['a'] == 7)
):
  print(i)

Output

{'a': 0, 'b': 1, 'c': 2, 'd': 3, 'e': 4}
{'a': 1, 'b': 2, 'c': 3, 'd': 4, 'e': 5}
{'a': 2, 'b': 3, 'c': 4, 'd': 5, 'e': 6}
{'a': 3, 'b': 4, 'c': 5, 'd': 6, 'e': 7}
{'a': 4, 'b': 5, 'c': 6, 'd': 7, 'e': 8}
{'a': 5, 'b': 6, 'c': 7, 'd': 8, 'e': 9}
{'a': 6, 'b': 7, 'c': 8, 'd': 9, 'e': 10}
{'a': 7, 'b': 8, 'c': 9, 'd': 10, 'e': 11}

merry pollen Jun 3, 2020, 2:03 AM

#

@brave badger if you want to preserve precednece order you should probably represent your function chaining as a tree or smth

#

and then resolve it at the end

#

actually no

brave badger Jun 3, 2020, 2:09 AM

#

I kinda cheated by using eval

>>> (_ + _ * _ * _)(5, 5, 5, 5)
130

merry pollen Jun 3, 2020, 2:09 AM

#

oh well, that works bloblul

brave badger Jun 3, 2020, 2:10 AM

#

I need to make it safe though because the func attribute can be tampered with

merry pollen Jun 3, 2020, 2:10 AM

#

i think you could do it with function chaining though

#

hmn

rich wharf Jun 3, 2020, 2:17 AM

#

alright

#

@tardy badge I think you should get a help channel for this

#

#❓｜how-to-get-help

tardy badge Jun 3, 2020, 2:19 AM

#

apologies

rich wharf Jun 3, 2020, 2:19 AM

#

no apologies needed, you're fine

paper echo Jun 3, 2020, 3:18 AM

#

havent people done things like that in e.g. macropy?

#

make _ mean something special

#

and hope your users dont overwrite it

#

but it causes a lot of issues because fundamentally there are certain things for which you cannot "defer" evaluation

#

whereas with actual macros (or arbitrarily overridable operators) that isnt a problem

#

i tried to implement something like this for Pandas a while ago, ended up just reimplementing Pyspark syntax... https://github.com/gwerbin/pandas-anaphora

GitHub

gwerbin/pandas-anaphora

Anaphora for Pandas DataFrame columns (because writing lambda x: sucks) - gwerbin/pandas-anaphora

#

its a fundamental python limitation. kind of frustrating if you're a tinkerer, but its also helpful to have certain limits

golden imp Jun 3, 2020, 4:03 AM

#

If you're not married to having _ as the placeholder, could probably implement something with

#

from fancylib import VAR1, VAR2, VAR3
func = VAR1 + VAR2 * VAR3

#

or the like

rich wharf Jun 3, 2020, 4:05 AM

#

you could do

#

from fancylib import args
func = args[0] + args[1] + arg[2]

deft pagoda Jun 3, 2020, 4:05 AM

#

you can use sympy for that too

golden imp Jun 3, 2020, 4:05 AM

#

oh that too

rich wharf Jun 3, 2020, 4:06 AM

#

print(
  iter(range(10))
  .replace(5, '5 is a number')
  .list()
)

Does anyone have other iterator suggestions?

golden imp Jun 3, 2020, 4:06 AM

#

or even

from fancylib import arg
func = arg() + arg() * arg()

#

but the implementation will get into #esoteric-python

twilit garnet Jun 3, 2020, 4:07 AM

#

doing it even without the parentheses would be easy enough i think.

rich wharf Jun 3, 2020, 4:07 AM

#

yeah

deft pagoda Jun 3, 2020, 4:08 AM

#

In [182]: from sympy import var

In [183]: x, y, z = var('x,y,z')

In [184]: x + y + z
Out[184]: 
x + y + z

rich wharf Jun 3, 2020, 4:16 AM

#

sympy can do that?

