compute_sigma_hat_jets
======================

.. py:function:: t3toolbox.backend.sampling_derivatives.compute_sigma_hat_jets(right_tt_cores, xi_jets, c, trs)

   .. code-block:: python

      def compute_sigma_hat_jets(
              right_tt_cores: typ.Sequence[NDArray],  # Q.  len=d, elm_shape=C+(rRi,nUi,rR(i+1))
              xi_jets:        typ.Sequence[NDArray],  # frame input jets, len=d, elm_shape=(2,)+W+C+(nUi,)
              c:              NDArray,                # residual jet (scalar), shape=(order+1)+W+K+C
              trs:            NDArray,                # binomial tensor, shape=(order+1,order+1,order+1)
      ) -> typ.Tuple[NDArray, ...]:                   # sigma_hats. len=d, elm_shape=(order+1)+W+K+C+(rR(i+1),)


   Propagation-only adjoint sweep via Q, seeded at the terminal bond by the residual jet ``c``.

   The apply-transpose analog of :py:func:`compute_sigma_tilde_jets`: no ``deta_tilde`` per-core source,
   and the init carry is the ``c``-seed on the terminal bond (``rR_d = 1``) rather than zeros.
   ``sigma_hats[i]`` is the adjoint of the after-core-``i`` perturbation carry; it carries the tangent
   stack ``K`` (from ``c``). Right-to-left via ``tt_reverse`` (mirroring the ``Q``-sweep there).