quadsv

Contents

quadsv#

quadsv: kernel-based spatial pattern detection and comparison for spatial omics.

The public top-level API is organised in four layers:

  1. Kernels — MatrixKernel (dense / sparse), FFTKernel (regular grid), NUFFTKernel (irregular 2D coordinates). The Kernel and MatrixKernelBase ABCs live in quadsv.kernels and are intended for backend authors.

  2. Statistical tests — spatial_q_test() and spatial_r_test(). A single entry point per test dispatches on the kernel type (matrix, FFT, or NUFFT). Signature: (x, kernel, null_params=None, return_pval=True, is_standardized=False).

  3. Detectors — DetectorIrregular consumes anndata.AnnData (irregular grids, matrix/NUFFT backends); DetectorGrid consumes spatialdata.SpatialData (regular grids, FFT backend).

  4. Comparators — cross-sample pattern comparison: ComparatorIrregular on a list of AnnData (NUFFT backend); ComparatorGrid on a list of SpatialData (FFT backend).

Submodules#

Contents