removed checking for MvNormal

Author: Michal-Novomestsky

pymc_extras/inference/INLA/inla.py

@@ -17,22 +17,6 @@ def fit_INLA(
 ) -> az.InferenceData:
     model = pm.modelcontext(model)
 
-    # Check if latent field is Gaussian
-    # if not isinstance(x.owner.op, MvNormal):
-    #     raise ValueError(
-    #         f"Latent field {x} is not instance of MvNormal. Has distribution {x.owner.op}."
-    #     )
-
-    # _, _, _, tau = x.owner.inputs
-
-    # # Latent field should use precison rather than covariance
-    # if not (tau.owner and tau.owner.op == matrix_inverse):
-    #     raise ValueError(
-    #         f"Latent field {x} is not in precision matrix form. Use MvNormal(tau=Q) instead."
-    #     )
-
-    # Q = tau.owner.inputs[0]
-
     # TODO is there a better way to check if it's a RV?
     # print(vars(Q.owner))
     # if isinstance(Q, TensorVariable) and "module" in vars(Q.owner):
@@ -48,7 +32,6 @@ def fit_INLA(
     marginal_model = marginalize(model, x, use_laplace=True, **marginalize_kwargs)
 
     # Sample over the hyperparameters
-    # marginal_model.logp().dprint()
     idata = pm.sample(model=marginal_model, **sampler_kwargs)
 
     if not return_latent_posteriors:

pymc_extras/model/marginal/marginal_model.py

@@ -191,10 +191,6 @@ def marginalize(
                 raise NotImplementedError(
                     "Marginalization for DiscreteMarkovChain with non-matrix transition probability is not supported"
                 )
-        # elif use_laplace and not isinstance(rv_op, MvNormal):
-        #     raise ValueError(
-        #         f"Marginalisation method set to Laplace but RV {rv_to_marginalize} is not instance of MvNormal. Has distribution {rv_to_marginalize.owner.op}"
-        #     )
 
         elif not use_laplace and not isinstance(rv_op, Bernoulli | Categorical | DiscreteUniform):
             raise NotImplementedError(

tests/inference/INLA/test_inla.py

@@ -26,50 +26,6 @@ def rng():
     return np.random.default_rng(seed)
 
 
-@pytest.mark.filterwarnings("ignore:Currently only valid for a nested normal model. WIP.")
-def test_non_gaussian_latent(rng):
-    """
-    INLA should raise an error if trying to use a non-Gaussian latent
-    """
-    n = 10000
-
-    mu_true = rng.random()
-    tau = 1
-    y_obs = rng.normal(loc=mu_true, scale=1 / tau, size=n)
-
-    with pm.Model() as model:
-        lam = pm.HalfNormal("lam", tau=tau)
-        x = pm.Exponential("x", lam=lam)
-        y = pm.Normal("y", mu=x, tau=tau, observed=y_obs)
-
-        with pytest.raises(ValueError):
-            pmx.fit(method="INLA", x=x)
-
-
-@pytest.mark.filterwarnings("ignore:Currently only valid for a nested normal model. WIP.")
-def test_non_precision_MvNormal(rng):
-    """
-    INLA should raise an error if trying to use a latent not in precision form
-    """
-    n = 10000
-    d = 3
-
-    mu_mu = np.zeros((d,))
-    mu_true = rng.random(d)
-    tau = np.identity(d)
-    cov = np.linalg.inv(tau)
-    y_obs = rng.multivariate_normal(mean=mu_true, cov=cov, size=n)
-
-    with pm.Model() as model:
-        mu = pm.MvNormal("mu", mu=mu_mu, tau=tau)
-        x = pm.MvNormal("x", mu=mu, cov=cov)
-        y = pm.MvNormal("y", mu=x, tau=tau, observed=y_obs)
-
-        with pytest.raises(ValueError):
-            pmx.fit(method="INLA", x=x)
-
-
-@pytest.mark.filterwarnings("ignore:Currently only valid for a nested normal model. WIP.")
 def test_3_layer_normal(rng):
     """
     Test INLA against a simple toy problem: