amazon-braket
Cannot run observables from non-indexed TensorProducts with ProgramSets
Describe the bug
device.run(Circuit().h(0).h(1).expectation(Z() @ Z(), target=[0,1])).result() works fine. However,
device.run(ProgramSet.zip(circuits = [Circuit().h(0).h(1)], observables= [Z() @ Z()]), shots = 1000).result() fails.
Non-indexed TensorProducts lead to some sort of errors in programs.inputs where there is incorrectly a None object.
To reproduce
See above snippets. Uses LocalSimulator.
Alternatively:
"device.run(ProgramSet( CircuitBinding( circuit=Circuit().h(0).h(1), observables= [Z() @ Z()] ), shots_per_executable=100)).result() "
Expected behavior
TensorProducts without indices should be able to run okay.
Screenshots or logs
251 return ParameterSets(program.inputs) 252 num_observables = len(observables) 253 return ParameterSets({ 254 k: v[::num_observables] --> 255 for k, v in program.inputs.items() 256 if not k.startswith(EULER_OBSERVABLE_PREFIX) 257 })
AttributeError: 'NoneType' object has no attribute 'items'
System information A description of your system. Please provide:
- Amazon Braket Python SDK version: 1.103.0
- Amazon Braket Python Schemas version: 1.26.1
- Amazon Braket Python Default Simulator version: 1.32.0
- Python version: 3.12.10
Additional context Add any other context about the problem here.
Definitely needs a clearer message; there's also the possibility of having implicit targets when the number of qubits in the observable matches that of the circuit.
