declare_config

This Python module allows a declarative syntax for defining the contents of a configuration file. It also enables easy support for features such as default configuration values, nested configuration settings (and nested defaults), and various other useful automatic sugars.

Example:

from pathlib import Path

from declare_config import Configuration,
    Setting, \
    configuration_source, \
    enable_nested_settings

@configuration_provider("$MYAPP_CONFIG")
@configuration_provider("./myapp.yaml")
@configuration_provider("~/init/myapp.yaml")
@configuration_provider("~/.myapp.yaml")
@enable_nested_settings
@enable_expanduser
class MyAppSettings(declare_config.Configuration):
    '''Settings for myapp.'''
    homepage = Setting('urls.root')
    api_root = Setting('urls.api_root', '${homepage}/api')
    timeout_ms = Setting('timeout_ms', 5000, int)
    log_file = Setting("log_location", ~/logs/myapp.log", Path)

settings = MyAppSettings.load()

The above code defines a configuration file for a sample application. It defines four locations that will be searched for the configuration file when MyAppSettings.load() is called: first it will check the value of the environment variable $MYAPP_CONFIG for a file location, and then if no such environment variable is defined it will check the other three file locations in order for a configuration file. Whenever it finds a valid file location, it will load the settings from that file.

Because of the @enable_nested_settings decorator, settings are allowed to reference one another (and default definitions are allowed to reference other settings as well) through the ${setting_name} syntax. All settings have type information,and the @enable_expanduser decorator ensures that for any setting declared to be a filesystem path, the ~/ directory will be expanded to the user's home directory.

Currently only yaml configuration files are supported, though there are plans to add support for JSON and INI syntaxes.

Development

Dependencies are managed with pipenv. To install dependencies for development, run pipenv install --dev from the project root.

To run tests once with coverage, use pipenv run pytest from the project root. To run tests continuously during development, run pipenv run ptw instead.

To upload to PyPi, update setup.py with the version and any new dependencies and run pipenv run python setup.py upload.