dev-contrib

[desert vessel]

im not able to commit nor stash changes nor pull

and i have an open pr

[tawdry vapor]

What happens when you try? And which tool are you using (e.g. git CLI, GitHub Desktop, through your IDE, or something else)

[desert vessel]

im using cli

the linitng broke the commit system that i had

i recloned and it works now 👍

[vocal wolf]

@vale ibex nice branch name lol

[severe tangle]

Lmao

[tawdry vapor]

Yes. A lack of bug fixes and support for breaking API changes may affect it

[desert vessel]

@vale ibex my existing pr is taking longer than i anticipated to be approved to rejected, can i wait for it to be resolved ?

so i can start on the issue you assigned me to

[gritty wind]

If you think you can manage it, you can work on multiple things at the same time, otherwise feel free to take as long as you need

[brisk brook]

Can you link it, I could take another look and review, might speed things along

[desert vessel]

Sure! Tysm!! here -> https://github.com/python-discord/sir-lancebot/pull/822

[brisk brook]

Ah I see, that is not very old. It usually takes a few days

@crude gyro where did you find the snakes introduced here? https://github.com/python-discord/bot/blob/9f5f0cef5000f741e22734d8537683d1c607995b/bot/utils/snakes/hatching.py (https://github.com/python-discord/bot/pull/78/files)

[cold island]

iirc it was part of the first code jam

And the top teams got their cogs added to the bot

[crude gyro]

that is correct

[frozen galleon]

May I ask here what will be happening with the bot since discord.py will no longer be maintained and was archived? Do you plan on switching to another working fork once it's made or another language rewrite? Or is there something else you plan on doing?

[cold island]

Currently discord.py is working, it just won't receive updates anymore. As time progresses, the featureset and the API we use will become outdated, but currently it's not urgent for us to switch.

[crude gyro]

people are getting a bit carried away after that announcement. There's no reason to panic.

[frozen galleon]

yeah, I do realize that, I'm just curious if you have any plans about it, since it will become necessary to do something eventually

[crude gyro]

by the time it actually affects us the market will look completely different. Putting together a plan right now would be a waste of time.

[cold island]

Well, there are lot of people who like both Python and discord bots, and will want to try their hand at writing a wrapper for the API, or at least maintain a fork of dpy. When the dust will settle we'll see.

[crude gyro]

we're not worried, you shouldn't be either.

[cold island]

Is there a way to have an optional string argument to a command, and have it not match in favor of an argument that comes afterwards?

For example if I have two arguments Optional[str] and Optional[TextChannel], if I input #algos-and-data-structs I want it to be in the second argument

[cold island]

Ugh, I guess I can force a specific pattern in the string

[brisk brook]

I don't think discord.py has that, though it has the inverse which is Greedy

[gritty wind]

I implemented something like this for the silence command

Basically, it’ll accept one of two data types as the first argument, both of which can function on their own. If someone wants to pass it both, there’s also a second argument which will the accept the other type

The command then tries to take in all the provided input, and create meaningful context from that

[cold island]

The issue is that a channel mention is a valid input for the string argument

[gritty wind]

What are you working on?

[cold island]

The clean command

[gritty wind]

What’s the desired interface?

[cold island]

I want to take both a regex pattern and a one of (a series of channels | an asterisk to denote all channels)

Since I want the asterisk, I can't use Greedy

So I used a positional only argument instead

So channels must be the last arg

And both the pattern and the channels are optional

[gritty wind]

So like
!clean [regex pattern] [channel or *]

?

[cold island]

yes

[gritty wind]

Hm

My recommendation is to drop the star, and instead make it its own command

!cleanall [regex]

I can’t think of a clean way to do this

[cold island]

I solved it by enforcing a pattern of r'<pattern>'

[cold island]

That already exists, but I created a command that combines all functionalities

None of this would be an issue if Greedy was a type instead of an object

Since then I would just do Union[Greedy[TextChannel], Literal["*"]]

But I'll open the PR and we can scratch our heads once you see everything in context

[brisk brook]

string = "A_B*[ <0-9 [#dev-contrib](/guild/267624335836053506/channel/635950537262759947/)"
regex = ""
converter = discord.ext.commands.TextChannelConverter()

for arg in string.split(" "):
    try:
        channel = converter.convert(arg)
    except BadArgument:
        regex += arg
    else:
        break
else:
    # We exited the loop without breaking, meaning we didn't resolve a channel
    raise BadArgument("Could not resolve a channel mention")

# We now have access to `channel` as well as `regex`:
# channel: "[#dev-contrib](/guild/267624335836053506/channel/635950537262759947/)"
# regex: "A_B*[ <0-9"

What about something like that?

@cold island ^

Hmm, right, we can't make a converter out of this because discord.py doesn't let you hook into the argument system like Greedy does

[desert vessel]

@brisk brook, is this good enough?

[brisk brook]

Hmmm, I am not sure. Do we want the text? Without it the logo won't make sense, but ehhh..

Can you post that in the PR?

[cold island]

@brisk brook

[molten perch]

I managed to make a converter(It's a wild idea, and implementation) , combined with Greedy, (Yes, I know it sounds insane) and it seems to work. Apart from two things. If in the case of!clean [regex] is provided then it'll be a List of 1 (It's always, assured, because of Greedy, but that shouldn't be a huge problem.). Also, another thing. It assumes that you provide only one regex string, otherwise it won't work as expected, but could be solved with additional checks.
https://paste.pythondiscord.com/ofojiwuyuf.py
(Also.. I apologise in advance 😄 )

[desert vessel]

done :)

we are following the py joke

[brisk brook]

Is there only one channel after that argument?

[desert vessel]

where baby python gives the py logo

[cold island]

This has the same issue with Bluenix's suggestion

[cold island]

No, there can be any number of channels

[desert vessel]

i can crop the pic to only the logo if you'd want

[brisk brook]

So wait, a channel is valid input for the regex but there can also be multiple channels?

[cold island]

yes

🙂

[desert vessel]

huhhhh, kinda doesnt make sense

how can there be multiple channels as an input tho-

[brisk brook]

Oh, well then the only option is "myregex[#the-pydis-mass-layoffs-lounge](/guild/267624335836053506/channel/464905259261755392/)" I see now

[desert vessel]

sup bluenix and zig btw

[brisk brook]

So you can search for channel mentions, that you tell to search in multiple channels

[cold island]

Yes, I went with this

[desert vessel]

ah-

[brisk brook]

!command <search that may contain a channel> <*channels>

!command fwefwefew #878694948810137621 #dev-log

Is that 2 channels, or channel in the regex?

[cold island]

Lol yes, the more we discuss it the more I see that there has to be something that separates the regex

[desert vessel]

2 channels i suppose,

let me try 🤔

[cold island]

Even if I could use Greedy on the channels it would have the same issue

[cold island]

Specifically this is two channels

Because a string argument breaks once a space is hit

You have to include quotes if you want spaces

[brisk brook]

Right, but in our imaginary convert situation

[cold island]

Undefined

[molten perch]

Which issue? Assuming that only one regex is provided?

[cold island]

That the regex itself can be a channel mention

[brisk brook]

Exactly, that's the issue I was explaining to @desert vessel

[molten perch]

So, in that case the channel mention would be the first and the second argument as well? 😄
I thought, when no regex is provided you want the channel mention to be the second argument.

[desert vessel]

ah

[cold island]

Yes but how do you decide that no regex is provided when you give a channel mention, if the regex can be the channel mention

[desert vessel]

can someone help with with git auth ;-;

i got a new computer

and now git got screwed

how do i login from cli-

[molten perch]

Well, in my implementation when you provide a channel mention the regex argument will return None, and the second argument will automatically be a TextChannel object, so you can just put a check if the regex argument is None, or not. !clean [channel]

[cold island]

Yes but that's not the desired behavior. The regex can be a channel mention.

Which is why I ended up enforcing a specific form the regex needs to be specified in

[molten perch]

Alright, I might have misinterpreted your request. So in case of a command of !clean [channel_mention] what is the desired behaviour? That the first argument [regex] automatically qualifier as a textchannel, and returns an object accordingly, but I'm not sure what happens with the channel second argument.

[cold island]

The request is not possible. You can't have two optional arguments that can have the same value and skip the first argument

So you have to make the possible values for each argument different

[fallen patrol]

@desert vessel are you trying to authenticate with github?

[clever chasm]

P

[desert vessel]

done :)

[slow steppe]

On line 56 of nominations.py (in @brisk belfry) why not just get a role object from constants.Roles.mod_team and use role.mention? Sure the current option works but it's usually not very good practice

[austere hornet]

Lol

[desert vessel]

true :( most people maybe confused with just the logo but- idk

what are your inputs ?

