Skip to main content

A hierarchical and highly sophisticated variant of argparse.

Project description

Hierarchical Argparse

PyPI PyPI - Python Version PyPI - License

Hiargparse is a hierarchical and highly sophisticated variant of argparse.

Minimal Code

from hiargparse import ArgsProvider, Arg, ChildProvider, ArgumentParser

child = ArgsProvider(
    args=[Arg('baz', default=42)]
root = ArgsProvider(
    args=[Arg('foo', default='bar')],
    child_providers=[ChildProvider(provider=child, name='child')]
parser = ArgumentParser()


Hiargparse automatically generates the command-line arguments with your all classes in your tree-like module structures with minimal codings.

Suppose you make a large module, and your hierarchically deep class, say, requires some arguments, say heights and widths. Foo makes an instance of Bar, Bar makes an instance of Baz, ..., and Spam makes an Egg which has the two arguments. When you want to pass them command-line options, you have to write the two, heights and widths, in the constructor of Foo, Bar, ..., and Spam. What if the Egg is updated and wants to require some new arguments, like depth or length? Hiargparse passes the arguments directly to the classes without any black magics.


With this module, you can

  • easily make hierarchical (tree-like) command-line arguments with argparse
  • Each argument is automatically help-texted and grouped into argparse.ArgumentGroup.
  • get a more useful Namespace object than the original
  • Accessing with dict-like key, getting the child Namespace, converting to/from dictionaries, and so on
  • write/read the arguments to/from some configure files with well known formats
  • Currently we supports yaml and toml.

Also, this module

  • is almost compatible with original argparse; you can gradually introduce it to your large projects.
  • works without command-line arguments; when you distribute your module with hiargparse, users still can select whether to feed arguments to it with a command-line or programatic way.

This module is inspired by rlpytorch.args_provider in FacebookAIResearch/ELF.


pip install hiargparse


  • python >= 3.6.0
  • typing_extensions (for typing_extentions.Protocol in python 3.6) >=
  • pyyaml >= 3.12 (only if you use yaml write/read)
  • toml >= 0.9.4 (only if you use toml write/read)



  • Any contribution is welcome!
  • Fork the repo, create a branch, add your awesome changes, and make a new Pull Request.
  • If you find some bugs, please report in issue.


  • Keigo Kawamura (Department of Electrical Engineering and Information Systems (EEIS), Graduate School of Engineering, The University of Tokyo)

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distribution

hiargparse-0.2.2-py3-none-any.whl (24.6 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page