Restrict Split2QUnitaries to run on UnitaryGates only

[Cryoris]

Summary

Fixes the remaining problems with #12970.

Details and comments

If we run Split2QUnitaries with an incomplete 1-q basis gate set, we can end up in the situation where a supported gate is split into 2 UnitaryGate objects that might not be representable in the current basis gate set. For example

from qiskit import QuantumCircuit, transpile
from qiskit.circuit.library import *

gate = PauliGate("XX")

qc = QuantumCircuit(2)
qc.append(gate, [0, 1])

basis_gates = ["x"]  # gate decomposes into XX
 
circuit = transpile(qc, basis_gates=basis_gates, optimization_level=2)  # fails -- cannot map U to the basis

This PR fixes the behavior by only running on UnitaryGate objects -- hence not changing the gate types existing in the circuit.

[qiskit-bot]

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

• @Qiskit/terra-core

[coveralls]

Pull Request Test Coverage Report for Build 10723089932

Details

• 4 of 4 (100.0%) changed or added relevant lines in 1 file are covered.
• 15 unchanged lines in 4 files lost coverage.
• Overall coverage increased (+0.02%) to 89.17%

---

Files with Coverage Reduction	New Missed Lines	%
crates/circuit/src/dag_node.rs	3	81.2%
crates/circuit/src/operations.rs	3	88.5%
crates/qasm2/src/lex.rs	4	91.73%
crates/circuit/src/circuit_instruction.rs	5	85.14%

Totals
Change from base Build 10721003464:	0.02%
Covered Lines:	72657
Relevant Lines:	81481

---

💛 - Coveralls

[mtreinish]

LGTM, I'll make sure to update #13025 to reflect this change too.

This could arguably be considered a behavior change but I think we're in the grey area because the pass was always named Split2QUnitaries (I totally missed the capital Q which is lowercase for everything else) and the intent was always for this to be run after consolidate blocks. Also since this is actively breaking use cases that worked in < 1.2 I would treat this as the lesser change.