Ouranosinc / xclim / 6563257718

90%

main: 92%

Build Type

push

github

Committed by web-flow

Commit Message

Adapt freq in training (#1407)

<!--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] Tests for the changes have been added (for bug fixes / features)
- [x] (If applicable) Documentation has been added / updated (for bug
fixes / features)
- [x] CHANGES.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?

* Frequency adaptation should sometimes be performed in the same step as
training so that adapted grouped block are preserved for the training
step. This PR allows this by invoking frequency adaptation directly in
training functions.
* nbstripout in pre-commit now removes the metadata.kernelspec
completely, which are annoying differences that come up when editing
notebooks with different software (I think). This might be a problem if
we want to mix different cell codes, e.g. python and R in a single
notebook. If our use case one day evolves towards this, we can revisit
this aspect.

### Does this PR introduce a breaking change?

Yes. The adaptation "in-training" is simply a new feature. But the
adaptation "out-of-training" (or "modular" approach), akin to the old
procedure, will also yield different results. This is because a
multi-dimensional rank is now used in "_processing.py".

### Other information:
I explain the motivation behind this PR (which is also explained below).

Using multi-dimensional rank changes the "out-of-training" adaptation
(which already is "more" in lined with the expected scientific
computation). But these results are "more" in-line with the scientific
computation expected. I say "more" in-line because if you really want
the e... (continued)

Run Details

27 of 31 new or added lines in 3 files covered. (87.1%)

1 existing line in 1 file now uncovered.

8431 of 9328 relevant lines covered (90.38%)

4.36 hits per line