TopologicalComputationGraph#

class tqec.compile.TopologicalComputationGraph(scalable_qubit_shape, observable_builder, observables=None)[source]#

Bases: object

Represents a topological computation with Block instances.

Methods

__init__(scalable_qubit_shape, ...[, ...])

Represents a topological computation with Block instances.

add_cube(position, block)

add_pipe(source, sink, block)

Add the provided block as a pipe between source and sink.

generate_crumble_url(k[, manhattan_radius, ...])

Generate the Crumble URL from the compiled graph.

generate_stim_circuit(k[, noise_model, ...])

Generate the stim.Circuit from the compiled graph.

to_layer_tree()

Merge layers happening in parallel at each time step.

Detailed methods

Parameters:

  • scalable_qubit_shape (PhysicalQubitScalable2D)

  • observable_builder (ObservableBuilder)

  • observables (list[AbstractObservable] | None)

__init__(scalable_qubit_shape, observable_builder, observables=None)[source]#

Represents a topological computation with Block instances.

Parameters:

  • scalable_qubit_shape (PhysicalQubitScalable2D)

  • observable_builder (ObservableBuilder)

  • observables (list[AbstractObservable] | None)

Return type:

None

add_cube(position, block)[source]#
Parameters:

  • position (BlockPosition3D)

  • block (Block)

Return type:

None

add_pipe(source, sink, block)[source]#

Add the provided block as a pipe between source and sink.

Raises:

  • TQECException – if source and sink are not neighbouring positions.

  • TQECException – if not source < sink.

  • TQECException – if there is already a pipe between source and sink.

  • TQECException – if block is not a valid pipe (i.e., has not exactly 2 scalable dimensions).

Parameters:

  • source (BlockPosition3D)

  • sink (BlockPosition3D)

  • block (Block)

Return type:

None

generate_crumble_url(k, manhattan_radius=2, detector_database=None, add_polygons=False)[source]#

Generate the Crumble URL from the compiled graph.

Parameters:

  • k (int) – scaling factor.

  • manhattan_radius (int) – Parameter for the automatic computation of detectors. Should be large enough so that flows canceling each other to form a detector are strictly contained in plaquettes that are at most at a distance of manhattan_radius from the central plaquette. Detector computation runtime grows with this parameter, so you should try to keep it to its minimum. A value too low might produce invalid detectors.

  • detector_database (DetectorDatabase | None) – existing database of detectors that is used to avoid computing detectors if the database already contains them. Default to None which result in not using any kind of database and unconditionally performing the detector computation.

  • add_polygons (bool) – whether to include polygons in the Crumble URL. If True, the polygons representing the stabilizers will be generated based on the RPNG information of underlying plaquettes and add to the Crumble URL.

Returns:

a string representing the Crumble URL of the quantum circuit.

Return type:

str

generate_stim_circuit(k, noise_model=None, manhattan_radius=2, detector_database=None)[source]#

Generate the stim.Circuit from the compiled graph.

Parameters:

  • k (int) – scale factor of the templates.

  • noise_models – noise models to be applied to the circuit.

  • manhattan_radius (int) – radius considered to compute detectors. Detectors are not computed and added to the circuit if this argument is negative.

  • detector_database (DetectorDatabase | None) – an instance to retrieve from / store in detectors that are computed as part of the circuit generation.

  • noise_model (NoiseModel | None)

Returns:

A compiled stim circuit.

Return type:

Circuit

to_layer_tree()[source]#

Merge layers happening in parallel at each time step.

This method considers all the layers contained in added blocks (cubes and pipes) and merges them into a sequence of LayoutLayer or BaseComposedLayer wrapping LayoutLayer instances.

Returns:

A tree representing the topological computation.

The root node of the returned tree is an instance of SequencedLayers. Each child of the root node represents the computation happening during one block of time.

Each child of the root node is also an instance of SequencedLayers.

Return type:

LayerTree