qsearch.post_processing

This module defines PostProcessor, a class used to modify circuits after they have been synthesized.

Several implementations are provided.

qsearch.post_processing.BasicSingleQubitReduction_PostProcessor

Attempts to remove single-qubit gates without sacrificing the quality of the solution in terms of eval_func

qsearch.post_processing.ParameterTuning_PostProcessor

Attempts to reduce eval_func simply by re-running the solver with stronger parameters.

qsearch.post_processing.LEAPReoptimizing_PostProcessor

Reduces the length of circuits produced using LEAP by re-running segments of the circuit.

Module Contents

Classes

PostProcessor	This class is used to modify circuits that have already been synthesized.
BasicSingleQubitReduction_PostProcessor	Attempts to reduce the number of single-qubit gates in a circuit by sequentially removing a gate, attempting to use a Solver on it, and keeping that gate removed if successful.
ParameterTuning_PostProcessor	Attempts to reduce the eval_func value of the circuit simply by tuning the parameters better using stronger Solver parameters.
LEAPReoptimizing_PostProcessor	A PostProcessor that re-optimizes LeapCompiler-compiled circuits via search.

class qsearch.post_processing.PostProcessor(options=opt.Options())

This class is used to modify circuits that have already been synthesized.

post_process_circuit(self, result, options=None)

Processes the circuit dictionary and returns a new one.

Parameters

result – A dictionary containing a synthesized circuit. Expect it to contain “structure” and “parameters”, but it may contain more, depending on what previous PostProcessors were run and on the compiler.

Returns

A dictionary containing any updates that should be made to the circuit dictionary, such as new values for “structure” or “parameters” or arbitrary other data.

Return type

dict

class qsearch.post_processing.BasicSingleQubitReduction_PostProcessor(options=opt.Options())

Bases: PostProcessor

Attempts to reduce the number of single-qubit gates in a circuit by sequentially removing a gate, attempting to use a Solver on it, and keeping that gate removed if successful.

post_process_circuit(self, result, options=None)

Processes the circuit dictionary and returns a new one.

Parameters

result – A dictionary containing a synthesized circuit. Expect it to contain “structure” and “parameters”, but it may contain more, depending on what previous PostProcessors were run and on the compiler.

Returns

A dictionary containing any updates that should be made to the circuit dictionary, such as new values for “structure” or “parameters” or arbitrary other data.

Return type

dict

class qsearch.post_processing.ParameterTuning_PostProcessor(options=opt.Options())

Bases: PostProcessor

Attempts to reduce the eval_func value of the circuit simply by tuning the parameters better using stronger Solver parameters.

post_process_circuit(self, result, options=None)

Processes the circuit dictionary and returns a new one.

Parameters

result – A dictionary containing a synthesized circuit. Expect it to contain “structure” and “parameters”, but it may contain more, depending on what previous PostProcessors were run and on the compiler.

Returns

A dictionary containing any updates that should be made to the circuit dictionary, such as new values for “structure” or “parameters” or arbitrary other data.

Return type

dict

class qsearch.post_processing.LEAPReoptimizing_PostProcessor(options=Options())

Bases: qsearch.compiler.Compiler, PostProcessor

A PostProcessor that re-optimizes LeapCompiler-compiled circuits via search.

This PostProcessor puts “holes” in the circuit where LEAP fixed prefixes and runs qsearch on those holes to reduce the total number of gates.

post_process_circuit(self, result, options=None)

Re-optimize a LEAP circuit. Pass “depth” to indicate the size to re-synthesize. It is recommended to call like: project.post_process(post_processing.LEAPReoptimizing_PostProcessor(), solver=multistart_solvers.MultiStart_Solver(8), parallelizer=parallelizers.ProcessPoolParallelizer, depth=7)

compile(self, options=Options())

Backwards compatible interface since this is technically a Compiler.

You should use LEAPReoptimizing_PostProcessor.post_process_circuit with the Project post_processing API.