Basic Pycord Help

[autumn gust]

a small bit

[errant trout]

i'll just throw out some example then

[autumn gust]

I have figured out autocomplete

[errant trout]

discord has no way of selecting multiple options in 1 arg

so instead of using a list of then, you make a separate bool arg for every spec

and have them all as default=False

it's kinda finicky, but that's more or less the only way

[autumn gust]

how does one make bool arguments again? just like uhh CPU=bool?

[errant trout]

no that's the type

i.e. where you put str

[autumn gust]

OH WAIT

[errant trout]

@option("CPU", bool, default=False)
@option("GPU", bool, default=False)```etc. etc.

[autumn gust]

ya, I just noticed

[errant trout]

ah, if im not mistaken

args have to be lowercase

on discord's side

[autumn gust]

got it

how do you define multiple embeds in discord.embeds?

or is it just discord.embed() x5 for me

[errant trout]

send/respond have an option for a list of embeds

up to 10, maximum 6000 characters across your entire message

[autumn gust]

so ctx.respond(embeds)?

but how does one define embeds

[errant trout]

embeds=[...]

which is a list of discord.Embed

[autumn gust]

ohhh wait so

[errant trout]

https://docs.pycord.dev/en/stable/api/data_classes.html#embed

[autumn gust]

embed1 = discord.Embed(...)
embed2 = discord.Embed(...)
...

embeds=[embed1, embed2, embed3, embed4, embed5]

[errant trout]

yep

[autumn gust]

got it

uuuuuuh, stupid question: how do you define color?

just throw in the hex code for it and call it a day?

[raw beacon]

discord.Embed(..., color=)

[autumn gust]

ya, but what to put inside of color

[raw beacon]

yes, the hex

like, discord.Embed(..., color=0x171515)

[autumn gust]

color=e62929 gets yellow underlined, color=#e62929 makes the rest of the code a comment

[raw beacon]

use 0x instead of #

[autumn gust]

ooooh, ty

[errant trout]

0x is how you do hexadecimal in python yeah

otherwise the discord.Color class also supports RGB and HSV values

[hard heath]

I have a question. I was experimenting with slash commands, but according to the official guide, in slash commands instead of using ctx.send () a ctx.respond is used. I would like to know if there is any difference, because apparently you can also use the ctx.send.

[raw beacon]

iirc .respond works for both undeffered and deffered

[lofty parcel]

.respond responds to the actual interaction received by discord

.send sends a plain message to the channel. it doesn't count towards responding the interaction

[pulsar ferry]

is there an easy way to only allow the user who called the function to use the buttons on a view? i know i can use ephemeral but want everyone to still see it

[raw beacon]

a check for author.id in the button callback

[pulsar ferry]

yeah so something like this?

        if interaction.user.id != self.user_id:
            # Optionally, you can send a message or disable the button
            await interaction.response.send_message("You're not authorized to interact with these buttons.", ephemeral=True)
            return

was wondering if there was just an out of box decorator

[raw beacon]

there could be, but im not aware of any

[raw beacon]

yes like that

[obsidian stratus]

How to change message cache size?

Specifically for delete and edit events

[errant trout]

max_messages when you create Bot or Client

default is 1000

[pulsar ferry]

how would i delete the message?

[errant trout]

interaction.delete_original_response

[pulsar ferry]

await interaction.delete_original_response() ?

gives discord.errors.NotFound: 404 Not Found (error code: 10015): Unknown Webhook

ultimatley what i am trying to do is give use a view that is ephmeral, then when they click the button update it but have it go public... is taht possible or is approach to delete and respond with new message?

[errant trout]

defer first, then delete

delete_original_response only works if you've already responded - defering counts as a response

[pulsar ferry]

interaction.defer() ?

[errant trout]

.response.defer()

(probably ephemeral too)

[pulsar ferry]

hmm something not right

i first have this

    @rps.command(name="unranked")
    async def rps_new(self, ctx):
        """Start a new unranked Rock-Paper-Scissors game."""
        view = RPSGameView(self, game_id=None, league="unranked", tip=None)
        await ctx.respond(view=view, embed=view.get_embed(), ephemeral=True)

so am doing ctx.respond with the view

then in the view

class RPSGameView(View):
    def __init__(self, rps_cog: RPS, game_id: int, league: str = None, tip: float = None):
        super().__init__(timeout=None)
        self.rps_cog = rps_cog
        self.game_id = game_id
        self.league = league
        self.tip = tip

        self.embed = discord.Embed(title=f"Choose your Move!", color=discord.Color.blue())
        # Add each detail as a separate field, with Game ID first
        self.embed.add_field(name="Game ID", value=f"{self.game_id or 'New'}", inline=True)
        self.embed.add_field(name="League", value=self.league.lower() if self.league else "None", inline=True)
        self.embed.add_field(name="Tip", value=f"💰 {self.tip if self.tip else 'No tip'}", inline=True)

        # Add the choices (buttons) to the view
        choices = [("🪨", "rock"), ("📄", "paper"), ("✂️", "scissors")]
        for emoji, choice in choices:
            button = Button(label=choice.capitalize(), style=discord.ButtonStyle.primary, emoji=emoji)
            button.callback = lambda interaction, choice=choice: self.handle_choice(interaction, choice)
            self.add_item(button)

    async def handle_choice(self, interaction: discord.Interaction, choice: str):
        """Handles user's game move choice."""

        await interaction.response.defer()
        await interaction.delete_original_response()

discord.errors.InteractionResponded: This interaction has already been responded to before

the delete actually seems to work but then the send message later doesnt

[pulsar ferry]

how can i send a message to the interaction after i delete it?

[errant trout]

eh

[errant trout]

what function are you using

interaction.respond normally handles it

[pulsar ferry]

        await interaction.response.defer()
        await interaction.delete_original_response()

        await interaction.response.send_message(embed=self.embed, view=self)

[errant trout]

then yeah just change it to interaction.respond, it chooses the appropriate function

[pulsar ferry]

ah that did it!

thanks

[errant trout]

(which is interaction.followup.send, the main point is convenience)

[echo wraith]

Yeah no but you can't block a user from going to the discord shop thing and subscribing to a tier

Like one way around it's easy

The other way I don't see how one would do it

[errant trout]

idk skill issue

[echo wraith]

I mean I don't really care cause the payment developer terms don't apply to me anyways

It would be weird to create a test entitlements for everyone that subscribes through your website

[nocturne jewel]

how to auto accept these kind of requests? when someone try to join

[raw beacon]

why would you want to auto accept them

their sole purpose is to be manually reviewed and accepted / rejected

[lapis dock]

I dont think join applications have been added to the API yet
Also it is unclear how they will change (or if they will exist) when guilds are discontinued

[maiden bloom]

When I release premium features I want to give the first x servers that purchase it an additional month free. Is there any way to do this without telling the user to make sure their membership doesn’t renew? I was planning on giving them a test entitlement after the first month has passed

[maiden bloom]

Wait guilds are already going away lmao

[lofty parcel]

Yeah, I don't think you can do it without telling the user not to renew

Monetization is still very simple

[sage tendon]

The experiment is being killed in favour of the features being added as regular server features

and good riddance imo

[torn barn]

yes they've been

those are member application forms that were used for clans

just, for normal guilds

[lapis dock]

Do you have a link for this? I am not finding any reference to them

[torn barn]

i know because i've made a clans pr for dpy and i had to do objects for it

[lapis dock]

just to clarify when I said "added to the API" I meant the documented bot accessible API

[torn barn]

lemme send it rq

https://docs.discord.sex/resources/guild#member-verification-form-field-structure

[sage tendon]

what the fuck is that link

[lapis dock]

for the userdoccers, unofficial community discord docs

[lapis dock]

This endpoint and the 2 endpoints below it might be of interest.
https://docs.discord.sex/resources/guild#action-guild-join-request
Note that these are unofficial docs so that stuff is not guaranteed to work.
Also it will not be implemented into pycord until it is added to the official docs (which some things may not)

[fresh sierra]

is this intended ?
💥┆Exception: can't subtract offset-naive and offset-aware datetimes

days_in_guild = (datetime.datetime.utcnow() - member.joined_at).days
days_account_age = (datetime.datetime.utcnow() - member.created_at).days

shouldnt they both be the same ?

[sage tendon]

do both cause that error

[fresh sierra]

nah, the joined_at only

thats why im a bit surprise, why would they be different

[sage tendon]

probably because one is from user and one from member

and joined_at says its an aware datetime object

created_at doesnt

[fresh sierra]

yes i know but why

i dont really see why one would ned to be aware and the other naive

[lapis dock]

This is what is seems like is happening. Discord is sending a timezone aware string for one of the fields but not the other. It is probably best to just get the UTC time of whatever is returned

[errant trout]

oversight, created_at uses utils.snowflake_time while joined_at uses utils.parse_time

[lapis dock]

which is prefered? Snowflake time I assume?

[errant trout]

well snowflake_time is for converting snowflakes, parse_time is for time strings

i.e. created_at just converts the user ID

parse_time uses fromisoformat which calculates the timezone from the included offset

[echo wraith]

-remindme 1d fix this datetime thing

[lofty vault]

Set a reminder in 1 day from now (<t:1740694004:f>)
View reminders with the reminders command

[shell radish]

ironic that it sends a datetime to fix a datetime

[echo wraith]

Yeah

[fresh sierra]

trying to fix for 1h a bug when there is no bug, you just forget that you fucking set a forbidden role and mandatry role features......

[lapis dock]

oh yeah that makes sense, I was not thinking

[silk spire]

use discord.utils.utcnow()

Also, datetime.utcnow() is deprecated and removed in 3.13(?)

[raw beacon]

its just .now() now

[echo wraith]

it is deprecated

from datetime import datetime, UTC

datetime.now(tz=UTC)

[silk spire]

still deprecated in 3.13?

Ah, deprecated since 3.12; I thought it was earlier. https://docs.python.org/3/library/datetime.html#datetime.datetime.utcnow

[echo wraith]

Both cause the error for me, and the issue is not on pycord's side afaik. It is datetime.utcnow() that is naive, not created_at or joined_at. You should use datetime.now(tz=UTC) as I exemplified above.

[echo wraith]

-reminders clear

[lofty vault]

Your reminders in this server:
113661592: #1132206148309749830: 'fix this datetime thing' - in 13 hours and 51 minutes from now (<t:1740694004:f>)

Remove a reminder with delreminder/rmreminder (id) where id is the first number for each reminder above.
To clear all reminders, use delreminder with the -a switch.

[echo wraith]

-delreminder -a

[lofty vault]

Cleared 1 reminders

[fresh sierra]

Yep I already did that, but did u check nelo’s message ?
I though he said 2 différents méthodes where used

[echo wraith]

Yeah but it makes sense

Like

joined at is most likely returned as an iso time string from discord's api

while joined at is not returned, it is the time embedded into the user id

[fresh sierra]

Ok I see

[sour widget]

I am a bit confused, when I do this

    async def reminder_cmd(self,
                           inter: discord.ApplicationContext,
                           name: str = discord.Option(input_type=str, required=True, description="The name of the reminder.", name="name"),
                           time: str = discord.Option(input_type=str, required=True, description="The time of the reminder, for better description use `/help time_formats`", name="time"),
                           description: str = discord.Option(input_type=str, required=False, description="The description of the reminder.", name="description"),
                           file: discord.Attachment = discord.Option(input_type=discord.Attachment, required=False, description="Any file", name="file"),
                           ):

the file looks like this (screenshot 1)

when I do this

file: discord.Attachment = discord.Attachment,

attachment works correctly, but it is said that description is not provided, so I think it is possible to do. And Attachment is a valid argument for the input_type parameter in Option() so I am confused what I am doing wrong

I can just use decorators, but maybe I made a silly mistake somewhere here

[sage tendon]

do you know how python works lol

Your issue is that you need to typehint the options, not set them as defaults
you are doing var = discord.Option(...) but you need to do var: discord.Option(...)

regardless i'd recommend using the decorator variant of defining options, personally
its much cleaner in the code, and less prone to errors imo

also, the first argument should be ctx, not inter (kind of conventional in pycord)

@sour widget

[little cobalt]

why is there a = and not a : ?

[sage tendon]

thats exactly the issue i just explained lol

[sage tendon]

also, just a tip, required is True by default, and you dont need to pass the name if its the same as the argument name (this is only true for the typehint version tho)

[lapis dock]

-# I dont know if it works for all types but I am pretty sure that arg = Option(...) is valid

[sage tendon]

it shouldn't be for all I know

[wheat tiger]

Is there any way to detect if the user's sent message is a forwarded msg or not? If yes, then how to fetch it's metadata?

Checked discord api docs, ig this is what I found?

https://discord.com/developers/docs/resources/message#message-reference-types

[silk spire]

There is a flag and stuff, but pycord doesn't support any of that yet
https://github.com/Pycord-Development/pycord/pull/2598

[nocturne jewel]

discord is having any issue with connection?

[sage tendon]

no

[wheat tiger]

Ye I saw that

but wasn't able to find that

[sage tendon]

find what

[wheat tiger]

To check if the message is a forwarded msg or not

Discord API provides the endpoint

(read the docs)

[sage tendon]

"but pycord doesnt support any of that yet"

[wheat tiger]

Currently in pycord if you fetch msg.content in a forwarded msg, just like pure embed msgs it returns "" / None

¯_(ツ)_/¯

that's what I meant

[sage tendon]

yea but like what do you want lol

pycord doesnt support it

[wheat tiger]

I am make a feature for my bot

https://github.com/swayam25/Square-Bot

[sage tendon]

i figured

[wheat tiger]

So the feature is somewhat like this. I am making a quotes channel setting, when someone forwards the msg (if it's a forwarded msg) it will fetch the msg metadata and will make a ss out of it by using Pillow

[sage tendon]

yea thats great and all but pycord doesnt support it

[wheat tiger]

Yep ik, that's just reply for ur q "but like what do you want lol"

[sage tendon]

you MIGHT be able to use https://docs.pycord.dev/en/stable/api/data_classes.html#discord.MessageFlags.value and this https://discord.com/developers/docs/resources/message#message-object-message-flags

but no guarantees

[sage tendon]

that was me asking what you're asking of us lol

try print(message.flags.value << 14)

if that works for a forwarded message tell me

[wheat tiger]

ok let's assume I got the status of the msg

how will I fetch the metadata?

the forwarded msgs' content, author data, time etc...

Only stress-free and possible way is to wait till pycord implements it. Or I gonna make pr

[sage tendon]

PR already exists. here.

and no, you wont be able to unless you make a raw api request i think

if message.flags.value & (1 << 14):
    # Message was forwarded
else:
    # Message wasn't forwarded

@wheat tiger ^^

[wheat tiger]

I want metadata too so..

I just need to wait ¯_(ツ)_/¯

[errant trout]

if you're comfortable with working with raw metadata, you can do py data = await bot.http.get_message(channel_id, message_id)

the forwarded message's data will be under the message_snapshots key

next release is not too far off, after which it'll be under message.snapshots

(and it will also include Message._raw_data if you ever need that in future; you could also add this yourself by updating message.py if you don't want to fetch every time)

[sage tendon]

why is it called snapshot?

[errant trout]

because it's a snapshot of the message at the time it was forwarded

[little mirage]

hello! is it acessible?

[sage tendon]

only the first one

[nocturne jewel]

how to use options with custom options for every server?

like "operator" for

server 1: add, remove
server 2: multiplicate, divide

[lofty parcel]

What

You mean the option displays different choices for each server?

[nocturne jewel]

yes!

[lofty parcel]

How do you determine such options for the server

[nocturne jewel]

To use options with custom choices for every server, the bot fetches options dynamically from an API (api.example.com/options?serverID=123) When a user runs the command, the bot makes an HTTP request to the API to retrieve the specific options for that server. These options are then displayed using dynamic autocomplete or a dropdown.

[lofty parcel]

Ok

You just explained how to do it

What's not working?

[nocturne jewel]

I should have thought about it first. I completely forgot, but after thinking it through (my bad), I remembered how it works. lol

nevermind

[wheat tiger]

Wait we can do this too, raw http req

Hmm I can use it

[lofty parcel]

You were always able to send http requests by yourself

[wheat tiger]

Ye?

But using pycord, I didn't notice doing that

[lofty parcel]

Yes...

[wheat tiger]

Ig till next release I can do this way

[lofty parcel]

How would the library make the api requests

[wheat tiger]

I though they using asyncio or requests when they need to fetch data

[lofty parcel]

That's what the library does...

.http is just the module that handles all the requests

[wheat tiger]

Ye ik (that the only way to make request) but never figured out they made some util functions

[lofty parcel]

I'm confused

[wheat tiger]

See, what I thought it, when they need to make some request they must be using the request module or asyncio module in that specific function, but I never realised that they made a unified class containing methods using which we can do http requests

Hope you understood

I was just lazy enough to dig in the package

[lofty parcel]

It still uses aiohttp to make the request

You're just seeing the abstraction of the request

.request is a method of the HTTPClient, if you dig through it, you will see it handles the request and uses aiohttp

https://github.com/Pycord-Development/pycord/blob/master/discord%2Fhttp.py#L288-L291

[rugged lodge]

discord%2Fhttp.py lines 288 to 291

try:
    async with self.__session.request(
        method, url, **kwargs
    ) as response:```

[lofty parcel]

self.__session being an aiohttp.ClientSession

[wheat tiger]

TQSM for you help

[errant trout]

allgood

[wheat tiger]

Using message_snapshots I can even check if it's forwarded msg or not

when not forwarded it returns None

else it returns the metadata

[errant trout]

pretty much yeah

[fresh sierra]

is ctx.user or ctx.author the dominant ?

[lofty parcel]

They're the same

[rugged lodge]

Searched in pycord

discord.ApplicationContext.user

[echo wraith]

I'd say user just becaue both are a shortcut for Interaction.user

but it's more a preference than anyhting

[errant trout]

ctx.author is because it came from ctx.message.author, while interaction has user instead; hence both exist

[lofty parcel]

✨ evolution✨

[fresh sierra]

ok thanks you

[shell radish]

backwards compatibility kinda

[maiden bloom]

Is the only way to determine if an entitlement is a test one, is to try to delete it?
I think both purchased and test entitlements have the application_subscription EntitlementType

[nocturne jewel]

@bot.slash_command(description="blablabla")
@option("choice", description="blabla", autocomplete=True)
async def choose(ctx: discord.ApplicationContext, product: str):
    await ctx.respond(f"you choose: {product}")


@choice.autocomplete("choice")
async def product_autocomplete(interaction: discord.Interaction, current: str):
    base_url = "https://api.example.com/"
    getshop_url = f"{base_url}?method=getlist"
    shop_response = await fetch_url(getshop_url)
    choices = []
    if shop_response is None or type(shop_response) is not list:
        return choices
    for item in shop_response:
        if current.lower() in item.lower():
            choices.append(discord.app_commands.Choice(name=item, value=item))
    return choices

Traceback (most recent call last):
  File "test.py", line 2891, in <module>
    @chooice.autocomplete("choose")
AttributeError: 'SlashCommand' object has no attribute 'autocomplete'

Hm

[lofty parcel]

That's not how it uh works

[nocturne jewel]

hmmmm

[lofty parcel]

Remove the decorator on the autocomplete func

Pass the function to the choice

autocomplete=product_autocomplete

[nocturne jewel]

lemme try

[sage tendon]

I've seen this syntax several times now
where do people get this from

and also, autocomplete takes AutoCompleteContext, not Interaction, and no other parameters

and the thing at the bottom with discord.app_commands Blabla, that's discord.py. This is pycord.

