Fix Lightning fallbacks to support PL 0.29 changes

[mlxd]

Before submitting

Please complete the following checklist when submitting a PR:

• All new features must include a unit test.
  If you've fixed a bug or added code that should be tested, add a test to the
  tests directory!

• All new functions and code must be clearly commented and documented.
  If you do make documentation changes, make sure that the docs build and
  render correctly by running make docs.

• Ensure that the test suite passes, by running make test.

• Add a new entry to the .github/CHANGELOG.md file, summarizing the
  change, and including a link back to the PR.

• Ensure that code is properly formatted by running make format.

When all the above are checked, delete everything above the dashed
line and fill in the pull request template.

---

Context: PR PennyLaneAI/pennylane#3585 updated how wires were handled for SparseHamiltonian types, where situations of device wires being different to operator wires required a reordering. The assumptions made in Lightning were that operator and device wires were always the same and in the same specific order --- this change broke that assumption. As such, when the two diverge Lightning does not apply the correct values to the appropriate locations.

In addition, due to the changes, fallback functionality in Lightning for non-CPP supported observables werre offloaded to _qubit_device with different assumptions made about the bases of said observables. In this case, where _qubit_device required the application of diagonalizing gate, Lightning by default did not, and so the results were inconsistent for said observables. This PR ensure all observable variants for prob, expval and var are offloaded to C++ functions where applicable, and raises an error where functionality support does not currently exist, rather than allowing the return of incorrect results.

Description of the Change: Update wire ordering of sparse matrix defined from Hamiltonian to be reflective of matching device and operator order. Ensure appropriate fallbacks are avoided, instead favouring C++ operations. Ensure error raised when attempting out of order wires for probs.

Benefits: Fixes edge case where the sparse matrix applied is not correct.

Possible Drawbacks:

Related GitHub Issues:

[codecov]

Codecov Report

Merging #416 (675e35e) into master (d6eb549) will decrease coverage by 0.07%.
The diff coverage is 84.21%.

@@            Coverage Diff             @@
##           master     #416      +/-   ##
==========================================
- Coverage   99.88%   99.82%   -0.07%     
==========================================
  Files          49       49              
  Lines        4505     4517      +12     
==========================================
+ Hits         4500     4509       +9     
- Misses          5        8       +3     

Impacted Files	Coverage Δ
pennylane_lightning/lightning_qubit.py	99.07% <76.92%> (-0.93%)	⬇️
pennylane_lightning/_serialize.py	100.00% <100.00%> (ø)
pennylane_lightning/_version.py	100.00% <100.00%> (ø)
pennylane_lightning/src/simulator/Measures.hpp	99.37% <100.00%> (+<0.01%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[AmintorDusko]

Great work, thanks!
I just made some suggestions, but I'm happy to approve.

[timmysilv]

I opened #419 (against this branch) to add more test cases to verify my latest comment. I think they wouldn't hurt to add, but this looks good to me either way. In other words, the tests are there should you want them - take em or leave em.

[vincentmr]

I'm not sure I understand properly what is being compared, so I left a comment about this line

(not np.all(list(wires)[:-1] <= list(wires)[1:]))

Otherwise, looks good! Thanks.