pyGCS

Repository containing implementation of Stochastic Grammar-based Classifier System in Python (pyGCS).

Requirements

To run library you have to had installed some dependencies:

• Python (>= 3.6.3) - major environemnt to run library
• Jupyter (>= 4.4.0) - notebook to run library and visualize results
• Networkx (>= 2.1) - python library for BSF tree
• Jsonpickle (>= 0.9.6)- serializing python objects to json objects
• Numpy (>= 1.14.1) - fundamental package for scientific computing

Running

To run pyGCS library you have to follow one of some methods. All library parameters are available from file app.ini

Grammar/app.ini

Console line without Jupyter

Go to main app directory

cd Grammar

Next run python3 app with following parameters

python3 main.py app.ini logger.ini

Console line with Jupyter

Go to main app directory

cd Grammar

Next run jupyter app with following parameters

jupter notebook

Jupyter notebook will be accesed on webpage

http://localhost:8888

All charts and graphs will be available on

http://localhost:8000

Docker

First you have to had installed

• Docker (>= 18.03) - virtual operational system
• Docker Compose (>= 1.21.1) - manager images and conatiners

From root directory run build command

docker-compose build gcs

To run pyGCS library run

docker-compose up -d gcs

Jupyter notebook will be access on webpage

http://localhost:8888

All charts and graphs will be available on

http://localhost:8000

Results

Library after induction saving results to data folder

cd ../Charts/data