Describe the bug
When creating a levels dataset from .zarr which contains bnds variables (i.e. lat_bnds, lon_bnds, time_bnds), the resulting .levels only keeps time_bnds as coodrinate variable and converts lat_bnds and lon_bnds into data variables.
To Reproduce
Convert a .zarr cube which contains bnds variables and compare .zattrs from the original dataset as well as the one for e.g. 0.zarr of the levels dataset.
Expected behavior
If not specified differently, the default of https://bcdev.github.io/zappend/config/#attrs_update_mode should be used, which is "keep" and therefor they should be identical.
Python Environment
operating system: linux
zappend version, output of zappend --version: 0.8.0.dev0
optional: packages and their versions, output of pip list or conda list:
Describe the bug When creating a levels dataset from .zarr which contains bnds variables (i.e. lat_bnds, lon_bnds, time_bnds), the resulting .levels only keeps time_bnds as coodrinate variable and converts lat_bnds and lon_bnds into data variables.
To Reproduce Convert a .zarr cube which contains bnds variables and compare .zattrs from the original dataset as well as the one for e.g. 0.zarr of the levels dataset. Expected behavior If not specified differently, the default of https://bcdev.github.io/zappend/config/#attrs_update_mode should be used, which is "keep" and therefor they should be identical.
Python Environment
zappend --version: 0.8.0.dev0pip listorconda list: