Move code from qml.utils.sparse_hamiltonian to Hamiltonian.sparse_matrix #3236

[12mB7693]

…trix, added a deprecation warning to qml.utils.sparse_hamiltonian

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
  test 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 doc/releases/changelog-dev.md file, summarizing the
  change, and including a link back to the PR.

• The PennyLane source code conforms to
  PEP8 standards.
  We check all of our code against Pylint.
  To lint modified files, simply pip install pylint, and then
  run pylint pennylane/path/to/file.py.

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

---

Context:

Description of the Change:

Benefits:

Possible Drawbacks:

Related GitHub Issues:

[codecov]

Codecov Report

Merging #3585 (be9ce6c) into master (1774fe3) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master    #3585   +/-   ##
=======================================
  Coverage   99.82%   99.82%           
=======================================
  Files         330      330           
  Lines       28943    28949    +6     
=======================================
+ Hits        28892    28898    +6     
  Misses         51       51           

Impacted Files	Coverage Δ
pennylane/optimize/lie_algebra.py	100.00% <ø> (ø)
pennylane/devices/default_qubit.py	99.69% <100.00%> (-0.01%)	⬇️
pennylane/ops/functions/matrix.py	100.00% <100.00%> (ø)
pennylane/ops/qubit/hamiltonian.py	100.00% <100.00%> (ø)
pennylane/utils.py	100.00% <100.00%> (ø)

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

[albi3ro]

Thanks for opening this PR @12mB7693 !

The tests for the original sparse_hamiltonian method in test_utils.py can be copied and repurposed to be added to tests/ops/qubit/test_hamiltonian.py.

We also need a changelog entry under "Deprecations". Don't forget to add your name to the list of contributors as well! The changelog can be found at: https://github.com/PennyLaneAI/pennylane/blob/master/doc/releases/changelog-dev.md

We will also need to add add entry to the doc/developement/deprecations.rst file. Since are now developing for v0.29, let's schedule qml.utils.sparse_hamiltonian for removal in v0.31.

I've left comments with suggestions to fix the various linting and formatting comments. If you want to learn more about running the various tools we use to standardize code formatting, you can read our pull request guide.

We can also do this is a later PR if necessary, but we can now update the three places in PennyLane that use qml.utils.sparse_hamiltonian:

• pennylane/pennylane/devices/default_qubit.py

  Line 590 in 8805172

  Hmat = qml.utils.sparse_hamiltonian(observable, wires=self.wires)

• pennylane/pennylane/ops/functions/matrix.py

  Line 125 in 8805172

  return qml.utils.sparse_hamiltonian(op, wires=wire_order).toarray()

• pennylane/pennylane/optimize/lie_algebra.py

  Line 62 in 8805172

  expm(-1j * t * qml.utils.sparse_hamiltonian(riemannian_gradient).toarray()),

[albi3ro]

Thanks for the updates. This looks in really good shape.

I'd just like to make a comment about handling the wire order at the end with expand_matrix, versus constructing the sparse matrix with the wire order from the start.

Even for a simple case like:

H = 2.0 * qml.PauliX(0) @ qml.PauliY(1) @ qml.PauliZ(2) @ qml.PauliX(3)
%timeit qml.utils.sparse_hamiltonian(H, wires = list(reversed(H.wires)))

You'll see that using that handling the wire order at the end is quite a bit slower. We want to construct to the specific wire order from the very beginning.

So instead of using self.wires inHamiltonian.sparse_matrix, we want to use wires = wire_order or self.wires.

Does that make sense?

[12mB7693]

@albi3ro Thank you for reviewing what I did and for your comments, that helped me a lot.
I think I get what you mean and I will push an update soon.
Shall I actually also update the tests, where the old function qml.utils.sparse_hamiltonian(..) is called?

[albi3ro]

@albi3ro Thank you for reviewing what I did and for your comments, that helped me a lot.
I think I get what you mean and I will push an update soon.
Shall I actually also update the tests, where the old function qml.utils.sparse_hamiltonian(..) is called?

Updating the tests where qml.utils.spase_hamiltonian is called (outside of unit tests) would be great! But I won't necessarily block approval based on that being accomplished.

[12mB7693]

@albi3ro Thank you for reviewing what I did and for your comments, that helped me a lot.
I think I get what you mean and I will push an update soon.
Shall I actually also update the tests, where the old function qml.utils.sparse_hamiltonian(..) is called?

Updating the tests where qml.utils.spase_hamiltonian is called (outside of unit tests) would be great! But I won't necessarily block approval based on that being accomplished.

I'm not sure if I understood you correctly regarding the updates of the tests. Do you mean with "outside of unit tests" that I should not update the unit tests?

[12mB7693]

Not sure why the documentation check fails, I ran "make doc" locally and I get the warning
releases/changelog-dev.md:391: WARNING: Definition list ends without a blank line; unexpected unindent.
However, I do not see how to fix it.

[albi3ro]

Thanks @12mB7693 ! 🎉

Looks good to me. Glad to finally have this tidied up :)

Sphinx may still be causing some problems, but I'll try and fix then if I can.

Excellent work :)

[12mB7693]

No problem, I'm glad that I could help. This task was definitely a good starting point as a new contributor.
Thank you for your support @albi3ro!