@nocturne jewel

your ide should've told you that that doesn't exist in Pycord

[silk spire]

Probably AI combining codes of multiple libs since it doesn't know better

[sage tendon]

i can guess stuff myself too

i asked them.

[nocturne jewel]

Copilot autocomplete :(

It gave me the idea

I also use discord.py and pycord in 2 different environments so it's a bit confusing

[sage tendon]

you really shouldn't do that
for your own sake

[ivory wind]

hello, i have a task that's supposed to go off at 4:20pm but it's 4min early

eastern = pytz.timezone('US/Eastern')
...
@tasks.loop(time=time(hour=16, minute=20, tzinfo=eastern))
...

system time is right. i don't know what it could be

[sage tendon]

what if you print time.now() in the task and check which time exactly the timestamp resolves to

[ivory wind]

2025-03-02 16:36:28,422 - cogs.four_twenty - INFO - 2025-03-02 16:36:28.422396

looks right to me

[sage tendon]

hm

[lapis dock]

are you sure your system time is correct? Does it show as 4:16 for others as well?

[ivory wind]

@tasks.loop(time=time(hour=16, minute=42, tzinfo=eastern))

[sage tendon]

check whether your system time is right on https://time.is/ just to rule that out completely

[ivory wind]

date
Sun Mar 2 16:41:53 EST 2025

[ivory wind]

vps is headless.

[sage tendon]

your system time

yours

[ivory wind]

it is

[sage tendon]

so python's time module seems to have the correct time based on this
your system time is right
and you say your VPS time is right (which i suppose it has to be if you matched it to your PC time)
hm..

[ivory wind]

it worked when I did it in utc. switched to eastern to handle daylight savings better

@tasks.loop(time=time(hour=20, minute=20, tzinfo=timezone.utc))

[lapis dock]

I think the timezone implementation is wrong

https://stackoverflow.com/questions/60996205/python-datetime-timezone-conversion-off-by-4-minutes

https://stackoverflow.com/questions/1398674/display-the-time-in-a-different-time-zone/63628816#63628816
this seems to be a better solution

[ivory wind]

thanks. i'll try it later

[lapis dock]

yeah pytz is not recommended to be used for python 3.9+

[ivory wind]

TIL

thank you, it worked

[errant trout]

wth LOL

[lapis dock]

would not be supprised if it had something to do with leap days/minutes/seconds/etc

[slow gazelle]

if i want to create a group of commands, for example /game trivia and /game akinator, but have the code of each command in a different file, what would be the best way to do it?

[echo wraith]

There is no "best way", but you can use multicog

.tag multicog

[sly karma]

Tag not found.

[echo wraith]

This: https://github.com/Dorukyum/pycord-multicog

[sage tendon]

pycord in itself does not support what you want, you need that extension

[echo wraith]

Or you'd do it manually but yeah

[sage tendon]

don't think that's easy

[echo wraith]

Yeah no it's not easy

[errant trout]

it would be if you disable syncing

i.e. if you subvert the library's command system, you can just setup interaction listeners

but you're actively fighting against the lib at that point

[shell radish]

skipping leap years is crazy 💀

[atomic fern]

is there an event for on guild boost?

[rugged lodge]

Searched in pycord

discord.on_guild_update
discord.on_guild_role_update
discord.on_guild_emojis_update
discord.on_guild_channel_update
discord.on_guild_stickers_update
discord.on_guild_channel_pins_update
discord.on_guild_integrations_update

[echo wraith]

afaik this

on_guild_update

but i'm not 100% cretain one would have to test

or wait for smbd else to confirm

[hazy turret]

Hello,

does anyone have any idea how long the Discord server-side cache lasts?
When I embed an image into an embed via URL, the image is cached with the URL from Discord. When I upload another image to the same URL, the old image is still displayed instead of the new one. The new one is only displayed when the URL changes.
Has anyone had any experience with this or knows how long the cache lasts? Or is there any way to delete or reset the cache?

[echo wraith]

Idk how long it last but last time it was for a very very long time

[silk spire]

1 day.

Oh wait that's for attachment urls

Not the whole cdn

[lapis dock]

A leap year is just a year with a leap day isn't it?

[echo wraith]

There's something with additional hours or minutes idk I forgot

[hazy turret]

Okay, thanks. There's no way around that or anything, is there?

[echo wraith]

not really I believe, maybe if you add www to the url it makes a new cache item or smth

[edgy nest]

just add a random query string to the end

imagecdn/image?wefhreuighkrejn

[round heart]

Since reloading slash commands doesn't change the behavior of the command function (vs text commands), how do you guys structure your Cogs? Do your slash command do nothing but respond, and you have some other function/layer that actually does whatever magic needed (including building the response e.g. embeds)?

[fresh sierra]

It does

If you reload a cog and then sync command the change will be reflected

[round heart]

meh. is there an atomic sync, or do I have to sync everything?

[sage tendon]

My commands contain all code they need unless its very unrelated to the command, duplicated, or stuff like DB handling

[round heart]

When developing, I might have to refresh things several times since there's often no way to simulate things without actually doing them

[fresh sierra]

You can sync the only command

As long as u sync it

[sage tendon]

why do you need to reload the cog tho, a restart is painless

[round heart]

Maybe for your bot. But not mine, which does a ton of stuff at startup

[fresh sierra]

Only issue is if ur edit the bot class or u need to add a shard

It takes a lonnnng time

[round heart]

What do you mean by this? How can I sync the single command I changed? You're saying sync_commands (or whatever it is) does this for me, or do I need to pass it some variable

Hm. Looks like I had this commented out.

    @commands.command(name='reload', hidden=True)
    @commands.is_owner()
    async def owner_cog_reload(self, ctx, *, cog: str):
        """Command which Reloads a Module.
        Remember to use dot path. e.g: cogs.owner"""

        try:
            self.bot.reload_extension(cog)
            # await asyncio.sleep(1)
            # await self.bot.sync_commands(force=True)
        except Exception as e:
            await ctx.send(f'**`ERROR:`** {type(e).__name__} - {e}')
        else:
            await ctx.message.add_reaction('👍')

Maybe it was broken in a previous version of pycord

[fresh sierra]

Yeah you can just sync it like since

But I feel like an update command which would reload all cog + sync command would be more efficient

Because you rarely edit only 1 command

[sage tendon]

i think you rarely edit more than one command at once lol

[fresh sierra]

Idk I have a test bot where k do all the change and then I merge it to the prod bot

So always some big update

[round heart]

When I'm iterating on a single task, yes I do.

For context, I was talking about reloading the cogs for the test bot. I don't mind restarting a prod bot for larger changesets.

[fresh sierra]

then do as you previsouly did

[opal swan]

hi everyone, can we send a dm to someone with await bot.fetch_user() ?

[silk spire]

Sure? As long as they share a server with the bot or have it installed to their profile

[lapis dock]

and they have DMs for that server turned on

[lapis dock]

Are you sure cog reloading does not update the code? I thought it did, it just does not reflect any changes to the app command on discords side unless you sync

[round heart]

Only if you explicitly sync commands, as Luma said. Otherwise, if I make any modifications, they are indeed not reflected.

[lapis dock]

afaik syncing commands has no effect on the actual callback of the command

But i trust you, it is probably a bug

[round heart]

The suck part for my current task is that when I reload my cog, this Web Client that I have loses its authentication for some reason. It calls out for the handshake and I never get a cookie the second time around. (Completely not Pycord related, just venting)

[marble thistle]

Does anyone know why when I call a bot.get_all_members then sort through member nick names the nicknames are missing on several people who indeed have nicknames on the server?

[lapis dock]

Never seen anyone use that function before 😅
Assuming it follows convention get_x returns cached things. So if the members nick changed after it was cached and the bot never received this update then it will be the old value

[raven breach]

When we use stop_recording() where does it save it?

[fresh sierra]

in the arg that you pass inside the start recording

[eternal kite]

can embeds have image and thumbnail at the same time?

