params-proto 0.5.5

A command line argument parsing utility using python class-based namespace for better IDE static auto-completion

Todo

Done

• [x] publish

• [x] add test

Installation

pip install params_proto

Usage

from .params_proto import cli_parse, is_hidden, ParamsProto, proto_signature


def test_cli_proto():
    @cli_parse
    class G(ParamsProto):
        """Supervised MAML in tensorflow"""
        npts: "number of points to sample from distribution" = 100
        num_epochs: "number of epochs to train" = 70000
        num_tasks: "number of tasks in the inner loop" = 10
        num_grad_steps: "number of gradient descent steps in the inner loop" = 1
        num_points_sampled: "effectively the k-shot" = 10
        fix_amp: "controls the sampling, fix the amplitude of the sample distribution if True" = False

    assert G.npts == 100
    G.npts = 10
    assert G.npts == 10
    print("showing the dictionary from G: ", vars(G))
    print("showing the original namespace class: ", G.__proto__)
    print('showing the properties of G: ', G.__props__())


def test_proto_signature():
    @cli_parse
    class G(ParamsProto):
        """some parameter proto"""
        npts: "number of points to sample from distribution" = 100

    @proto_signature(G.__proto__)
    def main_demo(**kwargs):
        print('npts = ', kwargs['npts'])
        pass

    # First way is to use proto_signature decorator. The dynamically generated signature
    # however does not show up in pyCharm. It does however, show during run time.
    import inspect
    print("main_demo<Function> signature:", inspect.signature(main_demo))
    main_demo(npts=10)

To Develop

git clone https://github.com/episodeyang/params_proto.git
cd params_proto
make dev

To test, run

make test

This make dev command should build the wheel and install it in your current python environment. Take a look at the https://github.com/episodeyang/params_proto/blob/master/Makefile for details.

To publish, first update the version number, then do:

make publish