py-params 0.8.3

A type safe dictionary in python

A type safe dictionary class in python.

LICENSE

MIT. See License File.

Install

params is on the Python Package Index (PyPI):

pip install py-params

Usage

Params instances are dict s for which default values could be provided as class level fields when subclassing Params. Those default values could than be overridden in instances of the class, by specifying the values in the constructor.

Accessing parameters not defined as class level variable would raise an AttributeError.

>>> from params import Params

>>> class TestParams(Params):
...     param_a = 'a'
...     param_b = True

>>> args=TestParams()                         ## using defaults
>>> args
{'param_a': 1, 'param_b': True}

>>> TestParams(param_a='c')                   ## override param_a
{'param_a': 'c', 'param_b': True}

>>> args.param_a = 2                          ## access as attribute or key
>>> args["param_a"] = 3
>>> args.param_a == args["param_a"]
True

>>> args.param_c
AttributeError: 'TestParams' object has no attribute 'test_c'

>>> args.param_c = 3
AttributeError: Setting unexpected parameter 'param_c' in Params instance 'TestParams'

>>> args["param_d"] = 4
AttributeError: Setting unexpected parameter 'param_d' in Params instance 'TestParams'

Resources

As an illustration of how Params could be used to reduce boilerplate code check:

• kpe/params-flow - utilities for reducing keras boilerplate code in custom layers

• kpe/bert-for-tf2 - BERT implementation using the TensorFlow 2 Keras API