GitLab

* add pmml to test

* refine pmml.py

* use ~n instead of -n-1

* change map to list comprehension

* fix check

* fix 'use ~n instead of -n-1'

* fix exception

We dont need it we aren't doing any computation with
float values. We print out whatever values are
read from the LightGBM_model.txt as a string.

file as input and output a directory with pmml files. Each file can be given as input to
pmml consumers. You can test the files using jpmml-evaluator. Please refer to this the
data mining group's website for the pmml specifcation (http://dmg.org/pmml/pmml-v4-3.html).
Also refer to the README.md file in the pmml directory to see how to test the pmml output
with jpmml-evaluator.