Skip to content

CHANGES.md

Latest commit

 

History

History
95 lines (76 loc) · 4.09 KB

CHANGES.md

File metadata and controls

95 lines (76 loc) · 4.09 KB
Raw

Version 0.3.4 (in development)

  • Remove the dependency on xcube.core.zarrstore by implementing the corresponding GenericZarrStore and GenericArray classes in this package under xcube_smos/zarrstore.
  • Restrict the supported xcube version to <1.14 to maintain compatibility with Zarr v2 until the migration to Zarr v3 is completed.

Version 0.3.3

*Fixed an issue with dataset_paths related to the updated EOData repository bucket name (eodata). Previously, paths were modified to replace eodata with EODATA. This workaround is no longer required, as the bucket name is now consistently eodata.

Version 0.3.2

Version 0.3.1

  • Updated endpoint URL from https://s3.cloudferro.com to https://eodata.cloudferro.com. Necessary due to changes by cloudprovider where the source data is stored.
  • Updated the Jupyter Notebooks in notebooks folder and applied black default code style.
  • After rename of GH organisation of dcs4cop into xcube-dev, updated all references in README.md and docs.
  • Pinned zarr package to 2.x.
  • Removed Anaconda defaults channel from requirements.yml.

Version 0.3.0

  • Added optional geographical bounding box parameter bbox. Using a STAC API to filter source files accordingly. bbox is None by default, which means global coverage. (#20)
  • Added optional parameter res_level in the range from 0 (default) to 4 to let users choose their desired spatial resolution. A parameter spatial_res is not suitable, because resolutions are provided at fixed levels using the DGG. (#21)
  • Added module smos_box.catalog.stac with new SmosStacCatalog that is now the default catalog used by the SMOS data store.
  • Removed module smos_box.catalog.index including SmosIndexCatalog and dropped package smos_box.nckcindex entirely. We are no longer using a Kerchunk reference file index.

Version 0.2.2

  • Fixed problem where package data in xcube_smos/mldataset/smos-dgg.levels did not include any hidden files prefixed with a dot such as .zattrs.

Version 0.2.1

  • Adjusted dependencies in environment.yml to fix the xcube-smos conda build.

Version 0.2.0

  • Added basic documentation. [#24]
  • Introduced open parameter bbox. [#20]
  • Introduced open parameter res_level, an integer value in the range 0 to 4. [#21]
  • The SMOS auxiliary dataset Discrete Global Grid (DGG) has been added as package data in xcube_smos/mldataset/smos-dgg.levels. Hence, the store parameter dgg_urlpath has been removed. [#9]
  • The NetCDF Kerchunk index can now also be a Zip archive. To use a Zip archive, pass an index path to the nckcidx CLI or to the NcKcIndexclass whose filename has a .zip extension. [#8]
  • Storage credentials and other text values in the JSON configuration file of the NetCDF Kerchunk index can now contain template variables of the form "$ENV" or "${ENV}". Such references will be replaced by the corresponding value of the environment variable named ENV. [#6]
  • Source NetCDF files to be indexed can now also be provided in the local filesystem instead of S3 only. [#4]
    • Changed JSON format of the index configuration to allow for any source filesystem. Its version number is now 2.
    • Changed interface of nckcindex CLI tool to allow for local source NetCDF files.
  • CLI nckcidx info has been renamed to nckcidx describe and now supports a --json flag for JSON output.
  • Argument index_path for NcKcIndex.create() and NcKcIndex.open() is now mandatory.

Version 0.1.0

Initial release featuring

  • Access to SMOS data via xcube data store interface.
  • nckcindex CLI tool for management of SMOS Kerchunk indices.