Combine 2q parallel XEB into one methods to simplify the XEB workflow

[NoureldinYosri]

This way the notebook https://quantumai.google/cirq/noise/qcvv/parallel_xeb becomes

>>> res = cirq.experiments.parallel_two_qubit_xeb(sampler)
>>> res.plot_heatmap()

[NoureldinYosri]

@eliottrosenberg this is now ready for review

[codecov]

Codecov Report

Attention: 1 lines in your changes are missing coverage. Please review.

Comparison is base (f2c6f3c) 97.81% compared to head (a91e175) 97.81%.
Report is 3 commits behind head on main.

❗ Current head a91e175 differs from pull request most recent head 64effd8. Consider uploading reports for the commit 64effd8 to get more accurate results

Files	Patch %	Lines
cirq-core/cirq/experiments/two_qubit_xeb.py	98.85%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff            @@
##             main    #6443    +/-   ##
========================================
  Coverage   97.81%   97.81%            
========================================
  Files        1111     1113     +2     
  Lines       97198    97326   +128     
========================================
+ Hits        95078    95204   +126     
- Misses       2120     2122     +2     

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

[eliottrosenberg]

I tested it and got an error:

https://colab.research.google.com/drive/1mPHLGtMnh2l602Pqh9cgh9wpdEnvYHaB?usp=sharing

[NoureldinYosri]

@eliottrosenberg please try again. I ran it several times and didn't run into this issue , though it can happen because of the random nature of operations. I assumed that qubit pairs are always ordered which turns out not to be true. I added a line to ensure results are accessed correctly.

[eliottrosenberg]

LGTM, thanks!

A couple of nits:

• It would be nice to have a helper method that feeds the xeb errors into cirq.integrated_histogram() or returns a dictionary of all the error rates that could then be used as the input for cirq.integrated_histogram.
• The strings that it prints out, "Generate circuit library", "Generate random two qubit combinations", "Compute fidelities", and "Fit exponential decays," aren't really needed. They're useful for debugging, but now that it is working, I don't think they are needed.

[NoureldinYosri]

@eliottrosenberg I added the histogram method and removed the print statements.

[eliottrosenberg]

@eliottrosenberg I added the histogram method and removed the print statements.

Thanks, Nour!

[pavoljuhas]

This does not seem to check if plot functions have any effect - perhaps we can add simple assertions once they return axes?

I am also getting some UserWarning-s in the pytest output such as below which we could avoid -

.../quantumlib/Cirq/cirq-core/cirq/experiments/two_qubit_xeb.py:88: UserWarning: FigureCanvasAgg is non-interactive, and thus cannot be shown
    fig.show()

Let's use @unittest.mock.patch("matplotlib.backend_bases.FigureCanvasBase.show" for the plotting tests which will avoid user warning and it would be also possible to check if the mocked show() got called as expected.

[NoureldinYosri]

this doesn't work. that method doesn't seem to exist. I can't find this being done in other places in Cirq.

[pavoljuhas]

OK, I will take a look later if warnings can be stopped in a separate PR.

[pavoljuhas]

LGTM with a few tweaks for plot functions.