[tawdry vapor]

Why did I do this... https://github.com/python-discord/site/pull/149/commits/007232649ed2b3eed2684a6291aa3bfa57c23027

Maybe the data that needed to be migrated had nullable reasons

The schema was nullable
https://github.com/python-discord/site/blob/flask-archive/pysite/views/api/bot/infractions.py#L137-L145

[stable mountain]

pysite/views/api/bot/infractions.py lines 137 to 145

CREATE_INFRACTION_SCHEMA = Schema({
    "type": lambda tp: tp in INFRACTION_TYPES,
    "reason": Or(str, None),
    "user_id": str,  # Discord user ID
    "actor_id": str,  # Discord user ID
    Optional("duration"): str,  # If not provided, may imply permanence depending on the infraction
    Optional("hidden"): bool,
    Optional("expand"): bool
})```

[tawdry vapor]

But the post request always expected a reason https://github.com/python-discord/site/blob/flask-archive/pysite/views/api/bot/infractions.py#L189-L195

[stable mountain]

pysite/views/api/bot/infractions.py lines 189 to 195

def post(self, data):
    deactivate_infraction_query = None

    infraction_type = data["type"]
    user_id = data["user_id"]
    actor_id = data["actor_id"]
    reason = data["reason"]```

[tawdry vapor]

Maybe it was to support data from rowboat?

Are there actually any infractions with null reasons in the db now?

[patent pivot]

!remind 8h this

[stable mountain]

Aye aye, cap'n!

Your reminder will arrive on <t:1630326005:F>!

ID: 2937

[gray geode]

error running snekbox on windows

should i try wsl

[short snow]

I don't see any error in the screenshot

And it should work perfectly fine on windows if i am not wrong

[gray geode]

gunicorn.error

didn't show up

i mean its running

but status

[short snow]

Umm, it does show gunicorn.error but its running, does the api work?

[gray geode]

umm ye it shows up

have you setup snekbox on windows ?

before

[short snow]

Nope, no I mean does the using the snekbox work like with requesting it

[gray geode]

umm , how can i test the api

http://localhost:8060/eval

this is the endpoint iirc

welp

imma try on my wsl anyway

it has some problems on windows ig

[short snow]

using simple requests.post, https://github.com/python-discord/bot/blob/b0ea616add8f6122ba5b98b348a4fa5d41774d21/bot/exts/utils/snekbox.py#L59..L64

[stable mountain]

bot/exts/utils/snekbox.py lines 59 to 64

async def post_eval(self, code: str) -> dict:
    """Send a POST request to the Snekbox API to evaluate code and return the results."""
    url = URLs.snekbox_eval_api
    data = {"input": code}
    async with self.bot.http_session.post(url, json=data, raise_for_status=True) as resp:
        return await resp.json()```

[short snow]

I think scal runs it on windows

[brisk brook]

I run it on Windows as well, though I haven't played around with it

[gritty wind]

Gunicon.error is the default logging handler

Which is a terrible name that I don’t understand

It’s not an error though

[green oriole]

@brisk belfry randomcommand

[brisk belfry]

Sorry, an unexpected error occurred. Please let us know!

CommandNotFound: Command "randomcommand" is not found

[green oriole]

We should fix that haha

[tawdry vapor]

@brisk belfry `@tawdry vapor

[brisk belfry]

Sorry, an unexpected error occurred. Please let us know!

CommandNotFound: Command "`@tawdry vapor" is not found

[tawdry vapor]

Okay at least that is fine

[desert vessel]

@tawdry vapor is there a repo for this bot that is maintained or is it closed source? i could help fix it while i wait for my PR to get approved or decline :)

[vale ibex]

&```@tawdry vapor

cool, that's fixed

[vale ibex]

&src

[brisk belfry]

Sir Threadevere's GitHub Repository

Repository

Go to GitHub

[desert vessel]

thanks!

[desert vessel]

btw im still waiting for something to happen to the pr, im really sorry chris :( i'll start as soon as something happens to that pr

[vale ibex]

There's no rush with any of our issues 🙂 Things take time

[desert vessel]

ah alright :)

i can work on another branch but im known for screwing the entire repo trying to make a branch

so nope

[vocal wolf]

@vale ibex it's dev time

I got my environment up and running on macos

you ready to review stuff?

[vale ibex]

Depends what you want reviewed lol I've got 30 minutes

[vale ibex]

macos?

[vocal wolf]

yes

with the power of the m1

[vale ibex]

Did manjaro burn you?

[vocal wolf]

nay, just needed a laptop

I still have arch/windows on my main rig

[vale ibex]

Ahh cool

[vocal wolf]

@vale ibex here https://github.com/python-discord/bot/pull/1790

das 4 u

[vale ibex]

60 lines

this'll take me hours

[vocal wolf]

[vale ibex]

[vocal wolf]

lmao

wtf

site not building because I don't have gcc?

[vale ibex]

lol what

Do you know which dep?

I'm guessing they don't have mac wheels

Also, when did GitHub change their approve colour?

These are definetely more faded that I'm used to

[vocal wolf]

@vale ibex they be changing stuff again

[short snow]

yeah

[vocal wolf]

@timid sentinel thanks for review

[short snow]

even discord i think, atleast i see colours being more darker today

[desert vessel]

i have an m1, i can help :D

call me anytime and i'll be there

heh

[vocal wolf]

what for?

[desert vessel]

^

[vocal wolf]

just general questions about the m1?

ah

[desert vessel]

oof i cant send a screenshot here

[vocal wolf]

I've re-figured out macos, I shall be fine. Thanks!

[desert vessel]

oh wait i can

[desert vessel]

oh, ciik

cool*

i reinstalled macOs today

a py model broke macOs yesterday

lmao

Chris, i can start with the pr now

@vale ibex are you guys accepting pr's for sir-threadaveare too ??

[vocal wolf]

@vale ibex should I close this PR lol it's been literally 2 years

https://github.com/python-discord/sir-lancebot/pull/281

[vocal wolf]

yes.

[desert vessel]

oh alright, ill pr a better help message lol

[vocal wolf]

not yet

[desert vessel]

ah

[vocal wolf]

we haven't approved your issue

[desert vessel]

kk

like ater that i meant-

[vale ibex]

Hmmmm, I think the recent timestamps may change this too

[desert vessel]

after*

[vale ibex]

oh, lol @vocal wolf

[vocal wolf]

?

[vale ibex]

I commented 10s after you

[vocal wolf]

lmfao

[vale ibex]

alr breakfast time

later

[vocal wolf]

cya

[desert vessel]

byee

i just realised that we are on diff parts of the world

its 5 pm for me heh

[brisk brook]

No Chris is just weird

[desert vessel]

xDD

[brisk brook]

It should be like 10 am there

[desert vessel]

ah

&help

[brisk belfry]

BotSource:
  source Display information and a GitHub link to the source code of a comman...
​No Category:
  help   Shows this message

Type &help command for more info on a command.
You can also type &help category for more info on a category.

[desert vessel]

yikes, wrong channel

[brisk brook]

Damm Daniel, back with the default help command

[desert vessel]

thats why i made an issue-

[vocal wolf]

k I'm closing this 2 year old PR

[desert vessel]

Finally-

[brisk brook]

Xith, things take time, stop rushing things

[vocal wolf]

lmfao

you have no idea

I once updated a PR and 3k commits merged into it

#dev-log message

that's when I found that the webhook caps at 1k commits

https://github.com/python-discord/bot/compare/9a00d703c1b8...01406f86ad9b

[vocal wolf]

@patent pivot

[desert vessel]

/Users/urmom/Library/Caches/pypoetry/virtualenvs/thread-bot-CQnFdfHJ-py3.9/bin/python: No module named thread_bot

help :)

im running poetry run task start in sir-threadevere

[gritty wind]

Which directory are you in?

[desert vessel]

inside the sir-threadevere dir?

[gritty wind]

Never mind, the task is just wrong

[desert vessel]

oof

[gritty wind]

Use this instead:
poetry run python -m bot

[desert vessel]

thanks :)

time to debug yikes

[vocal wolf]

@vale ibex @timid sentinel thanks y'all for the reviews and stuff. I go bed now.

[desert vessel]

discord.errors.HTTPException: 400 Bad Request (error code: 50035): Invalid Form Bod

[desert vessel]

good night :)

[vale ibex]

I think the issue was likely debug being false, could you pull from main and try now?

[desert vessel]

yeet, alright

[desert vessel]

yikes-

its something to do with env i believe

[vale ibex]

Did you update your config.yml to the right values for your server?

[desert vessel]

yee

[vale ibex]

the last 3 role IDs are the same

[desert vessel]

wait does that matter

its just me in the server, so im using the same role

but i'll change it and try again

[vale ibex]

yea, we make a list rather than a set, so it can contain duplicates

[desert vessel]

ah

kk

gg,it works

what are extensions exactly ??

