`qml.qchem.convert_observable` converts complex coefficients in openfermion `QubitOperator` to real without warning or error

[aemccoy]

[bug]

Issue description

• Expected behavior: Error when QubitOperator with complex coefficients are passed to qchem.convert_observable since vqe.Hamiltonian does not currently support complex coefficients.

• Actual behavior: No error arrises and complex part of QubitOperator is zeroed out.

• Reproduces how often: Always

• System information: (post the output of import pennylane as qml; qml.about()) Name: PennyLane Version: 0.16.0.dev0 Summary: PennyLane is a Python quantum machine learning library by Xanadu Inc. Home-page: https://github.com/XanaduAI/pennylane Author: None Author-email: None License: Apache License 2.0 Location: /opt/anaconda3/envs/pennylane/lib/python3.9/site-packages Requires: toml, semantic-version, networkx, autograd, appdirs, numpy, scipy Required-by: PennyLane-qiskit, PennyLane-Qchem Platform info: macOS-10.16-x86_64-i386-64bit Python version: 3.9.4 Numpy version: 1.20.1 Scipy version: 1.6.1 Installed devices:

  • default.gaussian (PennyLane-0.16.0.dev0)
  • default.mixed (PennyLane-0.16.0.dev0)
  • default.qubit (PennyLane-0.16.0.dev0)
  • default.qubit.autograd (PennyLane-0.16.0.dev0)
  • default.qubit.jax (PennyLane-0.16.0.dev0)
  • default.qubit.tf (PennyLane-0.16.0.dev0)
  • default.tensor (PennyLane-0.16.0.dev0)
  • default.tensor.tf (PennyLane-0.16.0.dev0)
  • qiskit.aer (PennyLane-qiskit-0.15.0)
  • qiskit.basicaer (PennyLane-qiskit-0.15.0)
  • qiskit.ibmq (PennyLane-qiskit-0.15.0)

Source code and tracebacks

import pennylane as qml
from openfermion.ops import QubitOperator

test_qop1=qml.qchem.convert_observable(QubitOperator("Y0 Y1", 1))
print(test_qop1)

returns

  (1) [Y0 Y1]

while

test_qop2=qml.qchem.convert_observable(QubitOperator("Y0 Y1", 1j))
print(test_qop2)

returns

  (0.0) [Y0 Y1]

[josh146]

Thanks @aemccoy! Just to confirm -- for your current workflow, would you prefer that attempting to convert a Hamiltonian with complex coefficients raises an error? Or, alternatively, would you like a way to represent a (potentially) non-Hermitian linear combination of Pauli terms?

[aemccoy]

The ability to represent a potentially non-Hermitian linear combination of Pauli terms would be ideal. Thanks!