Python Frameworks

Python guidelines and frameworks used for the Risk Engine

Python guidelines

• Always upgrade to the latest stable version and keep our code consistent with latest libraries maintained versions whenever possible.

• Pin dependencies to fixed versions in Dockerized containers for production.

• Use 2 spaces indenting.

• Always use virtual environments rather than try touching OS installation of Python.

• As standard as possible and exclusively open source dependencies.

Preferred libraries

The following are the team's selected libraries, recommended in general if you have multiple choices; but all contributors are free to introduce new ones and use their preferred options whenever there's a good rationale for it.

• Pandas: fast, powerful, flexible and easy to use open source data analysis and manipulation tool.

• Numpy: The fundamental package for scientific computing with Python.

• SK-Learn: Simple and efficient tools for predictive data analysis.

• XGBoost: optimized distributed gradient boosting library designed to be highly efficient, flexible and portable.

• Shap: SHapley Additive exPlanations is a game theoretic approach to explain the output of any machine learning model.

• Keras: For all things neural nets, deep learning, LSTM, etc.