WIP: DataTree slice generator function

[weiji14]

Description of proposed changes

Very experimental proof of concept to generate subsets out of a multi-resolution DataTree. Do we call the subsets DataBranch 🤔?

Initial implementation is using a standalone function rather than re-using anything from BatchSchema or BatchGenerator.

TODO:

• Add more unit tests, a lot more
• Refactor to use some of the BatchSchema code from https://github.com/xarray-contrib/xbatcher/blob/v0.3.0/xbatcher/generators.py?
  • Actually thinking it might be nice to separate the slicing from the batching, akin to torchdata's Slicer and Batcher , will open separate issue
• Provide an inline example/doctest
• etc

References (some of the complicated for-loops chip generators I've written over the years 🙈):

• https://github.com/weiji14/deepbedmap/blame/v1.1.0/data_prep.py#L497-L741
• https://github.com/weiji14/s2s2net/blob/4bd176a9a5eb002709a017186e81a8f02898e48b/s2s2net/data_chipper.py#L36-L37

Fixes #93

[codecov]

Codecov Report

Merging #171 (c2f9f01) into main (0f8c619) will increase coverage by 0.19%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main     #171      +/-   ##
==========================================
+ Coverage   96.04%   96.24%   +0.19%     
==========================================
  Files           6        6              
  Lines         329      346      +17     
  Branches       70       74       +4     
==========================================
+ Hits          316      333      +17     
  Misses          8        8              
  Partials        5        5              

Impacted Files	Coverage Δ
xbatcher/generators.py	95.16% <100.00%> (+0.48%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.