Create a command line interface with minimal setup.

Command line interface (Cli) with a schema (Ma) - Clima

Installation

pip install clima

Requires Python 3.9.2+.

Quick example

from clima import Schema

class S(Schema):
    place = 'Finland'

@S.cli
class Cli:
    def say_hi(self):
        print(f'Hi from {S.place}')

$ cli.py say_hi
Hi from Finland
$ cli.py say_hi --place 'Sweden'
Hi from Sweden

Why Clima?

Clima eliminates CLI boilerplate: just define a Schema dataclass and a Cli class. Opinionated for convenience; Unlike click, typer, or argparse, clima gives you a built-in config cascade (CLI args → env vars → .env file → config file → defaults) with zero extra code.

Features

Schema fields double as CLI flags, environment variables, and config file keys automatically while providing help printout, default values and type casting.

• Config cascade :: CLI args → env vars → .env file → config file → defaults, resolved automatically
• Type casting :: Schema field annotations are used to cast string CLI/env values to the right type
• Help from field comments :: Docstrings on Schema fields act as --help output
• IDE completions :: C.name is typed as str — completions and type checking work natively
• --verbose / --quiet logging :: Add verbose: bool or quiet: bool to Schema and get preconfigured logging
• Undefined param warnings :: Unknown --flags on the command line produce a clear warning
• version subcommand :: myscript version prints the package version automatically
• Config file support :: Declare defaults in toml .toml or INI-style .conf file keyed to your package name
• .env file and env var support :: Schema fields are also read from environment variables and .env files
• Optional gpg secrets :: Decrypt secrets via pass / gnupg if installed
• Optional shortened tracebacks :: Truncate python tracebacks into an opinionated format

Documentation

Full documentation at python-clima.readthedocs.io.

See and run the examples/ directory for more usage patterns.

AI Disclaimer

This work started in 2019 as my hobby project and continues as so. Claude was used to bridge gaps I had planned, but had not had the time to address.

Before merging to main and uploading to pypi the code is hand-vetted to find all code assistance related issues.