transpose wire ordering for state results after transpile

[albi3ro]

Context:

The transpile transform changes the wires of everything after the necessary SWAP gates. Unfortunately, we have classes of measurements in PennyLane that are sensitive to the wire order but do not store the wire order on the measurement process.

dev = qml.device("default.qubit", wires = 4)

coupling_map=[(1, 2), (0, 2), (1, 3)]

@qml.transforms.transpile(coupling_map=coupling_map)
@qml.qnode(dev)
def circuit():
  qml.Hadamard(wires = 0)
  qml.CNOT(wires = [0,1])
  return qml.state()

Description of the Change:

• The transpile transform now takes the device as a keyword argument
• If the device is default mixed and the measurement is StateMP, we convert it to DensityMatrixMP
• If the measurement process does not have wires and is not a StateMP, we update the measurement to give it the device wires

3. If any of the measurements are StateMP, the post-processing function transposes the state result

Benefits:

The output of the qnode will look the same with and without the transpile transform

Possible Drawbacks:

• Transposition of state result could potentially be classically expensive for larger systems

Related GitHub Issues:

[github-actions]

Hello. You may have forgotten to update the changelog!
Please edit doc/releases/changelog-dev.md with:

• A one-to-two sentence description of the change. You may include a small working example for new features.
• A link back to this PR.
• Your name (or GitHub username) in the contributors section.

[albi3ro]

[sc-49570]

[trbromley]

Thanks @albi3ro! Is a device-independent bugfix infeasible?

[albi3ro]

Thanks @albi3ro! Is a device-independent bugfix infeasible?

Probably not impossible, but I don't know how to do it. I'm willing to take suggestions. In my mind, a device independent fix would require deeper architectural changes to how we handle wire orders.

[timmysilv]

i will address the overlapping code in StateMP and simulate.py in a follow-up PR. this looks good!

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (deaf387) 99.64% compared to head (16c48c7) 99.63%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #4793      +/-   ##
==========================================
- Coverage   99.64%   99.63%   -0.02%     
==========================================
  Files         381      381              
  Lines       34336    34115     -221     
==========================================
- Hits        34214    33990     -224     
- Misses        122      125       +3     

Files	Coverage Δ
pennylane/transforms/transpile.py	100.00% <100.00%> (ø)

... and 43 files with indirect coverage changes

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