I'm happy to announce the release of confugue 0.1.1!

Confugue is a hierarchical configuration framework for Python. It provides a smart wrapper for nested configuration dictionaries (typically loaded from YAML files), which can automatically fill in function and class parameters with the configuration values. This eliminates the need for passing these values around or even explicitly referencing them.

Installation: pip install confugue Docs: https://confugue.readthedocs.io/ Tutorial for deep learning users: http://tiny.cc/confugue-colab GitHub: https://github.com/cifkao/confugue/ License: BSD-3-Clause

Confugue is very flexible, allowing for: - Arbitrary nesting of the configuration file, following the code structure - Specifying defaults and additional arguments at runtime - Omitting a whole section of the configuration file if defaults are provided - Using a list of dicts to configure a list of objects with different arguments for each - Overriding the type of an object while preserving the arguments passed by the caller - Detecting unused configuration values

Basic example (see docs for examples of the hierarchical machinery):

from confugue import Configuration cfg = Configuration({'my_obj': {'b': 'bar', 'c': 'baz'}}) # ...or load with Configuration.from_yaml()

obj = cfg['my_obj'].configure(MyClass, a='foo', b='moo') # Will instantiate MyClass(a='foo', b='bar', c='baz')

<P><A HREF="https://github.com/cifkao/confugue">Confugue 0.1.1</A>, a hierarchical configuration framework. (30-Apr-20)</P>