Pass kwargs to interpolate_over_time
#383
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #383 +/- ##
==========================================
+ Coverage 99.79% 99.80% +0.01%
==========================================
Files 14 14
Lines 969 1019 +50
==========================================
+ Hits 967 1017 +50
Misses 2 2 ☔ View full report in Codecov by Sentry. |
niksirbi
changed the title
niksirbi
changed the title
interpolate_over_time
There was a problem hiding this comment.
Thanks @niksirbi for the quick action!
I've briefly discussed this PR with @sfmig and we both agree that the interpolation method is rather important and should remain in the function signature.
I think, for clarity, **kwargs should be the exact **kwargs that xarray.DataArray.interpolate_na(dim=None, method='linear', limit=None, use_coordinate=True, max_gap=None, keep_attrs=None, **kwargs) expects to be passed to the underlying numpy.interp or scipy.interpolate.
Co-authored-by: Chang Huan Lo <[email protected]>
|
|
Thanks for the review @lochhh. I was anyway unsure about the removal of |
Description
What is this PR
Why is this PR needed?
See #379
What does this PR do?
**kwargsto theinterpolate_over_time()function, so that additional keyword arguments can be passed to underlying methods.fill_value="extrapolate"is no longer hard-coded, instead the user may alter this parameter via the aforementioned**kwargs. Henceworth, gaps at the edges of time series will not be filled by default.methodargument frominterpolate_over_time()function signature, as this is passed verbatim to underlying functions.max_gapargument, as its meaning is slightly different fromDataArray.interpolate_na().How has this PR been tested?
The aforementioned changes broke the filtering integration test, because that test was expecting
interpolate_over_time()to eliminate all missing values. I altered that test by explicitly passingfill_value="extrapolate"to it, which made it pass; this additionally acts as an implicit test that kwargs are indeed being forwarded.Is this a breaking change?
No, as previous syntaxes will still work, but it changes the default expected behaviour.
Does this PR require an update to the documentation?
Yes, the
filter_and_interpolate.pyand thesmooth.pyexamples have been updated accordingly. API docs are automatically refreshed due to the changes in the docstring.Checklist: