Revert "fixed shape calls to tensors"

This reverts commit 57177f1.
pnnl-predictive-phenomics · Jun 6, 2024 · ef5d2f7 · ef5d2f7
1 parent 57177f1
commit ef5d2f7
Show file tree

Hide file tree

Showing 3 changed files with 6 additions and 6 deletions.
diff --git a/src/emll/data_model_integration.py b/src/emll/data_model_integration.py
@@ -24,7 +24,7 @@ def create_noisy_observations_of_computed_values(name:str, computed_tensor:T.ten
         raise ValueError("Data shape does not match standard deviation shape!")
 
     # check shape of data == computed_tensor
-    if any(data.shape != computed_tensor.eval().shape):
+    if any(data.shape != computed_tensor.shape.eval()):
         raise ValueError(f"Data shape {data.shape} does not match computed tensor shape {pytensor.tensor.shape(computed_tensor)}!")
 
     # check that standard deviations are all positive

diff --git a/src/emll/linlog_model.py b/src/emll/linlog_model.py
@@ -135,7 +135,7 @@ def steady_state_pytensor(self, Ex, Ey=None, en=None, yn=None, method="scan"):
             en = np.atleast_2d(en)
             n_exp = en.shape[0]
         else:
-            n_exp = en.eval().shape[0]
+            n_exp = en.shape.eval()[0]
 
         if isinstance(yn, np.ndarray):
             yn = np.atleast_2d(yn)

diff --git a/tests/test_data_model_integration.py b/tests/test_data_model_integration.py
@@ -394,8 +394,8 @@ def test_create_pytensor_from_data():
         data_tensor = create_pytensor_from_data_naive(input_string, input_dataframe_mixed, input_stdev_dataframe_mixed, input_laplace_dataframe_mixed)
 
         # check that the n-dim and shape are correct
-        assert np.size(data_tensor.eval().shape) == np.size(input_dataframe_mixed.shape), f"Expected tensor n dim {np.size(input_dataframe_mixed.shape)}, found {np.size(data_tensor.eval().shape)}"
-        assert tuple(data_tensor.eval().shape) == input_dataframe_mixed.shape, f"Expected tensor shape {input_dataframe_mixed.shape}, found {data_tensor.eval().shape}"
+        assert np.size(data_tensor.shape.eval()) == np.size(input_dataframe_mixed.shape), f"Expected tensor n dim {np.size(input_dataframe_mixed.shape)}, found {np.size(data_tensor.shape.eval())}"
+        assert tuple(data_tensor.shape.eval()) == input_dataframe_mixed.shape, f"Expected tensor shape {input_dataframe_mixed.shape}, found {data_tensor.shape.eval()}"
 
         # Check that the steady_state_pytensor method runs
         error_occurred = False
@@ -422,8 +422,8 @@ def test_create_pytensor_from_data():
             y_n = pm.Normal("yn_t", 
                             mu=0, 
                             sigma=10, 
-                            shape=(data_tensor.eval().shape[0], lin_log.ny), 
-                            initval=0.1 * np.random.randn(data_tensor.eval().shape[0], lin_log.ny)
+                            shape=(data_tensor.shape.eval()[0], lin_log.ny), 
+                            initval=0.1 * np.random.randn(data_tensor.shape.eval()[0], lin_log.ny)
                             )
 
             chi_ss, vn_ss = lin_log.steady_state_pytensor(