add a NotImplementedError and catch _that_ instaead of waiting for un…

…yt to raise a UnitConversionError. This fixes the following test: yt/data_objects/tests/test_covering_grid.py::test_smoothed_covering_grid_2d_dataset

yt/fields/field_info_container.py

@@ -24,8 +24,6 @@
     standard_particle_fields,
 )
 
-# from unyt.exceptions import UnitConversionError
-
 
 def tupleize(inp):
     if isinstance(inp, tuple):
@@ -486,6 +484,9 @@ def check_derived_fields(self, fields_to_check=None):
                 # underlying issues locally.
                 fd = fi.get_dependencies(ds=self.ds)
             except (
+                # this one I added on purpose as a catch for stuff
+                # that was revealed to be unsupported
+                NotImplementedError,
                 # Yt errors
                 # those are probably fine (but should be checked)
                 YTFieldNotFound,

yt/fields/vector_operations.py

@@ -372,6 +372,7 @@ def _tangential_over_magnitude(field, data):
                 data[ftype, f"tangential_{basename}"]
                 / data[ftype, f"{basename}_magnitude"]
             )
+
             return np.abs(tr)
 
         registry.add_field(

yt/utilities/lib/misc_utilities.pyx

@@ -628,11 +628,14 @@ def obtain_relative_velocity_vector(
     cdef np.ndarray[np.float64_t, ndim=3] vzg
     cdef np.ndarray[np.float64_t, ndim=4] rvg
     cdef np.float64_t bv[3]
-    cdef int i, j, k
+    cdef int i, j, k, dim
 
     units = data[field_names[0]].units
     bulk_vector = data.get_field_parameter(bulk_vector).to(units)
-    if len(data[field_names[0]].shape) == 1:
+    dim = len(data[field_names[0]].shape) == 1
+    if dim == 2:
+        raise NotImplementedError
+    if dim == 1:
         # One dimensional data
         vxf = data[field_names[0]].astype("float64")
         vyf = data[field_names[1]].astype("float64")
@@ -652,7 +655,7 @@ def obtain_relative_velocity_vector(
             rvf[1, i] = vyf[i] - bv[1]
             rvf[2, i] = vzf[i] - bv[2]
         return rvf
-    else:
+    elif dim == 3:
         # Three dimensional data
         vxg = data[field_names[0]].astype("float64")
         vyg = data[field_names[1]].astype("float64")