[lapis dock]

yes

[eternal kite]

i attached two files with different names, and passed the arg files=[file1, file2] but it seems like file1 overwrites file2

[lapis dock]

can you show that snippet of your code please

[nocturne jewel]

async def finished_callback(sink, *args):
    text_channel = bot.get_channel(123)
    if text_channel is None:
        try:
            text_channel = await bot.fetch_channel(123)
        except Exception as e:
            print(f"Error: {e}")
            return

    recorded_users = [f"<@{user_id}>" for user_id in sink.audio_data.keys()]

    await sink.vc.disconnect()

    files = [
        discord.File(audio.file, f"{user_id}.{sink.encoding}")
        for user_id, audio in sink.audio_data.items()
    ]
    await text_channel.send(files=files)

@bot.event
async def on_voice_state_update(member, before, after):
    if member.bot:
        return

    if after.channel is not None and after.channel.category is not None:
        if after.channel.category.id == 123456:
            if member.guild.id not in connections:
                vc = await after.channel.connect()
                connections[member.guild.id] = vc
                sink_instance = discord.sinks.MP3Sink()
                vc.start_recording(sink_instance, finished_callback)
                print(f"Started recording in '{after.channel.name}'")
            else:
                vc = connections.get(member.guild.id)
                if vc.channel.id != after.channel.id:
                    await vc.move_to(after.channel)
                    #pass

    if before.channel is not None and before.channel.category is not None:
        if before.channel.category.id == 123456:
            vc = connections.get(member.guild.id)
            if vc is not None and vc.channel.id == before.channel.id:
                await asyncio.sleep(1)
                non_bot_members = [m for m in vc.channel.members if not m.bot]
                if len(non_bot_members) == 0:
                    vc.stop_recording()
                    del connections[member.guild.id]
                    print(f"Stopped recording in '{before.channel.name}'.")

Exception in thread Thread-43 (recv_audio): Traceback (most recent call last): File "/usr/lib/python3.10/threading.py", line 1016, in _bootstrap_inner self.run() File "/usr/lib/python3.10/threading.py", line 953, in run self._target(*self._args, **self._kwargs) File "/usr/local/lib/python3.10/dist-packages/discord/voice_client.py", line 863, in recv_audio self.unpack_audio(data) File "/usr/local/lib/python3.10/dist-packages/discord/voice_client.py", line 740, in unpack_audio data = RawData(data, self) File "/usr/local/lib/python3.10/dist-packages/discord/sinks/core.py", line 115, in __init__ self.decrypted_data = getattr(self.client, f"_decrypt_{self.client.mode}")( File "/usr/local/lib/python3.10/dist-packages/discord/voice_client.py", line 611, in _decrypt_xsalsa20_poly1305_lite return self.strip_header_ext(box.decrypt(bytes(data), bytes(nonce))) File "/usr/local/lib/python3.10/dist-packages/discord/voice_client.py", line 615, in strip_header_ext if data[0] == 0xBE and data[1] == 0xDE and len(data) > 4: IndexError: index out of range

is that normal?

i was trying to record a voice channel when for 10-20 seconds records it works, for longer like 2 minutes not

[upper slate]

Anyone know a way to jig it so that sending a voice channel invite to a user not already in the guild allows you to update overwrites for this user before they open the invite?

[errant trout]

no, best you could do is have a member_join listener (or use bot.wait_for) to update it after they join

[lapis dock]

This is really strange because it means discord is sending an empty packet (other than the nonce)
I guess the fix could be to just have the length check as the first of the 3 conditions but that would have to be something that is fixed on pycords side

[nocturne jewel]

Thanks for the suggestion! I will try tomorrow

[errant trout]

uhh isn't it because we didn't merge the encryption changes

[nocturne jewel]

Oh, there's any dev build I can try?

[lapis dock]

It should not get that far if that were the case. The mode selection would error out

The new system should work as an alternative but not actually fix this problem from what I can tell

[lapis dock]

https://github.com/Pycord-Development/pycord/pull/2651

[obsidian stratus]

I put 500k here but still the bot doesn’t log even 15 minute old messages which is deleted

😅

[sage tendon]

are you using on_message_delete ?

[errant trout]

it doesn't log past messages, only new messages from when the bot is started

[obsidian stratus]

Yes

[obsidian stratus]

Yeah I'm referring to the time the bot started

Uptime being 3 days+ but still failing to log 15 min old messages being deleted by a moderator manually (not purged)

[sage tendon]

and on_message works?

[wary silo]

does MessageType.application_command always have webhook id?

[ebon swift]

when using member.add_roles() , how do I use the snowflake feature to put multiple roles as parameters ?

def gives_necessary_roles(guild: Guild, member: Member):
    role1 : RoleTags|None = discord.utils.get(guild.roles, name="role1")
    role2 : RoleTags|None = discord.utils.get(guild.roles, name="role2")
    member.add_roles(???) #how to add both roles at once ?

[sage tendon]

await member.add_roles(role1, role2)

but getting roles by name is bad practice

if you already know the specific roles you want, just hardcode the IDs into the add_roles command

[ebon swift]

thank you

also about the role name I understand role ids are better, I'll switch to that soon

[sage tendon]

and then you should use guild.get_role

[ebon swift]

thank you for those better practice

[obsidian stratus]

Yes

On message works perfectly, I have intents.all enabled

[sage tendon]

and what if you send a message then delete it

[strange oracle]

I used pip install py-cord[voice] but cant seem to access anything from discord.sinks. Even in the example

from discord.sinks import MP3Sink```

I get

Exception has occurred: ModuleNotFoundError
No module named 'discord.sinks'
File "/home/mojobojo/dev/discoword/voice-recorder-test/vr.py", line 4, in <module>
from discord.sinks import MP3Sink
ModuleNotFoundError: No module named 'discord.sinks'```
Anyone have an idea why that could be?

[lapis dock]

can you make sure you have no other discord bot libraries installed. If you do uninstall them and reinstall pycord

[nocturne jewel]

@bot.group()
async def afk():
    pass

@afk.command()
async def set():
    ...

@afk.command()
async def remove():
    ...```

How to do this with ```python
@bot.slash_command```?

[lapis dock]

My instinct is telling me no, but you might want to ask in #api-chat on the discord developers server

[lapis dock]

Unfortunately slash commands do not allow this behavior.
An alternative could be to use /afk info or /afk show instead
Another alternative is to have set/remove as choices in an option instead of seperate commands if that works for your usecase

[nocturne jewel]

thank you

[ebon swift]

Hello

is there a way to make the listener, on_application_command execute and finish before the application command start ?

I made a listener so I can see who invoked a command and what command they used

@discord.Cog.listener()
    async def on_application_command(self, ctx: commands.Context):
        log.debug(f"Received command: {ctx.command} was invoked by {ctx.author}")

and any command like :

@commands.slash_command(name="ping", description="Check if the bot is alive.")
    async def ping(self, ctx: discord.ApplicationContext):
        log.info(f"LOGGING A MESSAGE")
        await ctx.respond(f"Pong! thanks for checking on me {ctx.author.mention} !")

when a user call the /ping command I get:

DATE TIME | INFO | LOGGING A MESSAGE
DATE TIME | DEBUG | Received command: pong was invoked by user

but it feels like its behaving as if I was using on_application_command_completion since this message always appears after a command completion.

[edgy nest]

if you want it to absolutely run before every time, you could make a global check

there might also be a before_application_command method

[lapis dock]

https://docs.pycord.dev/en/master/api/clients.html#discord.Bot.before_invoke
I believe this will do it globally

[wary silo]

got it, tyvm

[errant trout]

i mean, it definitely runs before ```py
self._bot.dispatch("application_command", ctx)
try:
...
await ctx.command.invoke(ctx)
...
except DiscordException as exc:
await ctx.command.dispatch_error(ctx, exc)
else:
self._bot.dispatch("application_command_completion", ctx)

(ironically, before_invoke is after on_application_command)

it might feel like the other way round because before_invoke and similar functions have to finish before the command runs, but the event doesn't

[edgy nest]

dispatch is async

runs out of order

[errant trout]

yeah i mean that was my point, the event is dispatched before invoke but it's possible to finish after unlike the other methods that actually block

[obsidian stratus]

It logs

Simple messages with just message content

It’s only not working when the message is like 15 mins old

It works sometimes but it’s unreliable

[strange oracle]

can confirm, no other discord libraries are installed. unless the remnants of one still exist in my virtual enviornment. I can try making a new one.

yep, so was just a messed up virtual environment. dunno why I didnt think of that. thanks for putting my brain in the right space.

[ebon swift]

halas it doesn't seem to

[ebon swift]

how would you do this please ?

I can't figure out how to do it globally

[lofty parcel]

This should still work even with the new changes?
https://github.com/fs86/discord-ext-ipc

[frail basin]

https://docs.pycord.dev/en/stable/api/clients.html#discord.Bot.add_check

Make sure to always return True, else the command wont run.

[ebon swift]

will try that rn !

works like a charm thank you so much !

if anyone ever needs the same feature here is my code

class Events(commands.Cog):
    def __init__(self, bot: discord:Bot):
        self.bot = bot
        self.bot.add_check(self.__log_command_call)


    async def __log_command_call(self, ctx: ApplicationContext):
        log.debug(f"Command {ctx.command} called by {ctx.author} in {ctx.guild}.")
        return True

[lofty parcel]

.

[lapis dock]

Saw that you got an answer!

[lapis dock]

what new changes?

[fresh sierra]

This seems really weird

[ebon swift]

does it ?

[fresh sierra]

Why are you not using an event like on_command, on_command_completion etc

[ebon swift]

does it work even if I'm using application commands ?

aren't commands and slash commands different ?

[fresh sierra]

If you use on_application_command it will work

[ebon swift]

as I've said here it doesn't

on applicaiton command shows after my command has been executed

[sage tendon]

so what about on_command

[lapis dock]

Does this not do it globally or does it not do it "in time"?

[fresh sierra]

It’s not likely to be after

It’s fire at the same time I think

[ebon swift]

yes

[fresh sierra]

Because completion is the one after

[ebon swift]

at the same time

[fresh sierra]

Or is it after and completion is only if there is no error

[ebon swift]

but I would rather have my log be always printed before the command is executed which makes the stack more readable

[sage tendon]

does it even matter lol

[ebon swift]

well to me it does ^^

[ebon swift]

as I've stated it's more of a time issue

[fresh sierra]

I understand the point for the logging

[sage tendon]

.

