Update Permute operation tests

[antalszava]

Context

There are test cases for the Permute operation that involves creating a QNode and checking that the underlying tape is expanded. Due to the tape expansion, the Permute operation is decomposed into SWAP operations.

These test cases call the QuantumTape.expand method explicitly. However, as the test cases use a QNode, and default.qubit, the device at hand doesn't support the Permute operation, tape expansion happens internal. Therefore, by the time the second call to the QuantumTape.expand method is placed, the tape has already been expanded.

This second call at the moment leaves the tape as is. However, defining a decomposition for qml.SWAP (as is done in #1317) would result in another depth of tape expansion and in causing errors to the current test cases which assume the tape contains qml.SWAP operations.

Changes

Updates the tests creating a QNode to not call QuantumTape.expand.

[github-actions]

Hello. You may have forgotten to update the changelog!
Please edit .github/CHANGELOG.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.

[codecov]

Codecov Report

Merging #1318 (f65252d) into master (acec10f) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #1318   +/-   ##
=======================================
  Coverage   98.16%   98.16%           
=======================================
  Files         148      148           
  Lines       11446    11446           
=======================================
  Hits        11236    11236           
  Misses        210      210           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update acec10f...f65252d. Read the comment docs.

[josh146]

Thanks @antalszava for both finding this and fixing this!

Does it make sense for Permute to be a native default.qubit operation in the future? If it is more efficient permuting wires all at once rather than using many swaps, that might potentially this might speed up simulations which use it.

[josh146]

I'm not so sure this belongs in 'improvements', as it is a bugfix to the tests suite (and not an improvement in PL directly). Perhaps worth moving to Bug fix, and adding a bit more detail here?

[antalszava]

Oh, good point! Updated

[antalszava]

Thank you for the comments @josh146! Created an issue for native support of Permute in default.qubit and updated the changelog.