continuing the campaign from #765
test_weight_broadcasting currently emits 420 numpy NaN-related RuntimeWarning due to the calculation on stochastic parameters on 100+ sparse triangles. These warnings are harmless in the context of the test because the stochastic parameters are not used. However, I'm not convinced that asserting them via pytest.warns is right approach.
alternatively, we can suppress stochastic parameter calculation. the stochastic parameters are currently calculated on every Development estimator. this is arguably wasteful. we can introduce a parameter, .e.g. stoch_params, to Development to skip the calc if the user doesn't intend to use them.
continuing the campaign from #765
test_weight_broadcastingcurrently emits 420 numpy NaN-relatedRuntimeWarningdue to the calculation on stochastic parameters on 100+ sparse triangles. These warnings are harmless in the context of the test because the stochastic parameters are not used. However, I'm not convinced that asserting them via pytest.warns is right approach.alternatively, we can suppress stochastic parameter calculation. the stochastic parameters are currently calculated on every
Developmentestimator. this is arguably wasteful. we can introduce a parameter, .e.g.stoch_params, toDevelopmentto skip the calc if the user doesn't intend to use them.