[vale ibex]

those are the cogs loaded

[desert vessel]

oh

last thing : we may have to clean this message up

its not making threads smh

[vale ibex]

What are you trying to do?

[vale ibex]

&ext load asdasd

[brisk belfry]

:x: Could not find the extension asdasd.

[vale ibex]

fixed

[short snow]

even your discord colours become darker? I thought there is some problem with my discord again 😛

[desert vessel]

Coil

Cool*

[short snow]

whats the zero width space code?

[vale ibex]

200B

[short snow]

Ok thanks, what I am doing wrong here?

question = question.replace(word, len(word) * "*\u200B")

[brazen charm]

What's the goal?

Looks like it could be the Markdown or something

[short snow]

to replace it would the same amount of *

so to escape the markdown, I added zero width spaces

[brazen charm]

Why zwsp instead of escaping it normally?

[short snow]

🤷‍♂️ I just went ahead with that

ok normal escaping works, nice

[dim pelican]

Regarding Sir-Lancebot#610, do I need to link a new branch to the PR?

Since the original PR has a different author I am thinking I need to set up a new remote based on their current branch, make a local branch / clone from that and then push my commits to my local branch. Then link my local branch to the PR. Does this sound sorta correct?

[dusky shore]

GitHub

Teapot support

[stable mountain]

It has arrived!

Here's your reminder: this
[Jump back to when you created the reminder](#dev-contrib message)

[brisk brook]

What you'll need to do is set up their fork as a remote yes, then create a local branch and pull their commits into it. Now you can fix what's necessary and open a new PR.

[dim pelican]

.bm #dev-contrib message Teapot PR

Ahh shucks

[desert vessel]

ye, whitelisting all the channels is in progress i believe

[dim pelican]

Yeah I guess I thought that was already merged lol

[desert vessel]

wth

[fervent sage]

hmmm

[desert vessel]

why is the server hype on the number in the namesss ?

hehe

[dim pelican]

sir-lancebot#816

[dusky shore]

GitHub

Whitelist .bm command in occupied help channels

[thorny obsidian]

It was us making fun of k8s during yesterday's staff meeting

[fervent sage]

im too lazy to change mine back

[desert vessel]

xD

[desert vessel]

lol

btw

did u guys take a look at sir-threadevere ?

[fervent sage]

@vale ibex slight issue with this being merged, it wasnt done correctly

akarys said to remove the check entirely so it could be used everywhere, but that never happened

[desert vessel]

https://github.com/python-discord/sir-lancebot/issues/827

thats this

[fervent sage]

ah

[desert vessel]

its open

im just waiting for the other dude to accept it or tell no

i cant pr cuz chris asked him first

[vale ibex]

Yea, the PR met the requirements so I merged it

no point pausing that one just to wait for the new reqs to be done

[desert vessel]

yep

[vale ibex]

just removing the check won't work, since the bot has a global check, so there's a bit more to it than that

[gritty wind]

I'll never understand how discord managed to make such a terrible API. Why is the type """null""". I can't actually get the API to give me that field, so what I'm assuming is you only get that field if the object has premium, otherwise you don't get it.

You know what the sane thing would've been? Just using a fucking boolean

Now I've gotta figure out how to represent this with pydantic models

[brazen charm]

could be a not so great way of cutting down the size

[brisk brook]

Yeah I remember asking someone about, and that's it yeah

[fervent sage]

well they send last_pin_timestamp as a date string in channel objects

so size wouldnt be the consideration

@celest charm and myself have been using the api docs a lot over the last 2 days and i think our collective conclusion is "wtf"

[brazen charm]

is a date string larger than a timestamp when text encoded?

[fervent sage]

no clue

but it also includes the id of the pin

and IDs are just timestamps on steroids

[brazen charm]

ah then yeah that just seems like bad design if you can get it anyway

[fervent sage]

user objects also have 2 values, i cant remember their names, which is the colour of their profile as an int, and then then the colour of their profile but a #hex string

including the #

[celest charm]

I can confirm that the Discord API is absolute nonsense and poop and bad and so on

okay maybe I exaggerated a bit

[fervent sage]

i mean

A large portion of it is next level "seriously why the fuck?"

[patent pivot]

yeah turns out there are

[brazen charm]

How often does the api change? I could see those as bad decisions early on and then keeping backwards compatibility but that could be removed when the api changes

[fervent sage]

we've gone from v6 to v9 in a year

with v10 soon™️

[patent pivot]

well, v7 was skipped

v8 didn't make a load of changes

[fervent sage]

it wasnt skipped

just in the client

bot devs still went "ooh shiny" and used it

[patent pivot]

it was skipped from a documentation point of view

[celest charm]

For example, this is the breakdown of the Channel object into subtypes. There's nothing wrong with the API per se, but the way it's documented (just a giant list of optional fields) is beyond horrible
https://paste.pythondiscord.com/raw/nujupuwoke

[fervent sage]

sometimes the api docs just completely neglect to include any sort of info about what an endpoint returns at all, so i guess we should be glad for what we do get

[brisk brook]

Wait that to the right is your notes?

[celest charm]

yeah

[brisk brook]

Ah okay

[thorny obsidian]

lmao wait, we have null notes and warnings?

[celest charm]

these are the contractions of these, minus group DMs

[patent pivot]

lmfao yeah apparently so

[cold island]

!note 212644551926611969

[stable mountain]

:ok_hand: applied note to @thorny obsidian.

[celest charm]

!note @cold island

[stable mountain]

:ok_hand: applied note to @cold island.

[fervent sage]

can i be null noted 🥺

[celest charm]

we don't joke infract helpers

[fervent sage]

its a note not an infract

[thorny obsidian]

we don't joke infract non-mods/admins

[fervent sage]

the note could just mean "h" but expressed via null-ness

[celest charm]

!h @fervent sage

[patent pivot]

oh yeah it could be joke infrs

[fervent sage]

wait we dont have an h command I must create an issue

[gritty wind]

here is another nice idiocy. Left is the name, right is the description. So I thought: cool, let's look at the hyperlink

all of these are in fact delivered

[cold island]

I find it weird that we can note outside of mod channels

[fervent sage]

class h(Cog):
  @command(name="h")
  async def h(self, ctx) -> None:
    resp = await session.get("https://h.pydis.com")
    await ctx.send(await resp.text())

[gritty wind]

Failed linting, banned

[brisk brook]

Always though 🤔

[celest charm]

Yeah, let's add pydis-style docstrings

class h(Cog):
  """The h cog"""

  @command(name="h")
  async def h(self, ctx: Context) -> None:
    """Perform the h"""
    resp = await session.get("https://h.pydis.com")
    await ctx.send(await resp.text())

[gritty wind]

There is no way to see where a model is used, but for this endpoint, yup

Still failed linting

Class name should start with a capital

[celest charm]

oh right

[celest charm]

class h(Cog):  # noqa
  """The h cog"""

  @command(name="h")
  async def h(self, ctx: Context) -> None:
    """Perform the h"""
    resp = await session.get("https://h.pydis.com")
    await ctx.send(await resp.text())

[cold island]

Also no period at the end of the docstring

[gritty wind]

Violated pre-commit

no blanket noqas

[celest charm]

I give up

[gritty wind]

lol

[vale ibex]

NameError: name 'session' is not defined

[gritty wind]

Pydis Roast™️ but it's just code

[vale ibex]

New event idea, spend an hour going through my PRs and roasting my code

[dim pelican]

I don't want to interrupt the discussion since its awesome, but how would I go about getting the .rp search command enabled in more channels than just #bot-commands ?

[timid sentinel]

Is it just me or have backticks for codeblocks started rendering really weirdly in github, e.g. typing ```hello``` looks like this

[gritty wind]

You can use the whitelist decorator, but have you gotten any other channels approved?

[brazen charm]

there should be a decorator to override the whitelist global check

[timid sentinel]

shows fine in the preview though

[gritty wind]

link wookie?

[dim pelican]

No, this is the first time I'm bringing it up

[timid sentinel]

it's only when i'm typing it, looks fine when I send it

[gritty wind]

Usually we don't want to enable these commands everwhere

For instance, SO command was deliberately disabled in help channels

[dim pelican]

Yeah that could get really spammy

[gritty wind]

Yeah, and because we don't want people just going into a help channel, copying the question into the command, and leaving

If someone wanted google, they could go to google

What channels did you have in mind?

