This guide will walkthrough the ways to create a custom help command by subclassing HelpCommand.

Skip ahead if you already know this.

In simple terms, a subclass is a way to inherit a class behaviour/attributes from another class. Here's how you would subclass a class in Python.

class A:
    def __init__(self, attribute1):
        self.attribute1 = attribute1

    def method1(self):
        print("method 1")

    def method2(self):
        print("method 2")


class B(A):
    def __init__(self, attribute1, attribute2):
        super().__init__(attribute1) # This calls A().__init__ magic method.
        self.attribute2 = attribute2

    def method1(self): # Overrides A().method1
        print("Hi")

Given the example, the variable instance_a contains the instance of class A. As expected, the output will be this.

>>> instance_a = A(1)
>>> instance_a.attribute1
1
>>> instance_a.method1()
"method 1"

How about B class? instance_b contains the instance of class B, it inherits attributes/methods from class A. Meaning it will have everything that class A have, but with an additional attribute/methods.

>>> instance_b = B(1, 2)
>>> instance_b.attribute1
1
>>> instance_b.attribute2
2
>>> instance_b.method1()
"Hi"
>>> instance_b.method2()
"method 2"

Make sure to look and practice more into subclassing classes to understand fully on what it is before diving into subclassing HelpCommand.

With that out of the way, let's get started. For subclassing HelpCommand, first, you would need to know the types of HelpCommand. Where each class has their own usage.

There are a few types of HelpCommand classes that you can choose;

1. DefaultHelpCommand a help command that is given by default.
2. MinimalHelpCommand a slightly better help command.
3. HelpCommand an empty class that is the base class for every HelpCommand you see. On its own, it will not do anything.

By default, help command is using the class DefaultHelpCommand. This is stored in bot.help_command. This attribute will ONLY accept instances that subclasses HelpCommand. Here is how you were to use the DefaultHelpCommand instance.

from discord.ext import commands
bot = commands.Bot(command_prefix="uwu ")
bot.help_command = commands.DefaultHelpCommand()

# OR

bot = commands.Bot(command_prefix="uwu ", help_command=commands.DefaultHelpCommand())
# Both are equivalent

Now, of course, this is done by default. I'm only showing you this as a demonstration. Don't scream at me

Let's do the same thing with MinimalHelpCommand next.

from discord.ext import commands
bot = commands.Bot(command_prefix="uwu ")
bot.help_command = commands.MinimalHelpCommand()

Now say, you want the content to be inside an embed. But you don't want to change the content of DefaultHelpCommand/MinimalHelpCommand since you want a simple HelpCommand with minimal work. There is a short code from ?tag embed help example by gogert in discord.py server, a sample code you can follow shows this;

import discord
from discord.ext import commands
bot = commands.Bot(command_prefix="uwu ")

class MyNewHelp(commands.MinimalHelpCommand):
    async def send_pages(self):
        destination = self.get_destination()
        for page in self.paginator.pages:
            emby = discord.Embed(description=page)
            await destination.send(embed=emby)

bot.help_command = MyNewHelp()

The resulting code will show that it have the content of MinimalHelpCommand but in an embed.

Looking over the MinimalHelpCommand source code, here. Every method that is responsible for <prefix>help <argument> will call MinimalHelpCommand.send_pages when it is about to send the content. This makes it easy to just override send_pages without having to override any other method there are in MinimalHelpCommand.

If you want to use HelpCommand class, we need to understand the basic of subclassing HelpCommand. Here are a list of HelpCommand relevant methods, and it's responsibility.

1. HelpCommand.send_bot_help(mapping) Gets called with <prefix>help
2. HelpCommand.send_command_help(command) Gets called with <prefix>help <command>
3. HelpCommand.send_group_help(group) Gets called with <prefix>help <group>
4. HelpCommand.send_cog_help(cog) Gets called with <prefix>help <cog>

1. HelpCommand.context the Context object in the help command.
2. HelpCommand.clean_prefix a cleanup prefix version that remove any mentions.

Seems simple enough? Now let's see what happens if you override one of the methods. Here's an example code of how you would do that. This override will say "hello!" when you type <prefix>help to demonstrate on what's going on.

We'll use HelpCommand.get_destination() to get the abc.Messageable instance for sending a message to the correct channel.

from discord.ext import commands
bot = commands.Bot(command_prefix="uwu ")

class MyHelp(commands.HelpCommand):
    async def send_bot_help(self, mapping):
        channel = self.get_destination()
        await channel.send("hello!")

bot.help_command = MyHelp()

Keep in mind, using HelpCommand class will require overriding every send_x_help methods. For example, <prefix>help jsk is a command that should call send_command_help method. However, since HelpCommand is an empty class, it will not say anything.

