Simplify DataTree handling in maybe_transpose_dims

Use single map_over_datasets call instead of two separate calls.
map_over_datasets supports tuple returns and unpacks them automatically,
making the explicit lambda extraction unnecessary.

Co-authored-by: Claude <[email protected]>

Author: max-sixty

xarray/testing/assertions.py

@@ -112,15 +112,8 @@ def _maybe_transpose_dims(a, b):
         return a, b
 
     if isinstance(a, DataTree):
-        # DataTree needs special handling - transpose both trees
-        # map_over_datasets applies a function over corresponding datasets
-        transposed_a = map_over_datasets(
-            lambda a_ds, b_ds: _maybe_transpose_dims(a_ds, b_ds)[0], a, b
-        )
-        transposed_b = map_over_datasets(
-            lambda a_ds, b_ds: _maybe_transpose_dims(a_ds, b_ds)[1], a, b
-        )
-        return transposed_a, transposed_b
+        # map_over_datasets supports tuple returns and unpacks them automatically
+        return map_over_datasets(_maybe_transpose_dims, a, b)
 
     return _maybe_transpose_dims(a, b)