Installation

Install sitsfeats.py from PyPI:

pip install sitsfeats

To also work with labelled data cubes, install the xarray extra, which brings in xarray and dask:

pip install "sitsfeats[xarray]"

The cube support is described in Working with data cubes.

What gets installed

The package requires Python 3.11 or newer and depends only on numpy.

Its operations are implemented in C++, so a binary wheel is used when one is available for your platform. When there is none, pip builds the extension from source. That build needs a C++ 17 compiler, but no system libraries: the extension uses nanobind and the header-only Eigen library, both resolved automatically at build time.

uv (development)

The project is developed with uv. Clone the repository and sync, which builds the C++ extension and installs the dev tools:

git clone https://github.com/m3nin0-labs/sitsfeats.py
cd sitsfeats.py
uv sync

Common tasks:

uv run pytest          # run the test suite
uvx ruff check .       # lint
uv run mkdocs serve    # preview these docs
uv build               # build wheel + sdist