develop python code for all java files in repo

your code should not be slower and produce the same results
https://github.com/canoalberto/ur-CAIM

1
select several data sets, at least 3 to test your code
2
guide me how to run this java code on windows to be able test your python code
3
develop python code for all java files
4
prove that python code output is the same as java
5
it is not wrote wrapper project , but rather to rerwite all java files to plain python - not use external software like weka , but use only plain python numpy code
all should be written in python including
https://github.com/canoalberto/ur-CAIM/blob/master/src/main/java/weka/filters/supervised/attribute/urCAIM.java

6
input to python code should be pandas data frame