[gritty wind]

I couldn't replicate, but for me it doesn't format at all, unless I look at the preview, or send it

[dim pelican]

Ot, help and general. But there is a lot of spam potential for 2/3 of those...so maybe its fine the way it is. Just wanted a soundboard to bounce the question off someone else

[gritty wind]

OT sounds fine, but I'd be against help channels and general

[thorny obsidian]

I'm going to advocate specifically for help channels, that's where RP could help the most

Also, I'd like to keep a "enable everywhere (except pygen) until proved noisy)" policy

[gritty wind]

I believe it'll have the same problems that we discussed when opting to not enable the stackoverflow command

[brisk brook]

I think that's different though?

[brisk brook]

StackOverflow caused this ^

[thorny obsidian]

real python is not a "just search your answer here", they are tailored articles to introduce subjects

It's a fundamentally different service

[brisk brook]

But I can see legitimate uses

For example

!decorators

[stable mountain]

Decorators

A decorator is a function that modifies another function.

Consider the following example of a timer decorator:

>>> import time
>>> def timer(f):
...     def inner(*args, **kwargs):
...         start = time.time()
...         result = f(*args, **kwargs)
...         print('Time elapsed:', time.time() - start)
...         return result
...     return inner
...
>>> @timer
... def slow(delay=1):
...     time.sleep(delay)
...     return 'Finished!'
...
>>> print(slow())
Time elapsed: 1.0011568069458008
Finished!
>>> print(slow(3))
Time elapsed: 3.000307321548462
Finished!

More information:
• Corey Schafer's video on decorators
• Real python article

[gritty wind]

But why would you need to search it if you are giving someone something specific

If you have an article you think someone should read, you can link it

[thorny obsidian]

Like, if someone wanted to ask for async help like "should I use threads or async?" what would make it faster?"
then I would do .realpython concurrency and bring up the articles already

.realpython concurrency

[dusky shore]

Search results - Real Python

Here are the top 5 results:

Speed Up Your Python Program With Concurrency

https://realpython.com/python-concurrency/

Python Concurrency

https://realpython.com/quizzes/python-concurrency/

Speed Up Python With Concurrency

https://realpython.com/courses/speed-python-concurrency/

Async IO in Python: A Complete Walkthrough

https://realpython.com/async-io-python/

Hands-On Python 3 Concurrency With the asyncio Module

https://realpython.com/courses/python-3-concurrency-asyncio-module/

Click the links to go to the articles.

[brazen charm]

what's the interface of the command? If it's a search command that you have to interact with then I don't think it fits into a help session

[thorny obsidian]

So it's right there, and they can see what's available

[gritty wind]

I don't see how this won't be abused

As the person that asked the question

This has gotten me no where

You've just given me 5 random links

No idea if you've vetted them or not

[thorny obsidian]

"Hey, check out this real python article like the 1st and 3rd, they'll explain what you need"

[gritty wind]

No idea if they will answer my question

If you already know the articles

Why do you need a search

[brisk brook]

I do agree that someone may ask to get a personal explanation

[thorny obsidian]

Because it's easier than exiting out of discord, doing the search on google, then coming back and linking it

[brazen charm]

Doing that in a separate channel and then sending the appropriate link themselves seems like it'd make more sense to me

[thorny obsidian]

Because why have this only be available in bot-commands. Then it's the same as not having it at all and just using google to me.

[gritty wind]

I don't think that's true, especially considering that your topic may not be in the top 5, especially for a really broad term

Someone said that exact same thing for SO

That exact same argument

[thorny obsidian]

RealPython is a fundamentally different service than SO. I don't think the same arguments hold.

[gritty wind]

But we still pushed on through with it, because we had people that wanted to implement it, and we weren't harmed by having it

I don't see how that's true. Sure, the format of info is clearly different, but fundamentally it's just giving you a bunch of links to random text

I can make your exact same argument for SO

[thorny obsidian]

I mean RealPython itself, not the command.

[gritty wind]

Q: "How do I solve file not found?"
A: SO link to answer that explains it

[thorny obsidian]

RealPython are tailored articles, written with a specific purpose in mind. So if you're searching through it to link something, it's far more than just a "plug in your question into SO and nothing else"

[brisk brook]

Q: How does decorators work?
A: Tag that explains it

[thorny obsidian]

It's a "hey, this service has written this better than I could, take a read through this"

[brazen charm]

half of the links from the search above aren't even articles that you could easily make sure are correct for the given question

[gritty wind]

Not sure that having "tailored articles" helps if the person sharing them didn't read and vet them. The same is also true of SO in the exact same way.

An SO answer can technically answer something much better than a helper here

[brisk brook]

Just noticed yeah wow. Gonna open an issue for that, I think we only want articles

[gritty wind]

And linking to that answer can help someone understand something

[gritty wind]

The difference here is you can't take a random keyword and throw it into a search command

Which both SO and RP are

[thorny obsidian]

The .so search command abuse was because people would just assume it's a normal stack overflow search and plug in the question wholesale. I don't think people would do that with RealPython because it doesn't strive to answer broad questions.

[gritty wind]

You need to know what the tags are, and you need to know their content

[thorny obsidian]

I don't think the avenue for using the command against the spirit of the server is there enough to warrant pre-emptively denying it in help channels.

[gritty wind]

We could put it up to a vote, but I believe the same fundamental problems that exist with SO, will exist here

[dim pelican]

I figured it would reduce the friction for looking for more reading / info on a given topic. But if every search returns 5 links I don't know if it is appropriate for a lot of channels. But I do like the discussion here, you guys are awesome

[gritty wind]

Where was that spirit for the SO command :P

Regardless, I'm not willing to run a trial by fire on help channels for no reason

[thorny obsidian]

Because as I've said above, I think SO and RP fulfill different needs. SO opens itself up for abuse while imo RP doesn't because of the difference of the services themselves.

[gritty wind]

and fwiw

.RP should I use threads or async?

what's the actual command?

.realpython should I use threads or async?

[dusky shore]

Search results - Real Python

Here are the top 5 results:

Speed Up Python With Concurrency › Speed Up Python With Concurrency (Overview)

https://realpython.com/lessons/python-concurrency-overview/

Speed Up Python With Concurrency › Speed Up Python With Concurrency (Summary)

https://realpython.com/lessons/python-concurrency-summary/

Speed Up Python With Concurrency › CPU Bound Workloads

https://realpython.com/lessons/cpu-bound-workloads/

Speed Up Python With Concurrency › Multi-Processing Version

https://realpython.com/lessons/multi-processing-version/

Hands-On Python 3 Concurrency With the asyncio Module › What Is Concurrency?

https://realpython.com/lessons/what-concurrency/

Click the links to go to the articles.

[gritty wind]

Just fed it that question

[thorny obsidian]

But who would think to do that with RP?

It's article-based, not question-answer-based

[brazen charm]

what does doing it in a help session offer over doing it in in bot commands and then posting a single link in the session offer?

There's no rush to get it to the person that's being helped, and is less confusing

[gritty wind]

Don't know why that distinction matters? RP has so many articles you can probably stick any question that you can in SO and get a result

[thorny obsidian]

Because why introduce that extra friction that would essentially mean it won't really see it being used because people don't know it exists. It's just burying this command.

[brisk brook]

But that doesn't even look like it will answer the question you gave it

[thorny obsidian]

Because of the expectations of the user. No one's going to think to do with Real Python because the service doesn't lend itself that way.

[gritty wind]

Right, which is the concern. People getting links thrown at them, which may or may answer their question, and which may or may not waste their time. The same concerns raised for SO

[gritty wind]

I don't think that's true

[green oriole]

What's the context?

[gritty wind]

The main RP site has a search bar, people must be using it, right?

[green oriole]

I actually never used it

Always google search

[dim pelican]

I started it: #dev-contrib message

Should the .rp command be used in more channels than #bot-commands

[gritty wind]

Enabling the realpython command in help channels. I believe it's problematic for the same reasons SO was.

[gritty wind]

Same here lol

[thorny obsidian]

I'll step out, but I just don't like us pre-emptively denying usage of useful commands without a strong reason. I don't think "people could abuse the .realpython search command by just plugging it in wholesale" is enough of one. When I search on RP, I search for general terms to find specific articles I have in mind.

[timid sentinel]

If the disagreement is just over whether people will abuse the command and we do think there's a valid use case, could we not just try it out and see if that's the case. Would be easy enough to revert?

[green oriole]

Hmm, that's a complex one since it is different from SO. I would say we can trial it

[fervent sage]

Yeah, you dont know how people will use it before they use it :P

[brazen charm]

even if it's not abused, imo the UX for the person that's being helped is not great

[gritty wind]

Because it can hurt the experience for people that do experience that abuse. I don't think it's utility is great enough to warrant doing that

[green oriole]

But I'm afraid it will be the same problem

[dim pelican]

I think we all agree that there is a potential for abuse though, just different strategies for how to avoid that?

