Automatic CLI¶
Every project built with pyrig gets a working command-line interface for free. The CLI is a Typer application that assembles itself at startup, so you get a runnable entry point without wiring anything up — and you can extend it with your own commands.
The entry point¶
A project gets its CLI from a console script that points to pyrig-runtime's
entry point. pyrig adds this for you, but it is just a normal script entry in
pyproject.toml:
Running the script builds the Typer application and dispatches the requested
command. The invoking project is detected from the name of the script that
launched it (my-project above); converted to snake_case (my_project),
that name is where the project's own commands are loaded from.
-v / -q adjust logging verbosity. Something like -vv or -qqq or
even -vqqvqvq are also possible.
Adding your own commands¶
Define commands in your project's my_project.rig.cli.subcommands module:
- every function defined directly there becomes a top-level command;
- every module-level
typer.Typerinstance becomes a command group, named after its variable in kebab-case.
# src/my_project/rig/cli/subcommands.py
def greet(name: str) -> None:
"""Greet someone."""
print(f"hello {name}")
Customizing the build¶
The application is built by a CLI class, which is itself a
discoverable plugin. Define a subclass in your project's
my_project.rig.cli.cli module to override any step of the build — the
most-derived subclass is used automatically.
Version¶
Every project's CLI provides a root-level --version option that prints
the invoking project's installed version and exits successfully:
The project must be installed. The option is eager: it exits before the normal
logging callback or command executes, and does not require a subcommand.
Application construction and plugin discovery still happen before option parsing.
Override CLI.version() to customize version lookup.