Just experimenting LTR algorithms in Python and Scala on the MSLR datasets.
Check my article for more details.
To run the Python notebook you need numpy, pandas, sklearn and xgboost (use the same version as the Scala app).
sbt -no-colors -mem 4000 run