#

that's interesting

deft pagoda Jun 3, 2020, 4:19 AM

#

sympy has a lot of nice things going on

rich wharf Jun 3, 2020, 4:22 AM

#

what new methods should I add to my iterators

boreal umbra Jun 3, 2020, 4:23 AM

#

Bogosort

rich wharf Jun 3, 2020, 4:23 AM

#

seriously

#

bogosort is||n't|| useful, but still

deft pagoda Jun 3, 2020, 4:24 AM

#

bogosort the most efficient sort ever

rich wharf Jun 3, 2020, 4:24 AM

#

but seriously, what should I add to it

#

seq = iter(range(10))

for i in (
  seq
  .context(count = 2)
  .fmap(
    'x, y, z',
    '''
      if True:
        return {'x': x, 'y': y, 'z': z}
      else:
        return {'x': None, 'y': None, 'z': None}
    '''
  )[::-2]
):
  print(i)

#

right now that works, and it outputs

#

{'x': 7, 'y': 8, 'z': 9}
{'x': 5, 'y': 6, 'z': 7}
{'x': 3, 'y': 4, 'z': 5}
{'x': 1, 'y': 2, 'z': 3}

#

but I want to make it useful

#

more useful

#

by adding more features and methods

grizzled vigil Jun 3, 2020, 5:58 AM

#

I've personally never been a fan of the "Fluent" style of programming that involves chaining together multiple methods like that. It can work for quick scripts and can be useful in something like a REPL where it never gets directly used again, but IMO it makes debugging more painful for no real gain. I'll admit that it looks decent from the perspective of concisely describing what the code does, I just don't find it to be worthwhile in "real-world" code, especially as the complexity of the chained operations become non-trivial.

golden imp Jun 3, 2020, 6:00 AM

#

spark lol

grizzled vigil Jun 3, 2020, 6:00 AM

#

That's just my opinion though; it's rather popular in C# (LINQ) and JS ecosystem.

spark parcel Jun 3, 2020, 6:00 AM

#

What do you mean with “fluent”?

grizzled vigil Jun 3, 2020, 6:01 AM

#

I'm referring to "Fluent interface". There's a few articles that describe it in-depth, but for a summary here's a wikipedia article: https://en.wikipedia.org/wiki/Fluent_interface

raven ridge Jun 3, 2020, 6:01 AM

#

SomeClass().withSomeAttribute(value).andSomeOther(value2)

flat gazelle Jun 3, 2020, 6:02 AM

#

I do prefer it over large expressions written with the usual fun() form. Generally, having sizable expressions is not ideal though.

spark parcel Jun 3, 2020, 6:02 AM

#

Okay, I’ll take a look

grizzled vigil Jun 3, 2020, 6:06 AM

#

@flat gazelle Yeah, I can see that. I don't think it's inherently a bad style of programming, I just try to avoid it for the most part since I find that it tends to make the code harder to change later or debug. If I'm performing too many operations in a single function or repeating them over and over, I'd rather create a separate function/method instead of chaining a bunch of calls together.

flat gazelle Jun 3, 2020, 6:08 AM

#

It does make sense in some cases in java where you would use kwargs in python/kotlin. But yes, it can get messy, especially in kotlin, which encourages that to an almost excessive extent

rich wharf Jun 3, 2020, 6:14 AM

#

how else would you suggest I do what I did

#

seq.map(lambda i: i ** 2).context(count = 2).collect(list)

#

without the attrs

#

maybe ```python
seq.manage(
map = lambda i: i ** 2,
context = 2,
collect = list
)

#

but ordered kwargs doesn't work in all languages

#

versions

vital vine Jun 3, 2020, 6:57 AM

#

Is it possible to add a getter-setter to a variable dynamically?
Without pre-defining it in the code

merry pollen Jun 3, 2020, 6:57 AM

#

wdym

#

a getter/setter for a like, standalone variable?

