Add new API for inserting multiple operations into the DAGCircuit in Rust.

[raynelfss]

Summary

Adds a new struct DAGCircuitConcat (name can be changed 😄) that enables chained addition of instructions both valid and invalid into the DAGCircuit. Also, adapts existing structures to those new changes.

Details and comments

• Add struct that allows continuous addition of instructions into the back of the DAGCircuit by performing edge modifications in the output nodes at the end of the additions.
• The new struct includes the following new methods:
  • push_operation_back which pushes a valid PackedInstruction into the back of the DAGCircuit. It behaves very similarly to push_back.
  • apply_operation_back which inserts a new operation into the back of the DAGCircuit. It has two variants which depend on the ownership of the bit indices (Qubit, Clbit).
  • pack_instruction: private method that does just that, it creates a valid instance of PackedInstruction based on other provided attributes.
• Modified DAGCircuit::extend to use this new API, without any considerable losses.

What about front additions?

This struct currently does not support adding instructions continuously to the front of the DAGCircuit because the methodology for that is a little different. It requires doing much more shifting than adding to the back, therefore I suggest this could be done in a follow-up.

Questions and comments?

Feel free to add these below as a review or comment 😸

[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 11619497041

Details

• 326 of 328 (99.39%) changed or added relevant lines in 3 files are covered.
• 7 unchanged lines in 3 files lost coverage.
• Overall coverage increased (+0.03%) to 88.764%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
crates/circuit/src/dag_circuit.rs	258	260	99.23%

Files with Coverage Reduction	New Missed Lines	%
crates/accelerate/src/two_qubit_decompose.rs	1	92.09%
crates/accelerate/src/unitary_synthesis.rs	1	92.44%
crates/qasm2/src/lex.rs	5	92.73%

Totals
Change from base Build 11617024653:	0.03%
Covered Lines:	76532
Relevant Lines:	86220

---

💛 - Coveralls

[raynelfss]

After rewriting the logic of the BasisTranslator pass to use this new API, there was some minor speedup within the pass:

Change	Before [f3bac5a]	After [c91d45b]	Ratio	Benchmark (Parameter)
-	17.0±0.1ms	15.3±0.2ms	0.9	passes.MultipleBasisPassBenchmarks.time_basis_translator(5, 1024, ['rz', 'x', 'sx', 'cx', 'id'])
-	15.1±0.3ms	13.0±0.2ms	0.86	passes.MultipleBasisPassBenchmarks.time_basis_translator(5, 1024, ['u', 'cx', 'id'])
-	42.8±0.8ms	36.4±0.4ms	0.85	passes.MultipleBasisPassBenchmarks.time_basis_translator(14, 1024, ['rz', 'x', 'sx', 'cx', 'id'])
-	36.7±0.3ms	30.9±0.2ms	0.84	passes.MultipleBasisPassBenchmarks.time_basis_translator(14, 1024, ['u', 'cx', 'id'])
-	51.1±0.4ms	42.9±0.3ms	0.84	passes.MultipleBasisPassBenchmarks.time_basis_translator(20, 1024, ['u', 'cx', 'id'])
-	21.6±0.1ms	17.7±0.1ms	0.82	passes.MultipleBasisPassBenchmarks.time_basis_translator(5, 1024, ['rx', 'ry', 'rz', 'r', 'rxx', 'id'])
-	59.3±0.9ms	48.0±0.2ms	0.81	passes.MultipleBasisPassBenchmarks.time_basis_translator(20, 1024, ['rz', 'x', 'sx', 'cx', 'id'])
-	81.3±0.3ms	64.0±1ms	0.79	passes.MultipleBasisPassBenchmarks.time_basis_translator(20, 1024, ['rx', 'ry', 'rz', 'r', 'rxx', 'id'])
-	59.2±1ms	45.8±0.6ms	0.77	passes.MultipleBasisPassBenchmarks.time_basis_translator(14, 1024, ['rx', 'ry', 'rz', 'r', 'rxx', 'id'])