PyArgWriter

PyArgWriter is a Python module that simplifies the generation of ArgumentParser setups for Python classes and their methods. It provides a convenient way to define and manage command-line arguments for your Python applications.

Installation

You can install PyArgWriter using pip:

pip install pyargwriter

Alternative from source code:

In root folder with pip:

pip install .

or with poetry:

poetry install

Command-Line Interface

PyArgWriter offers the following command-line commands:

• parse-code: Parse given files and create YAML structure with structural parser information.
• write-code: Read given parser YAML structure and create argument parser Python code.
• generate-argparser: Generate parser.py, which contains a setup_parser function to set up an appropriate parser.

For more detailed information on available command-line arguments and usage examples, refer to the official documentation.

Requirements

PyArgWriter depends strongly on information stated in the docstring. Your docstring should have a minimal structure like:

def add(a: int, b: int) -> int:
    """_summary_

    Args:
        a (int): _description_
        b (int): _description_

    Returns:
        int: _description_
    """
    return a + b

Usage

Usage PyArgWriter offers a command-line interface for generating and managing ArgumentParser setups. Below are the available commands and their usage:

parse-code

python -m pyargwriter parse-code --input file1.py file2.py --output output.yaml [--log-level LOG_LEVEL]

Parse given Python files and create a YAML structure with structural parser information.

write-code

python -m pyargwriter write-code --input input.yaml [--output OUTPUT_DIR] [--pretty] [--log-level LOG_LEVEL]

Read a given parser YAML structure and generate argument parser Python code.

generate-argparser

python -m pyargwriter generate-argparser --input file1.py file2.py [--output OUTPUT_DIR] [--pretty] [--log-level LOG_LEVEL]

Generate a parser.py file that contains a setup_parser function to set up an appropriate ArgumentParser. Common Options --log-level LOG_LEVEL: Sets the log level for command execution (default is "WARN"). --pretty (-p): If set, the generated code will be formatted with Black. For more detailed information on each command and additional options, run:

python -m pyargwriter <command> --help

Supported Argument Types

In the function of the process class you want to create the argument parser, following types of an argument are supported:

• int
• float
• str
• bool flags
• list[int]
• List[int] # from typing
• list[float]
• List[float] # from typing
• list[str]
• List[str] # from typing
• list[bool]
• List[bool] # from typing

Example

python -m pyargwriter generate-argparser --input examples/shopping.py examples/car.py --output examples --pretty

Hydra Integration

As an additional feature we support to combine your existing ArgumentParser with the Hydra framework. All you need to do is to use a decorator we provide with our framework. A short example can be viewed below.

from omegaconf import DictConfig
from pyargwriter.decorator import add_hydra

class Entrypoint:
    """ML training pipeline"""
    def __init__(self):
        """initiate pipeline"""
        pass
    @add_hydra("config", version_base=None)
    def train(config: DictConfig, device: str):
        """start training process
        
        Args:
            config (DictConfig): container which contains
            device (str): where to train on
        """

The decorator is indeed callable. Additionally to the usual hydra-arguments you have to pass in the config_var_name which specifies which argument in the train signature is the config object. After calling the pyargwriter tool you are now ready to go with a CLI which can now handle also almost all Hydra commands. Except the normal help message you get from Hydra. For further interest in this message please refer to their website

Documentation

The complete documentation for PyArgWriter, including detailed usage instructions and examples, can be found in the official documentation or at the documentation website.

Further you can have a look into the test coverage.

License

This project is licensed under the Apache 2.0 License - see the LICENSE file for details.

Contributing

If you would like to contribute to PyArgWriter, please read our Contributing Guidelines for details on how to get started.

Contact

For any questions, issues, or feedback, please open an issue on our GitHub repository.

Ideas for Naming

Here are some potential names for your tool, focusing on its functionality and the concept of automatically generating CLI interfaces:

Functional and Descriptive Names

1. AutoCLI
2. ClassCLI
3. MethodCLI
4. CLIBuilder
5. ArgParserGen
6. ClassParser
7. Method2CLI
8. CLIForge

Playful and Catchy Names

1. CLImate (a nod to "CLI" and "automate")
2. CLIonize (like ionizing a class into a CLI)
3. CLIckIt (emphasizing ease and speed)
4. Parse-o-Matic
5. Argonaut (exploring arguments and generating interfaces)

Tech-Inspired Names

1. CommandCrafter
2. InterfaceSmith
3. AutoArg
4. PyCLI
5. CodeCommander

Generalized and Creative Names

1. Argitect (argument architect)
2. CommandWeaver
3. ParseMaster
4. CLIMaker
5. MethodBridge

Each of these names is tailored to different branding styles. Do you have preferences about whether the name should be more functional, playful, or tech-inspired?

If you want the name to be short, easy to type, and terminal-friendly, here are some concise options:

1. cli-gen
2. argen (short for "argument generator")
3. clify
4. m2cli (short for "method to CLI")
5. pycli
6. cligen
7. argcli
8. cly (a super-short take on "CLI")
9. cmdgen
10. methcli (short for "method CLI") 11 . clipy

These names are minimal, straightforward, and easy to remember for developers. Let me know if any of these stand out!