uniform_probe_kind
==================

.. py:function:: t3toolbox.backend.uniform_fitting.uniform_probe_kind(x0_data, weight = None)

   .. code-block:: python

      def uniform_probe_kind(
              x0_data:  typ.Tuple,
              weight:   typ.Optional[typ.Any] = None,  # per-mode residual weight ω, (d,) / (d,1); None = 1
      ) -> bfit.SamplingKind:        # the uniform vector-valued `probe` sampling kind (optionally per-mode weighted)


   The uniform **probe** ``SamplingKind`` -- the twin of :py:func:`t3toolbox.backend.fitting.probe_kind`.

   The forward / residual are the **packed** probe output ``(d,)+W+C+(N,)`` (mode index ``d`` at axis 0, no
   order axis), so the per-mode weight ``ω`` is built with ``mode_axis=0`` (no order axis) and ``sumsq`` /
   ``transpose`` are overridden to weight it (``ω`` enters ``sumsq`` ×ω, ``transpose`` ×ω²).