URL: https://pypi.org/project/esig/

This package provides "rough path" tools for analysing vector time series. A python wrapper around libalgebra. It provides a clean but very limited interface for computing signatures and log signatures of multidimensional paths of moderate degree. A sequence of double precision samples from a multidimensional path, presented as a numpy array are converted into a truncated signature or logsignature suited for machine learning etc. The outputs are appropriately dimensioned numpy arrays and provide dense representations of these vectors.

https://coropa.sourceforge.io/ where CoRoPa stands for Computational Rough Paths is a software repository that includes libalgebra.

Originally described in the following arxiv paper:

https://arxiv.org/abs/1405.4537

Requires boost