Begin smoothing utility integration into JHU. #177
Conversation
dshemetov
force-pushed
the
dev-smoothing-jhu
branch
from
d9b35db to
7afba80
Compare
|
This should be good to go now. Needed fixes from #478, so this should be merged after that. |
dshemetov
force-pushed
the
dev-smoothing-jhu
branch
3 times, most recently
from
1d720ea to
d329ee2
Compare
| @@ -10,11 +11,26 @@ | |||
|
|
|||
|
|
|||
| class TestSmoothers: | |||
| def test_bad_inputs(self): | |||
| with pytest.raises(ValueError): | |||
| Smoother(smoother_name="hamburger") | |||
There was a problem hiding this comment.
I am but carrying on a tradition passed down to me by my advisor and his advisor before him for generations 🤣
| """Predict a single value using the savgol method. | ||
|
|
||
There was a problem hiding this comment.
this should fail linting -- we've got the utils package disabled in CI so you'll need to run it manually for now.
There was a problem hiding this comment.
Just opened #494, if that gets merged first then should probably merge it into this branch and let it run all checks.
Slight possibility when enabling these checks that an existing PR gets merged without the checks merged in and you end up with a failing main, but should be easy to resolve and won't happen once new PRs pull from the new main.
There was a problem hiding this comment.
FWIW, I now got these checks running automatically in my IDE, (thanks @chinandrew 🥇), so that will make things smoother.
| smoothed_signal = smoother.smooth(signal) | ||
| assert np.allclose(smoothed_signal, np.array([np.nan, 1, 1]), equal_nan=True) | ||
|
|
||
| # Test a range of cases where the signal size following a sequence of nans is returned |
jhu/tests/test_smooth.py
Outdated
| dates = [str(x) for x in range(20200303, 20200310)] | ||
| dates = [str(x) for x in range(20200304, 20200311)] |
There was a problem hiding this comment.
Ah, it was from this issue. I wasn't sure which date was the latest, I'll switch it back.
dshemetov
force-pushed
the
dev-smoothing-jhu
branch
from
ef3faeb to
f8af034
Compare
|
@krivard no value changes expected from the indicator. |
* remove local smoothing module * update and lint run.py * a few linter / formatting changes in jhu tests * a few linter / formatting changes in smoother util
dshemetov
force-pushed
the
dev-smoothing-jhu
branch
from
f8af034 to
dfee042
Compare
Starting this work here. Added a local-weighted regression option to the Savitzky-Golay smoother to allow more fine-grained control (needs a little more testing). Continuing to update the documentation IPython notebook.