Embeded Error Messages

[chanchantang]

I am trying to embed an error message in sfu.py, but I am getting an error trying to do so.
Here is the code in question:

e_obj = await embed(
    self.logger, interaction=interaction, title='SFU Courses Error',
    description=(
        f'Couldn\'t find anything for `{", ".join(departments)}, level {level}, term {term}, year {year}`'
        f'\n Maybe the course doesn\'t exist? Or isn\'t offered right now.'
    ),
    colour=WallEColour.ERROR,
)
if e_obj:
    msg = await interaction.followup.send(embed=e_obj)

The error I get is

If I had to guess, it would have to do something with this being a slash command but idk
Here is the entire code if needed

[modernNeo]

if I am not misremembering, you use interaction.followup.send(embed=e_obj) for an interaction where you have deferred it via await interaction.response.defer()

If you haven't deferred, you would only need interaction.response.send_message(embed=e_obj)

and just so you know, you only need to defer an interaction if you anticipate that the command you are running will take longer than 3 seconds to compute and respond to the command.

[chanchantang]

Hmm, I'm confused on how to use defer.
For example, in the bans() function, it uses defer in the beginning, then an await BanRecord query, then parsing it. I'm guessing this could take more than 3 seconds, so we use defer and followup.
But for the ban() function, it uses interaction.response.defer() in the beginning as well, then right after a interaction.followup.send if the purge window days are outside the range. Is it doing this because performing wall_e_ban() might take more than 3 seconds? If so, why doesn't the ban intercept() function need to use it?

But interaction.response.send_message(embed=e_obj) works perfectly, thanks a lot!

[modernNeo]

For example, in the bans() function, it uses defer in the beginning, then an await BanRecord query, then parsing it. I'm guessing this could take more than 3 seconds, so we use defer and followup.

Correct.

But for the ban() function, it uses interaction.response.defer() in the beginning as well, then right after a interaction.followup.send if the purge window days are outside the range. Is it doing this because performing wall_e_ban() might take more than 3 seconds? If so, why doesn't the ban intercept() function need to use it?

Can you point me to where the ban function uses interaction.response.defer right after the await interaction.followup.send?

[chanchantang]

Oh my apologies I miswrote, I meant it uses followup right after defer, ie defer() -> followup().

My main question was meant to ask how ban() uses defer() -> followup() for error -> wall_e_ban(),
but then within wall_e_ban() it does not use followup in its function

Also, if the main concern was the calling of wall_e_ban(), the function intercept(), which uses wall_e_ban(), does no use defer nor followup

Sorry if this does not make sense

[modernNeo]

Oh my apologies I miswrote, I meant it uses followup right after defer, ie defer() -> followup().

Are you referring to this bit here?

wall_e/wall_e/extensions/ban.py

Lines 404 to 418 in c0e6d39

	try:
	await interaction.response.defer()
	except discord.errors.NotFound:
	await interaction.channel.send(
	"Feeling a bit overloaded at the moment...Please try again in a few minutes"
	)
	return
	if purge_window_days <= 0 or purge_window_days > 14:
	e_obj = await embed(
	self.logger, interaction=interaction, title=f'{Ban.embed_title} Invalid Purge Specified',
	description=f'Window to purge message must be between 1 - 14 days instead of {purge_window_days}. '
	f'Using default of `1 day`'
	)
	if e_obj:
	await interaction.followup.send(embed=e_obj)

My main question was meant to ask how ban() uses defer() -> followup() for error -> wall_e_ban(),
but then within wall_e_ban() it does not use followup in its function

So, there are a few different things at play there.
given that ban is a slash command, that means that the command has to respond to the invoking message somehow [to use proper discord.py terminology, the interaction has to be responded to], this can be done either via

• interaction.response.send_message if you know the code will take <= 3 seconds to do its thing
• or via interaction.response.defer/interaction.followup.send if the response to the invoking message will take longer than 3 seconds.

However, the discord.py additionally allows us to send message in different contexts:

1. for example, we are given to option to send a stand-alone message just in the channel where the invoking message was called from but not as a response to the invoking message:
   

   wall_e/wall_e/extensions/ban.py

   Lines 407 to 410 in c0e6d39

   	await interaction.channel.send(
   	"Feeling a bit overloaded at the moment...Please try again in a few minutes"
   	)
   	return

2. in the wall_e_ban function, if the interaction object is not specified, then the message are just sent to the mod channel via this function:

   wall_e/wall_e/extensions/ban.py

   Lines 251 to 253 in c0e6d39

   	send_function_for_messages = self.mod_channel.send if interaction is None else \
   	interaction.edit_original_response
   	if banned_user.id in Ban.ban_list:

3. or we can aways send a DM directly to the user directly like this [assuming they haven't disabled DMs]:

   wall_e/wall_e/extensions/ban.py

   Line 350 in c0e6d39

   await banned_user.send(embed=e_obj)

But to more directly answer your question,

if you look back to

wall_e/wall_e/extensions/ban.py

Lines 251 to 252 in c0e6d39

	send_function_for_messages = self.mod_channel.send if interaction is None else \
	interaction.edit_original_response

you will notice that if the interaction parameter is not None, the function used is not interaction.followup.send, but interaction.edit_original_response.

What is going on is a user invoked the ban command, thereby creating an interaction that WALL_E has to respond to in some way. So the ban command first tells the discord API for that command that it needs to wait longer than 3 seconds for the reply via interaction.response.defer before eventually responding via interaction.followup.send. but even once the command has responded, while you cannot re-repond even if you have new information to relay to the user, you can either just send another non-interaction message tagging the original user or you can modify the response [edit the message that the bot responded back with] by using interaction.edit_original_response to include the new info.

Hope that clears it up.

[chanchantang]

Oooh I see, I think I get it now. Thanks for explaining it so thoroughly!