Skip to content

API Reference

The API mirrors the package layout. New here? Read What is a probabilistic ODE solver? and How to choose first — this reference is for looking things up, and assumes the notation.

Which function do I need?

Task Reach for
solve an ODE (fixed step) gaussian_filter + rts_smoother
solve with tolerance-based stepping gaussian_filter_adaptive (jit/vmap/grad-able)
pick the linearization (EK0 / EK1) TaylorCorrection
define the ODE / observations Measurement models
choose a prior IWP / MaternPrior / JointPrior
initialize the state taylor_mode_initialization
calibrate posterior uncertainty (diffusion sigma^2) calibration guide
fit ODE parameters to data (max-likelihood) ODEFilter / fit / marginal_loglik

Subpackages

  • Filters — filtering / smoothing loops, corrections, adaptive control.
  • Inference — ODE-parameter inference (fit, marginal_loglik, ODEFilter) and the square-root Gaussian algebra it builds on.
  • Measurement — ODE-information and observation models.
  • Priors — Gauss–Markov process priors.

Each page is generated with mkdocstrings and stays in sync with the source.