0.6.2.dev0

Changes in 0.6.2 (in development)

• The S3 data store xcube.core.store.stores.s3.S3DataStore now implements the describe_data() method.
  It therefore can also be used as a data store from which data is queried and read.

• The xcube gen2 data cube generator tool has been hidden from
  the set of "official" xcube tools. It is considered as an internal tool
  that is subject to change at any time until its interface has stabilized.
  Please refer to xcube gen2 --help for more information.

• Added coords property to DatasetDescriptor class.
  The data_vars property of the DatasetDescriptor class is now a dictionary.

• Removed function reproject_crs_to_wgs84() and tests (#375) because

  • it seemed to be no longer be working with GDAL 3.1+;
  • there was no direct use in xcube itself;
  • xcube plans to get rid of GDAL dependencies.

• CLI tool xcube gen2 may now also ingest non-cube datasets.

• Fixed unit tests broken by accident. (#396)

• Added new context manager xcube.util.observe_dask_progress() that can be used
  to observe tasks that known to be dominated by Dask computations:

  with observe_dask_progress('Writing dataset', 100):
      dataset.to_zarr(store)  

• The xcube normalisation process, which ensures that a dataset meets the requirements
  of a cube, internally requested a lot of data, causing the process to be slow and
  expensive in terms of memory consumption. This problem was resolved by avoiding to
  read in these large amounts of data. (#392)