Let's work our way to create a <prefix> help. Given the documentation, await send_bot_help(mapping) method receives mapping(Mapping[Optional[Cog], List[Command]]) as its parameter. await indicates that it should be an async function.

• Mapping[] is a collections.abc.Mapping, for simplicity’s sake, this usually refers to a dictionary since it's under collections.abc.Mapping.
• Optional[Cog] is a Cog object that has a chance to be None.
• List[Command] is a list of Command objects.
• Mapping[Optional[Cog], List[Command]] means it's a map object with Optional[Cog] as it's key and List[Command] as its value.

All of these are typehints in the typing module. You can learn more about it here.

Now, for an example, we will use this mapping given in the parameter of send_bot_help. For each of the command, we'll use HelpCommand.get_command_signature(command) to get the command signature of a command in an str form.

import discord
from discord.ext import commands
bot = commands.Bot(command_prefix="uwu ")

class MyHelp(commands.HelpCommand):
    async def send_bot_help(self, mapping):
        embed = discord.Embed(title="Help")
        for cog, commands in mapping.items():
           command_signatures = [self.get_command_signature(c) for c in commands]
           if command_signatures:
                cog_name = getattr(cog, "qualified_name", "No Category")
                embed.add_field(name=cog_name, value="\n".join(command_signatures), inline=False)

        channel = self.get_destination()
        await channel.send(embed=embed)

bot.help_command = MyHelp()

1. Create an embed.
2. Use dict.items() to get an iterable of (Cog, list[Command]).
3. Each element in list[Command], we will call self.get_command_signature(command) to get the proper signature of the command.
4. If the list is empty, meaning, no commands is available in the cog, we don't need to show it, hence if command_signatures:.
5. cog has a chance to be None, this refers to No Category. We'll use getattr to avoid getting an error to get cog's name through Cog.qualified_name.
6. Using str.join each command will be displayed on a separate line.
7. Once all of this is finished, display it.

I agree, this does not look pretty. But this demonstrate how to create a help command with minimal effort. The reason why it looks ugly, is that we're using HelpCommand.get_command_signature(command). Let's override that method to make it a little more readable.

We'll borrow codes from MinimalHelpCommand.get_command_signature. Optionally, we can subclass MinimalHelpCommand instead of copying codes. I'm doing this as a demonstration of overriding other methods.

We'll also use HelpCommand.filter_commands, this method will filter commands by removing any commands that the user cannot use. It is a handy method to use.

import discord
from discord.ext import commands
bot = commands.Bot(command_prefix="uwu ")

class MyHelp(commands.HelpCommand):
   def get_command_signature(self, command):
        return '%s%s %s' % (self.clean_prefix, command.qualified_name, command.signature)

    async def send_bot_help(self, mapping):
        embed = discord.Embed(title="Help")
        for cog, commands in mapping.items():
           filtered = await self.filter_commands(commands, sort=True)
           command_signatures = [self.get_command_signature(c) for c in filtered]
           if command_signatures:
                cog_name = getattr(cog, "qualified_name", "No Category")
                embed.add_field(name=cog_name, value="\n".join(command_signatures), inline=False)

        channel = self.get_destination()
        await channel.send(embed=embed)

bot.help_command = MyHelp()

This looks more readable than the other one. While this should cover most of your needs, you may want to know more helpful attribute that is available on HelpCommand in the official documentation.

Now that the hard part is done, let's take a look at <prefix>help [argument]. The method responsible for this is as follows;

1. send_command_help
2. send_cog_help
3. send_group_help

As a demonstration, let's go for send_command_help this method receive a Command object. For this, it's simple, all you have show is the attribute of the command.

For example, this is your command code, your goal is you want to show the help,aliases and the signature.

@bot.command(help="Shows all bot's command usage in the server on a sorted list.",
             aliases=["br", "brrrr", "botranks", "botpos", "botposition", "botpositions"])
async def botrank(ctx, bot: discord.Member):
   pass

Then it's simple, you can display each of the attribute by Command.help and Command.aliases.

