v0.1-nn

First release of project resilience implementation of ESP. Has a neural network representation implemented for prescriptors. Simple evaluation with no constraint violation (yet). Uses the NSGA-II algorithm with optional elitism. The user simply implements their prescriptor and evaluator and passes it into the evolution object to run ESP.

What's Changed

• Added linting, unittests, and requirements for github actions by @danyoungday in #1
• Added pyproject.toml so we can build the project by @danyoungday in #2
• Neural net prescriptor by @danyoungday in #3
• Added nn prescriptor to init by @danyoungday in #4
• Fix imports and saving. Fix no elites case by @danyoungday in #5
• Cartpole example and end to end unittest by @danyoungday in #6
• Modified fast nondominated sort to work in-place. Added nsga2 utils tests by @danyoungday in #7
• Parallelization by @danyoungday in #9
• Added some unittests for parent selection, population creation, and crossover/mutation by @danyoungday in #8
• Changed n jobs to not be set to pop size when -1, but rather stay at -1 by @danyoungday in #10
• Changed remove pop pct to be more testable then added unit test by @danyoungday in #11
• Cartpole and NN Implementation Unit Tests by @danyoungday in #12
• Changed how the NN prescriptor initializes to match ESP by @danyoungday in #13
• Parallel/Sequential Test by @danyoungday in #14
• Cleanup prescriptor/NN, move mutation by @danyoungday in #15
• Nsga refactor by @danyoungday in #16
• Extensible NNPrescriptorFactory by @danyoungday in #17
• Added readme explaining project by @danyoungday in #18
• Kw nn by @danyoungday in #19
• fix mps init issue and kwargs order by @danyoungday in #20
• Added prints to show seeding is happening by @danyoungday in #21

Full Changelog: https://github.com/Project-Resilience/presp/commits/v0.1