Welcome to Stone Soup’s documentation!
Stone Soup is a software project to provide the target tracking and state estimation community with a framework for the development and testing of tracking and state estimation algorithms.
As Stone Soup is focused on development and testing of algorithms, and such components may not be the most optimised implementations, instead focusing on being flexible. Its also intended to aid choice of component/algorithms to tackle real world problems.
Stone Soup is under active development, where feedback and contributions are welcomed to grow the number of components and features available.
Please see the Stone Soup Tutorials for learning about tracking and using Stone Soup, Examples for examples of Stone Soup features, and Demonstrations for demonstrations of using Stone Soup.
For community support, head over to the Stone Soup room on Gitter.
Installation
Stone Soup requires Python 3.10 or later.
To install Stone Soup from PyPI with pip:
python -m pip install stonesoup
To install Stone Soup from Conda-Forge with conda:
conda config --add channels conda-forge
conda install stonesoup
Stone Soup is currently in active development. To install the latest version from the GitHub repository:
python -m pip install git+https://github.com/dstl/Stone-Soup.git#egg=stonesoup
Optional dependencies
Some Stone Soup features use dependencies that are not required by the core
package. These can be installed with pip extras. Multiple extras can be
installed together, for example:
python -m pip install "stonesoup[video,optuna]"
The currently available feature extras are:
Extra |
Feature |
Optional dependencies |
|---|---|---|
|
Video reading, processing and related demonstrations |
ffmpeg-python, moviepy, OpenCV, yt-dlp |
|
TensorFlow integrations |
TensorFlow |
|
Ultralytics object detection |
ultralytics |
|
Multi-frame assignment data association |
OR-Tools |
|
Efficient hypothesis management data association |
pyehm |
|
Optuna-based sensor management |
optuna |
|
ODE-based functionality and examples |
PyTorch |
|
Road-network functionality |
GeoPandas, NetworkX |
|
Architecture graph visualisation |
Graphviz, NetworkX, pydot |
The dev extra installs the dependencies used for development and testing,
and the docs extra installs additional dependencies required when building
the documentation. The authoritative list of extras and packages is maintained
in pyproject.toml.
Developing
If you are looking to carry out development with Stone Soup, you should first clone from GitHub and install with development dependencies by doing the following:
git clone "https://github.com/dstl/Stone-Soup.git"
cd Stone-Soup
python -m pip install -e ".[dev]"
Please also see our Contributing page.