For the signature, instead of using the previous get_command_signature, we're going to subclass MinimalHelpCommand.

import discord
from discord.ext import commands
bot = commands.Bot(command_prefix="uwu ")

class MyHelp(commands.MinimalHelpCommand):
    async def send_command_help(self, command):
        embed = discord.Embed(title=self.get_command_signature(command))
        embed.add_field(name="Help", value=command.help)
        alias = command.aliases
        if alias:
            embed.add_field(name="Aliases", value=", ".join(alias), inline=False)

        channel = self.get_destination()
        await channel.send(embed=embed)

bot.help_command = MyHelp()

As you can see, it is very easy to create <prefix>help [argument]. The class already handles the pain of checking whether the given argument is a command, a cog, or a group command. It's up to you on how you want to display it, whether it's through a plain message, an embed or even using discord.ext.menus.

Let's say, someone is spamming your help command. For a normal command, all you have to do to combat this is using a cooldown decorator and slap that thing above the command declaration. Or, what about if you want an alias? Usually, you would put an aliases kwargs in the command decorator. However, HelpCommand is a bit special, It's a god damn class. You can't just put a decorator on it and expect it to work.

That is when HelpCommand.command_attrs come to the rescue. This attribute can be set during the HelpCommand declaration, or a direct attribute assignment. According to the documentation, it accepts exactly the same thing as a command decorator in a form of a dictionary.

For example, we want to rename the help command as "hell" instead of "help" for whatever reason. We also want to make an alias for "help" so users can call the command with "hell" and "help". Finally, we want to put a cooldown, because help command messages are big, and we don't want people to spam those. So, what would the code look like?

from discord.ext import commands

attributes = {
   'name': "hell",
   'aliases': ["help", "helps"],
   'cooldown': commands.Cooldown(2, 5.0, commands.BucketType.user)
}

# During declaration
help_object = commands.MinimalHelpCommand(command_attrs=attributes)

# OR through attribute assignment
help_object = commands.MinimalHelpCommand()
help_object.command_attrs = attributes

bot = commands.Bot(command_prefix="uwu ", help_command=help_object)

1. sets the name into "hell" is refers to here 'name': "hell".
2. sets the aliases by passing the list of str to the aliases key, which refers to here 'aliases': ["help", "helps"].
3. sets the cooldown through the "cooldown" key by passing in a Cooldown object. This object will make a cooldown with a rate of 2, per 5 with a bucket type BucketType.user, which in simple terms, for every discord.User, they can call the command twice, every 5 seconds.
4. We're going to use MinimalHelpCommand as the HelpCommand object.

As you can see, the name of the help command is now "hell", and you can also trigger the help command by "help". It will also raise an OnCommandCooldown error if it was triggered 3 times in 5 seconds due to our Cooldown object. Of course, I didn't show that in the result, but you can try it yourself. You should handle the error in an error handler when an OnCommandCooldown is raised.

What happens when <prefix>help command fails to get a command/cog/group? Simple, HelpCommand.send_error_message will be called. HelpCommand will not call on_command_error when it can't find an existing command. It will also give you an str instead of an error instance.

import discord
from discord.ext import commands
bot = commands.Bot(command_prefix="uwu ")

class MyHelp(commands.HelpCommand):
    async def send_error_message(self, error):
        embed = discord.Embed(title="Error", value=error)
        channel = self.get_destination()
        await channel.send(embed=embed)

bot.help_command = MyHelp()

error is a string that will only contain the message, all you have to do is display the message.

Indeed, we have it. HelpCommand.on_help_command_error, this method is responsible for handling any error just like any other local error handler.

import discord
from discord.ext import commands
bot = commands.Bot(command_prefix="uwu ")

class MyHelp(commands.HelpCommand):
    async def send_bot_help(self, mapping):
        raise commands.BadArgument("Something broke")

    async def on_help_command_error(self, ctx, error):
        if isinstance(error, commands.BadArgument):
            embed = discord.Embed(title="Error", description=str(error))
            await ctx.send(embed=embed)
        else:
            raise error

bot.help_command = MyHelp()

Rather basic, just raise an error that subclasses commands.CommandError such as commands.BadArgument. The error raised will cause on_help_command_error to be invoked. The code shown will catch this commands.BadArgument instance that is stored in error variable, and show the message.

To be fair, you should create a proper error handler through this official documentation. Here.

There is also a lovely example by Mysty on error handling in general. Here. This example shows how to properly create a global error handler and local error handler.

I hope that reading this walkthrough will assist you and give a better understanding on how to subclass HelpCommand. All the example code given are to demonstrate the feature of HelpCommand and feel free to try it. There are lots of creative things you can do to create a HelpCommand.

Here's my code if you want to take a look. my help command.

Now, of course, any question regarding HelpCommand should be asked in the discord.py server because I don't really check this gist as much, and because there is a lot of helpful discord.py helpers if you're nice enough to them. :D