[gritty wind]

And like Numerlor mentioned here and earlier

[gritty wind]

^

[green oriole]

There is a difference between "Hey, look at that article: $link" and dropping a search command in front of the user

Same problem as SO

[thorny obsidian]

Did we stop SO from being used in help channels because of bad UX or because we think people will just drop a "I dunno, search SO <search and leave>"

[gritty wind]

We could put it up for a vote, but I'm not in favor

[gritty wind]

Both

People can do both

[thorny obsidian]

Where is that documented?

[gritty wind]

It isn't, because we don't document everything

It's in this channel

[green oriole]

But we don't know that's actually what will happen. Our policy is to trial more perms and reduce if it problematic.

[timid sentinel]

The main good use case I would see is when you see a common topic, for which you know a realpython article exists, and just want to reference it, e.g. "for more on that you can check out this realpython article on lists, !realpython lists, take a look at the first link there".

[gritty wind]

^

[thorny obsidian]

Because I thought our policy was a "we'll enable everywhere then limit if it becomes problematic" and this feels like a significant departure from that policy. So it'd be nice if that reasoning and discussion was documented somewhere.

[gritty wind]

My response to that:
<#dev-contrib message>

[green oriole]

The problem is that it will give a lot of hits, there isn't a single article talking about lists but many of them.

[gritty wind]

I'm willing to trial a lot of things, but in my mind I don't see all that many great use cases for this command, and I think it can be a significant harm.

That combination, along with one of our most core systems

Yeah... don't like it

[dim pelican]

Just spitballing, what if it returned a single link based on which channel it is in? Could lead to more problems though

[timid sentinel]

.realpython lists

[dusky shore]

Search results - Real Python

Here are the top 5 results:

Linked Lists in Python: An Introduction

https://realpython.com/linked-lists-python/

Reverse Python Lists: Beyond .reverse() and reversed()

https://realpython.com/python-reverse-list/

Lists and Tuples in Python

https://realpython.com/python-lists-tuples/

Using Python Optional Arguments When Defining Functions

https://realpython.com/python-optional-arguments/

Python Lists and Tuples

https://realpython.com/quizzes/python-lists-tuples/

Click the links to go to the articles.

[timid sentinel]

Third one there works

[green oriole]

Anyway, this conversation seems quite heated and we aren't very much listening to each other, how about we solve this with a poll? We can always go back and change it later

[gritty wind]

With the ever changing nature of articles (new articles come, old go), wouldn't it make more sense for you to vet them first and return the one you want

[dim pelican]

I'm good with keeping it where it is, just figured I'd ask how to go about having it elsewhere 😅 First contribution so I guess I would like to use it more places haha

[timid sentinel]

From what i've seen of the command the results have been pretty relevant to the search query. Most of the problems being described here are ones that could be confirmed or (whatever the opposite of confirmed is) by trying it out. I think the harm of trying it out has been overstated here

[patent pivot]

@short snow I believe Dorsan is waiting for Sebastiaan

#dev-contrib message

[short snow]

for his approval?

he has approved it iirc, lemme find the link

https://github.com/python-discord/api/issues/7#issuecomment-894690425 counts as a approval?

[molten perch]

Well, I don't think so. I never got assigned :/

I've set up the environment but never generated the models.

[short snow]

Ayy wookie, fast boy

[dim pelican]

@brisk brook just made a PR for the issue you opened

sir-lancebot#830

[dusky shore]

GitHub

Rp article fix (Issue #828)

[tawdry vapor]

Thanks

[brisk brook]

Wait I don't get it, why does it add a completely new file?

[dim pelican]

Probably because I suck at git

[brazen charm]

looks like an old base branch

[dim pelican]

I tried to make a new branch for the fix, since I read that was a good way to handle separate issues

[brisk brook]

You may have based on the old branch then

[short snow]

wookie, should I make the hidden * bold?

[timid sentinel]

@short snow wdym by "replacing all punctuations would space could spoil words I guess", I think it would be OK if we just replace it in the title, maybe soemthing like

for word in re.split("\W+", title):
    extract = re.sub(word, "\*"*len(word), extract, flags=re.IGNORECASE)

[timid sentinel]

doing that sounds fine

[short snow]

like if any of the title has ' in the title, and then we remove it, like can't becomes cant but i think rapidfuzz can detect that

[timid sentinel]

oh hmm in that case my solution just there would remove all cans and ts from the results though which isn't great lol. I guess we could just not split by some characters like ' or ", lemme try something out.

[green oriole]

I mean, this is the rabbit hole we probably don't want to fell into haha

[short snow]

Ok rapidfuzz can guess that nice

we can ignore those

[green oriole]

This is natural language processing, fuzzy matching will of course not provide any great result

[timid sentinel]

true lol, would be hard to get something perfect 😄

[green oriole]

That, or we lock recruit @trim cradle in a cave to make the algorithm perfect

[trim cradle]

Wa

Natural language processing. Yes, I nominally know about that

[short snow]

wookie mail, co-authoring you

[brazen charm]

you can check a commit from them

[trim cradle]

Why is there wookie mail

[green oriole]

When co-authoring I just go to a random commit from the user and copy paste their author line

[timid sentinel]

I think we can just go for, split by characters that separate words, e.g. whitespace or -, then strip punctuation from the edges, and then sub

[green oriole]

For real though, @trim cradle, would you know any tips and tricks we could use to make the quiz command detect correct answers better than blindly fuzzy matching the message?

[dim pelican]

I think I figured it out. Tried to do the first one from the command line only, but the new one is just from "editing" the file from the github web page

[green oriole]

Is there like a better algorithm than fuzzy matching more adapted to this?

[timid sentinel]

for word in re.split("[\s-]", title):
    word = word.strip(string.punctuation)
    extract = re.sub(word, "\*"*len(word), extract, flags=re.IGNORECASE)

Ok that's perfect nobody can convince me otherwise 😎

[trim cradle]

@green oriole I can look into it later

[green oriole]

Sounds good, thank you

Have fun!

[timid sentinel]

All we really want to do is split a sentence into words so we can substitute them out of a text

[short snow]

how is \W different from \s

[timid sentinel]

hmm actually we should only sub out full words rather than parts of words

[short snow]

[timid sentinel]

\W is the same as [^a-zA-Z0-9_], \s is only whitespace

so basically \W contains punctuation, \s doesn't

[short snow]

oh

[green oriole]

W stands for word characters

[timid sentinel]

Ok,

for word in re.split("[\s-]", title):
    word = word.strip(string.punctuation)
    extract = re.sub(rf'\b{word}\b', "\*"*len(word), extract, flags=re.IGNORECASE)

[short snow]

@timid sentinel done

[desert vessel]

Heyyy wookie 😀

[vocal wolf]

alright

I'm awake

@trim cradle https://github.com/python-discord/bot/pull/1784#discussion_r697234127

[trim cradle]

8 is a literal, not a constant. But what feedback were you interested in for linking me to this?

[vocal wolf]

I'm not too sure, https://github.com/python-discord/bot/pull/1786#issuecomment-908505064

[trim cradle]

Unless they think anything is inaccurate about my proposed changes, I think they should just substitute them in verbatim.

"By adding variables as attributes to your bot" isn't misleading per se, but the variable itself is something that can be deleted after the object it refers to is given another reference via bot, but the actual object will still be there.

[vocal wolf]

I don't know their username on Discord unfortunately so we can't talk to them directly

unless

[brisk brook]

It's both haha

[vocal wolf]

oh right

[brisk brook]

@severe tangle

[vocal wolf]

@severe tangle discussion of the reviews on the previous version of your PR.

[severe tangle]

Hm?

[trim cradle]

I guess? is "Hello world!" a constant?

[brisk brook]

No ```py

a = "Hello world!"
b = "Hello world!"
a is b
False
x = 8
y = 8
x is y
True

[trim cradle]

8 is 8 might work coincidentally because of how certain integers are pre-allocated when the interpreter starts, but that's just an optimization. It's not part of the spec.

[brazen charm]

it is a constant in the same way 8 is, but constants is not really a term for objects in python beyond the few builtin constants

[trim cradle]

Yes, thank you for your contribution! Would you be open to substituting the changes I suggested verbatim?

[severe tangle]

sure

[brisk brook]

ah

[trim cradle]

!e print(10_000_000 is 10_000_000)

[stable mountain]

@trim cradle :white_check_mark: Your eval job has completed with return code 0.

001 | <string>:1: SyntaxWarning: "is" with a literal. Did you mean "=="?
002 | True

[trim cradle]

huh

[brazen charm]

immutable literals compiled together are fetched from code constants

[trim cradle]

dank

[severe tangle]

so, what do u want me to change in the md file now?

