xclim 0.48 compatibility

RondeauG commented 9 months ago

Pull Request Checklist:

[x] This PR addresses an already opened issue (for bug fixes / features)
- No
[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] CHANGES.rst has been updated (with summary of main changes).
- [x] Link to issue (:issue:number) and pull request (:pull:number) has been added.
[x] Maybe... Add a workaround to convert the old xrfreq nomenclature to the new one upon loading an existing catalog or parsing a directory.

What kind of change does this PR introduce?

Changes to be compatible with recent versions of pandas/xarray/xclim

Does this PR introduce a breaking change?

xscen now requires pandas >= 2.2
Functions that output a dict with keys as xrfreq (extract_dataset, compute_indicators) will now return the new nomenclature.
The frequency --> xrfreq/timedelta will now only result in the new nomenclature for hours (lower case h).

Other information:

RondeauG commented 9 months ago

Let's wait for https://github.com/Ouranosinc/xclim/pull/1664 before we finish this PR.

Zeitsperre commented 9 months ago

The problem with Python3.12 is quite silly. Pushing a fix in a few mins.

aulemahal commented 9 months ago

I can push something that ensures the new xrfreq is used when we open catalogs (and parse them).

aulemahal commented 9 months ago

Last commit added a ensure_new_xrfreq function to utils. It is used when opening a catalog and in parse_directory. This means that legacy frequencies either in the paths or in the catalogs shouldn't be a problem.

Running on MRCC5-disponible, our largest catalog, my n=1 test finds that the opening time has increased by 6%. I decided that this was small enough not to require further optimization or skipping options.

Zeitsperre commented 8 months ago

I'll be releasing xclim v0.48.2 today!

Zeitsperre commented 8 months ago

This PR is waiting on the newest patch release from conda-forge. Should be available later today.

RondeauG commented 8 months ago

xclim>=0.48.0 ou xclim>=0.48.2 ?

review-notebook-app[bot] commented 8 months ago

Check out this pull request on

See visual diffs & provide feedback on Jupyter Notebooks.

Powered by ReviewNB

Ouranosinc / xscen