#

or like, an instance variable?

vital vine Jun 3, 2020, 6:59 AM

#

I want to detect when a variable gets changed without defining a getter setter at least by default

class Foo:
  bar = None

And the program makes bar a property so it can detect when it gets called or set to another value

rich wharf Jun 3, 2020, 7:00 AM

#

use property

#

class Foo:
  @property
  def val(self):
    return self._boo
  @property.setter
  def val(self, value):
    self._boo = value

merry pollen Jun 3, 2020, 7:01 AM

#

class Foo:
  ...
a = Foo()
a.boo = property(getter_function)
a.boo.setter(setter_function)

would work, but

#

why

#

or at least i hazard it would work

brave badger Jun 3, 2020, 7:02 AM

#

That would require writing decorating each attribute though, in terms of not defining a getter/setter by default, you could probably do some automation with metaclasses/plain parent classes

vital vine Jun 3, 2020, 7:02 AM

#

Thanks guys, I'll try them 😄

merry pollen Jun 3, 2020, 7:05 AM

#

oh nop does not seem to work

#

oh you add it to the class

#

huh

dire hull Jun 3, 2020, 7:09 AM

#

so in the docs of pickle it says JSON can only represent a usbset of python build-in types and no custom classes

merry pollen Jun 3, 2020, 7:09 AM

#

descriptors.. wacc

dire hull Jun 3, 2020, 7:10 AM

#

does this mean that it also cannot serialize class instances created from custom classes?

merry pollen Jun 3, 2020, 7:10 AM

#

i mean, json certainly can't

dire hull Jun 3, 2020, 7:10 AM

#

but it says json is used everywhere

merry pollen Jun 3, 2020, 7:10 AM

#

it indeed is

dire hull Jun 3, 2020, 7:10 AM

#

so if i pickle objects to binary then i can onyl read it from python

merry pollen Jun 3, 2020, 7:11 AM

#

yup

dire hull Jun 3, 2020, 7:11 AM

#

also pickle is not human readable welp

merry pollen Jun 3, 2020, 7:11 AM

#

also yup

dire hull Jun 3, 2020, 7:11 AM

#

this is like very suc

merry pollen Jun 3, 2020, 7:11 AM

#

you can try jsonpickle

#

which will pickle the bits that have to be pickled and json the rest

dire hull Jun 3, 2020, 7:12 AM

#

so if there are only custom class isntances then its just all piclkle?

#

lol i feel like this is slightly better but welp

merry pollen Jun 3, 2020, 7:13 AM

#

ThinkShrug

undone hare Jun 3, 2020, 7:15 AM

#

Pickle isn't safe too

#

By modifying it, you can make it run any code

#

It can be really dangerous in some cases

dire hull Jun 3, 2020, 7:18 AM

#

yeah i read about it but since im still learning the ropes ill ignore that part for now

#

lol

#

actually come to think of it i should probbly just store the inputs and create objects reading those inputs to avoid this and make the serialized file readable

#

welp

undone hare Jun 3, 2020, 7:35 AM

#

Well, pickle is easily readable, even if that's not a plain text format

#

Someone could just fire up an interpreter and load the file

dire hull Jun 3, 2020, 7:57 AM

#

i see

undone hare Jun 3, 2020, 8:01 AM

#

If you don't share your source code, the best way to do it is to drop all your attributes as raw hexadecimal numbers in a file, so without the source code it is almost impossible to know what is what, and it should be pretty easy to parse

vital vine Jun 3, 2020, 9:12 AM

#

Why is that when I use the property() function to set a variable to a property in a function and not in the root of a class it'll be just a property class and doesn't actually work as a property? (I mean I can only get its value with var.__get__(self) and not just with var)

dire hull Jun 3, 2020, 9:24 AM

#

ill look into that thx!@undone hare

molten onyx Jun 3, 2020, 9:34 AM

#

@vital vine how exactly are you using it? I don't quite understand

#

>>> class C:
...     ...
...
>>> def getter(self):
...     return 10
...
>>> def setter(self, val):
...     print(val)
...
>>> C.x = property(getter, setter)
>>>
>>> c = C()
>>> c.x
10
>>> c.x = 11
11```

