Installation#

From PyPI#

pip install quadsv

Two optional extras are available for development and documentation:

pip install 'quadsv[dev]'    # tests, linting, jupyter, matplotlib
pip install 'quadsv[docs]'   # Sphinx + theme + autoapi

From source#

git clone https://github.com/JiayuSuPKU/EquivSVT.git
cd EquivSVT
pip install -e '.[dev,docs]'

Requirements#

  • Python 3.10+.

  • Runtime dependencies (installed automatically): scanpy, spatialdata, finufft, joblib, tqdm. Through scanpy you also get anndata, numpy, scipy, scikit-learn and pandas.

spatialdata is needed by DetectorGrid and ComparatorGrid. finufft is needed by NUFFTKernel, by DetectorIrregular when you set backend="nufft", and by ComparatorIrregular.

Verify the install#

import quadsv

print(quadsv.__version__)
print(sorted(quadsv.__all__))

You should see 14 public names organised into four layers (see Quick Start for what each layer does). The top-level package is the user-facing surface. The canonical submodule paths (quadsv.kernels.*, quadsv.detectors.*, quadsv.comparators.multisample, quadsv.statistics) are documented under quadsv.