tud_lbm.operators.boundary
==========================

.. py:module:: tud_lbm.operators.boundary

.. autoapi-nested-parse::

   Boundary condition operators — composite builder.

   Public API: build_bc(), BCMasks, build_bc_masks

   Implementation modules (_bounce_back.py, _periodic.py, _symmetry.py) are internal.

   .. rubric:: Example

   from operators.boundary import build_bc, build_bc_masks

   bc_fn = build_bc({"top": "symmetry", "bottom": "bounce-back"}, lattice)
   masks = build_bc_masks((64, 64))
   f = bc_fn(f_stream, f_col, masks)



Classes
-------

.. autoapisummary::

   tud_lbm.operators.boundary.BCMasks


Functions
---------

.. autoapisummary::

   tud_lbm.operators.boundary.build_bc_masks
   tud_lbm.operators.boundary.build_bc


Package Contents
----------------

.. py:class:: BCMasks

   Bases: :py:obj:`NamedTuple`


   Pre-computed boundary-condition masks — valid JAX pytree.

   Each mask is a boolean ``jax.Array`` of shape ``(nx, ny, nz, 1, 1)``
   that is ``True`` on the corresponding edge row/column/plane.

   .. attribute:: top

      Mask for the top boundary (y = ny-1).

   .. attribute:: bottom

      Mask for the bottom boundary (y = 0).

   .. attribute:: left

      Mask for the left boundary (x = 0).

   .. attribute:: right

      Mask for the right boundary (x = nx-1).

   .. attribute:: front

      Mask for the front boundary (z = nz-1).

   .. attribute:: back

      Mask for the back boundary (z = 0).


   .. py:attribute:: top
      :type:  jax.numpy.ndarray


   .. py:attribute:: bottom
      :type:  jax.numpy.ndarray


   .. py:attribute:: left
      :type:  jax.numpy.ndarray


   .. py:attribute:: right
      :type:  jax.numpy.ndarray


   .. py:attribute:: front
      :type:  jax.numpy.ndarray | None
      :value: None



   .. py:attribute:: back
      :type:  jax.numpy.ndarray | None
      :value: None



.. py:function:: build_bc_masks(grid_shape: tuple[int, Ellipsis]) -> BCMasks

   Construct pre-computed boundary-condition masks.

   Each mask is a boolean array of shape ``(nx, ny, nz, 1, 1)`` that is
   ``True`` on the corresponding edge row/column/plane.

   :param grid_shape: Spatial dimensions ``(nx, ny, nz)``.

   :returns: A :class:`BCMasks` NamedTuple with 3D-shaped masks for all 6 faces.


.. py:function:: build_bc(bc_config: dict[str, Any] | None, lattice: tud_lbm.lattice.lattice.Lattice) -> tud_lbm.operators.protocols.BoundaryOperator

   Build a composite BC closure from a bc_config dict.

   The returned function applies boundary conditions in order:
   bottom → top → left → right.  Edges not present in *bc_config*
   or mapped to ``"periodic"`` are no-ops.

   :param bc_config: Mapping ``{_edge: bc_type, ...}``, e.g.
                     ``{"top": "symmetry", "bottom": "bounce-back", "left": "periodic", "right": "periodic"}``.
                     ``None`` means all-periodic.
   :param lattice: :class:`~setup.lattice.Lattice`.

   :returns: A callable satisfying the BoundaryOperator protocol,
             ``bc_fn(f_stream, f_col, bc_masks) → f``.