quimb.tensor.belief_propagation.sparse_ops ========================================== .. py:module:: quimb.tensor.belief_propagation.sparse_ops .. autoapi-nested-parse:: Specialized kernels for belief propagation with sparse COO tensors and dense vector messages. Both ``sparse.COO`` and n-dimensional ``scipy.sparse.coo_array`` are supported. Functions --------- .. autoapisummary:: quimb.tensor.belief_propagation.sparse_ops.parse_coo quimb.tensor.belief_propagation.sparse_ops.to_dense quimb.tensor.belief_propagation.sparse_ops._compute_all_tensor_messages_coo quimb.tensor.belief_propagation.sparse_ops._contract_all_messages_coo quimb.tensor.belief_propagation.sparse_ops._prepare_coo_ms_for_numba quimb.tensor.belief_propagation.sparse_ops.compute_all_tensor_messages_coo quimb.tensor.belief_propagation.sparse_ops.contract_tensor_messages_coo quimb.tensor.belief_propagation.sparse_ops.sum_all_but_axis_coo Module Contents --------------- .. py:function:: parse_coo(x) Get the ``(coords, data)`` pair of a COO array, with ``coords`` as a tuple of one flat coordinate array per dimension. Returns ``None`` if ``x`` is not a COO array with zero fill value. Duplicate coordinates are allowed, since every kernel here accumulates additively. .. py:function:: to_dense(x) Convert a possibly sparse array to a dense numpy array. .. py:function:: _compute_all_tensor_messages_coo(coords, data, ms, out) Numba kernel that computes all outgoing messages for COO tensor. .. py:function:: _contract_all_messages_coo(coords, data, ms, out) Numba kernel that fully contracts a COO tensor with all incoming messages. .. py:function:: _prepare_coo_ms_for_numba(coo, ms) Make sure COO format tensor and messages are ready for numba kernels. .. py:function:: compute_all_tensor_messages_coo(coo, ms) Given messages ``ms`` incident to a sparse tensor, compute all the outgoing messages, each the contraction of the tensor with every incident message but one. :param coo: The ``(coords, data)`` pair, as returned by ``parse_coo``. :type coo: (tuple[array], array) :param ms: The dense vector messages, one per dimension. :type ms: sequence of array :rtype: list[array] .. py:function:: contract_tensor_messages_coo(coo, ms) Contract a sparse tensor with a dense vector message on every dimension, to a scalar. :param coo: The ``(coords, data)`` pair, as returned by ``parse_coo``. :type coo: (tuple[array], array) :param ms: The dense vector messages, one per dimension. :type ms: sequence of array :rtype: scalar .. py:function:: sum_all_but_axis_coo(x, axis) Sum the sparse tensor ``x`` over every dimension but ``axis``, returning a dense vector. This is the BP 'uniform message' initialization step.