[Release] 0.14

[jhamman]

#3358 is going to make some fairly major changes to the minimum supported versions of required and optional dependencies. We also have a few bug fixes that have landed since releasing 0.13 that would be good to get out.

From what I can tell, the following pending PRs are close enough to get into this release.

• tests for arrays with units tests for arrays with units #3238
• map_blocks map_blocks #3276
• Rolling minimum dependency versions policy Rolling minimum dependency versions policy #3358
• Remove all OrderedDict's (OrderedDict --> dict, some python3.5 cleanup too #3389)
• Speed up isel and __getitem__ Speed up isel and __getitem__ #3375
• Fix concat bug when concatenating unlabeled dimensions. Fix concat bug when concatenating unlabeled dimensions. #3362
• Add hypothesis test for netCDF4 roundtrip Add hypothesis test for netCDF4 roundtrip #3283
• Fix groupby reduce for dataarray Add groupby.dims & Fix groupby reduce for DataArray #3338
• Need a fix for Changed behavior for replacing coordinates on dataset. #3377

Am I missing anything else that needs to get in?

I think we should aim to wrap this release up soon (this week). I can volunteer to go through the release steps once we're ready.

[dcherian]

Added #3338

[crusaderky]

Nuke away all OrderedDicts! (blocked by #3358)

[dcherian]

This looks like it may have fallen out of the indexing refactor; definitely should be fixed: #3377

[keewis]

I don't know if I will have the time to finish #3238 this week, even if I limit it to only testing DataArray.

[jhamman]

Nuke away all OrderedDicts! (blocked by #3358)

@crusaderky - have you started on this yet?

[crusaderky]

@jhamman no - I should have it ready by the end of the weekend.

[jhamman]

@jhamman no - I should have it ready by the end of the weekend.

I plan to make the release on Friday. I can do this task before then unless you really want to do it 😉.

[shoyer]

Should we also make assign and assign_coords no longer sort their arguments? We can do that now that dictionary are always ordered

[crusaderky]

@jhamman never release to production on a Friday! 🏖
Jokes aside, I'm happy to leave the burden to you

[jhamman]

See #3389 for the PR removing OrderedDicts.

[jhamman]

Need a fix for #3377

See #3393

[jhamman]

I think we're feature complete here. There are three PRs that are close (#3238, #3283, #3285) but probably don't need to hold up a release. I have some time Monday morning to issue the release so we'll go with that. Nice work all!

[dcherian]

#3364 could be useful for the hackathon but definitely not necessary.

Thanks for pushing this along Joe.

[crusaderky]

I think #3238 has somewhat little value until __array_function__ is properly implemented in pint master. @keewis what's your opinion?

The other two only improve the tests so users of the release version don't benefit from them.

I'd say let's close it!

[keewis]

I agree that until __array_function__ is at least in pint master (which may take a while, especially since hgrecco/pint#764 seems to be on hold), having #3238 in a release is not really important.

[jhamman]

Okay all. Making the release now.

[jhamman]

0.14 is on github/pypi/conda-forge/readthedocs now. Announcement emails/tweets are pending. Closing this now.

[dcherian]

🎊

[max-sixty]

Thanks a lot @jhamman !