Confuse: painless YAML config files

https://github.com/beetbox/confuse/workflows/Build/badge.svg?branch=master http://img.shields.io/pypi/v/confuse.svg

Confuse is a configuration library for Python that uses YAML. It takes care of defaults, overrides, type checking, command-line integration, human-readable errors, and standard OS-specific locations.

What It Does

Here’s what Confuse brings to the table:

  • An utterly sensible API resembling dictionary-and-list structures but providing transparent validation without lots of boilerplate code. Type config['num_goats'].get(int) to get the configured number of goats and ensure that it’s an integer.
  • Combine configuration data from multiple sources. Using layering, Confuse allows user-specific configuration to seamlessly override system-wide configuration, which in turn overrides built-in defaults. An in-package config_default.yaml can be used to provide bottom-layer defaults using the same syntax that users will see. A runtime overlay allows the program to programmatically override and add configuration values.
  • Look for configuration files in platform-specific paths. Like $XDG_CONFIG_HOME or ~/.config on Unix; “Application Support” on macOS; %APPDATA% on Windows. Your program gets its own directory, which you can use to store additional data. You can transparently create this directory on demand if, for example, you need to initialize the configuration file on first run. And an environment variable can be used to override the directory’s location.
  • Integration with command-line arguments via argparse or optparse from the standard library. Use argparse’s declarative API to allow command-line options to override configured defaults.

Installation

Confuse is available on [PyPI](https://pypi.org/project/confuse/) and can be installed using pip:

pip install confuse

Using Confuse

Confuse’s documentation describes its API in detail.

Credits

Confuse was made to power beets. Like beets, it is available under the MIT license.