[fresh sierra]

So why are you not using bot.before invoke

[fresh sierra]

Will be the same, still not fire before

[sage tendon]

yes it will

read the fucking docs

[lapis dock]

sorry, I was talking about the bot.before_invoke. I am just curious what went wrong when you tried that

[ebon swift]

is on command sure to be called before the execution of the command ? I may have mis understood the difference between a command and an application command ? I thought application command = slash command and command is the hold <PREFIX> command

[sage tendon]

https://docs.pycord.dev/en/stable/ext/commands/api.html#discord.discord.ext.commands.on_command

read and try

i cant tell you more than that

[fresh sierra]

It will not wait for the listener to be done before invoking the command

[lapis dock]

^

[fresh sierra]

So if the thing takes more time it will likely break

[ebon swift]

for the bot before_invoke, I feel like I miss coded it, does it also requires to have a return True at the end ? cause that might have been my mistake

[sage tendon]

its one line of non async code..

[fresh sierra]

I don’t think so it’s just a Coro that is run before

[ebon swift]

I'll try all the solutions and let you know

[fresh sierra]

According to his exemple it was not working

And dispatch is async

So it might run out of order and so after the command has been invoked

[ebon swift]

class Events(commands.Cog):
    def __init__(self, bot: discord.Bot):
        self.bot = bot
        self.bot.add_check(self.__log_every_command)

    async def __log_every_command(self, ctx: ApplicationContext):
        log.trace(f"Command {ctx.command} called by {ctx.author}. ADD_CHECK_METHOD")
        return True

    @discord.Cog.listener()
    async def on_command(self, ctx: ApplicationContext):
        log.trace(f"Command {ctx.command} called by {ctx.author}. ON_COMMAND")

    @discord.Cog.listener()
    async def on_application_command(self, ctx: ApplicationContext):
        log.trace(f"Command {ctx.command} completed by {ctx.author}. ON_APPLICATION_COMMAND")
    
    # gives me : Extension 'cogs.events' raised an error: AttributeError: module 'discord.bot' has no attribute 'before_invoke'
    # @discord.bot.before_invoke
    # async def before_invoke(self, ctx: ApplicationContext):
    #    log.trace(f"Command {ctx.command} called by {ctx.author}. BEFORE_INVOKE")
...
class Fun(commands.Cog):
    def __init__(self, bot: discord.Bot):
        self.bot = bot

    @commands.slash_command(name="ping", description="Check if the bot is alive.")
    async def ping(self, ctx: ApplicationContext):
        log.info("Pong!")
        await ctx.respond(f"Pong! thanks for checking on me {ctx.author.mention} !")

09/03/2025 15:52:44:44 | TRACE | Command ping called by lolozen (ZEN). ADD_CHECK_METHOD
09/03/2025 15:52:44:44 | INFO | Pong!
09/03/2025 15:52:44:44 | TRACE | Command ping completed by lolozen (ZEN). ON_APPLICATION_COMMAND

so before invoke doesn't work, but I assume I'm not using it right
my slash command isn't considered as a command of on_command()
and on_applicaiton_command shows up after my log aeverytime (might be because it's ran in parallel)

[sage tendon]

hm

did you try https://docs.pycord.dev/en/stable/api/clients.html#discord.Bot.before_invoke ?

[ebon swift]

it's in my exemple right above I commented it, it gives me an error but it might be because I'm not using it correctly

[sage tendon]

yea lol

you need to use your botvar, not the discord.bot module lol

discord.bot just refers to the bot.py file inside pycord

[ebon swift]

hooo I need to use it like I did with the add_check ?

[lapis dock]

yes, i think so

[ebon swift]

trying it rn, but just out of curiosity, why is it better than add check ?

[sage tendon]

just

@bot.before_invoke()
async def logging(...):

must be in your main bot file i believe

[sage tendon]

because logging isnt a check, you're kinda just abusing the checks feature for logging

before invoke is for all kinds of tasks that have to run before a command is invoked

[ebon swift]

well this works perfectly

thank you

I didn't put it in my main since I like to keep it in my event cog but it works perfectly still

[sage tendon]

yea i wasnt sure if you can use self.bot in the decorators in a cog

show how you did it

[ebon swift]

class Events(commands.Cog):
    def __init__(self, bot: discord.Bot):
        self.bot = bot
        self.bot.before_invoke(self.__log_every_command)

    async def __log_every_command(self, ctx: ApplicationContext):
        log.trace(f"Command {ctx.command} called by {ctx.author}. BEFORE_INVOKE")

[sage tendon]

ah

in the docs it says its a decorator but ig you can use it like that too

[ebon swift]

when my cog is initiated it sets the before invoke to the bot

[sage tendon]

also you shouldn't put a double underscore infront of it, single underscore is convention for private stuff

[ebon swift]

isn't one underscore for protected ?

repalcing all my double underscore ^^

[sage tendon]

we aren't in java lol

[ebon swift]

😭

I'm jsute trying to be a good dev

lol

[sage tendon]

dunder methods are things like obj.__string__

[ebon swift]

okay thank you 🙂

[sage tendon]

and those are afaik the only methods that should use double underscores

[ebon swift]

alringt !

did a gloabl replacement ^^

well thank you all for your reactivity and for your help !

[sage tendon]

np

[lapis dock]

__ only in front is used in some non-dunder places, idk that much tho

[lapis dock]

I think all a decorator is is a function that takes a function as the first argument

[ebon swift]

I'll dig in and make research, but will stick to simple underscore for private from now on

[sage tendon]

yea true name mangling

but thats kind of a weird thing

[lapis dock]

Ah yeah, forgot about that it feels so gross I dont know why you would ever use it

[fresh sierra]

Shouldn’t it be only async def before_invoke() ?

Without anything else ?

[fresh sierra]

Mb didn’t saw that, if it’s working it’s good

[ebon swift]

we figured it out yes ^^

and it works perfectly

[fresh sierra]

Perfect

[ebon swift]

https://tenor.com/view/boredmemes-cultonape-notacult-cult-cult-on-ape-gif-9889342624274105559

[sage tendon]

also just a note, unless im wrong, this will still log every command globally
not sure if you want that or not but thats what'll happen

[ebon swift]

ho yes you're right

for now it's okay caus emy bot will onyl be on one server

but I will have to think about a guild filter

[lofty parcel]

2.7. It hasn't been updated dor 3 years.

[quartz umbra]

is it possible to make a bot which can be added to a user profile using pycord?

[sage tendon]

yes

[rugged lodge]

Here's the slash users example.

[quartz umbra]

is it any different from developing a normal guild-installed bot?

[sage tendon]

you get a lot less info in slash commands, like basically no guild attribute access etc

[quartz umbra]

oh so i added this bot I made to my profile but I can't use any of its slash commands in dms?

[sage tendon]

did you look at this?

[quartz umbra]

yeah I think I just overlooked the fact you need to set the integration types

[lapis dock]

Coding wise it is not that much different, but like toothy said you just get a lot less information sent by discord

[lofty parcel]

bump :p

[quartz umbra]

Hi! is it possible for a bot to upload emojis to a guild? If so, how do you do it in py-cord because I only see create_sticker while the argument is called emoji so it's a bit confusing

[echo wraith]

That argument refers to the emoji linked to the sticker

[rugged lodge]

Searched in pycord

discord.Guild.create_custom_emoji

[echo wraith]

use this

[quartz umbra]

tysm

[echo wraith]

np

[quartz umbra]

is Attachment.to_file().fp a byte-like object?

[sage tendon]

what does your IDE say?

[quartz umbra]

I use vscode so nothing 0-0

[rugged lodge]

Searched in pycord

discord.Asset
discord.Asset.is_animated
discord.Asset.key
discord.Asset.read
discord.Asset.replace
discord.Asset.save
discord.Asset.url
discord.Asset.with_format
discord.Asset.with_size
discord.Asset.with_static_format
discord.asset.Asset
discord.Activity.assets
discord.RawMessageDeleteEvent
discord.RawMessageDeleteEvent.cached_message
discord.RawMessageDeleteEvent.channel_id

Searched in pycord

discord.Attachment
discord.Attachment.content_type
discord.Attachment.description
discord.Attachment.duration_secs
discord.Attachment.ephemeral
discord.Attachment.expires_at
discord.Attachment.filename
discord.Attachment.flags
discord.Attachment.height
discord.Attachment.hm
discord.Attachment.id
discord.Attachment.is_spoiler
discord.Attachment.issued_at
discord.Attachment.proxy_url
discord.Attachment.read

[echo wraith]

btw it's await

so

(await attachment.to_file).fp

[sage tendon]

vscode can do that too.

[frail basin]

wait, thats valid syntax?

[echo wraith]

yes

afaik

[frail basin]

so I could, in theory (await fetch_...()).attr ?

Thats useful to know

[sage tendon]

That's the only way to do such things in one line

because if not for the brackets you would await the .attr

[lofty parcel]

Python moment

[sage tendon]

no, completely logical

[lapis dock]

One thing I wish python had was bot?.user?.id which I think just returns None if bot or user are None instead of throwing a NoneType error

[frail basin]

Is the ? even an operator in python?

[frail basin]

As in, would adding this overlap with any current functionality?

