Development

The Python package is built with scikit-build-core and uses nanobind for the Python bindings, while the numerical C++ code is available as the standalone quantlop_core CMake target.

Project setup

Clone the repository, then install the package in editable mode together with the development and documentation dependencies:

python -m pip install -e .[dev,docs]

Warning

If you are on macOS, you need first to install libomp with either Homebrew or MacPorts:

brew install libomp
# or
sudo port install libomp +top_level

Run the Python test suite with:

python -m pytest -v

To configure and build the native C++ target directly:

cmake -S . -B build
cmake --build build

Format code

Format Python files and run the Ruff checks with:

python -m ruff format .
python -m ruff check .

Install the pre-commit hooks to format Python and C++ files automatically:

python -m pre_commit install

Build docs

Check the examples in API docstrings, then build the documentation with warnings treated as errors:

python -m sphinx -b doctest -W docs site
python -m sphinx -W docs site