t3_concatenate
==============

.. py:function:: t3toolbox.backend.t3_operations.t3_concatenate(xx)

   .. code-block:: python

      def t3_concatenate(
              xx: typ.Sequence[
                  typ.Tuple[typ.Sequence[NDArray], typ.Sequence[NDArray]]
              ],  # T3 data tuples, each (tucker_cores, tt_cores); TT ranks must match at each seam
      ) -> typ.Tuple[
          typ.Tuple[NDArray, ...],  # tucker_cores (concatenated over modes)
          typ.Tuple[NDArray, ...],  # tt_cores
      ]:


   Concatenate a sequence of T3 segments into one T3 (inverse of :py:func:`t3_segment`).

   At each seam the TT ranks must match -- the trailing TT rank of one segment equals the leading
   TT rank of the next -- otherwise a ``ValueError`` is raised. The join is structural (core-tuple
   concatenation); no re-orthogonalization.