updating pennylane version

[bdoolittle]

Updating Pennylane dependencies and associated code.

[bdoolittle]

@trbromley Do you want to take a quick review before merge? There is nothing breaking in here, but I did need to make some changes to accommodate various dependency updates.

[trbromley]

@trbromley Do you want to take a quick review before merge? There is nothing breaking in here, but I did need to make some changes to accommodate various dependency updates.

No problem, I'll take a look hopefully today 👍

[trbromley]

Ok I might not have capacity to give a full review, but it looks like there were no challenges with working with the latest PL outside of dependency/version/environment issues? I.e., nothing from the latest PL UI caused an issue 👍

[bdoolittle]

Ok I might not have capacity to give a full review, but it looks like there were no challenges with working with the latest PL outside of dependency/version/environment issues? I.e., nothing from the latest PL UI caused an issue 👍

No worries, I'll proceed with the merge. I noted two things from the PennyLane update:

1. The syntax/API for tensorflow optimization using the Keras solver changed. This was easy to resolve because the PennyLane docs showed the correct syntax.
2. I saw a strange behavior in a test that might suggest that PennyLane fixed a bug involving midcircuit measurements. More specifically, in the LOCC (midcircuit measurement) tests, operations got added to the Quantum Tape that weren't there in previous tests. As far as I can tell, the current behavior is correct and the old behavior was wrong. The high-level functionality seems to work in both cases, so I'm not to concerned about digging into what changed here.