DiscreteLayerGenerator

class DiscreteLayerGenerator(gateset=[HGate, TGate, CNOTGate], double_headed=False, dividing_gate_type=<class 'bqskit.ir.gates.parameterized.pauliz.PauliZGate'>)[source]

Bases: LayerGenerator

The DiscreteLayerGenerator class.

Expands circuits using only discrete gates. This is a non-reinforcement learning version of diagonalizing in https://arxiv.org/abs/2409.00433.

__init__(gateset=[HGate, TGate, CNOTGate], double_headed=False, dividing_gate_type=<class 'bqskit.ir.gates.parameterized.pauliz.PauliZGate'>)[source]

Construct a DiscreteLayerGenerator.

Parameters:

  • gateset (Sequence[Gate]) – A sequence of gates that can be used in the output circuit. These must be non-parameterized gates. (Default: [HGate, TGate, CNOTGate])

  • double_headed (bool) – If True, successors will be generated by both appending and prepending gates. This lets unitaries be diagonalized instead of inverted. (Default: False)

  • dividing_gate (Callable[[int], Gate]) – A gate that goes between the two heads of the discrete searches. If double_headed is False, this gate simply goes at the beggining of the circuit. (Default: PauliZGate)

Raises:

  • ValueError – If the gateset is not a sequence.

  • ValueError – If the gateset contains a parameterized gate.

  • ValueError – If the radices of gates are different.

Methods

cancels_something(circuit, gate, location)

Ensure applying gate at location does not cancel a previous gate.

count_repeats(circuit, gate, qudit)

Count the number of times the last gate is repeated on qudit.

gen_initial_layer(target, data)

Generate the initial layer, see LayerGenerator for more.

gen_successors(circuit, data)

Generate the successors of a circuit node.

hash_circuit_structure(circuit)