Change domain names to standard scheme (additional)

[nvaytet]

In addition to #229 , we rename some more types (see commit messages for details).

I am not so sure about the WavelengthScaledQ -> NormalizedQ rename as some scaling happens for the Denominator but there isn't really a normalization going on for the Numerator...
Maybe we should just use a different name altogether?
I did feel that WavelengthScaledQ was not a good name anyway.

I was also unsure about the BinnedQ name. We go from WavelengthDetector to QDetector (after computing the Q coord but keeping detector dims). Then we bin in Q, so in principle the Detector part goes away, and we are left with just Q as the dim. In other techniques, the data is called Corrected<SOMETHING> at this point, so we still have the Corrected prefix, but we don't have it here.
So I just came up with BinnedQ.

[nvaytet]

I have a flaky test on my local machine:
tests/loki/iofq_test.py::test_pipeline_can_compute_IofQ[False-UncertaintyBroadcastMode.drop].

Full error is

=================================================== FAILURES ====================================================
______________________ test_pipeline_can_compute_IofQ[False-UncertaintyBroadcastMode.drop] ______________________

uncertainties = <UncertaintyBroadcastMode.drop: 1>, qxy = False

    @pytest.mark.parametrize(
        'uncertainties',
        [UncertaintyBroadcastMode.drop, UncertaintyBroadcastMode.upper_bound],
    )
    @pytest.mark.parametrize('qxy', [False, True])
    def test_pipeline_can_compute_IofQ(uncertainties, qxy: bool):
        pipeline = make_workflow(no_masks=False)
        pipeline[UncertaintyBroadcastMode] = uncertainties
        pipeline = sans.with_pixel_mask_filenames(
            pipeline, loki.data.loki_tutorial_mask_filenames()
        )
        pipeline[BeamCenter] = sans.beam_center_from_center_of_mass(pipeline)
        if qxy:
            result = pipeline.compute(BackgroundSubtractedIofQxy)
            assert result.dims == ('Qy', 'Qx')
            assert sc.identical(result.coords['Qx'], pipeline.compute(QxBins))
            assert sc.identical(result.coords['Qy'], pipeline.compute(QyBins))
            assert result.sizes['Qx'] == 90
            assert result.sizes['Qy'] == 77
        else:
            result = pipeline.compute(BackgroundSubtractedIofQ)
            assert result.dims == ('Q',)
            assert sc.identical(result.coords['Q'], pipeline.compute(QBins))
            assert result.sizes['Q'] == 100
        if uncertainties == UncertaintyBroadcastMode.drop:
            test_dir = os.path.dirname(os.path.abspath(__file__))
            name = Path(f'reference_IofQ{"xy" if qxy else ""}_{uncertainties}.hdf5')
            reference = sc.io.load_hdf5(test_dir / name)
>           assert_identical(result, reference)

tests/loki/iofq_test.py:87: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
../../path/scipp/testing/assertions.py:99: in assert_identical
    return _assert_similar(_assert_identical_impl, a, b)
../../path/scipp/testing/assertions.py:104: in _assert_similar
    impl(*args, **kwargs)
../../path/scipp/testing/assertions.py:127: in _assert_identical_impl
    _assert_identical_data(a, b)
../../path/scipp/testing/assertions.py:204: in _assert_identical_data
    _assert_identical_variable_data(a.data, b.data)
../../path/scipp/testing/assertions.py:217: in _assert_identical_variable_data
    _assert_identical_dense_variable_data(a, b)
../../path/scipp/testing/assertions.py:222: in _assert_identical_dense_variable_data
    np.testing.assert_array_equal(
../../../miniforge3/lib/python3.12/site-packages/numpy/_utils/__init__.py:85: in wrapper
    return fun(*args, **kwargs)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

args = (<built-in function eq>, array([4.2163105 , 4.616472  , 4.709668  , 4.601788  , 4.390278  ,
       4.0894866 , 3.74352...8204, 0.2738136 , 0.2516059 ,
       0.26882824, 0.2726485 , 0.28334573, 0.25686255, 0.25695458],
      dtype=float32))
kwds = {'err_msg': 'when comparing values', 'header': 'Arrays are not equal', 'strict': False, 'verbose': True}

    @wraps(func)
    def inner(*args, **kwds):
        with self._recreate_cm():
>           return func(*args, **kwds)
E           AssertionError: 
E           Arrays are not equal
E           when comparing values
E           Mismatched elements: 11 / 100 (11%)
E           Max absolute difference among violations: 4.7683716e-07
E           Max relative difference among violations: 1.1717833e-07
E            ACTUAL: array([4.216311, 4.616472, 4.709668, 4.601788, 4.390278, 4.089487,
E                  3.743527, 3.373496, 3.044424, 2.733558, 2.482377, 2.240014,
E                  2.070589, 1.893589, 1.749089, 1.633279, 1.532102, 1.421638,...
E            DESIRED: array([4.216311, 4.616472, 4.709669, 4.601788, 4.390278, 4.089487,
E                  3.743527, 3.373496, 3.044424, 2.733558, 2.482377, 2.240015,
E                  2.070589, 1.893589, 1.749089, 1.633279, 1.532102, 1.421638,...
E           in values

../../../miniforge3/lib/python3.12/contextlib.py:81: AssertionError

[jl-wynen]

I have a flaky test on my local machine:
tests/loki/iofq_test.py::test_pipeline_can_compute_IofQ[False-UncertaintyBroadcastMode.drop].

Is there randomness in the workflow? Or is this due to multi threading?

[nvaytet]

My guess is multi-threading... The tests passes if I replace identical with allclose.

[jl-wynen]

I am not so sure about the WavelengthScaledQ -> NormalizedQ rename as some scaling happens for the Denominator but there isn't really a normalization going on for the Numerator...
Maybe we should just use a different name altogether?
I did feel that WavelengthScaledQ was not a good name anyway.

I agree. Why do we have the same type for numerator and denominator? It seems odd to use a generic here because the two represent different things conceptually.

I was also unsure about the BinnedQ name. We go from WavelengthDetector to QDetector (after computing the Q coord but keeping detector dims). Then we bin in Q, so in principle the Detector part goes away, and we are left with just Q as the dim. In other techniques, the data is called Corrected at this point, so we still have the Corrected prefix, but we don't have it here.
So I just came up with BinnedQ.

You could carry the Corrected prefix through from CorrectedDetector. But I don't love the idea. Is there a reason this workflow first corrects and then computes wavelength and Q? Other workflows do it the other way around.

[nvaytet]

Is there a reason this workflow first corrects and then computes wavelength and Q? Other workflows do it the other way around.

This is the question I spent a lot of time on during the retreat... but for now I decided I don't want to touch it because I am bound to break something somewhere...

[jl-wynen]

Remove?