— turbfpe

This project is a fast, highly optimised Python implementation of the analysis pipeline of OPEN_FPE_IFT, the open-source MATLAB toolkit created by the research group Turbulence, Wind energy and Stochastics (TWiSt) at the Carl von Ossietzky University of Oldenburg (https://uol.de/en/physics/twist). It is based on the original package and provides the same end-to-end pipeline for turbulent time-series: classical statistics, Markov-property tests, Kramers–Moyal/Fokker–Planck coefficient estimation, and entropy/fluctuation-theorem analysis. In addition, it includes:

• Support for 2-D data (multiple independent realisations), with the option of concatenating them into a single trajectory.
• Entropy computation with the intermediate steps of each cascade trajectory.
• Optimisation of the Kramers–Moyal coefficients with the detailed fluctuation theorem (DFT), besides the integral fluctuation theorem (IFT).
• Isolated runs, identified by a prefix, that can be executed in parallel.

For the scientific background and reference algorithms see:

• A. Fuchs et al., “An open source package to perform basic and advanced statistical analysis of turbulence data and other complex systems” Phys. Fluids 34, 101801 (2022). https://doi.org/10.1063/5.0107974.
• B. L. Español et al., “Non-equilibrium formulation for inertial particles in turbulent swirling flows” J. Fluid Mech. 1030, A10 (2026). https://doi.org/10.1017/jfm.2026.11202.

Installation

git clone https://github.com/bersp/turbfpe
cd turbfpe
python -m pip install -e .

The -e flag installs the package in editable mode, meaning Python imports the code directly from this folder. Any edits you make are therefore picked up instantly, which is convenient if you want to browse the source, modify functions, or extend the library without reinstalling after every change.

Usage

The analysis is configured in a TOML parameter file and run with a Python script. The analysis_example/ directory contains an example: the filtered velocity data set from the reference study (Renner_8000Hz_filter.npy), its parameter file (Renner_params.toml), and the script that runs the full pipeline (analysis.py). To run it:

cd analysis_example
python analysis.py

The figures and processed data are saved in analysis_example/Renner/.

Before running the analysis on your own data, it is highly recommended to read the documentation.