Create a gauge for idle moments

[NoureldinYosri]

Adding a gauge compiler that surrounds empty sequences by randomly chosen gates ... this gauge is technically half way between a gauge and dynmical decoupling

[zlatko-minev]

Awesome! Some small docstring changes

[zlatko-minev]

Awesome this looks really nice and useful

[NoureldinYosri]

Awesome this looks really nice and useful

great I will add tests and send for review

[zlatko-minev]

Great, looking forward to having it in the codebase

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.37%. Comparing base (1fa73c7) to head (f3dbbbd).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff            @@
##             main    #7648    +/-   ##
========================================
  Coverage   99.37%   99.37%            
========================================
  Files        1078     1080     +2     
  Lines       96212    96378   +166     
========================================
+ Hits        95609    95775   +166     
  Misses        603      603            

☔ 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.

[zlatko-minev]

LGTM

[pavoljuhas]

I think we can remove the gauges_inverse and compute it from the gauges list instead. Otherwise LGTM after addressing inline comments.

[pavoljuhas]

Nit - here and below in case named gate tuples are created from a sequence.

Suggested change

	if gauges is _PAULIS:
	if gauges == _PAULIS:

[NoureldinYosri]

I prefer to do both

[pavoljuhas]

That makes a little difference for small tuples, but why not. :-)
There was a typo in variable name - fixed in 365e6c6
(in cached_property it is not worth to check identity)

[pavoljuhas]

Here we rely on np.random.default_rng(0).choice(4) returning always the same value at

Cirq/cirq-core/cirq/transformers/gauge_compiling/idle_moments_gauge.py

Line 223 in e8557b5

gate_index = rng.choice(len(self.gauges))

This can break if numpy generator changes in the future.

We could make this more secure by adding a local mock for rng.choice() which would return preset values to get reproducible results. Your pick if it is worthwhile to do now.

[NoureldinYosri]

if a change in the generated cases breaks the test then the code is buggy ... the only reason I'm adding a seed is so that if it breaks we would know exactly which case it is

[NoureldinYosri]

actually you are right, I changed to using a mock

[pavoljuhas]

LGTM after pushing in eb2e865 to avoid overlap with the builtin gc module.

Thanks for adding this!