Release v0.2.0

[maxrjones]

What is your issue?

There's been some good progress on xbatcher since the last release over a year ago. @weiji14 mentioned a second release at the Pangeo ML meeting on Monday. I'm opening this issue to suggest a revised target date and track progress on the remaining tasks. Please comment with any items that should be considered for the release.

Proposed release date: October 26, 2022

To-do:

• Use defaults for Black code formatting #103
• Specify build backend and project metadata in pyproject.toml #100
• Let TensorFlow accessor and data loader handle either xarray.DataArray or xarray.Dataset inputs (https://github.com/xarray-contrib/xbatcher/issues/83 part 1)
• Add GitHub workflow for PyPI releases #108
• Lazily generate batches #112

[maxrjones]

Just a note that since the 0.1.0 tag wasn't accompanied by a GitHub release, we'll need to be careful when publishing the release notes to make sure it's associated with a new 0.2.0 tag rather than the existing 0.1.0 tag.

[weiji14]

Since we're overdue on the release already by a week, is it possible to wait for #112 to be merged first 🥺?

[maxrjones]

Since we're overdue on the release already by a week, is it possible to wait for #112 to be merged first 🥺?

Yes, I agree that it would be best to wait for that. In particular since the _gen_batches call was actually moved to __init__ since the last release, so that PR addresses an unreleased performance regression

[dhruvbalwada]

I would really like if #43 can be resolved too.
Would be happy to contribute an example for this, if there is some guidance on how to tackle this.

[maxrjones]

I would really like if #43 can be resolved too. Would be happy to contribute an example for this, if there is some guidance on how to tackle this.

Thanks for your interest in this. I recognize that this is an important component to document and have self-assigned #43 to at least provide guidance on creating an example. That said, given that this is a documentation issue and the default rendering for https://xbatcher.readthedocs.io is off latest rather than stable, I think we could create a v0.2.0 prior to adding the example.

[maxrjones]

Just need to verify that #119 didn't break the long description rendering for PyPI and then I think we're ready for the release. Thanks for your patience!