vital vine Jun 3, 2020, 9:36 AM

#

class Foo:
  bar = None

class A:
  def __init__(self):
    def _getter(self):print("Getter called")
    def _setter(self, v):print("Setter called")
    Inst = Foo()
    Inst.bar = property(_getter, _setter)

molten onyx Jun 3, 2020, 9:37 AM

#

ah, this is because your getter and setter are bound to A already

#

wait no that's in __init__

vital vine Jun 3, 2020, 9:38 AM

#

yeah

molten onyx Jun 3, 2020, 9:38 AM

#

It's because you set the property to the instance instead of the class

vital vine Jun 3, 2020, 9:40 AM

#

ah I see

#

Thanks 🙂

celest juniper Jun 3, 2020, 3:28 PM

#

How difficult is python GUI and what would be the best route to take for learning it? I remember I stumbled upon one GUI library once upon a time but it barely had a KB / documentation

rich wharf Jun 3, 2020, 3:30 PM

#

@celest juniper There's a few options for it, and Python isn't the best for it but Python can do it decently well

#

Tkinter, PyQt, WtPython and etc are a few options if I remember

#

I'm making a simple GUI library as well

#

Kivy as well

#

https://docs.python.org/3/library/othergui.html Here's a list of a few other ones other then the built-in default, Tkinter

fossil jetty Jun 3, 2020, 3:39 PM

#

why is it faster to write out CSVs iteratively vs. with multiprocessing.pool? As far as I can tell, my RAM, CPU, and disk are are far below 100%... Shouldn't multiprocessing make it faster by writing on a separate process?

i.e. I should be able to write as many CSVs as pool executors in 1/n the time

celest juniper Jun 3, 2020, 3:40 PM

#

What language would you say bests supports GUI interfaces? @rich wharf

rich wharf Jun 3, 2020, 3:40 PM

#

I'm not sure

#

Python has some good options but it's not the absolute best

#

It isn't bad either though

wide shuttle Jun 3, 2020, 3:43 PM

#

why is it faster to write out CSVs iteratively vs. with multiprocessing.pool? As far as I can tell, my RAM, CPU, and disk are are far below 100%... Shouldn't multiprocessing make it faster by writing on a separate process?

i.e. I should be able to write as many CSVs as pool executors in 1/n the time
@fossil jetty

What's limiting your writing speed? Probably not your CPU, but your disk. Adding more CPU overhead won't make your harddisk write data faster.

fossil jetty Jun 3, 2020, 3:43 PM

#

But SSD reports 2% usage?

#

the constraint is SSD write speed, but my disk is showing ~1MB/s write

#

random writes should be in the ~100MB/s range?

wide shuttle Jun 3, 2020, 3:48 PM

#

I'm not sure what's limiting your writing speeds, but unless you do a lot of preprocessing on the spot during your write iteration, I doubt your task is CPU bound

spiral willow Jun 3, 2020, 4:19 PM

#

because disk is likely bottleneck but Windows won't always show it at 100% even when it's effectively bottlenecked

#

esp if it's a spinny disk

#

++, no, multithreading can make it worse as multiple applications keep sending interrupt requests for disk time

#

esp it's line by line

#

at my company, we write out large CSVs from DB results, we found single threaded to be much better for performance so all CSV create jobs are loaded into ServiceBus where exporter process runs SQL command to execute the CSV write then moves the files to proper place.

sturdy timber Jun 3, 2020, 4:49 PM

#

What was the change that meant you could you not assign to True or False? I'm sure you used to be able to but now it gives a SyntaxError?

flat gazelle Jun 3, 2020, 4:50 PM

#

I would guess they became literals instead of builtins

wide shuttle Jun 3, 2020, 5:00 PM

#

The change was made in the transition to 3.0, if you want to dig into it. True and False were probably given what was already implemented for None

sturdy timber Jun 3, 2020, 5:02 PM

#

Ah yes, I found it here https://docs.python.org/3.0/whatsnew/3.0.html#changed-syntax
True, False, and None are reserved words. (2.6 partially enforced the restrictions on None already.). Thanks

wide shuttle Jun 3, 2020, 5:04 PM

#

In Python 2.2, True and False were introduced, but just as "constants" assigned to 1 and 0. The bool type was introduced in 2.3.

