Guenp/18375 syndrome

[guenp]

Add PseudoSyndrome sample based on @apaetz's version in the private repo. Modified to use a Python script as application layer and CLI script instead of C#.

This is still a Draft PR - one way to enrich this sample would be to introduce random errors and show how this affects the auxiliary qubit. @cgranade @rmshaffer @msoeken I'm curious to hear your thoughts on this, based on how we know these samples are being used and by whom, it may or may not be useful.

[rmshaffer]

I know this PR is still a draft; just a handful of quick comments! Of course I defer to you, @cgranade, and @msoeken on all things samples-related...

[cgranade]

This is still a Draft PR…

Makes sense, I've provided some preliminary feedback keeping that in mind, then. Thanks for opening this early, looks great so far!

…one way to enrich this sample would be to introduce random errors and show how this affects the auxiliary qubit. @cgranade @rmshaffer @msoeken I'm curious to hear your thoughts on this, based on how we know these samples are being used and by whom, it may or may not be useful.

This is one place where an open systems simulator would be really nice to have, as that would allow for demonstrating the sample locally without running on hardware. Similarly, a noise injection simulator that randomly performs operations on the pure-state simulator would also be able to model random overrotations such as those described in your comments.

@bettinaheim may know more about plans from the runtime, but at the moment, all of our simulators simulate noiseless systems such that this sample should always return trivial syndromes when simulated locally. In other samples (e.g. the tomography sample used for Python interop), we've moved that randomness into Q#, but that then has the undesirable side-effect of making Q# programs do something even more noisy when run on noisy hardware.

In general, the discussion of noise in simulation is likely beyond the scope of this sample, such that I think that's a great discussion to have but not one that I want to block your progress here over.

[msoeken]

Looks good to me. Just left some suggestions.

[efratshabtai]

Can this sample run on IonQ or Honeywell?
What will be the result in that case?

[guenp]

Thanks @cgranade @msoeken @efratshabtai @rmshaffer for the thorough code review! The purpose of the draft was really to discuss if we need to extend the sample to include random simulated errors (as to make the Syndrome usage more clear), however it looks like we prefer to leave the sample as-is and revisit after we make a decision on implementing noisy simulators. I'll convert this to a PR now, would appreciate a final pass from one of you.
@cgranade based on your comments it looks like there are some additions to make to the standard libraries. Where do you suggest to follow up on this? would this be a topic for our api review meeting or shall I just create a GH issue?

[guenp]

Can this sample run on IonQ or Honeywell?
What will be the result in that case?

@efratshabtai yes we should be able to run this on hardware (see details in DM). The expected result is a measurable error on the syndrome, depending on the noise in the system.

[cgranade]

I think this is looking pretty close! I've left a few comments, but I think the only other thing I'd suggest is to have a README specific to this sample so that it can be onboarded to docs.microsoft.com/samples. Please let me know if I can help with that. Thank you!

[guenp]

Thanks, @cgranade ! I've implemented your suggestions. I would appreciate a read of the README when you have a moment.

[cgranade]

Looks good, just a few more suggestions on the README. Thanks!