ufv_leaf_structure
==================

.. py:function:: t3toolbox.backend.ufv_operations.ufv_leaf_structure(d, n_supercores)

   .. code-block:: python

      def ufv_leaf_structure(
              d:            int,  # number of modes (length of the int-tuple `shape`)
              n_supercores: int,  # 4 for a UT3Frame .data leaf, 2 for a UT3Variations one
      ):  # -> leaf-structure template for stacking.apply_func_to_leaf_subtrees


   Template marking one bv ``.data`` leaf ``(*supercores, shape, masks)`` for the tree machinery.

   The int-tuple ``shape`` has ``d`` int leaves, so the template must spell its length out: a bare
   ``None`` there fails to match (an int tuple is a ``Sequence``, unlike the ndarray leaves, so the
   walker would recurse into it). Mirrors ``ut3_operations.ut3_leaf_structure``.