Fix quadratic dependency in QuantumCircuit.add_bits

[jakelishman]

Summary

This is in the Python space component only. When add_bits is called many times, it constructs as temporary set of the union of all qubits and clbits present in the circuit in order to calculate an intersection with the input. This causes the method to be linear in the number of bits already present in the circuit, whereas it should be amortised linear in the to be added.

This commit fixes the intersection to be a manual calculation that does not construct three temporary set objects, removing the quadratic cost.

Details and comments

This has been around for ages, but I noticed it when seeing excessive runtimes for parsing OQ2 and OQ3 programs constructed as

qubit q0;
qubit q1;
qubit q2;
// ...
qubit q999999;

Adding jillions of registers is still quadratic in the number of registers because there's no constant-time lookup for them in QuantumCircuit (still), but that's a known problem.

[qiskit-bot]

One or more of the the following people are requested to review this:

• @Qiskit/terra-core

[coveralls]

Pull Request Test Coverage Report for Build 7491854130

Warning: This coverage report may be inaccurate.

We've detected an issue with your CI configuration that might affect the accuracy of this pull request's coverage report.
To ensure accuracy in future PRs, please see these guidelines.
A quick fix for this PR: rebase it; your next report should be accurate.

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage decreased (-0.02%) to 89.278%

---

Totals
Change from base Build 7490243021:	-0.02%
Covered Lines:	59512
Relevant Lines:	66659

---

💛 - Coveralls

[kevinhartman]

LGTM.

I'd be surprised if we don't make the same mistake in other places where we consider the set of all wires in the circuit.