[lapis dock]

dont think so

not in that context at least

[fresh sierra]

bot.user.id if bot and bot.user and bot.user.id else none ;)

make so much sens ;)

[errant trout]

that was proposed in PEP 505 , but it's a pretty controversial topic amongst the maintainers and ended up being deferred

[lapis dock]

yeah I just found that now

[fickle salmon]

so redundant though

[fresh sierra]

yeah i was joking

[fickle salmon]

oh

[silk spire]

Or just bot and bot.user and bot.user.id lol

[fallow crescent]

Hi py-cord friends. Is there a way to access a guild member's signals as it's displayed in the discord members window? I want to notify myself when a member's suspicious behavior or excessive DMs signals go up.

I thought it might be some undocumented bit in the public_flags field but that doesn't seem to always be the case

[silk spire]

There are undocumented fields on the Member object that discord doesn't want bots to use iirc yes

[fallow crescent]

Undocumented fields that make it to py-cord's Member representation?

[silk spire]

No, libraries usually don't add undocumented fields as they can be unstable

[echo wraith]

I'm not sure, I don't think on 2.6.1 but maybe there was a pr that did it on master

[tiny shard]

Hello, I have an issue with my bot that is in 1 guild with 350k members. I want to receive on member join event only. I enabled intents.members however this makes my bot's RAM usage jump from ~50MB to almost 500MB. Disabling cache with bot = discord.Bot(intents = intents, member_cache_flags = discord.MemberCacheFlags.none()) didn't decrease the ram usage. Any help?

[fresh sierra]

Just enabling it ? Or there is a function which use this intents ?

Or else it can just be normal, if you have to receive intents from 300k members that’s a lot

[little cobalt]

member intents can use a lot of RAM

[errant trout]

could you try setting chunk_guilds_at_startup=False in your Bot

[sage tendon]

yes

but only the spammer flag

[sage tendon]

you can still access it.

[fallow crescent]

is the spammer flag the "excessive DMs" signal?

[sage tendon]

yes

[silk spire]

Ah til

[round heart]

Not pycord-specific, but I've always been curious about this. What's the "correct" way to define a property on your class without the ide whining that you typed it but it's initially None?

It's not really optional, I just have to pre-define before it actually gets used it so I don't get errors when I do.

[errant trout]

i mean

that's up to your typechecker, no?

[lofty parcel]

Just turn it off

Boom

No more issues

[lapis dock]

I just dont until it is set and leave a typehint at the class level

class A:
  hi: int
  def set(self, h: int):
    self.hi = h
a1 = A()
a1.hi # This returns an attribute error
a1.set(4)
a1.hi

[round heart]

Yeah it's stupid that you can typehint it with no value but IDE thinks it's not defined.

[errant trout]

i haven't opened pycharm like a year, all my pycord work is done on github.dev (vsc) with basic autocomplete and my bot work is in my server's web panel (lol)

[silk spire]

Just hi: int = None # type: ignore

it technically isn't defined. Without a value it's just an annotation that's added to the dict (cls.__annotations__)

[echo wraith]

class Foo:
    _hi: int | None = None

    @property
    def hi(self) -> int:
        if self._hi is None:
            raise AttributeError("bleh")
        return self._hi

    def set(self, v: int) -> None:
        self._hi = v

smth like that

[silk spire]

That's how not-lazy people do it yeah

hi could have a setter instead of the separate method but eh

[echo wraith]

yeah I did that at first

but then saw Wolfy's example with set

so I wanted it to be similar

[silk spire]

Ah

[round heart]

I have like a dozen props that need to be added this way; I’m not defining getters for them all, that’s silly

It’s stupid that you must define properties in init, and yet throw a warning if you assign to None initially because it’ll be updated later

[silk spire]

Do you perhaps want a dataclass

[silk spire]

How exactly is that stupid?

What if the value is never assigned? The type checker would be lying atp

And a: int = None being invalid makes sense too

None is NOT an integer

[round heart]

Because it makes sense that not every property will be assigned at init. For instance a response from an api that the runtime of the class relies on

[silk spire]

It doesn't need to be assigned at the init

It can be in a method too that's called by you

The type checker will know that

[round heart]

But it does because if you set it outside of init, that’s technically wrong because it doesn’t exist. At least according to PyCharm

Plus I need to know it if it’s referenced in other classes. I imagine dynamic setting wouldn’t autocomplete

[silk spire]

Well yeah defining it outside without a value means it doesn't exist at runtime

Sounds like you want a dataclass

[rugged lodge]

Searched in python

dataclasses
dataclasses.Field
dataclasses.FrozenInstanceError
dataclasses.KW_ONLY
dataclasses.MISSING
dataclasses.__post_init__
dataclasses.asdict
dataclasses.astuple
dataclasses.dataclass
dataclasses.field
dataclasses.fields
dataclasses.is_dataclass
dataclasses.make_dataclass
dataclasses.replace
label:dataclasses

[silk spire]

You can assign the attrs all in the class and it will assign them when you construct the class

[round heart]

But I don’t? In my cogs, I want to be able to reference self.bot.object_from_api. There are several of them, and they’re collected at different times (or at least need to be accessible before others are set)

[echo wraith]

I mean

wait youre saiying if you only type them pycharm will complain ?

[silk spire]

Oh god

[echo wraith]

class E:
    hi: int

[silk spire]

Is pycharm that stupid

[echo wraith]

Idk I use basedpyright so I don't really know

[silk spire]

Does that take over the built in checker?

[echo wraith]

yes

[round heart]

Well that’s a class var rather than an instance var. is that what I need to change?

[echo wraith]

what happens if you do that ?

[silk spire]

Oh pog

[round heart]

I think that works (not at my computer). It’s when I define self.hi in init that it complains

[echo wraith]

what does it say ?

from typing import reveal_type

class Foo:
    bar: int # Instance variable "bar" is not initialized in the class body or __init__ method  (reportUninitializedInstanceVariable)

foo = Foo()

reveal_type(foo.bar) # Type of "foo.bar" is "int"

this is w/ basedpyright link

[round heart]

If I self.hi: int = None it complains on type. If I self.hi: int, then it acts like it’s not an available instance var when you try to reference it

[echo wraith]

well that's really pycharm doing weird stuff imo

cause it shouldn't

[round heart]

When I’m back at my computer I’ll write out a test class with all the different ways

[silk spire]

self.hi: int isn't valid and complaining about assigning int to None makes sense tho?

Oh not available

[round heart]

How is it not valid? wtf

[echo wraith]

I think if you don't assign a value to a classvar pycharm will ignore it that's the thing here

[silk spire]

thing.thing: T isn't valid, at least with pyright

[echo wraith]

yeah no that no

[silk spire]

x: T = thing.thing is

[errant trout]

nah py @hi.setter def hi(self, v)to do property assignment

(func name doesn't matter)

(i did not read the rest but idk why you would bother doing .set instead of overloading)

[echo wraith]

I wouldn't, I changed it before sending to match Wolfy's example

But that's what my original example did

Whatever

[round heart]

Weird. If I set a simple class

class TestClass:
    def __init__(self):
        self.b: int

    def test(self):
        self.b = 2

There doesn't seem to be any issue with defining it this way; Pycharm does not complain.

But it treats my actual code differently...

class MyBot(commands.Bot):
    def __init__(self):
        self.http_session: aiohttp.ClientSession
        ....

    async def set_http_clients():
        self.http_session = aiohttp.ClientSession(...)

Setting it whines Instance attribute http_session defined outside __init__ . It doesn't if I make it a class attribute, but I admit I don't really understand where you'd use one or the other.

[sage tendon]

I think class attributes apply globally to the entire class and all of it's objects

[round heart]

Which for the bot would only be once, so I assume that shouldn't matter in this specific case.

But I still don't get why instance properties are being treated differently here. In my TestClass I can reference it fine outside of init, but not in the bot

[sage tendon]

because the bot class uses slots, i think

[ebon swift]

I need some help, when I do a guild.members, some user on my server are not listed, does anyone have any idea where this might come from ?

[echo wraith]

cache stuff doing weird things

but no idea

[ebon swift]

is there a way for me to go around that ?

[lofty parcel]

Fetch the users

[ebon swift]

I should use guild.chunk() then ?

[sage tendon]

just fetch the members

[ebon swift]

Sorry if I'm not getting this correctly I want to fetch all the user from the guild

I don't get their id in the first place

[sage tendon]

generally everyone should be in cache already
how long after startup is this

[ebon swift]

just restarted it

[sage tendon]

give it like a minute or two

[ebon swift]

I see

it's disturbing me a bit since I'm trying to do a db sync on startup

[sage tendon]

no i mean try that, im not sure if thats the issue

[ebon swift]

I'll try to sync it ig

[ebon swift]

very strange

[errant trout]

how many members are we talking

[ebon swift]

this:

@commands.Cog.listener()
    async def on_ready(self):
        """
        Event that is called when the bot is ready.
        Returns:
        """
        self.bot.main_guild = self.bot.guilds[0]
        members = await self.bot.main_guild.fetch_members(limit=None).flatten()
        user1 = self.bot.main_guild.get_member(48<...>00)
        user2 = self.bot.main_guild.get_member(33<...>25)
        user3 = self.bot.main_guild.get_member(82<...>72)
        user4 = self.bot.main_guild.get_member(69<...>58)
        loguru.logger.debug(f"{user1} | {user2} | {user3} | {user4}")

gives me:
11/03/2025 21:33:21:56 | DEBUG | None | None | None | None

4

[errant trout]

i mean in the entire guild

[ebon swift]

on a server of less than 50

[errant trout]

hm

[ebon swift]

I can't find something in common between all of them

my bot is also administrator

and I gave it : discord.Intents.all()

(yes not best practice its for the tests)

[errant trout]

fetch_members doesn't fill cache so it's pointless there, but if you have all intents it should cache all members at startup anyway

(unless you disabled it in your Bot class)

[ebon swift]

doesn't feel like I did

[errant trout]

seems fine

does your bot have a custom init

also pretty normal...

