Sorts gate array to have deterministic gate selection

[amruthpremjith]

AI/LLM disclosure

• I didn't use LLM tooling, or only used it privately.
• I used the following tool to help write this PR description:
• I used the following tool to generate or modify code:

Fix #16138

Sort the gate array before selecting from it to return the same gates for the same set of inputs.

Tested with and without changes.

from qiskit.circuit.random.utils import random_clifford_circuit

cliff_circuit = random_clifford_circuit(num_qubits=10, num_gates=100, gates="all", seed=0)
print(cliff_circuit.count_ops())

Without change

(venv) amruthpremjith@MacBookPro qiskit % python3 test-6.py 
OrderedDict({'x': 10, 'h': 10, 'iswap': 9, 'cz': 9, 'z': 9, 's': 8, 'cy': 8, 'sx': 6, 'ecr': 6, 'cx': 6, 'dcx': 4, 'swap': 4, 'sdg': 4, 'y': 3, 'id': 3, 'sxdg': 1})
(venv) amruthpremjith@MacBookPro qiskit % python3 test-6.py
OrderedDict({'h': 10, 'sxdg': 10, 'iswap': 9, 'cz': 9, 'sdg': 9, 'z': 8, 'cy': 8, 'x': 6, 'ecr': 6, 'cx': 6, 'dcx': 4, 'swap': 4, 'sx': 4, 'id': 3, 'y': 3, 's': 1})

With change

(venv) amruthpremjith@MacBookPro qiskit % python3 test-6.py
OrderedDict({'z': 10, 'x': 10, 'ecr': 9, 'cy': 9, 'h': 9, 'sx': 8, 'cz': 8, 'id': 6, 'iswap': 6, 'cx': 6, 'swap': 4, 'dcx': 4, 'sxdg': 4, 's': 3, 'y': 3, 'sdg': 1})
(venv) amruthpremjith@MacBookPro qiskit % python3 test-6.py
OrderedDict({'z': 10, 'x': 10, 'ecr': 9, 'cy': 9, 'h': 9, 'sx': 8, 'cz': 8, 'id': 6, 'iswap': 6, 'cx': 6, 'swap': 4, 'dcx': 4, 'sxdg': 4, 's': 3, 'y': 3, 'sdg': 1})

[qiskit-bot]

Thank you for opening a new pull request.

Before your PR can be merged it will first need to pass continuous integration tests and be reviewed. Sometimes the review process can be slow, so please be patient.

While you're waiting, please feel free to review other open PRs. While only a subset of people are authorized to approve pull requests for merging, everyone is encouraged to review open pull requests. Doing reviews helps reduce the burden on the core team and helps make the project's code better for everyone.

One or more of the following people are relevant to this code:

• @Qiskit/terra-core

[alexanderivrii]

Thanks! That was fast. Do you think there is any meaningful test we could add? Could you also add a (short) bugfix release note?

[amruthpremjith]

Sure, maybe we can add a test case that runs it twice and compares the gate count?

[jakelishman]

The randomisation of set with string keys is fixed by PYTHONHASHSEED, which is only read/set once during initialisation of the Python interpreter, so just doing the same thing twice in the same process would pass without this change. We can test either by comparing a decent-sized circuit to a "blessed" value, or by spawning separate interpreters with different PYTHONHASHSEEDs and ensuring they come back the same.

[jakelishman]

sorted is perhaps a bit unnecessarily heavy as a normalisation step; dict is already consistently ordered (including the dict_keys iterator), so the sort of gates_2q isn't needed, and the top line can be something along the lines of

ignore_1q = {"v", "w", "id", "iden", "sinv"}
gates_1q = [gate for gate in _BASIS_1Q if gate not in ignore_1q]

and avoid the extra $\lg n$. (Not that this would be performance critical, but still.)

[amruthpremjith]

I updated it to the snippet you mentioned, I was thinking it was straightforward to just sort it it to be consistent across any changes in the source dictionary as well.

[alexanderivrii]

Lol, Jake beat me at replying :)

[coveralls]

Coverage Report for CI Build 26029779428

Coverage decreased (-0.002%) to 87.694%

Details

• Coverage decreased (-0.002%) from the base build.
• Patch coverage: 1 of 1 lines across 1 file are fully covered (100%).
• 16 coverage regressions across 5 files.

Uncovered Changes

No uncovered changes found.

Coverage Regressions

16 previously-covered lines in 5 files lost coverage.

File	Lines Losing Coverage	Coverage
crates/qasm2/src/lex.rs	6	91.52%
crates/qasm2/src/parse.rs	6	97.63%
crates/circuit/src/parameter/symbol_expr.rs	2	73.93%
crates/circuit/src/parameter/parameter_expression.rs	1	91.04%
crates/qasm2/src/expr.rs	1	93.82%

---

Coverage Stats

Relevant Lines:	123047
Covered Lines:	107905
Line Coverage:	87.69%
Coverage Strength:	952443.89 hits per line

---

💛 - Coveralls

[amruthpremjith]

I added a test case that passes in different seeds to a subprocess, without the change the test case failed and it was passing with it.

[alexanderivrii]

Thanks @amruthpremjith. I think the fix and the test look ok now (taking the comparing to a "blessed" value approach from Jake's comment, but I would like to have @jakelishman take another look as well.

[alexanderivrii]

Some of the tests fail because we seem to get different blessed values on different platforms. @amruthpremjith, consider changing the test to check that all the results come back the same.

[amruthpremjith]

@alexanderivrii, I've updated the test to first generate the circuit and then use that with varying hashes.

[alexanderivrii]

Thanks! I have retriggered CI, let's see what happens. The fix itself looks good, but I believe formatting might complain, please run black if you have not done so already.

[alexanderivrii]

You should remove this comment now.

[amruthpremjith]

@alexanderivrii, I ran lint and added a bugfix note, could you please check now?

[alexanderivrii]

Thanks for remembering to add the release note. I have left a few minor suggestions.

[alexanderivrii]

You can delete this comment (but pay attention not to delete the initial "---" line).

[alexanderivrii]

I think "deterministic" might be a bit more precise than "consistent", maybe something like:

Suggested change

	Always return consistent clifford circuits while calling ``random_clifford_circuit()`` with same inputs
	Fixed :func:`.random_clifford_circuit` to produce deterministic results across multiple runs given the same ``seed``.

[alexanderivrii]

You need the two trailing underscores for this to render nicely:

Suggested change

	Fixes `#16138 <https://github.com/Qiskit/qiskit/issues/16138>`.
	Fixes `#16138 <https://github.com/Qiskit/qiskit/issues/16138>`__.

Also you want an empty trailing line (see the circled minus sign when you look at the changed files).

[amruthpremjith]

updated the comments, pls check

[alexanderivrii]

Thanks!