diff --git a/test/runtests.jl b/test/runtests.jl index 81266f2..b64f4e4 100644 --- a/test/runtests.jl +++ b/test/runtests.jl @@ -513,22 +513,41 @@ function run_tests_for_optimal_proposal_model_interface( end @testset ( - "Optimal proposal model interface unit tests - $(parentmodule(typeof(model)))" -) for model in ( - # Use sigma != 1. to test if covariance is being scaled by sigma correctly - LLW2d.init(Dict("llw2d" => Dict("sigma" => [0.5, 1.5, 1.5]))), - Lorenz63.init(Dict()), - LinearGaussian.init(LinearGaussian.stochastically_driven_dsho_model_parameters()) + "Optimal proposal model interface unit tests - $(parentmodule(typeof(config.model)))" +) for config in ( + (; + # Use sigma != 1. to test if covariance is being scaled by sigma correctly + # Reduce mesh dimensions to keep test run time reasonable + model = LLW2d.init( + Dict( + "llw2d" => Dict( + "sigma" => [0.5, 1.5, 1.5], + "nx" => 11, + "ny" => 11, + "x_length" => 100e3, + "y_length" => 100e3, + "station_boundary_x" => 30e3, + "station_boundary_y" => 30e3, + ) + ) + ), + estimate_n_samples = [10, 100], + ), + (; model = Lorenz63.init(Dict()), estimate_n_samples = [10, 100, 1000]), + (; + model = LinearGaussian.init( + LinearGaussian.stochastically_driven_dsho_model_parameters() + ), + estimate_n_samples = [10, 100, 1000] + ) ) seed = 1234 - # Number of samples to use in convergence tests of Monte Carlo estimates - estimate_n_samples = [10, 100, 1000] # Constant factor used in Monte Carlo estimate convergence tests. Set based on some # trial and error to keep tests relatively sensitive while avoiding too high # probability of false failures but may require tweaking for each model estimate_bound_constant = 12.5 run_tests_for_optimal_proposal_model_interface( - model, seed, estimate_bound_constant, estimate_n_samples + config.model, seed, estimate_bound_constant, config.estimate_n_samples ) end