Reimplement two_qubit_decompose.num_basis_gates in rust (#11019)

* Use newer rust

* Adding weyl chamber code in rust

* Finish num_basis_gates in rust, but calculated values wrong

* Fix bugs in weyl_coordinates

* Upgrade faer from 0.12 to 0.13

* This will allow copyless convert of numpy complex matrix to faer
* abs2 disappeared probably an error, so we work around here.

* Use copyless conversion of complex matrix from numpy to faer

* Fix thinko bugs so that existing tests pass

* Add comment

* Remove allocation

* Replace loop with iterator

* Silence clippy complaints

* refactor

* Run rustfmt

* Slight refactor and format

* Reorganize new code in crate accelerate

* Moved `eigenvalues` into module `utils`.
* Access num_basis_gates via module two_qubit_decompose

* Remove some allocations

* Run rustfmt

* Raise faer version to 0.13.5, lower rustc to 1.67.0

faer 0.13.5 introduces a specific MSRV

* Fix some conflicts with main branch

* Fix more conflicts

* More conflicts

* More conflicts

* Update crates/accelerate/src/two_qubit_decompose.rs

Co-authored-by: Matthew Treinish <mtreinish@kortar.org>

* Apply suggestions from code review

Apply several suggested edits made in review

Co-authored-by: Matthew Treinish <mtreinish@kortar.org>

* Increase faer version and remove some cruft

* We introduced to_num_complex, but this required upgrading faer. We chose 0.15
* Upgrade our code, in turn, for the upgrade to faer 0.15
* Remove a function exported via pyo3 that was only used for testing
* Remove `def old_num_basis_gate`

* Make __num_basis_gates more idiomatic

* Explain use of `min_by` when `max_by` is expected

* Replace myabs2 with faer_abs2

We recently increased the version of the faer dependency.
The newer version has an abs2 in the api named `faer_abs2`.

So we can delete `myabs2`.

* Use alternative mod implementation following review requests

* Make pyo3 function eigenvalues return ndarray rather than list

Following review suggestion.

* Bump faer version to latest 0.16

This was previously at 0.15 for accidental reasons.

* Use rem_euclid for remainder

* Fix black complaint

* Fix pylint complaints

* Apply suggestions from code review

Co-authored-by: Matthew Treinish <mtreinish@kortar.org>

* Remove nalgebra feature from faer

* Use constants for fractions of pi

---------

Co-authored-by: Matthew Treinish <mtreinish@kortar.org>