quri_parts.tensornetwork.operator package¶

class TensorNetworkOperator(input_edges, output_edges, container, layer_tensor_map)¶

Bases: TensorNetworkLayer

Tensor network representation of a operators.

This class subclasses TensorNetworkLayer and provides, in addition to input and output edges for the operator, also a list of indices that the operator acts on. These indices are defined with respect to some quantum state that the operator is intended to act on. The intent is to allow for certain optimizations with tensor contraction.

Parameters:

input_edges (Sequence[Edge])
output_edges (Sequence[Edge])
container (Union[set[AbstractNode], list[AbstractNode]])
layer_tensor_map (list[dict[int, AbstractNode]])

copy()¶

Returns a copy of itself.

Return type:: TensorNetworkOperator

contract_with(state)¶

Parameters:: state (TensorNetworkState)
Return type:: AbstractNode

get_observable_data(pauli_list, dim=2, n=None)¶

Take a list of matrices that represent pauli operators and calculate their tensor-product.

Parameters:

pauli_list (Sequence[Sequence[Sequence[complex]]])
dim (int)
n (int | None)

Return type:

ndarray[Any, dtype[complex128]]

pauli_label_to_array(pauli, index_list=None)¶

Convert a PauliLabel to a numpy array.

If this function is used to convert a PauliLabel belonging to an Operator, then an index_list, describing all of the indices acted on by the Operator, should be passed as an argument. The returned array will then be padded with identity operators where needed.

Parameters:

pauli (PauliLabel)
index_list (Collection[int] | None)

Return type:

ndarray[Any, dtype[complex128]]

tensor_to_mpo(operator, max_bond_dimension=None, max_truncation_err=None, qubits_per_node=1)¶

Perform the singular value decomposition on an operator represented using TensorNetworkLayer and return the matrix product operator (MPO) as a TensorNetworkLayer

Parameters:

operator (TensorNetworkOperator) – Input operator
max_bond_dimension (int | None) – Optional specification of MPO bond-dimension
max_truncation_error – Optional specification of truncation error tolerance
qubits_per_node (int) – Number of physical qubits attached to each node, defaults to 1
max_truncation_err (float | None)

Return type:

TensorNetworkOperator

Outputs:: MPO as a TensorNetworkOperator

operator_to_tensor(operator, convert_to_mpo=True, backend='numpy', *args, **kwargs)¶

Convert an Operator or a PauliLabel to a TensorNetworkOperator.

Parameters:

operator (Operator | PauliLabel) – Input operator
convert_to_mpo (bool) – Whether or not to convert the operator to an MPO
backend (str) – Backend to use for tensor operations
args (Any)
kwargs (Any)

Keyword Arguments:

max_bond_dimension – if convert_to_mpo is true, this sets the maximum bond dimension
max_truncation_err – if convert_to_mpo is true, this sets the maximum truncation error

Return type: