Skip to content

Latest commit

 

History

History
142 lines (114 loc) · 4.62 KB

CONTRIBUTING.md

File metadata and controls

142 lines (114 loc) · 4.62 KB
Raw

Contributing

Thank you for your interest in contributing! Here are some guidelines:

Installation

Install the package locally using pip and in editable mode. This way, you can immediately test your changes to the codebase.

pip install -e .[dev,docs]

Updating the documentation

The documentation is built using MkDocs. To build the documentation locally, run:

mkdocs serve

Please make sure that all markdown is rendered correctly, and all links are working.

Contributing denoising algorithms

  1. Fork and clone this repository
  2. Create a branch git checkout -b method/fancy-method for your new method.
  3. Create a folder for the new method in ldctbench/methods. The folder must contain the following files
    • __init__.py
    • argparser.py: Should implement a method add_args() that takes as input an argparse.ArgumentParser, adds custom arguments and returns it. If your method has an argument fancy_arg, then argparser.py should look like this: 
      from argparse import ArgumentParser


def add_args(parser: ArgumentParser) -> ArgumentParser:
    parser.add_argument(
        "--fancy_arg", 
        type=float, 
        help="My fancy argument."
    )
    return parser
    • network.py: Should implement the model as class Model(torch.nn.Module).
    • Trainer.py: Should imeplement a Trainer class. This class should be initialized with Trainer(args: argparse.Namespace, device: torch.device) and implement a fit() method that trains the network. A base class is provided in methods/base.py.
  4. Add the method to METHODS in argparser.py.
  5. Add the method to docs/denoising_algorithms.md.
  6. Add a fancy-method.yaml config file containing all hyperparameters to configs/.

Contributing other features or bug fixes

  1. Fork and clone this repository

  2. Create a new branch based on your change type:

    • fix/some-fix for bug fixes
    • feat/some-feature for adding new features
    git checkout -b <your-branch-name>

Code Formatting and Unit Tests

Make sure your code is:

  1. Readable
  2. Nicely documented with docstrings in numpy format
  3. Follows the style guide enforced by black, isort, and flake8

For new methods or features, always implement some unit tests and put them in tests/module/test_new_feature.py. The test function names should be descriptive. A good format would be something like test_functionname_when_given_arguments_return_x(). A bad format would be something like test123().

Examples

Don't do! 
# The new "feature" we implemented
def divide(a, b):
    # This function divides stuff
    if b == 0:
        raise ZeroDivisionError("You can't divide by zero!")
    return a / b

# Test functions
def test1():
    assert divide(1, 2) == 0.5
    assert divide(5, 5) == 1.

def test2():
    assert divide(5, -2) == -2.5
    assert divide(-2, 5) == -0.4
Do! 
# The new "feature" we implemented
def divide(a: int, b: int) -> float:
    """Function to divide two numbers

    Parameters
    ----------
    a : int
        The numerator
    b : int
        The denominator

    Returns
    -------
    float
        Result of division operation

    Raises
    ------
    ZeroDivisionError
        If the denominator (b) is zero
    """
    if b == 0:
        raise ZeroDivisionError("You can't divide by zero!")
    return a / b

# Test functions
import pytest
def test_divide_when_given_positive_numbers_returns_correct_result():
    assert divide(1, 2) == 0.5
    assert divide(5, 5) == 1.

def test_divide_when_given_negative_numbers_returns_correct_result():
    assert divide(5, -2) == -2.5
    assert divide(-2, 5) == -0.4

def test_divide_when_given_zero_denominator_raises_error():
    with pytest.raises(ZeroDivisionError):
        divide(1, 0)

Create a pull request

Before starting a pull request:

  1. Ensure all tests pass: poe test
  2. Ensure that the code is formatted correctly: poe format_check
  3. Fix any formatting errors with: poe format

Alternatively, you can run 1 & 2 in one command using poe verify

Once ready, start a pull request on GitHub and include:

  • A clear description of what you're fixing/adding
  • Any relevant context or background information

Our pipeline will automatically check your pull request for formatting errors and test results. If you contributed a novel method, it would be great if you can also provide pretrained weights.