fv_weights_consistent
=====================

.. py:function:: t3toolbox.backend.fv_operations.fv_weights_consistent(variations, weights)

   .. code-block:: python

      def fv_weights_consistent(
              variations: typ.Tuple[typ.Sequence[NDArray], typ.Sequence[NDArray]],  # (V, H)
              weights:    typ.Tuple[typ.Sequence[NDArray], typ.Sequence[NDArray],
                                    typ.Sequence[NDArray], typ.Sequence[NDArray]],   # (up, down, left, right)
      ) -> bool:                                                                     # True iff shape-consistent


   True iff the four weight families (each len=d) can be absorbed into ``variations`` (non-raising).

   Ranks: up<->H.nU (axis -2), down<->V.nD (axis -2), left<->H.rL (axis -3), right<->H.rR (axis -1).

   Stacks -- the **trailing rule**: a weight is **frame-like**, carrying the frame stack ``C``, while the
   variations carry ``K + C`` (a ``K``-batch of tangents at one frame shares the one metric). So the
   weight's stack must be the **trailing (inner) part** of the variation stack -- exactly the rule
   ``check_fv_pair`` applies to a (frame, variations) pair, with the weight playing the frame's role.
   Absorption then broadcasts ``C`` over ``K + C`` for free through the leading ``'...'`` (which works
   because ``C`` is innermost). ``K`` may be empty, the common case.

   Like the variations themselves, this predicate is **blind to the frame**: a weight whose stack is the
   whole variation stack also passes, reading as ``C_w = K + C`` (that many base points, one tangent each)
   -- a legitimate absorption. Whether the weight is the metric of *this* tangent's frame needs the frame,
   and is checked by ``frame_variations_format.check_fw_pair`` at the tangent level.