Add 2q fractional gates to the ConsolidateBlocks transpiler pass

[ShellyGarion]

Summary

With the introduction of new 2q-fractional gates, such as RZZGate as part of the QPU, we add them to the ConsolidateBlocks transpiler pass.
https://www.ibm.com/quantum/blog/fractional-gates

This PR should close #13428

Details and comments

Note that this PR may be eventually superseded by #13419

[qiskit-bot]

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

• @Qiskit/terra-core

[ShellyGarion]

The only test currently failing is:

qiskit/test/python/transpiler/test_consolidate_blocks.py

Line 567 in 682e2d3

def test_no_kak_gates_in_preset_pm(self, opt_level):

This test was added in #13450 to handle issue #13438, and I wonder if we should remove it.
@mtreinish @ElePT - what's your opinion?

the original circuit is:

        ┌────┐   
q_0: ─■─┤ √X ├─■─
      │ ├────┤ │ 
q_1: ─■─┤ √X ├─■─
        └────┘   

the circuit after basis translation is:

global phase: 7π/4
               ┌──────────┐   ┌────┐            ┌──────────┐
q_0: ─■────────┤ Rz(-π/2) ├───┤ √X ├───■────────┤ Rz(-π/2) ├
      │ZZ(π/2) ├──────────┤┌──┴────┴─┐ │ZZ(π/2) ├──────────┤
q_1: ─■────────┤ Rz(-π/2) ├┤ Rx(π/2) ├─■────────┤ Rz(-π/2) ├
               └──────────┘└─────────┘          └──────────┘

The circuit after optimization_level 2 or 3 transpilation is:

global phase: 3π/2
     ┌──────────┐┌───────┐           ┌──────────┐┌────────┐           »
q_0: ┤ Rz(-π/2) ├┤ Rx(π) ├─■─────────┤ Rx(-π/2) ├┤ Rz(-π) ├─■─────────»
     ├─────────┬┘└───────┘ │ZZ(-π/2) ├──────────┤├────────┤ │ZZ(-π/2) »
q_1: ┤ Rz(π/2) ├───────────■─────────┤ Rx(-π/2) ├┤ Rz(-π) ├─■─────────»
     └─────────┘                     └──────────┘└────────┘           »
«     ┌──────────┐
«q_0: ┤ Rz(-π/2) ├
«     ├─────────┬┘
«q_1: ┤ Rz(π/2) ├─
«     └─────────┘ 

The reason that this circuit is being consolidated is that it originally contains CZ gates. If it had contained only RZZ gates, then it would not had been consolidated.

I think that it would be better handled in #13419 (which should take into account the 1-qubit gates).

[ShellyGarion]

Note that due to the TwoQubitControlledUDecomposer synthesis algorithm, the synthesized circuit is:

q_0: ──────────■──────────────────
     ┌───────┐ │ZZ(-0.3) ┌───────┐
q_1: ┤ Rx(π) ├─■─────────┤ Rx(π) ├
     └───────┘           └───────┘

and not just RZZGate(0.3)

[coveralls]

Pull Request Test Coverage Report for Build 13636391692

Warning: This coverage report may be inaccurate.

This pull request's base commit is no longer the HEAD commit of its target branch. This means it includes changes from outside the original pull request, including, potentially, unrelated coverage changes.

• For more information on this, see Tracking coverage changes with pull request builds.
• To avoid this issue with future PRs, see these Recommended CI Configurations.
• For a quick fix, rebase this PR at GitHub. Your next report should be accurate.

Details

• 28 of 28 (100.0%) changed or added relevant lines in 4 files are covered.
• 3509 unchanged lines in 162 files lost coverage.
• Overall coverage decreased (-1.0%) to 87.105%

---

Files with Coverage Reduction	New Missed Lines	%
crates/accelerate/src/basis/basis_translator/basis_search.rs	1	99.31%
crates/qasm2/src/expr.rs	1	94.23%
qiskit/circuit/classical/types/ordering.py	1	98.25%
qiskit/circuit/controlflow/switch_case.py	1	95.35%
qiskit/circuit/library/standard_gates/r.py	1	97.62%
qiskit/circuit/library/standard_gates/rxx.py	1	97.78%
qiskit/circuit/library/standard_gates/rzx.py	1	97.78%
qiskit/circuit/library/standard_gates/u2.py	1	96.88%
qiskit/circuit/random/utils.py	1	94.87%
qiskit/circuit/store.py	1	96.77%

Totals
Change from base Build 13411473274:	-1.0%
Covered Lines:	75855
Relevant Lines:	87085

---

💛 - Coveralls

[ElePT]

Thanks Shelly! The PR looks good, I just left two small comments that should be easy to look into.

[ElePT]

I don't see any public uses of DecomposerType, so the enum can probably be kept private, right? Similarly, I didn't find where is the FromPyObject trait used, but I might have missed it.

[ShellyGarion]

FromPyObject is used to call either the TwoQubitBasisDecomposer or the TwoQubitControlledUDecomposer from python, see this comment:
#13884 (comment)

[ShellyGarion]

removing pub from the enum produces this warning:
warning: type consolidate_blocks::DecomposerType is more private than the item consolidate_blocks

[ElePT]

I remember a discussion we had in unitary synthesis where we agreed not to use pop because of result reproducibility. I think the same logic would apply in this case.

[ShellyGarion]

done in 9095358

[ElePT]

LGTM! thanks for answering my questions.

[mtreinish]

For those worried that this bounced off the merge queue twice, looking at the log those failures were caused by issues with the latest aer release and not issues in this pr itself.