Contributing

The project welcomes contributions in any form! They could take the form of reporting bugs, feature requests, updates to documentation and even enhancements to the code.

To help smooth out the contributions, here are some guidelines of conventions used in the code (the list will evolve over time):

General

  • If you find anything not in line with these guidelines, it is because this repository takes the approach of “if youchange it please align it to the current standards”. Please correct the code to follow these guidelines.

Code

  • The maximum line length is 89 in Python files.

  • Import modules as namespaces. For example, instead of from enum import auto use import enum and then enum.auto.

  • Use the following pattern for docstring:

    def func(arg1: int, arg2: str = "hi") -> float:
        """
        Single line explaining the function.
    
        More detailed information (such as exception that are raised, algorithms,
        ...) if required.
    
        Args:
            arg_1: Description 1.
            arg_2: Description 2.
    
        Returns:
            Information about the return value.
    
        """
    

    The Args are not needed if there are no arguments and Returns is not needed if None is returned.

  • Use type hints, avoid using typing.Any where reasonable.

  • 100% code coverage is required (except as configured with a documented reason).

  • In error messages, explain what has gone wrong and how to fix it.

  • If an argument is optional in a function, any functions that it calls should no longer make the argument optional.

  • Take note of the tools used in the pipeline for code quality checks. Any disabling of any check must be documented with a reason.

  • Note changes in the changelog.

  • Any raised exceptions should inherit from open_alchemy.exceptions.BaseError. Exception handling should specify which exception is being handled (e.i. more specific than Exception.

  • Update the readme with any significant new features

  • Update the documentation and examples, if appropriate

  • Functions that are intended for re-use across the code base (e.g. helper functions) should use keyword only arguments for increased clarity. Private module/ class functions may optionally use positional arguments for enhanced execution speed.

Tests

  • In tests, preferable use parameterization instead of a for loop in the test body.

  • Test docstrings should use the following structure:

    """
    GIVEN <preconditions>
    WHEN <an action is taken>
    THEN <this outcome is expected>
    """
    
  • Separate the code for GIVEN, WHEN and THEN in tests using a single blank line to indicate where each piece starts

  • For parameterization, use pytest.param and provide a value for the id argument

  • If the number of parameters for a test becomes large, take the list out of the decorator and make it a module variable.

Documentation

  • Use passive language in documentation.

  • The maximum line length is 80 in the documentation.