[trim cradle]

I made two suggestions in the original PR--if you can just copy them in, I will approve it.

Though it might be later today that I take another look

[severe tangle]

ah, ok. I will just accept them. That is fine with you?

[trim cradle]

I only left them as comments in the original PR. I can't go back and add them as proper suggested changes because that PR is closed.

[severe tangle]

Ah, I will manually do it then

[trim cradle]

Sorry about that. I don't really do code reviews very often except for language-related stuff.

[severe tangle]

ah that is fine

there

I have fixed that

[trim cradle]

I think I can take a look this evening

[severe tangle]

Sure

[vocal wolf]

@vale ibex lmfao that PR timed out

I tried trolling you and I got hit with karma

the god damn coverage report timed out

[vale ibex]

lmfao

leave the trolling to the professionals

[vocal wolf]

@vale ibex ok we're good

[vale ibex]

Nice

Now I have one for you @vocal wolf bot#1794

[dusky shore]

GitHub

Add concurrency rules to all GitHub workflows

[vocal wolf]

you didn't review for request

[vale ibex]

fixed

[vocal wolf]

@vale ibex have you tested this?

[vale ibex]

Yea, we've been testing it for weeks in lance too 😄

[vocal wolf]

ah

nais

then I see no reason to not approve

[vale ibex]

https://github.com/python-discord/sir-lancebot/blob/main/.github/workflows/build.yaml#L11-L13

[stable mountain]

.github/workflows/build.yaml lines 11 to 13

concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true```

[vocal wolf]

yah

approved

[vale ibex]

👌

[dim pelican]

What are your opinions on the .gitmessage templates? I could see them being useful for newbies like me

[brazen charm]

is that a template for commit messages?

[dim pelican]

Yup! I am trying one out locally that is kinda nice, includes the 50 and 72 character limit lines, place to include an issue etc

[patent pivot]

is it a standardised format?

ah I see, it pops up with git commit?

[dim pelican]

Yes

[patent pivot]

interesting

[dim pelican]

And the format can be standardized based on what you save it as. Just have to make sure that the config is set that way

https://alex-wasik.medium.com/create-a-custom-git-commit-template-84468232a459

This is mine right now:

# 50 characters                                 #
# Subject 
# (If implemented this commit will ___)

# 72 characters                                                       #
# Problem
# Problem, Task, Reason for Commit
# Solution
# Solution or List of Changes
# Note
# Special instructions, testing steps, rake, etc

Issue #

git config --global commit.template ~/.gitmessage.txt

[molten perch]

Hey! Could someone take a look at site#369 ? It seems to be a high priority issue. 🙂

[dusky shore]

GitHub

User's role field needs to have a cascade delete

[green oriole]

Is the fix you talked about one you developed during the Django codejam?

[molten perch]

I believe so, yes.

[green oriole]

Alright, sure, that can work

I don't have access to github from here though, do you wanna get assigned?

[molten perch]

Yes, I do! 😄

[green oriole]

Cool cool, if there's a staffer around who can do it

[brisk brook]

Just grab the GitHub Mobile app©®™ :p

[green oriole]

I need to log onto that device for at least three months by now

I just always forget about it

[fallen patrol]

can I get approval to remove fuzzywuzzy from here and switch to rapidfuzz?
https://github.com/python-discord/snekbox/blob/main/deployment.yaml

[tawdry vapor]

Something to note is that the GPL licence is most likely of no concern since we're not distributing snekbox with it. It's only installed in the production env.

But if you just prefer rapidfuzz's features over fuzzywuzzy then sure

[patent pivot]

hm, I think fuzzywuzzy is the wider used one

I'm fine with it as is — as Mark mentions there isn't a licensing concern

[fallen patrol]

so, would we switch it, or just add rapidfuzz?

[tawdry vapor]

You can just add both

I am personally not strict at all with adding new packages unless it's just for a meme or something, or would have no practical use given snekbox's restrictions (e.g. something for making network requests)

[fallen patrol]

right

[tawdry vapor]

Also there is something to be said about the development activity on rapidfuzz vs fuzzywuzzy

[fallen patrol]

yeah?

[patent pivot]

someone proposed adding scapy today which I'm for as well

[tawdry vapor]

I don't know what that does but if it doesn't go against what I wrote above then sure

[patent pivot]

building network packets

[fallen patrol]

seeing that toml is in there, tomli too :))

[thorny obsidian]

I'm not sure what the code does, but apparently it works without connection: #community-meta message

[patent pivot]

yeah it's just packet building with a pythonic interface

you use / to encapsulate IP in an eth frame and whatnot, neat for teaching networking

[patent pivot]

what does tomli grant that toml doesn't have?

[fallen patrol]

is compatiable with the 1.0 spec

toml only is compatiable with 0.5

you can get more info from pip and black devs, having had to move since 1.0 toml spec pyproject.tomls were breaking their tools on toml IIRC

[patent pivot]

yeah having a look, not opposed to adding that

[tawdry vapor]

So is there any point in keeping toml around too?

[fallen patrol]

well, tomli needs a seperate module if you want to make toml

[tawdry vapor]

I kind of doubt anyone actually uses toml in their eval anyway, but whatever

[fallen patrol]

yeah lol

[patent pivot]

I wonder if we can remove packages lol

[fallen patrol]

I personally use atoml 🙃

[patent pivot]

dep caching might make that weird

[tawdry vapor]

Doesn't look like it would remove it.

[brisk brook]

What? But didn't we go through this? If the library is used (imported or linked) the project also has to be GPL?

[tawdry vapor]

snekbox code never imports it

It's just there available in the eval env. So it is the user code that imports it

[fallen patrol]

^ you commented on using it for the tools in python which are never imported by the code

ex flake8 plugins

pydis code never imports it, yet you're still worried about their licenses

[brisk brook]

Ooh, makes sense.

[tawdry vapor]

@fallen patrol I comment on that? I do not remember that.

[fallen patrol]

uh a specific pr

bot#1779

[dusky shore]

GitHub

CI: check for dependency licenses

[patent pivot]

@plush narwhal yeah, but I think this is distinct from that, that has a much tighter tie to our stack, but snekbox evaluations don't, I'm not sure where the actual line is drawn though

[fallen patrol]

wrong user ;-;

[tawdry vapor]

Yeah I didn't forget that but I didn't specifically mention tools like flake8

[patent pivot]

ah, whoops, one char off

[tawdry vapor]

I don't licences of tools generally matters, but I am no expert on this matter

Well, they do matter but things aren't typically in such manner that restricts use of the binary

[tawdry vapor]

Snekbox only exposes it through a server rather than distributing the software and I think AGPL was created to ensure that also counts

[patent pivot]

hmmmm yeah AGPL is just if you use a service you're entitled to source code right?

[tawdry vapor]

Maybe something like that. After yesterday's conversation I don't really wanna dig into licences anymore.

[vocal wolf]

@patent pivot join us in vc

wait crap nvm I've forgotten what I've needed to tell you

[patent pivot]

lol, neither

[patent pivot]

am gaming

[fallen patrol]

oh nice dev-voice got opened to all

[vocal wolf]

yeah but you can't unmute yourself unless you're a contrib or staff person

[fallen patrol]

right

[vocal wolf]

https://github.com/python-discord/sir-lancebot/pull/808

[patent pivot]

sure

don't reimplement a lot – just needs to call the command

we will need to get that command in the container though

bind9-dnsutils

[clever wraith]

.\reverse hasn't been changelogged yet huh

Damn couldn't escape it

[tawdry vapor]

Hello I have been experimenting with an OOP infractions refactor. Any thoughts on this so far? https://paste.pythondiscord.com/mewogidoli.py

I have my own comments but I think it's better for them not to influence you yet.

[cold island]

Why is it necessary for the classes to have names?

[tawdry vapor]

It's a way to map them to their infraction types. The goal is, when having some infraction data from the API, to avoid a big if-else to instantiate the right class.

Though that reminds me that for create(), that should be changed so it's able to infer the type rather than requiring it to be passed explicitly

[cold island]

It's a bit odd to me to be able to search all user's infractions with Ban.search

[slim widget]

Mind if I join VC to see what you're working on? 👀

[patent pivot]

for sure xith is doing some programming

[tawdry vapor]

Should the class methods just be module-level functions instead then? I felt like that would be a bit inconsistent with some of the other CRUD methods that are instance specific

[slim widget]

Joe, have you tried Kakoune? 😄

http://kakoune.org

Erm, similar to vim.

One nice feature is it has a client-server model. So you can use your own window manager, instead of having to learn two sets of shortcuts.

