ufv_weights_from_ut3_weights
============================

.. py:function:: t3toolbox.backend.ufv_operations.ufv_weights_from_ut3_weights(weights)

   .. code-block:: python

      def ufv_weights_from_ut3_weights(
              weights: typ.Tuple[NDArray, NDArray, typ.Tuple[NDArray, NDArray]],  # UT3Weights .data
      ) -> UT3FrameWeightsData:  # (up, down, left, right, masks) -- a tangent metric


   Build uniform frame weights (a tangent metric) from uniform base-point edge weights -- the twin of
   ``fv_weights_from_t3_weights``, and the same slicing, applied to supercores **and** masks:

   ``up = down = tucker``; ``left = tt[:-1]``, ``right = tt[1:]``. The TT slicing encodes the ``H_i`` bond
   convention (``H_i``'s left bond is TT bond ``i``, its right bond is bond ``i+1``), which turns the
   ``d+1`` bond supercore into the two ``d``-length families -- non-obvious, hence a named function.

   The result pairs with a **minimal-rank** tangent (where the complement rank ``nD`` equals the Tucker
   rank ``nU``, as for ``ut3svd`` output). A non-minimal tangent has ``nD < nU`` and would mismatch the
   ``down`` family at use -- caught by ``ufv_weights_consistent`` rather than silently absorbed.