[ebon swift]

I know

I mean

[echo wraith]

does your bot override any metjopds

or what methods does it implement

[ebon swift]

that's what is making me really lost

[errant trout]

if you really want, await guild.chunk() should fill your cache

but if even that isn't working discord might just be really weird about it...?

do the missing members appear in your fetch_members list

[echo wraith]

are yo u sure that's the right guild ?

[ebon swift]

yup it's only in one

might check twice still

[echo wraith]

yeah I mean I trust you but who knows atp

[ebon swift]

...

[errant trout]

oh

[ebon swift]

sorry to ahve bothered you all

[errant trout]

LOL

[ebon swift]

[redacted but there was two servers]

BUT

WAIT

ITS NOT APPEARING IN THE FIRST SERVER WTF

[errant trout]

burn him at the stake

[echo wraith]

maybe doesn't have the perms to show up ?

in the online tab ?

[errant trout]

uhhh don't think a bot would need them

[echo wraith]

I'm pretty sure if @ everyone disbles view channels and the bot doesn't have that in its default role it doens't show up

[ebon swift]

okay well now I need to udnerstand how to kick him

or jsut force with guild id

[echo wraith]

/kick <bot_id> should work

[errant trout]

code-wise, you can kick any member with guild.kick(discord.Object(user_id))

(or fetch the user if you really want)

[ebon swift]

I'm very sorry for that 😦

thank you guys

[errant trout]

nah all good LOL

[ebon swift]

I feel so stupid

anyway thank you 🙏 as always, awesome community

[echo wraith]

nw and feel free to ask if there's anything

[lofty parcel]

i know im awesome, ty

[fierce seal]

Does anyone know how to start a task that's in a cog when the bot starts up?

[sage tendon]

self.task.start()

lol

[errant trout]

erm actually there's no await

[sage tendon]

true

[fierce seal]

didn't seem to work but i'll check again

[sage tendon]

in the cog init

not in your main file

[little cobalt]

Also add a await self.bot.wait_until_ready() to the first line at the task function

[sage tendon]

damn zervy typed faster

[little cobalt]

Im at my phone ;3

[fierce seal]

that's probably it

[sage tendon]

i was held back by pycharm loading

[echo wraith]

Curious if starting in a once=True on_ready is also good

[errant trout]

theoretically yeah

well

you generallyyyy don't need to add once to on_ready, but i guess you can

[fierce seal]

yup that was it

thx

[echo wraith]

I do it casue I saw some people doing it but yeah make sense

[errant trout]

fairfair

[upper slate]

why did my post in #creations get deleted?

it was an lfg bot

[echo wraith]

idk ask lala or some mod

probably has to do with #creations message

[upper slate]

ah I see, sorry. if anybody has any tips for advertising a bot feel free to dm me

[sage tendon]

top.gg

its a 2 seconds google search

which is why you chose the several-minute path that this is

[lethal loom]

I have the following problem, somehow I get all members who are on the server at member_ids and not only from the specific channel and no these are not admins or so

    async def user_callback(self, select, interaction):
        users_ids = [user.id for user in select.values]
        print(users_ids)
        users = []
        users_errors = []
        channel = interaction.channel
        print(channel)
        member_ids = [member.id for member in channel.members]
        print(member_ids)
        print(len(member_ids))```

and the channel prints the right channel

[sage tendon]

thats normal

[lethal loom]

why?

[sage tendon]

because thats how user selectors work

[lethal loom]

Ah nonono

[sage tendon]

oh yes

[lethal loom]

but i use the itneraction?

and how can i do it 0only become the channel members?

[sage tendon]

you cant limit user selects to only show specific people

it always shows everyone afaik

and if you mean this, that should work if the docs are right

[lethal loom]

So i want to make a suer select and there are all users inside that works and is correct but i want now wnen i can make the dropdown in a private only a lsite of all suer who are in the channel

class UserSysAddUser(discord.ui.View):
    @discord.ui.user_select(placeholder="Wähle einen user aus", min_values=1,max_values=3)
    async def user_callback(self, select, interaction):
        # Get the user IDs from the select
        users_ids = [user.id for user in select.values]
        print(users_ids)
        users = []
        users_errors = []
        # Get the channel where the dropdown was used
        channel = interaction.channel
        print(channel)
        # Get all users who are in the channel where the dropdown was used
        member_ids = [member.id for member in channel.members]
        print(member_ids)
        print(len(member_ids))```

i added coms.

[lofty parcel]

I'm sure user selects only show relevant users who can see the channel.

[lethal loom]

no XD

[sage tendon]

No, they show all

[lethal loom]

I have a dropdown with all users, this is right

[sage tendon]

all user selectors work like that, in slash commands as well

[lethal loom]

BUT

[sage tendon]

I dont think there's an easy way to achieve what you want

[lofty parcel]

Well then create the dropdown yourself

[sage tendon]

but then you are limited to 25? options

[lethal loom]

You don't understand that

[lofty parcel]

You won't believe it: multiple selects. Boom.

It's ugly?

[lethal loom]

Everything is ok with the userselector

[lofty parcel]

Yes

It works?

Also yes

[lethal loom]

but after i have selected one or more users i would like to get all users who are in the channel in a list

[sage tendon]

well again the docs do say channel.members returns only members who can see the channel

so if the docs lie no clue

[lethal loom]

but i don't get the users from the channel but everyone who is on the whole server

I get a site with 28 users but in the channel only 10 people are authorised to see and read it

[sage tendon]

make sure they dont have read messages in that channel

deny them that specifically

that should fix it

[sage tendon]

It's really a dumb way of checking it not gonna lie

[lethal loom]

but i become all members with this
member_ids = [member.id for member in channel.members]

[sage tendon]

what did i just say?

[lethal loom]

look no body can see this

[sage tendon]

that isn't the permission that matters as i just showed

[sage tendon]

.

[sage tendon]

.

[sage tendon]

.

[lethal loom]

i have desable all perms but i become all suers 😦

[sage tendon]

In the channel overrides, does @ everyone have X for read messages?

[sage tendon]

Okay yea looking at the library code again I saw that my assumption was wrong, but now I have even less of a clue why your stuff doesnt work

[lethal loom]

and this is my code to create this channelpy permissions = { interaction.user: discord.PermissionOverwrite(send_messages=True, read_messages=True, view_channel=True), interaction.guild.default_role: discord.PermissionOverwrite(view_channel=False, read_messages=False) }

[sage tendon]

can you do print("Test " + len(channel.members))

and show the output

just to make sure its what i expect

[lethal loom]

print("Test " + len(channel.members)) ~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~ TypeError: can only concatenate str (not "int") to str

[sage tendon]

bruh then do an f string

[lethal loom]

ah XD

[sage tendon]

god python is so dumb sometimes

[lofty parcel]

print("test", len(channel.members))?

[sage tendon]

I choose to disagree with that syntax

[lofty parcel]

Python doesn't

[echo wraith]

Looking at it but might be too tired to do a full repro, rn can't repro

[lethal loom]

i become 28 😦

[sage tendon]

and how many members do you see in the members list

[lethal loom]

9

[sage tendon]

also its "get", just to clear out confusion
become means something different

Then can you try this

for member in channel.members:
    print(channel.permissions_for(member).read_messages)

[lofty parcel]

It's weird, it works for me

[sage tendon]

2 v 1 then lol

and i have to agree i dont see anywhere this could go wrong, the internal code seems solid

Can you show a full screen recording of the entire process as well
I think we're missing something

[lethal loom]

i become 28 x ture its bad

yeah

[echo wraith]

@lethal loom just to be clear, you create a channel with that code, and its .members returns members that aren't in the membvers list ?

[lethal loom]

yes

[echo wraith]

Oh

[sage tendon]

.

[lethal loom]

yeah wait

[echo wraith]

thx toothy

can you try with pycord master ?

