Intake Take2 is about to release

[martindurant]

Please consider pinning intake<2 to keep this package working, if it breaks with the new code.

[andersy005]

thank you for the heads-up, @martindurant! Looking forward to trying intake v2 it out

[mgrover1]

Thanks for the heads up - and thanks @andersy005 for handling the PR/releases here.

[charles-turner-1]

We've tested upgrading to intake take2 here at ACCESS-NRI and so far haven't discovered any issues.

• All tests on intake-esm passing when upgrading to take2 (branch), diff:

git diff main -- ci/environment.yml

diff --git a/ci/environment.yml b/ci/environment.yml
index 2f0b192..0bcfcd5 100644
--- a/ci/environment.yml
+++ b/ci/environment.yml
@@ -9,7 +9,7 @@ dependencies:
   - fsspec>=2022.11.0
   - gcsfs >=2022.11.0
   - h5netcdf>=0.8.1
-  - intake<2.0
+  - intake>=2.0
   - ipython
   - matplotlib
   - netcdf4>=1.5.5
@@ -26,7 +26,7 @@ dependencies:
   - pytest-mock
   - s3fs >=2022.11.0
   - scipy
-  - xarray>=2022.06
+  - xarray>=2022.06, !=2024.10
   - xarray-datatree
   - zarr>=2.12
   # - pytest-icdiff

NB. xarray 2024.10 seems to be causing some issues - it looks like they did a lot of work related to datatree in this version - I'm looking into it further now.

ImportError: cannot import name 'HybridMappingProxy' from 'xarray.core.utils' 

• All of our tests are passing on both intake<2.0 and intake >=2.0: see Intake Take2 ACCESS-NRI/access-nri-intake-catalog#233 and tox file.

We haven't run any E2E tests yet, so I'm not 100% sure that nothing breaks (in our use case), but as of right now I'm pretty confident that upgrading to intake take2 is unlikely to cause any issues for intake-esm.

[andersy005]

thank you for looking into this, @charles-turner-1! when you get a chance, do you mind opening a PR with this change?