boreal umbra Jun 3, 2020, 5:18 PM

#

>>> PathLike = t.Union[os.PathLike, str]
>>> isinstance("hi", PathLike)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/lib/python3.7/typing.py", line 711, in __instancecheck__
    return self.__subclasscheck__(type(obj))
  File "/usr/local/lib/python3.7/typing.py", line 719, in __subclasscheck__
    raise TypeError("Subscripted generics cannot be used with"
TypeError: Subscripted generics cannot be used with class and instance checks

#

Is there anything that supports this behavior, or am I just abusing the type checker?

grave jolt Jun 3, 2020, 5:18 PM

#

I thought this was supported 🤔

boreal umbra Jun 3, 2020, 5:19 PM

#

this is 3.7

#

not sure if that matters

grave jolt Jun 3, 2020, 5:20 PM

#

!e

from typing import Union
from os import PathLike
PathLike = Union[PathLike, str]
print(isinstance("hi", PathLike))

fallen slateBOT Jun 3, 2020, 5:21 PM

#

@grave jolt :x: Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 4, in <module>
003 |   File "/usr/local/lib/python3.8/typing.py", line 766, in __instancecheck__
004 |     return self.__subclasscheck__(type(obj))
005 |   File "/usr/local/lib/python3.8/typing.py", line 774, in __subclasscheck__
006 |     raise TypeError("Subscripted generics cannot be used with"
007 | TypeError: Subscripted generics cannot be used with class and instance checks

boreal umbra Jun 3, 2020, 5:21 PM

#

😦

peak spoke Jun 3, 2020, 5:21 PM

#

Why do you need str and path like union?

boreal umbra Jun 3, 2020, 5:21 PM

#

Because one could use either sometimes.

peak spoke Jun 3, 2020, 5:21 PM

#

a str path is path like

grave jolt Jun 3, 2020, 5:21 PM

#

nope

#

📎 unknown.png

boreal umbra Jun 3, 2020, 5:22 PM

#

nope

#

I discovered this recently

#

I had just finished a new release version of my package and it crashed when I used the CLI

unkempt rock Jun 3, 2020, 7:17 PM

#

what is that font?

grave jolt Jun 3, 2020, 7:18 PM

#

Hasklig Black

#

regular Hasklig might be even better

📎 unknown.png

unkempt rock Jun 3, 2020, 7:21 PM

#

ok, this one I can read

#

not like I understand it or anything

grave jolt Jun 3, 2020, 7:23 PM

#

Hasklig Semibold

📎 unknown.png

#

I like this one better, actually

rich wharf Jun 3, 2020, 8:10 PM

#

I'm making a package with modules that make Python "easier".

astparse - Allows you to use a helper class to more easily parse ast.
map - Allows you to create dictionaries with more methods and allows . item setting and getting
decorators - Makes creating decorators that attach information to functions easier, as well as making general decorators
iter - Adds plenty of methods to make iteration easier
timespan - A convenient way to define and manage timespans that shows you the years, days, hours, minutes, seconds, etc of the timespan.

chains - Allows you to define anonymous functions with attribute and method chaining
lambdatools - Allows you to define anonymous functions using functions that take in lambdas

Those are some of the existing ones I have, does anyone have suggestions for new ones I should make?

grave jolt Jun 3, 2020, 8:31 PM

#

Maybe you should post that in open-source-collaboration/show-your-projects and leave a permanent link to some repo, so that people can post some suggestions

rich wharf Jun 3, 2020, 8:33 PM

#

Huh, maybe.

