Simplifies configuration stacking. Primarily, this was written to allow stacking of default, global, and local settings. It allows for validation so that you can enforce contraints on supplied options.

Using conflib is just a matter of importing it and giving it some dictionaries to stack:

>>> import conflib
>>> Default = {'hello': 'world', 'alpha': 5}
>>> Global = {'wat': 'wut', 'fancy': (20, 'fish')}
>>> Local = {'hello': 'everybody', 'beta': 'qwerty'}
>>> Config = conflib.Config(Default, Global, Local)
>>> print(Config.options)
{'alpha': 5, 'wat': 'wut', 'beta': 'qwerty', 'hello': 'everybody', 'fancy': (20, 'fish')}

If you need to stack on new configs later, go for it:

>>> Config.stack({'wat': 'new_value', 'extra': 10})
>>> print(Config.options)
{'alpha': 5, 'wat': 'new_value', 'beta': 'qwerty', 'hello': 'everybody', 'fancy': (20, 'fish'), 'extra': 10}

Pass in a validation dictionary if you want to check the provided arguments:

Validator = {
    'alpha': lambda x: x < 10,
    'fancy': tuple,
    'beta': [('asdf', 'qwerty'), ('fizz','buzz')]
}
Config.validate(Validator)

Alternately, you can pass in a validation_dict when creating a Config object. Validation failures raise ValueError with a message indicating the specific failure. There are several available options for the Validation dictionary:

• bool: Converts 'y', 'yes', '1', 1, and True to True, and 'n', 'no', '0', 0 and False to False
• int: Casts provided value as an integer
• A list of tuples: looks for the provided value as a member of any of the tuples, and returns the first item in that tuple
• A list: looks for the provided value in the list
• Any type (str, float, etc): validates that the provided value is of the specified type (does not try to cast it)
• A callable: Runs the provided function, which should either return the validated value or raise TypeError

pip install conflib

conflib is released under the MIT License. See the bundled LICENSE file for details.