a debugging and utility extension for discord.py bots
Read the documentation online.
jishaku is an extension for bot developers that enables rapid prototyping, experimentation, and debugging of features for bots.
One of jishaku's core philosophies is to be dynamic and easy-to-use. Here's the two step install:
- Download jishaku on the command line using pip:
- Load the extension in your bot code before it runs:
You can also import the module to use the command development utilities.
jishaku [py|python] <argument>
The Python commands execute or evaluate Python code passed into them.
It supports simple expressions:
jishaku [pyi|python_inspect] <argument>
> jsk py 3+4
It also supports async expressions:
> jsk py await _ctx.pins()
[<Message id=123456789012345678 ...>, ...]
You can pass in codeblocks for longer blocks of code to execute, and you can use
yield to return intermediate results within your processing.
The inspect variant of the command will return a codeblock with detailed inspection information on all objects returned.
The variables available by default in all execution contexts are:
The Context that invoked the command.
The running Bot instance.
Shortcuts for attributes on
The underscore prefix on the provided variables is intended to help prevent shadowing when writing large blocks of code within the command.
If you decide that you don't want the prefix, you can disable it by setting the
JISHAKU_NO_UNDERSCORE environment variable to
Each Python command is individually scoped. That means variables you create won't be retained in later invocations.
You can use
jishaku retain on to change this behavior and retain variables, and
jishaku retain off if you change your mind later.
This command compiles Python code in an asynchronous context, and then disassembles the resulting function into Python bytecode in the style of >
jishaku [dis|disassemble] <argument>
This allows you to quickly and easily determine the bytecode that results from a given expression or piece of code. The code itself is not actually executed.
This command compiles Python code into its Abstract Syntax Tree, and then displays it as a formatted tree of AST blocks.
jishaku ast <argument>
The shell command executes commands within your system shell.
jishaku [sh|shell] <argument>
If you're on Linux and are using a custom shell, jishaku will obey the
SHELL environment variable, otherwise, it will use
On Windows, jishaku will use PowerShell if it's detected, otherwise, it will use Command Prompt.
The results from the commands you pass in are returned through a paginator interface live as the command runs. If you need to stop a command, you can press the stop button reaction, or use
The execution will terminate automatically if no output is produced for 120 seconds.
jishaku git <argument>
These commands act as shortcuts to the shell command, so you can save typing a word if you use these frequently.
jishaku pip <argument>
jishaku [load|reload] [extensions...]
These commands load, reload, or unload extensions on your bot.
jishaku unload [extensions...]
You can reload jishaku itself with
jsk reload jishaku.
jsk reload ~ will reload all extensions on your bot.
You can load, reload, or unload multiple extensions at once:
jsk reload cogs.one cogs.two
This command gracefully shuts down your bot.
This command calculates > Round-Trip Time for your bot to the API. It does this by calculating response time samples, so you can tell if your bot is being slow or not.
This command syncs the app_command tree for a list of guilds, or globally if none are given, and then reports the results.
jishaku sync [guild_ids...]
This command reads a file from your file system, automatically detecting encoding and (if applicable) highlighting.
jishaku cat <file>
You can use this to read things like bot logs or source files in your project.
This command reads text from a URL and attempts to detect encoding and language, similar to >
jishaku curl <url>
You can use this to display contents of files online, for instance, the
message.txt files created when a message is too long, or raw files from paste sites.
jishaku exec [member and/or channel...] <command string>
jishaku debug <command string>
These commands serve as command control for other commands.
jishaku repeat <times> <command string>
jishaku exec allows you to execute a command as another user, in another channel, or both. Using aliases with a postfix exclamation mark (such as
jsk exec! ...) executes the command bypassing checks and cooldowns.
jishaku debug executes a command with an exception wrapper and a timer. This allows you to quickly get feedback on reproducable command errors and slowdowns.
jishaku repeat repeats a command a number of times.
This command allows you to investigate the source of expressed permissions in a given channel. Targets can be either a member, or a list of roles (to simulate a member with those roles).
jishaku permtrace <channel> [targets...]
It will read all of the guild permissions and channel overwrites for the given member or roles in the channel, and provide a breakdown containing whether the permission is granted, and the most fundamental reason why.
Installing development versions
If you'd like to test the latest versions of jishaku, you can do so by downloading from the git hosts instead of from PyPI.
pip install -U "jishaku @ git+https://github.com/Gorialis/jishaku@master"
pip install -U "jishaku @ git+https://gitlab.com/Gorialis/jishaku@master"
Please note that the new 2020 dependency resolver now no longer discounts git package sources from reinstall prevention,
which means that if you are installing the jishaku development version multiple times within the same version target you may run into pip just discarding the update.
If you run into such a problem, you can force jishaku to be reinstalled like this:
pip install -U --force-reinstall "jishaku @ git+https://github.com/Gorialis/jishaku@master"
pip install -U --force-reinstall "jishaku @ git+https://gitlab.com/Gorialis/jishaku@master"
You must have installed jishaku with one of the commands above before doing this else you will probably end up with a broken installation.
The documentation and this README uses icons from the Material Design Icon library, which is licensed under the Apache License Version 2.0.