Fixes issue #12977

[Ak-ash22]

Summary

Updates the synthesis methods to use the PhaseGate instead of the IBM-legacy U1Gate.

Details and comments

Substituted the CU1/MCU1Gate for the CPhase/MCPhaseGate in the qiskit/synthesis/multi_controlled/mcx_synthesis.py file.

[Cryoris]

Thanks for opening the draft 🙂 The tests are failing because they were expecting the u1 gates (or cu1/mcu1) and now the phase gates are used. You can just update those tests to now expect the phase gates.

Could you include a releasenote which mentions this gate change for certain MCX gates under the upgrades section?

[Ak-ash22]

Hey @Cryoris ... can you help me navigate through these tests that are breaking? For ex:

1. One of the test that breaks is test_decompose_mixture_of_names_and_labels() , and the only possible line of code I could track here is this self.assertEqual(dag.op_nodes()[5].name, "cu1"), am I just supposed to change "cu1" to "cphase"?
2. Also in test_mcx_gate_variants(), you would expect me to change the expected_qasm() line, to expect cphase instead, right?

I just wanted to confirm before I went ahead with these, as this is my second time contributing.

[alexanderivrii]

Hi @Ak-ash22, you are correct: you need to modify the failing tests to refer to CPhase instead of CU1.

For test_decompose_mixture_of_names_and_labels(): the name of the CPhase gate is "cp", so you should change the test to self.assertEqual(dag.op_nodes()[5].name, "cp").

For test_mcx_gate_variants(): you should indeed alter the expected_qasm string.

In case you do not know this already, you can click on "Details" right next to the "Tests/macOS-..." line above, taking you to the "Run tests" log, which shows both how many and which tests have failed, and also the specific error messages (right at the bottom of the log). For test_decompose_mixture_of_names_and_labels() you see

    AssertionError: 'cp' != 'cu1'
- cp
+ cu1

which indeed indicates that the gate name's is "cp".

[Ak-ash22]

Hello @alexanderivrii @Cryoris .. Can you help me solve this particular issue?... I have modified the failed tests to use cp instead of cu1 which was successful for 2 of the tests.
The other 2 tests, where I have to alter the expected_qasm string is not being solved.

The expected_qasm string has cu1 and also mcu1. I tried changing both to cp and mcp which didnt work. I also tried changing only cp to cu1, which is not working as well.

[Cryoris]

These QASM output files can change substantially because it now dumps the decomposition via the MC-Phase path. So we don't only have to change the definition of mcu1 to mcp but also the entire decomposition of the mcp, which is not a native gate in QASM. Since all other tests pass, you can just print the QASM output of the current circuit and update the expected QASM accordingly.

[coveralls]

Pull Request Test Coverage Report for Build 13102387676

Details

• 1 of 1 (100.0%) changed or added relevant line in 1 file are covered.
• 13 unchanged lines in 3 files lost coverage.
• Overall coverage decreased (-0.01%) to 88.816%

---

Files with Coverage Reduction	New Missed Lines	%
crates/accelerate/src/unitary_synthesis.rs	1	92.97%
crates/qasm2/src/lex.rs	6	91.73%
crates/qasm2/src/parse.rs	6	97.61%

Totals
Change from base Build 13072040346:	-0.01%
Covered Lines:	79762
Relevant Lines:	89806

---

💛 - Coveralls

[Ak-ash22]

Hello @alexanderivrii @Cryoris .. It took me a while but I managed to correct the expected_qasm string.
The main tests all seem to be running successfully, but now there is an issue with the lint test, although it rates 10/10 for the code.

The issue seems to be that some lines of expected_qasm strings are too long. Can you help me fix this one last issue?