Compile github project (https://github.com/barbua/FSA) and demonstrate usage for feature selection

I need to compile the above github project on my my windows and on my ubuntu machine and need working sample code how to use the code in python:
In python, I have a DataFrame x_df with 100-200 features (column-name=feature) and values in features are floats.
Requirements to the sample code:
- demonstrate use of different parameters
- demonstrate feature selection on full dataframe
- demonstrate feature selection with cross-validation
- working sample must be repeatable (seed-value?) with exact same results if repeated on the same dataframe.
- if the code allows for usage of GPU - I also need working code with a NVidia RTX 2060 on Windows

I also need a documentation (step by step in depth) how to recomile the code from the repositiory.