Fix classical control diagram symbols and svg

[daxfohl]

Fixes #5688. Fixes #5689.

A quick image search for circuit diagrams shows that the convention seems to be arrows for measurement, which I represent here in diagrams with V. I followed the suggestion from #5688 to use the quantum control symbol @ to represent classical control as well.

For SVG, now the circuit from the teleportation example renders in a more standard fashion, with double-lines for classical registers and measurements, and arrows representing classical data flow:

(previously looked like this, which is obviously misleading):

I avoided the suggestion from #5688 to represent measurement as X in diagrams. The rationale is that X implies a bit flip. However in Cirq, multiple measurements to the same register will be stored independently in an array. So, representing those as X, which implies a bit flip, would be a misrepresentation.

I'll leave it open to discussion on whether this is correct and/or better. It's a three-character change in the source and then a global find-and-replace to patch up the tests and docs, so it's easy to accommodate whatever y'all decide here.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.65%. Comparing base (02941bf) to head (eaa0367).
Report is 67 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #7233   +/-   ##
=======================================
  Coverage   98.65%   98.65%           
=======================================
  Files        1106     1106           
  Lines       95869    95882   +13     
=======================================
+ Hits        94581    94594   +13     
  Misses       1288     1288           

☔ 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.

[mhucka]

I tagged this to be discussed at the next Cirq Cynq.

A change like this strikes me as a breaking change. I see in the comments on #5688 this concerns was brought up a couple of years ago, but not resolved at the time.

[daxfohl]

Yeah I understand the concern. If it's decided that this is breaking, we should also probably revisit other issues involving presentation of circuit diagrams (e.g. #1245, #3597, #3612, #3913, etc), SVG (#5689), 3D view, Bloch sphere, and so on, and close anything else that would be considered breaking.

[mhucka]

Discussed in Cirq Cynq 2025-05-14: it's not clear if this is a breaking change. In general, the feeling is that improving the diagrams and making them more standard and readable would be a good thing. (A breaking change would be more of a concern for function signatures, data output, etc. Diagrams are more often used for human consumption.)

[daxfohl]

An old comment from a previous TL states that diagrams are not intended to be used as inputs: #3597 (comment). Also the intermediate data structure "CircuitDiagramInfo" isn't used for anything but generating diagrams, so the change is unlikely to result in any production breakages. So, obviously biased, but I don't think I'd call it a breaking change.

One remaining concern is if other teams have unit test cases that check diagram output of their circuit. Ideally test cases shouldn't use them that way, but sometimes it's convenient and/or more didactic to put the circuit diagram in the test case rather than a list of operations. I know we do that a fair amount internally, but I have a harder time imagining consumers doing the same. Also this only affects classical controls, which is a fairly niche feature, and the fix can be done in two global find-and-replaces: s/=@=/=V= and s/=^=/=@= in any test files that have circuit diagrams.