Python Versions and Package Support on Blueshift®¶
We use a single engine for both backtesting and live trading. This engine support only Python 3.6+ and is at present Python 3.7 in the lastest release.
Supported Python packages¶
We have a white-list of packages, listed below, which you can use. Please reach out to us if you need any other (public and well-known) packages to be added. We will review the use case and security aspects and update the white-list accordingly. If you need any personal package, consider using the workspace feature to implement the same on the site. We support only pure Python code on the site (no c-extensions etc.).
Deprecated package support.
johansen is no longer supported. Use the
|An useful array sorting package.
|Provides access to mathematical functions for complex numbers.
|Package for convex optimization.
|A "nice and disciplined" interface to cvxopt.
|For manipulating dates and times in both simple and complex ways.
|Higher-order functions and operations on callable objects.
|For unsupervised learning and inference of Hidden Markov Models.
|for analysing random walks and evaluating the Hurst exponent.
|ARCH and other tools for financial econometrics.
|A deep learning API running on top of TensorFlow.
|Provides access to the mathematical functions defined by the C standard.
|Package for scientific computing with Python.
|High-performance, easy-to-use data structures and data analysis tools.
|Implements Kalman filter and Kalman smoother in Python.
|Allows accurate and cross platform timezone calculations.
|Random number generators for various distributions.
|Efficient numerical routines for scientific computing.
|For machine learning in Python.
|For statistics in Python.
|For technical analysis in Python.
|XGBoost library interface for Python.