osmose.calibration.uq

Surrogate-based Bayesian UQ for OSMOSE calibration.

Thin subpackage boundary: no eager re-exports and no heavy imports here, so later phases’ optional dependencies (emcee, dynesty, arviz) stay lazy. Import from submodules directly, e.g. from osmose.calibration.uq.emulator import GPEmulator.

Modules

design	Seeded LHS design executed through the Python engine, reduced for the UQ emulator.
emulator	Single-output GP emulator for one OSMOSE output stat.
gate	Probabilistic emulator-calibration gate (Bastos & O'Hagan 2009 style).
keying	Map a calibration target to its distinct UQ emulator output-stat key.
likelihood	Per-target log-likelihoods over the natural-log biomass residual.
output_stats	Reduce one OsmoseResults to a per-species UQ stat dict.
posterior	Compose prior + per-target likelihood into log_posterior(theta).
predictive	Posterior-predictive diagnostic (Goal 2): per-species marginal emulator-predictive ranges from the posterior over theta.
prior	Uniform prior over the FreeParameter sampling-space box.
run	End-to-end surrogate-Bayesian UQ orchestrator.
sampler	Posterior sampling from Phase 2a's log_posterior via nested sampling (dynesty).