optmanage: A library to create flexible option managers.

Flexible option managers, supporting options with default values, static type hints, runtime type checking, and custom runtime validation logic.

Contents

• Install
• Usage
• API
• Contributing
• License

Install

You can install the latest release from PyPI as follows:

$ pip install --upgrade optmanage

Usage

Custom option manager classes can be created by subclassing OptionManager and using the Option descriptor to set options. An option manager object can then be obtained by instantiating the option manager class:

from optmanage import Option, OptionManager

class MyOptions(OptionManager):
    """ Options of some library. """

    validate = Option(bool, True)
    """ Whether to validate arguments to functions and methods. """

    eq_atol = Option(float, 1e-8, lambda x: x >= 0)
    """ Absolute tolerance used for equality comparisons."""

    scaling = Option(
        Mapping[Literal["x", "y", "z"], float],
        {"x": 1.0, "y": 2.0, "z": 1.0},
        lambda scaling: all(v >= 0 for v in scaling.values())
    )
    """ Scaling for coordinate axes used in plots.  """

options = MyOptions()

Each option takes a default value, a type, and an optional validator function:

validate = Option(bool, True)
#     option type ^^^^  ^^^^ default value

eq_atol = Option(float, 1e-8, lambda x: x >= 0)
#           optional validator ^^^^^^^^^^^^^^^^

Any type supported by the typing-validation library can be used for options, including PEP 484 type hints:

scaling: Option(
    Mapping[Literal["x", "y", "z"], float], # <- type hints supported
    {"x": 1.0, "y": 2.0, "z": 1.0},
    lambda scaling: all(v >= 0 for v in scaling.values())
)

Options can be accessed and set like attributes of the options object:

print(options.scaling)  # {'x': 1.0, 'y': 2.0, 'z': 1.0}
options.scaling = {"x": 2.5, "y": 1.5, "z": 1.0}
print(options.scaling) # {'x': 2.5, 'y': 1.5, 'z': 1.0}

It is possible to set multiple options simultaneously using the set method of the options object:

options.set(validate=False, eq_atol=1e-3)
print(options.validate) # False
print(options.eq_atol)  # 0.001

It is also possible to use the options object as a context manager, for temporary option setting:

with options(validate=False, eq_atol=1e-3):
    print(options.validate) # False
    print(options.eq_atol)  # 0.001
print(options.validate) # True
print(options.eq_atol)  # 0.00000001

All options can be reset to their default values by using the OptionManager.reset method of the options object:

options.set(validate=False, eq_atol=1e-3)
print(options.validate) # False
print(options.eq_atol)  # 0.001
options.reset()
print(options.validate) # True
print(options.eq_atol)  # 0.00000001

An individual option can be reset to its default value by using the Option.reset method of the Option object, accessed from the option manager class:

options.set(validate=False, eq_atol=1e-3)
print(options.validate) # False
print(options.eq_atol)  # 0.001
MyOptions.eq_atol.reset(options) # resets 'eq_atol' on the 'options' object
print(options.validate) # True
print(options.eq_atol)  # 0.001

API

Full documentation is available at https://optmanage.readthedocs.io/

Contributing

This project is currently in private development. Public contribution guidelines are available at CONTRIBUTING.md.

License

MIT © Hashberg Ltd.