Combine RB and XEB to compute inferred errors

[NoureldinYosri]

This PR:

• Adds the option to specify the qubits to the XEB workflow
• Adds pauli error computation for 2q XEB
• Adds InferredXEBResult class which
  • computes inferred errors: pauli, XEB, and decay constant
  • has a convenience method to plot heatmap for each of the errors
  • has a convenience method to plot histogram of each of the erros.

[NoureldinYosri]

@eliottrosenberg I added the ability to plot both single and two qubit pauli errors

---

The issue with a single function that runs both 2Q XEB and 1Q RB is the number of arguments which is not the best thing to have. That method will have 10+ arguments. alternatively I could group them into an RBOptions and XEBOptions classes, what do you think?

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Comparison is base (c3de706) 97.82% compared to head (9f7d628) 97.82%.

:exclamation: Current head 9f7d628 differs from pull request most recent head f14bf7e. Consider uploading reports for the commit f14bf7e to get more accurate results

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #6455 +/- ## ======================================== Coverage 97.82% 97.82% ======================================== Files 1115 1115 Lines 97337 97448 +111 ======================================== + Hits 95216 95327 +111 Misses 2121 2121 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[eliottrosenberg]

@eliottrosenberg I added the ability to plot both single and two qubit pauli errors

The issue with a single function that runs both 2Q XEB and 1Q RB is the number of arguments which is not the best thing to have. That method will have 10+ arguments. alternatively I could group them into an RBOptions and XEBOptions classes, what do you think?

I'm thinking that most of the parameters would be shared. It could be something like this:

Shared parameters:

• sampler
• qubits
• repetitions
• num_circuits
• random_state (seed for rng)

RB:

• num_clifford_range (call this depths_rb or something)

XEB:

• entangling gate
• cycle_depths (call this depths_xeb or something)

We can get rid of the use_xy_basis option for RB here, since we will always want that to be its default value, I think. So that's 8 parameters, and all of them will have default values except sampler, so it will be easy to use.

[NoureldinYosri]

@eliottrosenberg sg, I will add that method in a followup PR