Installing

EvoKit can be installed with pip. If an error occurs during installation, please update pip then try again.

If you do not wish to install EvoKit globally, consider using a virtual environment. An official tutorial can be found here.

The build script uses the build module. Please install this module or update it by running the following script:

pip install build --upgrade

Install from PiPy

To install EvoKit directly from PyPi, run the following command:

pip install evokit

To install all optional dependencies, run the following command:

pip install evokit[full]

Install from Source

To install EvoKit directly from source, run the following command at the root directory:

pip install .

To install all optional dependencies, run the following command:

pip install '.[full]'

Build Documentation

The docs/ directory has everything that build the documentation:

  • Run make.bat to rebuild the documentation.

  • Run update.bat to update API references in source/ then rebuild the documentation.

  • source/ contains all configuration files, including conf.py.

Quick Start

After installing the package, you can run a simple OneMax experiment to see if it works.

Command line:

python -m evokit.evolvables.bitstring

Python code:

import evokit.evolvables.bitstring as bitstring
bitstring.trial_run()

The output should be a string of tuples with non-decreasing (and hopefully increasing) values:

  • If the values are non-decreasing, then the elitist selector has successfully retained the best individual.

  • If the values are increasing, then the variator has successfully improved the population. If this does not happen, try rerunning the program times.