ut3frameweights_to_t3frameweights
=================================

.. py:function:: t3toolbox.backend.ufv_conversions.ut3frameweights_to_t3frameweights(weights)

   .. code-block:: python

      def ut3frameweights_to_t3frameweights(
              weights: typ.Tuple[
                  NDArray, NDArray, NDArray, NDArray,             # up, down, left, right supercores, (d,)+C+(size,)
                  typ.Tuple[NDArray, NDArray, NDArray, NDArray],  # the four edge masks
              ],
      ) -> typ.Union[
          typ.Tuple[typ.Tuple[NDArray, ...], ...],  # (up, down, left, right) ragged families, if unstacked
          typ.Tuple,                                 # else a nested tree (shaped like C) of those
      ]:


   Convert uniform frame-weight supercores + masks back to ragged ``T3FrameWeights`` families.

   The frame-weight twin of :py:func:`ut3frame_to_t3frame`. As there, a **stacked** weight returns a
   *tree* of ragged weights (a varying-rank stack has no single ragged representation), and the real slots
   are selected *through the masks* rather than by slicing a prefix, since a mask may be gappy after
   ``+``/``x`` (``docs/uniform_masks_vs_ranks.md``).