Ouranosinc / xscen / 37715224835

90%

push

github

blur weights (#784)

<!-- Please ensure the PR fulfills the following requirements! -->
<!-- If this is your first PR, make sure to add your details to the
AUTHORS.rst! -->
### Pull Request Checklist:
- [ ] This PR addresses an already opened issue (for bug fixes /
features)
    - This PR fixes #xyz
- [x] (If applicable) Documentation has been added / updated (for bug
fixes / features).
- [x] (If applicable) Tests have been added.
- [x] This PR does not seem to break the templates.
- [x] CHANGELOG.rst has been updated (with summary of main changes).
- [x] Link to issue (:issue:`number`) and pull request (:pull:`number`)
has been added.

### What kind of change does this PR introduce?

* Adds `blur_weights` which is similar to `creep_weights` but instead of
a creep fill, it performs a blurring (i.e. rolling mean on two or more
axes, i.e. a convolution with a square uniform window of sum 1)
* Renames `creep_fill` to `apply_weights`, the same application function
can be used for both creep and blur.

Doing it this way, is incredibly more performant than a rolling mean in
2D and much more than scipy's ndimage filter. Of course it is limited to
cases where the mask of invalid values is the same for all 2D planes
(i.e. timesteps).

### Does this PR introduce a breaking change?
No.


### Other information:
I'll need this for PC, so I'll merge this in `upd-intake-esm` too.

48 of 55 new or added lines in 1 file covered. (87.27%)

4510 of 5005 relevant lines covered (90.11%)

9.74 hits per line