♻️ compute parallelism feature as described in arXiv:2202.11045

[burgholzer]

As discussed in #895, this PR refactors the implementation of the parallelism feature to align with the description from the original paper https://arxiv.org/abs/2202.11045. The implementation takes extra care to avoid corner cases where divisions by zero could happen.

Fixes #895

[vtomole]

LGTM!

[vtomole]

You'll need to add or modify an already existing test to get through the coverage check.

[stephanielee9]

Hi @burgholzer, do you have any updates on this PR?

[burgholzer]

Hi @burgholzer, do you have any updates on this PR?

Sorry for the long wait here. Somehow the message above got lost in a flood of emails.
Just added the necessary tests and rebased on main.

[vtomole]

@burgholzer Looks like 1 more line needs to be covered by the test.

[burgholzer]

@burgholzer Looks like 1 more line needs to be covered by the test.

Yeah. That's on me for writing a stupid test 😓 (the cirq.Circuit() one also triggers the num_qubits<=1 condition)
I am having a bit of trouble creating a 0-depth circuit with more than 1 qubit using Cirq. Do you happen to know if there is any way to do that?
In Qiskit, I would just do QuantumCircuit(2) and that would be it and it would cover the remaining missing line.

[vtomole]

Create a test in converters_test.py that passes a Qiskit circuit into compute_parallelism_with_qiskit to get that line covered.

[burgholzer]

Create a test in converters_test.py that passes a Qiskit circuit into compute_parallelism_with_qiskit to get that line covered.

Alright. That should do it. Didn't wanna add tests in that file at first as there were basically no related tests in that file.