Use type inference of pydantic settings class to create ArgumentParser

[bbugyi200]

• The following client code should work (see pydantic.BaseSettings for more information on pydantic's settings management):

### dummy.py

import clap


class Context(clap.Context):
    foo: str
    bar: int = 4


def run(ctx: Context) -> int:
    assert isinstance(ctx.foo, str)
    assert isinstance(ctx.bar, int)
    return 0


main = clap.main_factory(run)

• The foo and bar arguments should be derived from the following sources, ordered from least-preferred to most-preferred:
  • Default value (if specified) in Context class.
  • Global YAML configuration file (e.g. ~/.config/clap.yml).
  • Project-specific YAML configuration file (e.g. dummy.yml).
  • Environment variables (e.g. DUMMY_FOO or DUMMY_BAR) or .env file(s).
  • Command-line arguments (e.g. -f|--foo or -b|--bar).