Qiskit / qiskit / 28781210717

88%

Build Type

push

github

Committed by web-flow

Commit Message

fixes #11393 sampled_expectation_value returning incorrect results for complex coefficients (gave 0.0 rather than 1j) (#16443)

* Fix sampled_expectation_value dropping imaginary part (#11393)

    sampled_expval_complex and sampled_expval_sparse_observable computed
    the full Complex64 result but returned only Ok(.re). Complex coefficients
    (e.g. SparsePauliOp with -1j) produced 0.0 instead of 1j.

    - crates/accelerate: return Complex64, not f64 (2 functions)
    - sampled_expval.py: return float | complex, updated docstring
    - bit_array.py: dtype=complex128, removed misleading note
    - test: 5 regression tests across both modules
    - releasenotes: release note

    Co-authored-by: zeb33n <zeb33n@users.noreply.github.com>
    Fixes #11393

* added np.real_if_close so real dont come back as complex. Added test and ammended one test verify real comes back real, and complex comes back complex.

* Apply suggestion from @Cryoris

changes to complex or float here, instead of waiting for the return

Co-authored-by: Julien Gacon <gaconju@gmail.com>

* Apply suggestion from @Cryoris

shorten release notes

Co-authored-by: Julien Gacon <gaconju@gmail.com>

* Update qiskit/result/sampled_expval.py

complex->complex-valued for clarity

Co-authored-by: Julien Gacon <gaconju@gmail.com>

* Clarify fix for sampled expectation value functions

Updated the fix description for sampled expectation value functions to clarify the change in return types.

* Modify expectation_values return type to float64

Update expectation_values method to always return float64 dtype.
internal complex because gate blocks it anyway in ObservablesArray.coerce(observables)

* revert bit_array.py changes — allow_non_hermitian is a separate feature PR

* Simplify and reduce the tests

That floats/reals are returned is already
sufficiently tested in the existing tests,
we can reduce to only checking the
previously faulty behavior.

---------

Co-authored-by: Julien Gacon <gac... (continued)

Coverage Stats

6 of 6 new or added lines in 2 files covered. (100.0%)

6 existing lines in 1 file now uncovered.

112875 of 128783 relevant lines covered (87.65%)

999064.81 hits per line