[cold island]

I'm wondering if it should maybe stay as a cog command and keep the infraction classes to apply and pardon. Although I understand your goal was abstracting away the API calls

But otherwise yeah, the search function doesn't seem actually tied to the class

[tawdry vapor]

Yes, in practice working with dictionaries has not turned out to be nice. I suggested years ago that we should wrap the API but people weren't feeling it.

[tawdry vapor]

What about the other classemethods

[cold island]

get is the same I think

I'm a bit fuzzy on class method inheritance and not in front of a computer. Won't calling Ban.from_data set cls to Ban?

[tawdry vapor]

Yes it will

I didn't really intend for it to be used that way

[cold island]

Why not? If you call Ban.create that'll call Ban.from_data

[tawdry vapor]

I wasn't using Infraction instead of cls because I was playing around with an abstract subclass for expanded versions of expansions, which may have needed their own set of concrete subclasses. I didn't really like how it was turning out but I never changed it back to Infraction.from_data

[cold island]

Hmm. cls seems fine here though, you don't need to look for the right class in from_data then

[dim pelican]

So I tried to take over another's PR that was assigned to me...I somehow added all of my previous commits to the PR. The changed file is there though, and is ready to start the review process again. Sir-Lancebot#832
@brisk brook I think I got it half way right, I was able to pull his remote branch into my local

[dusky shore]

GitHub

Teapot Support - Continue Changes

[tawdry vapor]

Maybe. My concern is that it is using the _types defined on Infraction. I guess it should be since the subclasses don't override that

If the point you're trying to make is that the _types is redundant cause the cls gets set, I was trying to say that if you receive a response from the API you won't know which class to use so from_data takes care of creating the correct class.

[cold island]

Sorry I'm not sure I'm following. Reading the code, my impression is that the ban command will mainly become

infraction = Ban.create(...)
infraction.apply(...)

In that case I'm not sure why cls can't be used

[tawdry vapor]

Think about what needs to happen when you get a response from the API and need to create a class (e.g. get or search)

Yes, these are classmethods that are accessible on the concrete types, but that was never the intention.

[fallen patrol]

you needed to change to main-- git switch main

[tawdry vapor]

The idea is that if you just have an infraction ID, no idea what the type is, you get a response and it has a "type" as a string, and now you need to find the class that corresponds to that "type" string.

[cold island]

Hmm I see

It can be optional in that case, but that doesn't really matter

But yeah, create and from_data seem fine as class methods

[short snow]

hellllo

oh

right

hmm

yeah

maybe add a another embed for the edited embed

to see the two differences?

or just add another field to the embed

ah yeah

no edits?

yeah

idts

def shorten_text(text: str) -> str:
    """Truncate the text if there are over 3 lines or 300 characters, or if it is a single word."""
    original_length = len(text)
    lines = text.count("\n")
    # Limit to a maximum of three lines
    if lines > 3:
        text = "\n".join(line for line in text.split('\n')[:3])
    # If it is a single word, then truncate it to 50 characters
    if text.count(" ") < 1:
        text = text[:50]
    # Truncate text to a maximum of 300 characters
    if len(text) > 300:
        text = text[:300]
    # Add placeholder if the text was shortened
    if len(text) < original_length:
        text += "..."

    return text

haha yeah

okay mr. webscale

[tawdry vapor]

Alright, thanks. I agree with the feedback and will make those adjustments

[short snow]

any other functionality to be removed?

[tawdry vapor]

Maybe from_data is still confusing as a classmethod since it doesn't return the same type

[short snow]

xithhh which icons extension do you use

yeah

incidents webhook

you made it

lol

[cold island]

Doesn't the response say what type of infraction that is?

You could get the right class in the search and get functions

Instead of inside from_data

[tawdry vapor]

What I mean is that if you do Ban.from_data it doesn't necessarily return a Ban. It could be any type. It makes sense when used as Infraction.from_data but it technically works on any subclass too.

Maybe not that confusing to use it, but if you're reading it, it might lead you to believe it's a ban that's returned

And there's no way to enforce that it's only called from Infraction

[cold island]

If you use cls and this, then it'll be guaranteed to return the same type

[tawdry vapor]

Yeah I guess it can be removed entirely or just be made private. I don't really see a reason for someone outside the class to be dealing with raw API data.

[cold island]

Ah yeah, if you do that then it's just an init wrapper

[slim widget]

Tbh, that would be useful. I use fan-noise to tell if there's an infinite loop in my code

@patent pivot

[cold island]

There's an xkcd for that

[tawdry vapor]

Well that's what it's intended to be currently.

[cold island]

Yeah I mean if you export the class fetching then it doesn't do anything but call the init

At which point you can just call the init

[short snow]

what, are these some new stuff

[tawdry vapor]

Right. I was speaking in terms of either doing your idea of fetching the glass in get and friends, or keeping it in from_data but making it private

[short snow]

nope we don't ignore

in tox.ini

[cold island]

@slim widget https://xkcd.com/1172/

xkcd: Workflow

There are probably children out there holding down spacebar to stay warm in the winter! YOUR UPDATE MURDERS CHILDREN.

[short snow]

lemme see

whose venquatil on discord?

Mr.Webscale your task has finished

[vocal wolf]

lol

classic

[cold island]

Out of curiosity what did you notice yourself?

[dim pelican]

I was thinking the same thing

[short snow]

your volcy has a pr up mr.webscale

[tawdry vapor]

• Using an overload so I can easily get kwargs has the disadvantage that it defers validation completely to the API.
• Type annotations don't have the Unspecified type, but it's so ugly if it's added 😦
• The time parameters could accept anything that arrow.get accepts, which can be implemented once my time refactor PR is merged
• Not sure how to add some of the other infraction metadata (e.g. the icon, the pardon text)
• Don't know how to implement support for expanded infractions
• The dataclass has to be separate from the abstract class because mypy complains (I don't remember if it otherwise works or if Python also crashes)
• Converting timestamps to arrow.Arrow in the dataclass is hacky since the dataclass is frozen. It also makes the __init__ a bit confusing since it's supposed to take strings but that isn't clear from type annotations

[slim widget]

Ohh nice Xith. I use Amethyst.

It's great.

[short snow]

Alright cya, bye guys

[cold island]

Ah yeah, I wondered about the unspecified part, but thought that was maybe some practice I'm not familiar with 😅

It's an interesting problem though

I guess you want kwargs but be able to tell the user what the possible keywords are

[tawdry vapor]

Yes. Otherwise I'd have to check each parameter for being unspecified so I can avoid adding it to the payload

Plus I'd have to manually construct the dict to send to the request

[cold island]

Constructing the dict doesn't seem bad

[tawdry vapor]

I had to do it for create. It feels redundant especially when there are many params

[cold island]

Wait if you're doing the overload anyway why can't the args all be optionals?

[outer oasis]

Is the top free memory?

Activity monitor?

[tawdry vapor]

For which function?

[cold island]

Search. Just not sure why you need default values for them. Oh wait I do

[slim widget]

Yeah, on macos it's running an entire linux virtual machine @twilit juniper

[tawdry vapor]

To get the right behaviour in IDEs like PyCharm

[cold island]

Yeah I don't see a great way to do that

I'd probably just build the dict manually

[tawdry vapor]

How is that an improvement, besides being able to validate params a bit earlier?

[cold island]

Just seems a bit cryptic atm

The infraction icon is probably a class attribute for each subclass

[twilit juniper]

lol, well doesn’t windows do the same thing?

With WSL

[outer oasis]

Using a WSL backend is optional, as I remember it

[twilit juniper]

For docker no

[cold island]

You could potentially iterate over the locals

[outer oasis]

You can uncheck Use WSL to use Hyper-V instead

[outer oasis]

Then you have to remove unrelated locals

[tawdry vapor]

Yeah I was thinking that. The one caveat is that abstractness cannot be enforced, for what that's worth. There is a trick where the abc defines @abstractproperty but the subclasses just use regular class attributes. However, concrete classes are only validated when instantiated, yet class attributes don't require instantiate to be accessed. IIRC I played around with a way to "statically" validate that but yeah I don't think it's necessary.

[cold island]

That would be done at the beginning of the function

[outer oasis]

Fair

[tawdry vapor]

Interesting idea. I might try it to get a sense for how hacky it feels

[cold island]

Hmm yeah. Maybe you could raise an error with a metaclass but that seems overly complex, already having a separate dataclass and abstract class

[brazen charm]

@patent pivot arcade still broken

[patent pivot]

ah yeah did't get around to that

having a look

[brazen charm]

I believe it should be https://api.arcade.academy/en/latest/

and the pd and np packages which do nothing apart from adding duplicate symbols

[patent pivot]

!doc refreshdoc

[stable mountain]

Inventories refreshed

- np, pd```

[patent pivot]

!d arcade.AnimatedTimeBasedSprite

[stable mountain]

arcade.AnimatedTimeBasedSprite

class arcade.AnimatedTimeBasedSprite(filename: Optional[str] = None, scale: float = 1, image_x: float = 0, image_y: float = 0, image_width: float = 0, image_height: float = 0, ...)```
Sprite for platformer games that supports animations. These can be automatically created by the Tiled Map Editor.

