Speed up doctest of calibrate_denoiser by ~30 s

[lagru]

Description

denoise_tv_chambolle (34.01 s) is a lot slower than denoise_wavelet (2.14 s). denoise_wavelet, was previously replaced because PyWavelets got turned into an optional dependency and handling this in doctests is awkward. So let's revert this change.

denoise_tv_bregman would be another faster option (5.23 s) but I am not sure if using something different from denoise_wavelet actually makes sense from an algorithm standpoint.

Other top slowest tests in our suite (use --duration=X with pytest to get X slowest tests):

========================================== slowest durations ==========================================
34.01s call     restoration/j_invariant.py::skimage.restoration.j_invariant.denoise_invariant
32.39s call     data/tests/test_data.py::test_download_all_with_pooch
16.65s call     registration/tests/test_tvl1.py::test_3d_motion
15.12s call     restoration/tests/test_rolling_ball.py::test_ndim
12.85s call     graph/tests/test_rag.py::test_reproducibility
7.16s call     restoration/_rolling_ball.py::skimage.restoration._rolling_ball.rolling_ball
4.00s call     filters/tests/test_ridges.py::test_3d_cropped_camera_image
3.95s call     registration/_optical_flow.py::skimage.registration._optical_flow.optical_flow_tvl1
3.31s call     registration/tests/test_ilk.py::test_3d_motion[True-True]
3.10s call     registration/tests/test_ilk.py::test_3d_motion[True-False]
2.84s call     measure/tests/test_blur_effect.py::test_blur_effect_3d
2.65s call     segmentation/tests/test_random_walker.py::test_spacing_1
2.60s call     morphology/tests/test_footprints.py::test_nsphere_series_approximation[100-ball]
2.58s call     feature/sift.py::skimage.feature.sift.SIFT
...

Checklist

• A descriptive but concise pull request title
• Docstrings for all functions
• Unit tests
• A gallery example in ./doc/examples for new features
• Contribution guide is followed

Release note

Summarize the introduced changes in the code block below in one or a few sentences. The
summary will be included in the next release notes automatically:

...

[lagru]

Ideally, we could hide this line altogether.

@pllim does doctest-plus support something like that? I don't see an directive at a first glance.

[pllim]

Maybe you want .. doctest-requires:: documented in https://github.com/scientific-python/pytest-doctestplus#doctest-dependencies ?

[lagru]

That seems like a good option and in my book it's an argument in favor of using doctest-plus! Thanks for the quick reply. :)