1.0.1

• Build/install now uses pyproject.toml instead of the deprecated setup.py
• unittest -> pytest
• Remove Azure pipelines config
• Add GitHub workflows for PyPI publishing, unit testing

1.0.0 Release

• Overhaul of backend algorithms, API

Fitness rework, property additions

• rework of fitness (standardized)
• added properties for:
  • average, median, best cost function return values
  • best fitness, parameters
• best population member determined after each new generation
• updated documentation

Fix to multiprocessing threshold

• Set threshold for number of processes to 1 (> 1 means multiprocessing is enabled)

Bug fixes

Bug fixes for 0.5.0 release

Documentation, imports, next_generation() arguments, and more

• Updated docstrings to Google format
• Reduced imports to only functions/objects used from packages
• Removed "num_survivors" argument from next_generation()
• Added "log_base" argument to next_generation(), used to scale the reverse log space in parent selection (higher == more skewed towards choosing a selected member towards the front of the list returned by the selection function)
• Changed names of class properties for Members and Population (param_vals to parameters)
• Updated README to reflect changes, add additional documentation on mutation

Arbitrary argument pass-through to cost_fn

• Added support for passing arbitrary arguments to a cost function with cost_fn_args

Multiprocessing Support

• Added support for multiprocessing in Member generation

Pep8 and documentation update

• Updated code to pep8 styling
• Updated README with installation instructions, usage information

Update to docstrings

0.2.2

Update to docstrings