t3frame_to_ut3frame#

t3toolbox.backend.ufv_conversions.t3frame_to_ut3frame(frame_data, N=None, nU=None, nD=None, rL=None, rR=None)#
def t3frame_to_ut3frame(
        frame_data: typ.Tuple[
            typ.Tuple[NDArray, ...],  # up_tucker_cores,   len=d, elm_shape=C+(nUi, Ni)
            typ.Tuple[NDArray, ...],  # down_tt_cores,     len=d, elm_shape=C+(rLi, nDi, rR(i+1))
            typ.Tuple[NDArray, ...],  # left_tt_cores,     len=d, elm_shape=C+(rLi, nUi, rL(i+1))
            typ.Tuple[NDArray, ...],  # right_tt_cores,    len=d, elm_shape=C+(rRi, nUi, rR(i+1))
        ],
        N:  typ.Optional[int] = None,   # padded mode dim   (default max(shape))
        nU: typ.Optional[int] = None,   # padded up rank    (default max(up_ranks))
        nD: typ.Optional[int] = None,   # padded down rank  (default max(down_ranks))
        rL: typ.Optional[int] = None,   # padded left rank  (default max(left_ranks))
        rR: typ.Optional[int] = None,   # padded right rank (default max(right_ranks))
) -> typ.Tuple[
    NDArray,                          # up_tucker_supercore,  (d,)+C+(nU, N)
    NDArray,                          # down_tt_supercore,    (d,)+C+(rL, nD, rR)
    NDArray,                          # left_tt_supercore,    (d,)+C+(rL, nU, rL)
    NDArray,                          # right_tt_supercore,   (d,)+C+(rR, nU, rR)
    typ.Tuple[int, ...],              # shape = (N0,...,N(d-1))
    typ.Tuple[NDArray, NDArray, NDArray, NDArray],  # (up, down, frame_left, frame_right) masks
]:

Pack a ragged T3Frame core-tuple into uniform frame .data (supercores + shape + masks).

Inverse of ut3frame_to_t3frame(). A single ragged frame has ranks shared across its C stack, so the masks come out uniform across the stack (varying-rank uniform batches arise only by stack-ing a heterogeneous tree). Pads each family to common dims (default: max over modes; pass N/nU/nD/rL/rR to force larger) and records the real extents as prefix masks (ufv_make_frame_masks; the real content lands in the upper-left, so the prefix masks are correct).

Parameters:

  • frame_data (t3toolbox.backend.common.typ.Tuple[t3toolbox.backend.common.typ.Tuple[NDArray, Ellipsis], t3toolbox.backend.common.typ.Tuple[NDArray, Ellipsis], t3toolbox.backend.common.typ.Tuple[NDArray, Ellipsis], t3toolbox.backend.common.typ.Tuple[NDArray, Ellipsis]])

  • N (t3toolbox.backend.common.typ.Optional[int])

  • nU (t3toolbox.backend.common.typ.Optional[int])

  • nD (t3toolbox.backend.common.typ.Optional[int])

  • rL (t3toolbox.backend.common.typ.Optional[int])

  • rR (t3toolbox.backend.common.typ.Optional[int])

Return type:

t3toolbox.backend.common.typ.Tuple[NDArray, NDArray, NDArray, NDArray, t3toolbox.backend.common.typ.Tuple[int, Ellipsis], t3toolbox.backend.common.typ.Tuple[NDArray, NDArray, NDArray, NDArray]]