murefi.symbolic

class murefi.symbolic.IntegrationOp(solver, keys_y)

Bases: object

This is a PyTensor Op that becomes a node in the computation graph. It is not differentiable, because it uses a ‘solver’ function that is provided by the user.

Methods

grad
infer_shape
make_node
perform

grad(inputs, outputs)

infer_shape(fgraph, node, input_shapes)

make_node(y0: list, t, theta: list)

perform(node, inputs, output_storage)

murefi.symbolic.make_hash_sha256(obj)

Computes a sha256 hash for the object.

murefi.symbolic.make_hashable(obj)

Makes tuples, lists, dicts, sets and frozensets hashable.

murefi.symbolic.named_with_shapes_dict(vars: Dict[str, ndarray | Any], names: Sequence[str]) → Dict[str, Tuple[ndarray | Any, Tuple[int, ...]]]

murefi.symbolic.solve_sunode(dydt, independent_keys, y0, t, ode_parameters, parameter_names) → dict