Commands¶

We have seen how to create a CLI program with possibly several CLI options and CLI arguments.

But Typer allows you to create CLI programs with several commands (also known as subcommands).

For example, the program git has several commands.

One command of git is git push. And git push in turn takes its own CLI arguments and CLI options.

For example:

// The push command with no parameters
$ git push

---> 100%

// The push command with one CLI option --set-upstream and 2 CLI arguments
$ git push --set-upstream origin master

---> 100%

Another command of git is git pull, it also has some CLI parameters.

It's like if the same big program git had several small programs inside.

Tip

A command looks the same as a CLI argument, it's just some name without a preceding --. But commands have a predefined name, and are used to group different sets of functionalities into the same CLI application.

Command or subcommand¶

It's common to call a CLI program a "command".

But when one of these programs have subcommands, those subcommands are also frequently called just "commands".

Have that in mind so you don't get confused.

Here I'll use CLI application or program to refer to the program you are building in Python with Typer, and command to refer to one of these "subcommands" of your program.

Explicit application¶

Before creating CLI applications with multiple commands/subcommands we need to understand how to create an explicit typer.Typer() application.

In the CLI options and CLI argument tutorials you have seen how to create a single function and then pass that function to typer.run().

For example:

Python 3.8+

import typer


def main(name: str):
    print(f"Hello {name}")


if __name__ == "__main__":
    typer.run(main)

But that is actually a shortcut. Under the hood, Typer converts that to a CLI application with typer.Typer() and executes it. All that inside of typer.run().

There's also a more explicit way to achieve the same:

Python 3.8+

import typer

app = typer.Typer()


@app.command()
def main(name: str):
    print(f"Hello {name}")


if __name__ == "__main__":
    app()

When you use typer.run(), Typer is doing more or less the same as above, it will:

Create a new typer.Typer() "application".
Create a new "command" with your function.
Call the same "application" as if it was a function with "app()".

@decorator Info

That @something syntax in Python is called a "decorator".

You put it on top of a function. Like a pretty decorative hat (I guess that's where the term came from).

A "decorator" takes the function below and does something with it.

In our case, this decorator tells Typer that the function below is a "command".

Both ways, with typer.run() and creating the explicit application, achieve almost the same.

Tip

If your use case is solved with just typer.run(), that's fine, you don't have to create the explicit app and use @app.command(), etc.

You might want to do that later when your app needs the extra features, but if it doesn't need them yet, that's fine.

If you run the second example, with the explicit app, it works exactly the same:

// Without a CLI argument
$ python main.py

Usage: main.py [OPTIONS] NAME
Try "main.py --help" for help.

Error: Missing argument 'NAME'.

// With the NAME CLI argument
$ python main.py Camila

Hello Camila

// Asking for help
$ python main.py  --help

Usage: main.py [OPTIONS] NAME

Options:
  --install-completion  Install completion for the current shell.
  --show-completion     Show completion for the current shell, to copy it or customize the installation.
  --help                Show this message and exit.

CLI application completion¶

There's a little detail that is worth noting here.

Now the help shows two new CLI options:

--install-completion
--show-completion

To get shell/tab completion, it's necessary to build a package that you and your users can install and call directly.

So instead of running a Python script like:

$ python main.py

✨ Some magic here ✨

...It would be called like:

$ magic-app

✨ Some magic here ✨

Having a standalone program like that allows setting up shell/tab completion.

The first step to be able to create an installable package like that is to use an explicit typer.Typer() app.

Later you can learn all the process to create a standalone CLI application and Build a Package.

But for now, it's just good to know that you are on that path. 😎

A CLI application with multiple commands¶

Coming back to the CLI applications with multiple commands/subcommands, Typer allows creating CLI applications with multiple of them.

Now that you know how to create an explicit typer.Typer() application and add one command, let's see how to add multiple commands.

Let's say that we have a CLI application to manage users.

We'll have a command to create users and another command to delete them.

To begin, let's say it can only create and delete one single predefined user:

Python 3.8+

import typer

app = typer.Typer()


@app.command()
def create():
    print("Creating user: Hiro Hamada")


@app.command()
def delete():
    print("Deleting user: Hiro Hamada")


if __name__ == "__main__":
    app()

Now we have a CLI application with 2 commands, create and delete:

// Check the help
$ python main.py --help

Usage: main.py [OPTIONS] COMMAND [ARGS]...

Options:
  --install-completion  Install completion for the current shell.
  --show-completion     Show completion for the current shell, to copy it or customize the installation.
  --help                Show this message and exit.

Commands:
  create
  delete

// Test them
$ python main.py create

Creating user: Hiro Hamada

$ python main.py delete

Deleting user: Hiro Hamada

// Now we have 2 commands! 🎉

Notice that the help text now shows the 2 commands: create and delete.

Tip

By default, the names of the commands are generated from the function name.

Show the help message if no command is given¶

By default, we need to specify --help to get the command's help page.

However, by setting no_args_is_help=True when defining the typer.Typer() application, the help function will be shown whenever no argument is given:

Python 3.8+

import typer

app = typer.Typer(no_args_is_help=True)


@app.command()
def create():
    print("Creating user: Hiro Hamada")


@app.command()
def delete():
    print("Deleting user: Hiro Hamada")


if __name__ == "__main__":
    app()

Now we can run this:

// Check the help without having to type --help
$ python main.py

Usage: main.py [OPTIONS] COMMAND [ARGS]...

Options:
  --install-completion  Install completion for the current shell.
  --show-completion     Show completion for the current shell, to copy it or customize the installation.
  --help                Show this message and exit.

Commands:
  create
  delete

Sorting of the commands¶

Note that by design, Typer shows the commands in the order they've been declared.

So, if we take our original example, with create and delete commands, and reverse the order in the Python file:

Python 3.8+

import typer

app = typer.Typer()


@app.command()
def delete():
    print("Deleting user: Hiro Hamada")


@app.command()
def create():
    print("Creating user: Hiro Hamada")


if __name__ == "__main__":
    app()

Then we will see the delete command first in the help output:

// Check the help
$ python main.py --help

Usage: main.py [OPTIONS] COMMAND [ARGS]...

Options:
  --install-completion  Install completion for the current shell.
  --show-completion     Show completion for the current shell, to copy it or customize the installation.
  --help                Show this message and exit.

Commands:
  delete
  create

Click Group¶

If you come from Click, a typer.Typer app with subcommands is more or less the equivalent of a Click Group.

Technical Details

A typer.Typer app is not a Click Group, but it provides the equivalent functionality. And it creates a Click Group when calling it.

It is not directly a Group because Typer doesn't modify the functions in your code to convert them to another type of object, it only registers them.

Decorator Technical Details¶

When you use @app.command() the function under the decorator is registered in the Typer application and is then used later by the application.

But Typer doesn't modify that function itself, the function is left as is.

That means that if your function is simple enough that you could create it without using typer.Option() or typer.Argument(), you could use the same function for a Typer application and a FastAPI application putting both decorators on top, or similar tricks.

Click Technical Details

This behavior is a design difference with Click.

In Click, when you add a @click.command() decorator it actually modifies the function underneath and replaces it with an object.