Design

Installation

The installation is easily possible by using pip install . or directly from git (pip install "optimal_splitk @ git+https://github.com/mborn1/optimal_splitk.git"). There are two additional options: dev and examples (installed by pip install .[dev] and pip install .[examples] respectively). The first adds extra dependencies to update the documentation, the second add some dependencies (like pandas) to run the examples.

Folder structure

The project contains a couple folders:

• src: This contains the source code of the package

• examples: Contains some example scripts for using the package

• docs: Contains the documentation

• assets: Contains possible assets (like figures) for the readme.

Code design

The code is designed using a couple modules under src/optimal_splitk. The doe module contains the main entry point for generating designs. Besides the main function, the user could add new optimization criteria by adding a new Python file in the optim directory. Please inspect the other (D-optimal and I-optimal closely). Finally the other modules contain helper functions and code, they are all documented in the API section.