[sage tendon]

wat

[echo wraith]

I think I pred something abotu broken permissions overwrites

and it got merged

[echo wraith]

this

didnt see it

what

[sage tendon]

oh neat

[echo wraith]

But I'm not sur if it applies here

[lofty parcel]

I don't use master and it works for me

[sage tendon]

the screen rec will show us the issue

100%

[echo wraith]

yeah but who knows atp

[sage tendon]

this seems an issue with your backend code tho? lol

like where do you print that "already in the ticket"

the logic deciding that is busted

[lethal loom]

this is my complet funcvtion

class UserSysAddUser(discord.ui.View):
    @discord.ui.user_select(placeholder="Wähle einen user aus", min_values=1,max_values=3)
    async def user_callback(self, select, interaction):
        # Get the user IDs from the select
        users_ids = [user.id for user in select.values]
        print(users_ids)
        users = []
        users_errors = []
        # Get the channel where the dropdown was used
        channel = interaction.channel
        print(len(channel.members))
        print(channel)
        for member in channel.members:
            print(channel.permissions_for(member).read_messages)
        # Get all users who are in the channel where the dropdown was used
        member_ids = [member.id for member in channel.members]
        print(member_ids)
        print(len(member_ids))
        for u in users_ids:
            user = interaction.guild.get_member(int(u))
            if user is None:
                users_errors.append(f"\n - <@{u}> `Invaid User`")
                continue
            elif user.id == interaction.user.id:
                users_errors.append(f"\n - <@{u}> `Du kannst dich nicht selber zu dem Ticket hinzufügen`")
                continue
            elif user.id in member_ids:
                users_errors.append(f"\n - <@{u}> `der Nutzer ist schon im Ticket vorhanden`")
                continue
            overwrite = discord.PermissionOverwrite()
            overwrite.read_messages = True
            await channel.set_permissions(user, overwrite=overwrite)
            users.append(f"<@{u}>")
        user_string = ", ".join(users)
        users_errors_string = ", ".join(users_errors)
        await interaction.response.defer()
        embed = discord.Embed(description=f"{interaction.user.mention} hat {user_string} zum Ticket hinzugefügt",color=TICKET_COLOR_GREEN)
        embed_error = discord.Embed(description=f"Folgende User konnten nicht hinzugefügt werden: {users_errors_string}", color=TICKET_COLOR_RED)
        if len(user_string) == 0:
            await channel.send(embed=embed_error)
        elif len(users_errors_string) == 0:
            await channel.send(embed=embed)
        else:
            await channel.send(embeds=[embed, embed_error])```

[sage tendon]

i have to admit i have no clue why you're extracting the IDs from the full objects, then later you need to jump through hoops because of that

[lofty parcel]

Long ass callback

[sage tendon]

yea a lot going on lol

[lethal loom]

yeah but why channel.members give me 28 😭

[echo wraith]

this

idk if it's that or smth

[sage tendon]

i was just intently looking at the "interaction.channel" line

lmao

yea please try on master branch

[echo wraith]

@lethal loom try with master

.tag master

[sly karma]

pip install -U git+https://github.com/Pycord-Development/pycord
-# Note: The Master Branch May Have The Newest Features But It May Also Have More Bugs

[lofty parcel]

If its that

I'll fucking lose it

[echo wraith]

its that 100%

cause

member_ids = [member.id for member in channel.members]

channel.members uses permission_for

which was broken

[lofty parcel]

Yeah

[sage tendon]

the code itself looks fine, but i heavily recommend splitting it up into several functions

and importantly, from the full user and member objects you get, you just extract the ID. Then later you construct a mention manually beacuse, well, you only have the ID. What's the point of reducing the objects to their ID?

[lofty parcel]

You can also improve that code lol

Or I'd just do it different

[echo wraith]

https://github.com/Pycord-Development/pycord/issues/2281

[sage tendon]

2.7 is really fucking overdue now

if for nothing but bug fixes

[lethal loom]

with master its work 😦

[echo wraith]

alright there we go

[lofty parcel]

Why u sad

It works

[lethal loom]

yeah thx for all helpers 🙂

[sage tendon]

because we wasted a long time on an issue outside of our control, which should have gone into a stable version ages ago

that fucking PR was merged 2 years ago

[echo wraith]

what

[sage tendon]

wait

[echo wraith]

my pr ?

[sage tendon]

ugh okay im tierd

no the issue, i thought it was a pR

[echo wraith]

this is the pr

[sage tendon]

well nevermind that, we need 2.7 now seriously

[echo wraith]

the issue btw
InteractionPayload["channel"] is partial, thus, when creating a GuildChannel object from it, some attributes were missing/incorrectly set. With this pr, InteractionPayload["channel"] is used only in the following cases:

1. When the interaction happens in private channels.
2. When the channel is missing from the bot's cache.

In all other cases, it uses the full channel object from the cache.

[sage tendon]

ill pretend i understand

[echo wraith]

Basically with every interaction discord sends a channel object so we tought it was a good idea to use it instead of the cached one but that one was missing permissions

whatever

gn chat

[raw beacon]

For slash command interactions, how would i get the supplied values of the command arguments?

[lofty parcel]

It works exactly the same way. You just use the function arguments you set.

[lapis dock]

You should have it setup like arguments to the callback function

@bot.slash_command()
async def warn(ctx, user, reason):
  ...

/warn user:@Ice Wolfy reason:bad :3

[lofty parcel]

OK wolfy beat me with the example

[raw beacon]

Oh i guess i wasnt clear, apologies, i meant if i have the interaction object of a command that was executed

just like how i would get the command's name with interaction.command.name

[little cobalt]

dont forget the ctx

[lapis dock]

You can access the raw data of in interaction in the Interaction.data property.
To make it more friendly I think you could turn the interaction into ApplicationContext

context = discord.ApplicationContext(bot, my_interaction)

This should work, but I have never done it myself

Actually, this might not work

[raw beacon]

AttributeError: 'ApplicationContext' object has no attribute 'data'

[lapis dock]

oh, you already have an application context object, not interaction

[lofty parcel]

Slash commands receive an ApplicationContext, not a "plain" interaction object like discord.py

[lapis dock]

Then you can just access with ctx.command ctx.selected_options

[round heart]

So if you have ctx, it'd just be ctx.command which is the SlashCommand object (and the ctx.selected_options as Wolfy mentions)

[raw beacon]

command.options only returned the args it has, not values that were supplied

[lapis dock]

^ selected options includes the user provided values

[raw beacon]

Yep! thank you for your help

[wintry pelican]

hullo... i tried searching first, but i couldn't find an answer for my question...

i am running a bot that maintains a set of threads for different work issues; we use Jira which posts an event to our bot on every issue update, then the bot checks to see if that issue should have a Discord thread and if so creates/updates it. we are hitting this rate limit after three or four updates in a minute or so:

[2025-03-13 20:27:13] [WARNING ] discord.http: We are being rate limited. PATCH https://discord.com/api/v10/channels/1349184233557590076 responded with 429. Retrying in 518.31 seconds.

i am trying to figure out which rate limit we are hitting (global? that endpoint? some grouping of endpoints?) and what the limit is. i know from the docs some of that is available in the http response headers, but i don't know how to get at those using discord.py. does anyone have any advice? ty in advance!

[sage tendon]

Are you ever renaming the channel

[silk spire]

oh no do threads have the same limits now nvm that's a channel url

[lapis dock]

Also note that this is the support server for py-cord not discord.py

[wintry pelican]

right sorry, i meant py-cord. 🙂

[wintry pelican]

yes that happens fairly often.

[sage tendon]

renaming channel has an extremely low rate limit

[wintry pelican]

i bet that's it, ty!

[silk spire]

channel edit name/topic has a ratelimit of 2 changes per 10 mins, discord does not want you to use channels as some sort of counters of stuff

[sage tendon]

yea prolly

[silk spire]

threads shouldn't have this low limit

use and rename those instead, if possible

[wintry pelican]

will do, thank you so much.

[maiden bloom]

Can an event listener be added for when entitlements are cancelled and expire?

[lofty parcel]

.rtfm on_entitlement

[sly karma]

Target not found, try again and make sure to check your spelling.

[nimble matrix]

Searched in pycord

discord.on_entitlement_create
discord.on_entitlement_delete
discord.on_entitlement_update

[lofty parcel]

You listen to the update event

[maiden bloom]

Ah so how do I know if the entitlement is being renewed or cancelled

[lofty parcel]

Per the Discord docs:

The start of a subscription is determined by its ID. When the subscription renews, its current period is updated.

If the user cancels the subscription, the subscription will enter the ENDING status and the canceled_at timestamp will reflect the time of the cancellation.

[maiden bloom]

Oh, sweet. Do you also know how I can tell a test entitlement apart from a purchased one? They both have the application_subscription EntitlementType

[lofty parcel]

uhh

not really, i have never worked with entitlements

but they should differ somehow, check the discord docs themselves

[tulip verge]

hey so when i try to install pycord it comes up with this

[lofty parcel]

remove the $

[tulip verge]

yeah i did then it comes up with the same error for install

then pip

see

[lofty parcel]

Where are you running this?...

[tulip verge]

in pyhton

[lofty parcel]

You don't run it in the python idle

You use your terminal

[tulip verge]

where tf i get that lmao

[lofty parcel]

You open a Command Prompt

[tulip verge]

yeah

[lofty parcel]

or simply type cmd in the search bar of your apps

[tulip verge]

yeah i got it up

[lofty parcel]

ok now install the library

consider learning python before trying to create a discord bot

[tulip verge]

just need it for a account gen lol

[lofty parcel]

a what?

[tulip verge]

dw

btw

it still dont weork

[lofty parcel]

No, it does matter, because if it breaks the ToS, I won't help you and remove you from here.

[tulip verge]

what lmao

what breaks tos

doesnt say anywhere in rules about it

[lofty parcel]

"account gen"

[tulip verge]

so idk what u on about

[lofty parcel]

What

Following the Discord ToS is common sense

you dont need to read it in a rules channel

[tulip verge]

💀

noone finna read discord tos

ibsr

[lofty parcel]

Not reading it does not make you excempt from it.

[tulip verge]

this makse sense now

[round heart]

It's funny when people who don't even understand that Windows is not for development are trying to do stupid shit like auto-generating accounts.

[flat wind]

Hello !
is it possible to know if a timestamp is sent in a message ? like a mention I mean.
f.e a message containing a mention will have <@someID> but also message.mentions, is it possible to get the same information for the <t:123:F> ?

if not, I will fetch it manually using a regex

[silk spire]

You need to regex it yourself yes

[flat wind]

okay thanks, it will not difficult but if discord provided the information 🤷‍♂️

[fresh sierra]

There might be something in discord.utils