Create a new implementation of parallel XEB and its methods

[NoureldinYosri]

This is a new implemenation of parallel XEB that doesn't rely on legacy code. This module reimplements almost all of the what is needed to run parallel XEB. The reason for a completly new implementation is that the old implementation (also written by me) relied on old code that don't support the new use case of running XEB on circuits that depenend on the qubit pair and the effort needed to update the old code is more than the effort needed to write a new implementation.

>> import cirq
>> import cirq.experiments.benchmarking.parallel_xeb as xeb
>> sampler = cirq.DensityMatrixSimulator(
    seed=0, noise=cirq.ConstantQubitNoiseModel(cirq.depolarize(1e-2))
)
>> qubits = cirq.q(0, 0), cirq.q(0, 1), cirq.q(0, 2)
>> parameters = xeb.XEBParameters(
    n_repetitions = 10**4,
    n_combinations = 20,
    n_circuits = 30,
    cycle_depths = (5, 25, 50, 100, 200, 300),
)
>> circ = cirq.Circuit(cirq.X(cirq.q(0)), cirq.Y(cirq.q(1)), cirq.CZ(cirq.q(1), cirq.q(0)))
>> xeb.parallel_two_qubit_xeb(sampler, qubits=qubits, target={(cirq.q(0, 0), cirq.q(0, 1)): circ})
TwoQubitXEBResult(fidelities=                           a  layer_fid                cycle_depths  \
pair                                                                  
(q(0, 0), q(0, 1))  0.948672   0.958812  [5, 25, 50, 100, 200, 300]   

                                                           fidelities  \
pair                                                                    
(q(0, 0), q(0, 1))  [0.7693233439521296, 0.3291275172915138, 0.118...   

                       a_std  layer_fid_std  
pair                                         
(q(0, 0), q(0, 1))  0.007211       0.000565  )

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.67%. Comparing base (8c5753d) to head (b24f317).
Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff            @@
##             main    #6940    +/-   ##
========================================
  Coverage   98.66%   98.67%            
========================================
  Files        1106     1109     +3     
  Lines       96448    96826   +378     
========================================
+ Hits        95165    95542   +377     
- Misses       1283     1284     +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[eliottrosenberg]

Actually, in order to do parallel XEB, we will want to be able to input something like a set of circuits (that each act on two qubits) or a dictionary from pairs to circuits because, for example, we may have different z-phase corrections on different pairs of qubits.

[NoureldinYosri]

I reimplemented parallel XEB to support this use case. ptal

[eliottrosenberg]

Thanks, Nour! I will take a look soon.

[eliottrosenberg]

I tried using it and got an error (colab). Also, maybe we can call the folder benchmarking instead of error_mitigation.

[eliottrosenberg]

I tried it again inputting a dictionary and am still getting an error when I do result.pauli_error() (colab)

[pavoljuhas]

LGTM. Please fix type annotations for the new return value.