make vi (posterior) mean and std accessible as a structured xarray

pymc/tests/test_variational_inference.py

@@ -499,6 +499,63 @@ def simple_model(simple_model_data):
     return model
 
 
+@pytest.fixture
+def hierarchical_model_data():
+    group_coords = {
+        "group_d1": np.arange(3),
+        "group_d2": np.arange(7),
+    }
+    group_shape = tuple(len(d) for d in group_coords.values())
+    data_coords = {"data_d": np.arange(11)} | group_coords
+
+    data_shape = tuple(len(d) for d in data_coords.values())
+
+    mu = -5.0
+
+    sigma_group_mu = 3
+    group_mu = sigma_group_mu * np.random.randn(*group_shape)
+
+    sigma = 3.0
+
+    data = sigma * np.random.randn(*data_shape) + group_mu + mu
+
+    return dict(
+        group_coords=group_coords,
+        group_shape=group_shape,
+        data_coords=data_coords,
+        data_shape=data_shape,
+        mu=mu,
+        sigma_group_mu=sigma_group_mu,
+        sigma=sigma,
+        group_mu=group_mu,
+        data=data,
+    )
+
+
+@pytest.fixture
+def hierarchical_model(hierarchical_model_data):
+    with pm.Model(coords=hierarchical_model_data["data_coords"]) as model:
+        mu = pm.Normal("mu", mu=0, sigma=10)
+        sigma_group_mu = pm.HalfNormal("sigma_group_mu", sigma=5)
+
+        group_mu = pm.Normal(
+            "group_mu",
+            mu=0,
+            sigma=sigma_group_mu,
+            dims=list(hierarchical_model_data["group_coords"].keys()),
+        )
+
+        sigma = pm.HalfNormal("sigma", sigma=3)
+
+        pm.Normal(
+            "data",
+            mu=(mu + group_mu),
+            sigma=sigma,
+            observed=hierarchical_model_data["data"],
+        )
+    return model
+
+
 @pytest.fixture(
     scope="module",
     params=[
@@ -571,6 +628,27 @@ def test_fit_oo(inference, fit_kwargs, simple_model_data):
     np.testing.assert_allclose(np.std(trace.posterior["mu"]), np.sqrt(1.0 / d), rtol=0.2)
 
 
+def test_fit_data(inference, fit_kwargs, simple_model_data):
+    fitted = inference.fit(**fit_kwargs)
+    mu_post = simple_model_data["mu_post"]
+    d = simple_model_data["d"]
+    np.testing.assert_allclose(fitted.mean_data["mu"].values, mu_post, rtol=0.05)
+    np.testing.assert_allclose(fitted.std_data["mu"], np.sqrt(1.0 / d), rtol=0.2)
+
+
+def test_fit_data_coords(hierarchical_model, hierarchical_model_data):
+    with hierarchical_model:
+        fitted = pm.fit(1)
+
+    for data in [fitted.mean_data, fitted.std_data]:
+        assert set(data.keys()) == {"sigma_group_mu_log__", "sigma_log__", "group_mu", "mu"}
+        assert data["group_mu"].shape == hierarchical_model_data["group_shape"]
+        assert list(data["group_mu"].coords.keys()) == list(
+            hierarchical_model_data["group_coords"].keys()
+        )
+        assert data["mu"].shape == tuple()
+
+
 def test_profile(inference):
     inference.run_profiling(n=100).summary()
 

pymc/variational/opvi.py

@@ -52,6 +52,7 @@
 import aesara
 import aesara.tensor as at
 import numpy as np
+import xarray
 
 from aesara.graph.basic import Variable
 
@@ -1109,16 +1110,47 @@ def __str__(self):
 
     @node_property
     def std(self):
+        """Standard deviation of the latent variables as an unstructured 1-dimensional Aesara variable"""
         raise NotImplementedError
 
     @node_property
     def cov(self):
+        """Covariance between the latent variables as an unstructured 2-dimensional Aesara variable"""
         raise NotImplementedError
 
     @node_property
     def mean(self):
+        """Mean of the latent variables as an unstructured 1-dimensional Aesara variable"""
         raise NotImplementedError
 
+    def var_to_data(self, shared):
+        """Takes a flat 1-dimensional Aesara variable and maps it to an xarray data set based on the information in
+        `self.ordering`.
+        """
+        # This is somewhat similar to `DictToArrayBijection.rmap`, which doesn't work here since we don't have
+        # `RaveledVars` and need to take the information from `self.ordering` instead
+        shared = shared.eval()
+        result = dict()
+        for name, s, shape, dtype in self.ordering.values():
+            dims = self.model.RV_dims.get(name, None)
+            if dims is not None:
+                coords = {d: np.array(self.model.coords[d]) for d in dims}
+            else:
+                coords = None
+            values = np.array(shared[s]).reshape(shape).astype(dtype)
+            result[name] = xarray.DataArray(values, coords=coords, dims=dims, name=name)
+        return xarray.Dataset(result)
+
+    @property
+    def mean_data(self):
+        """Mean of the latent variables as an xarray Dataset"""
+        return self.var_to_data(self.mean)
+
+    @property
+    def std_data(self):
+        """Standard deviation of the latent variables as an xarray Dataset"""
+        return self.var_to_data(self.std)
+
 
 group_for_params = Group.group_for_params
 group_for_short_name = Group.group_for_short_name