Noise Scaling for LRE

[purva-thakre]

Fixes #2307

Description

---

License

• I license this contribution under the terms of the GNU GPL, version 3 and grant Unitary Fund the right to provide additional permissions as described in section 7 of the GNU GPL, version 3.

Before opening the PR, please ensure you have completed the following where appropriate.

• I added unit tests for new code.
• I used type hints in function signatures.
• I used Google-style docstrings for functions.
• I updated the documentation where relevant.
• Added myself / the copyright holder to the AUTHORS file

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.34%. Comparing base (b58f029) to head (bfe07bd).
Report is 73 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2347      +/-   ##
==========================================
+ Coverage   98.22%   98.34%   +0.11%     
==========================================
  Files          87       89       +2     
  Lines        4056     4105      +49     
==========================================
+ Hits         3984     4037      +53     
+ Misses         72       68       -4     

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

[purva-thakre]

@cosenal @vprusso This is ready for a review up to using layerwise folding on cirq circuits.

I am working on figuring out why layerwise folding on a pyquil circuit raises errors. Once this is fixed, laywerise folding for non-cirq circuits will be added.

[vprusso]

Side note: Are we still supporting versions of Python that wouldn't allow -> list[Circuit] vs. List[Circuit]?

[purva-thakre]

This is a good question. Which versions of python did you have in mind? I know there's an open issue to drop 3.9 once we add support for 3.12.

The validate workflow only runs for python 3.11 which is used for type checking by mypy.

mitiq/.github/workflows/build.yml

Lines 25 to 32 in 837ea83

	python-version: "3.11"
	- name: Install Python dependencies
	run: |
	python -m pip install --upgrade pip
	make install requirements
	- name: Check types with mypy
	run: make check-types

From what I could find, 3.9 uses list[Circuit] as a type alias and List[Circuit] for type annotations. Same with 3.10.

[cosenal]

TIL that type aliases List and Dict are deprecated starting from Python 3.9 https://docs.python.org/3/library/typing.html#deprecated-aliases

[purva-thakre]

I think it's better to limit this PR to Cirq circuits only because I need to spend some more time understanding how the Mitiq interface decorators work. I have used the default args for the decorators in the past without any issues. This PR requires using the optional arguments.

Two different ways to implement the optional args led to errors. Either the one_to_many argument was ignored (first code block) or _pop_measurement was used on a Qiskit circuit instead of converting the circuit to a cirq circuit first (second code block).

@accept_qprogram_and_validate
def multivariate_layer_scaling( input_circuit, degree, ............, one_to_many = True)

Similar to how the optional arg is used in PEC, I also tried changing multivariate_layer_scaling to _cirq_multivariate_layer_scaling such that the new multivariate_layer_scaling only accepted a QPROGRAM as input.

 noise_scaling_func = accept_qprogram_and_validate(
      _cirq_multivariate_layer_scaling,
      one_to_many=True,
  )
  scaled_circuits = noise_scaling_func(
      input_circuit, degree, fold_multiplier, num_chunks, folding_method
  )
 return scaled_circuits

[jordandsullivan]

Sounds good Purva, we can limit this PR for Cirq for now.

[purva-thakre]

Except for 1 unresolved comment #2347 (comment), this PR is ready for another review.

[cosenal]

TIL that type aliases List and Dict are deprecated starting from Python 3.9 https://docs.python.org/3/library/typing.html#deprecated-aliases

[cosenal]

This can raise an UnfoldableCircuitError, it should be documented in the docstring of this caller function as well.

[purva-thakre]

I had included it previously. #2347 (comment)

We should agree on style guidelines for this project and update the docs style guidelines section on docstrings.

Couple of things that confused me:

• Since I am relying on private functions, a user won't be able to read the source of the error in the API. Which is why I felt it's better to document the raised errors in the public function.
• Referring to Vincent's comment on raising errors from a function in a different module, the API-doc page only mentions UnfoldableCircuitError once as a defined exception. Since _check_foldable is a private function, the error raised by the function does not show up publicly.
• Most of the raised errors in the public-facing function are specified in a paragraph.
  
  As we rely on google-style docstrings, there is no convention to allow lists in the description of the different sections of a docstring. Portions of our API-doc have run-on sentences because someone used a list in the section description which appear as badly formatted sentences publicly.

Intended	in API-doc

[purva-thakre]

This PR is ready to be merged once all checks pass.

As of 1:45 pm CST, not all checks are being run after I pushed the latest commit to this PR.

Looks like Github is aware of the issues and they are working on fixing it. https://www.githubstatus.com/

[cosenal]

LGTM

To re-run all checks, you will have to push another commit (even an empty one). I re-run it manually here and everything looks fine.

[cosenal]

LGTM

[vprusso]

(Minor comments only--but aside from those, LGTM!)