ancli

Building argument parser from a function annotation. A simple utility inspired by Fire and docopt. Ad-hoc solution for someone who often writes scripts with a single entry point.

How?

The process of building CLI with ancli is very simple.

1. Write a plain Python function with annotated parameters.
2. Wrap it with make_cli.
3. Run your script.

Examples

1. Function with annotated parameters

The function run has explicitly annotated parameters and its signature is used to instantiate argparse.ArgumentParser instance that accepts parameters with specific types and default (if any) parameters. If default value is not provided, then the parameter is considered to be required.

from ancli import make_cli

def run(path: str, flag: bool = True, iterations: int = 1):
    print(f'run: path={path}, flag={flag}, iterations={iterations}')

if __name__ == '__main__':
    make_cli(run)

Now this snippet can be used as follows.

$ python script.py --path file.txt --flag 0
run: path=file.txt, flag=False, iterations=1

2. Function without annotations

The functions without type annotations try to infer the parameters types based on their default values.

from ancli import make_cli

def run(a, b=2, c=3.0):
    for param in (a, b, c):
        print(type(param))

if __name__ == '__main__':
    make_cli(run)

The parameters without default values are considered as strings.

$ python script.py --a 1 --b 2 --c 3.0
<type 'str'>
<type 'int'>
<type 'float'>

3. Running ancli as a module

Running package as a module allows to dynamically build a CLI from some function. You just need to specify a path to the module, and function which should be treated as an entry point.

$ python -m ancli examples.functions:compute --a 2 --b 6
42