[thorny obsidian]

@patent pivot can you change your name to something longer or shorter? Right now your name is causing me to think you're hemlock >_>

Mr. Webscale K8s Man perhaps

[patent pivot]

OH sorry sorry sorry

[hoary haven]

Mrs

[patent pivot]

fixedd fixed

[thorny obsidian]

big thanks! :D

[vocal wolf]

lmfao

[thorny obsidian]

I ... apparently don't read names anymore and match on approximate shape, length, and color

[foggy tiger]

@twilit juniper which video

[dim pelican]

I'm learning so much about github by failing so hard at it....Finally got everything set up to work on the actual file that needs changes

sir-lancebot#833

[dusky shore]

GitHub

Teapot support - Continues PR #610

[desert vessel]

where did the issue template dissapear ????

[brazen charm]

to lancebot

[desert vessel]

isnt it in the py bot ?

[brazen charm]

no

[vocal wolf]

There are no templates for the python bot

[desert vessel]

ah alright

thanks for letting me know :D

[dim pelican]

Speaking of templates, just opened up sir-lancebot#834 for commit templates

[dusky shore]

GitHub

Add .gitmessage to allow for consistent commit messages

[static canyon]

@vocal wolf you mentioning on sir-lancebot#814 that it's preferred to only assign a PR when it's not the author working on it. Does this also apply to the bot repo?

[dusky shore]

GitHub

Add coinflip command

[vocal wolf]

imo it applies to any repo

but yes

[static canyon]

👍

My logic was that it distinguished between "does this have anyone still working on it?" and "the person working on it is the author"

But good to know

[static canyon]

I suppose it doesn't matter too much, but can anyone think of a branch name for bot#1780? All I can think of right now is ignore-errors-from-infr-attempts-on-mods which seems rather long lol

[dusky shore]

GitHub

Suppress exceptions for infraction attempts on mods+

[static canyon]

Then again, we do have longer ones like Only-check-domain-filters-against-URL-like-parts-of-a-message, so guess length just doesn't matter in a branch name

[static canyon]

                    if hasattr(user, 'roles') and any(
                        moderation_role in user.roles
                        for moderation_role in MODERATION_ROLES
                    ):
                        confirm_msg += " as target is a moderator or higher"
                    else:```is this the correct over-hang indentation for the `any` statement? pycharm indents an extra level but not sure if that's inline with pydis styling or not

[brisk brook]

ignore-infra-mods-errors

Doesn't need to be long enough, or explain what it does (because you will have a pull request attached to it), just needs to be unique.

[static canyon]

👍

[brisk brook]

Hell, you can even use fix-1234 etc.

[static canyon]

True lol

[static canyon]

This passes linting so assuming it's fine

[brisk brook]

Can you send a screenshot?

Doesn't render nicely on my phone

[static canyon]

It's the any overhang I'm not sure about (lines 4 and 5)

[brisk brook]

I have a faint memory of such an error, something like "continuation line on the same level as next logical line"

I think that looks fine so as long as passes linting I'd say I am happy

[static canyon]

Yeah, it passes linting so I'll leave it 👍

[static canyon]

async def infraction_group(self, ctx: Context, infr_id: int = None) -> None:```should the typehint for `infr_id` be changed to `Optional[int]`? I was told in the dpy server that you shouldn't do this when the argument is the last of the command (or in this case, group), but not doing it does technically make the type annotation incorrect

[brisk brook]

It's not a big deal and I don't think we want to mess with discord.py too much

It's fine

[desert vessel]

Hey peopl!

[timid sentinel]

will changelog that now

[austere hornet]

@clever wraith Is there an unreverse command for @dusky shore ? Cuz he just made my avatar so weird and I want my old one back. Or would I just have to manually change it?

Oh wait I thought it saves it to my actual profile

Nvm

Sorry for the ping

[short snow]

just reverse your reversed avatar to make it normal

[austere hornet]

Right

Yeah mb

I guess I haven't fully waken up yet

[fervent sage]

lol, i know the feeling

[austere hornet]

I literally woke up 10 min ago

[green oriole]

Folks, it is happening https://github.com/python-discord/sir-lancebot/pull/835

If we can have many people trying to break this version of the bot, that'd be perfect

[short snow]

tests :help_me:

for clean cog, zig's rewrite PR, why don't we use the metricity for getting messages from channels? I would be much faster IMO?

[cold island]

Metricity doesn't store message contents. I'm also not sure speed is a concern here

As in, it shouldn't be unbearably slow, but that's not the case we're in

[clever wraith]

no dw

it doesnt save

[austere hornet]

Yeah I realized that only after I sent that

[clever wraith]

lmao

[austere hornet]

Mb

[short snow]

Ah right, I thought metricity also stores the message contents

also really nicely documented 👌 PR and code

@patent pivot This is about the enhanced incidents message PR, do I show that the message was deleted in the embed if the message gets deleted after the incident gets filed

[patent pivot]

nope — no retroactive updating of message stuff, we can do that via existing logs

[brisk brook]

Spring cleanup?

[vale ibex]

hardly, it's just 4 nicely separated commits

[green oriole]

why spring cleanup lol

there are like a few dozens actual changes, rest is lock file

[brisk brook]

Autumn cleanup 😤

[desert vessel]

@vale ibex mind taking a look at https://github.com/python-discord/sir-lancebot/issues/819 when you are free ?

[vale ibex]

Done

[austere hornet]

@vale ibex Could you also take a look at sir-lancebot#823 as well please? Thanks so much!

[dusky shore]

GitHub

Add Mastermind game command

[desert vessel]

tysm <3

[vale ibex]

As wookie said, it looks very similar to the snake antidote game we already have

which is essentially mastermind but with a different theme

[dim pelican]

Is it ask for review time?

I have two lol

[austere hornet]

Ok, so my guess is I can't do it then?

[vale ibex]

Yea, I don't think it's worth having the same game in the bot twice

you could take a look at the snake antidote game source and see how it compares to your implementation

You could always see if you'd do it differently or if there are improvements to be made

or, you can wait a few days for the discord py 2.0 PR to be merged, and make a game with buttons

or re-write some of the games with buttons

[austere hornet]

Ok, sounds good, tysm!

[desert vessel]

btw

are you guys planning to switching to drop down menu's

and buttons ?

[vale ibex]

Yes

we have a PR up to migrate sir lance to discord py 2.0

then we can try out the new features there, and do a dry run of 2.0

[thorny obsidian]

Buttons! Buttons everywhere for all the events!

[vale ibex]

if all goes well, we'll migrate bot to 2.0 too, and see what features can be migrated to interactions

we're not going to do one huge PR to re-write everything with interactions tho

[dim pelican]

Sorry for the silly question, but do we clone a local sir-lancebot from the PR to test it out on 2.0?

Or how would that work

[thorny obsidian]

Not a silly question! But essentially, you clone the sir-lancebot and then switch to the branch that Akarys is PRing

[dim pelican]

Ahh branches. I've been learning breaking a bunch of things with branches recently

[austere hornet]

@vale ibex Just to make sure, this is the snake antidote source, correct? https://github.com/python-discord/sir-lancebot/blob/main/bot/exts/evergreen/snakes/_snakes_cog.py

[vale ibex]

I've got a nice alias I use for git to check out a PR, since contribs can't make public branches

cpr = "!cd ${GIT_PREFIX:-.} && git fetch origin refs/pull/${1}/head && git checkout FETCH_HEAD #"

then you can do git cpr 1234 to checkout PR 1234

[thorny obsidian]

A Pull Request is essentially saying "hey, I'd like to merge this branch and its commits into the main branch". So to do the actual evaluation of the PR, you review the branch

[vale ibex]

you'd need to change origin to upstream though, to be able to checkout via your fork

[thorny obsidian]

... or just feel free to use GitHub desktop and do it with buttons

[vale ibex]

.src snake antidote

[dusky shore]

Command: snakes antidote

Antidote! Can you create the antivenom before the patient dies?

Source Code

Go to GitHub

bot/exts/evergreen/snakes/_snakes_cog.py:445

[dim pelican]

Right. And I keep running into the issue where a make a new branch, but it includes all of the commits from my main branch

[austere hornet]

Thanks

[thorny obsidian]

If you have a fork, you can pull down the changes to your fork, then switch to the branch

[dim pelican]

I'll bookmark Chris's command there for a reference

[desert vessel]

configuring the help command in sir-threadevere is impossible hehe

like-