whole rune Jun 3, 2020, 8:51 PM

#

HI there! i need help dealing with browsers, if anyone available can help me pls pm me

rich wharf Jun 3, 2020, 8:54 PM

#

@whole rune Hi there! Maybe ask in #❓｜how-to-get-help if it's Python related

#

This isn't the channel for that

magic python Jun 3, 2020, 9:14 PM

#

@rich wharf what's an example from iter

rich wharf Jun 3, 2020, 9:16 PM

#

for c, i in (
  iter([1, 2, 3])
  .map(lambda i: i ** 2)
  .enumerate()
):
  print(f'count {c}: {i}')
>>> count 0: 1
>>> count 1: 4
>>> count 2: 9

#

@magic python

#

iter is compatible with the builtin iter function

magic python Jun 3, 2020, 9:18 PM

#

is the 0,1,2 intentionally shifted from 1,4,9?

rich wharf Jun 3, 2020, 9:19 PM

#

yeah

#

[1, 2, 3] and list(range(3)) aren't equal

#

was just showing that as an example

#

there's more methods like .sample, .shuffle, .filter, .peek, __next__, __len__, you can use __getitem__ for items or slices, etc

whole rune Jun 3, 2020, 10:21 PM

#

thank you @rich wharf

rich wharf Jun 3, 2020, 10:37 PM

#

@whole rune no thanks necessary

stoic bone Jun 4, 2020, 2:01 AM

#

I have a math question, sort of, to do with random floats between [0.0, 1.0). So I'm looping through an array of strings, and I have a function that I call that returns True or False based on a random float between [0.0, 1.0). I want to make it so that the closer it is to the end of the array of strings, the more and more likely it is that the function will return True. But I need to do so in a way that's exponential, not linear. So until you get to the last 10-20% of the array, the likelihood only increases slightly but once it gets to that area the likelyhood increases drastically until the last few entries it's almost a certainty that it will return True

prime estuary Jun 4, 2020, 2:10 AM

#

@stoic bone Well the way you'd do that is produce an exponential equation that is 1 at the end, and 0.2 (or whatever) at the start. Then compare if random() < equation(pos). That would trigger 20% of the time at the start, 100% at the end.
To create the equation, what you need is the "standard" equation y = a*exp(b*x). a and b are constants, which you can figure out with some algebra. First feed in x=0, your desired y (0.2?) and find a. Then add x=1 (or whatever), the desired y=1 and solve for b. Rearranging produces b = logn(y/a)/x.

stoic bone Jun 4, 2020, 2:11 AM

#

y = a ** (b * x) you mean?

prime estuary Jun 4, 2020, 3:08 AM

#

That would also apply, I was assuming you meant ae^bx.

cedar latch Jun 4, 2020, 3:42 AM

#

anyone used textgenrnn

shut anchor Jun 4, 2020, 3:48 AM

#

I've done textgen with weighted random predictive stuff. Not an rnn, but the results aren't that different.

#

It wrote goose poetry.

#

and ribald Grimm's fairytales.

rich wharf Jun 4, 2020, 4:29 AM

#

I prefer GPT2 now, but yes I have

#

gpt-2-simple

#

made by the same guy as textgenrnn

#

not the model, but the library

cedar latch Jun 4, 2020, 4:35 AM

#

it it better and easy to implement

humble finch Jun 4, 2020, 4:44 AM

#

How to use future annotations?

#

PEP 563

boreal umbra Jun 4, 2020, 5:29 AM

#

@humble finch I believe it's just from __future__ import annotations

#

As the first line of the file.

north root Jun 4, 2020, 5:37 AM

#

the future annotations are quite nice

boreal umbra Jun 4, 2020, 5:41 AM

#

Speaking of which, I remember reading that annotations were originally introduced to store arbitrary data rather than type hints. Why is that?

cedar latch Jun 4, 2020, 5:42 AM

#

gpt-2-simple
@rich wharf is it pretrained?

rich wharf Jun 4, 2020, 5:42 AM

#

@cedar latch yep

#

but it can be finetuned

cedar latch Jun 4, 2020, 5:43 AM

#

its it better?