1.8.2

Changes in 1.8.2

• Bundled xcube Viewer 1.4.1 with fixes regarding the share feature.

• No longer logging a TypeError if xcube server's GET viewer/ext/contributions is called without any viewer extensions configured. (#1116)

1.8.1

Resolved an issue where the resample_in_space function no longer worked with irregular grid mappings, such as Sentinel-3 data, due to changes introduced in version 1.8.0. (#1114)

1.8.0

Changes in 1.8.0

Enhancements

• The method xcube.core.GridMapping.transform now supports lazy execution. If
  computations based on actual data are required—such as determining whether the
  grid mapping is regular or estimating the resolution in the x or y direction—only a
  single chunk is accessed whenever possible, ensuring faster performance.

• The function xcube.core.resampling.rectify_dataset now supports xarray.Datasets
  containing multi-dimensional data variables structured as var(..., y_dim, x_dim).
  The two spatial dimensions (y_dim and x_dim) must occupy the last two positions
  in the variable's dimensions.

• Added a new preload API to xcube data stores:

  • Enhanced the xcube.core.store.DataStore class to optionally support
    preloading of datasets via an API represented by the
    new xcube.core.store.DataPreloader interface.
  • Added handy default implementations NullPreloadHandle and ExecutorPreloadHandle
    to be returned by implementations of the prepare_data() method of a
    given data store.

• A xy_res keyword argument was added to the transform() method of
  xcube.core.gridmapping.GridMapping, enabling users to set the grid-mapping
  resolution directly, which speeds up the method by avoiding time-consuming
  spatial resolution estimation. (#1082)

• The behaviour of the function xcube.core.resample.resample_in_space() has
  been changed if no tile_size is specified for the target grid mapping. It now
  defaults to the tile_size of the source grid mapping, improving the
  user-friendliness of resampling and reprojection. (#1082)

• The "https" data store (store = new_data_store("https", ...)) now allows
  for lazily accessing NetCDF files.
  Implementation note: For this to work, the DatasetNetcdfFsDataAccessor
  class has been adjusted. (#1083)

• Added new endpoint /viewer/state to xcube Server that allows for xcube Viewer
  state persistence. (#1088)

  The new viewer API operations are:

  • GET /viewer/state to get a keys of stored states or restore a specific state;
  • PUT /viewer/state to store a state and receive a key for it.

  Persistence is configured using new optional Viewer/Persistence setting:

  Viewer:
   Persistence:
     # Any filesystem. Can also be relative to base_dir.
     Path: memory://states
     # Filesystem-specific storage options   
     # StorageOptions: ...

• The get_data_ids() method in DataStore has an enhanced include_attrs parameter.
  Previously accepting only Container[str], it now also supports a bool value.
  Setting include_attrs to True retrieves all attributes of the data_ids.

• Updated dependency urllib3 to be >=2.0.

• Bundled xcube-viewer 1.4.0

Fixes

• The function xcube.core.resample.resample_in_space() now supports the parameter
  source_ds_subset=True when calling rectify_dataset. This feature enables
  performing the reprojection exclusively on the spatially congruent subset of
  the dataset.
• The function xcube.core.resample.resample_in_space() now always operates
  lazily and therefore supports chunk-wise, parallel processing. (#1082)
• Bug fix in the has_data method of the "https" data store
  (store = new_data_store("https", ...)). (#1084)
• Bug fix in the has_data method of all filesystem-based data store
  ("file", "s3", "https"). data_type can be any of the supported data types,
  e.g. for .tif file, data_type can be either dataset or mldataset. (#1084)
• The explanation of the parameter xy_scale in the method
  xcube.core.gridmapping.GridMapping.scale has been corrected. (#1086)
• The spurious tileserver/viewer warning "no explicit representation of
  timezones available…" (formerly "parsing timezone aware datetimes is
  deprecated…") is no longer generated. (#807)

Other changes

• Added experimental feature that allows for extending the xcube Viewer
  user interface with server-side panels. For this to work, users can now
  configure xcube Server to load one or more Python modules that provide
  xcube.webapi.viewer.contrib.Panel UI-contributions.
  Panel instances provide two decorators layout() and callback()
  which are used to implement the UI and the interaction behaviour,
  respectively. The functionality is provided by the
  Chartlets Python library.
  A working example can be found in examples/serve/panels-demo.

• The xcube test helper module test.s3test has been enhanced to support
  testing the experimental server-side panels described above:

  • added new decorator @s3_test() for individual tests with timeout arg;
  • added new context manager s3_test_server() with timeout arg to be used
    within tests function bodies;
  • S3Test, @s3_test(), and s3_test_server() now restore environment
    variables modified for the Moto S3 test server.

Full Changelog: v1.7.1...v1.8.0

1.7.2.dev1

What's Changed

• Update dataaccess.md by @b-yogesh in #1090
• Allow server-side panel code on S3 by @forman in #1089

Full Changelog: v1.7.2.dev0...v1.7.2.dev1

1.7.2.dev0

What's Changed

• Allow lazy execution of the reprojection method by @konstntokas in #1082
• Allow to open netcdf file from remote HTTPS server by @konstntokas in #1083
• Error in documentation of GridMapping.scale method corrected by @konstntokas in #1087
• Server-side panels by @forman in #1085
• Bug fix in has_data method for "https" data store by @konstntokas in #1084

Full Changelog: v1.7.1...v1.7.2.dev0

1.7.1

Changes in 1.7.1

Enhancements

• Level creation now supports aggregation method mode to aggregate to the value which is most frequent. (#913)

Fixes

• The time query parameter of the /statistics endpoint of xcube server has now been made optional. (#1066)
• The /statistics endpoint now supports datasets using non-WGS84 grid systems, expanding its compatibility with a wider range of geospatial datasets. (#1069)
• Bug fix in resampling_in_space when projecting from geographic to non-geographic projection. (#1073)
• Bug fix of the extent field in the single item collection published by the xcube server STAC API so that it follows the collection STAC specifications. (#1077)

New Contributors

• @b-yogesh made their first contribution in #1070

Full Changelog: v1.7.0...v1.7.1

1.7.1.dev0

What's Changed

• Adjustments to resample_in_time() so that xcube now supports xarray=2024.7 by @konstntokas in #1068

Full Changelog: v1.7.0...v1.7.1.dev0

1.7.0

Changes in 1.7.0

Full Changelog: v1.6.0...v1.7.0

Enhancements

• Bundled xcube-viewer 1.3.0.

• xcube server can now deal with "user-defined" variables. Endpoints
  that accept a {varName} path parameter in their URL path can now be
  called with assignment expressions of the form <var_name>=<var_expr>
  where <var_name> is the name user defined variable and <var_expr>
  is an arbitrary band-math expression,
  see xcube-dev/xcube-viewer#371.

• xcube server now allows for configuring new dataset properties
  GroupTitle and Tags . This feature has been added in order to support
  grouping and filtering of datasets in UIs,
  see xcube-dev/xcube-viewer#385.

• Added server endpoint GET /statistics/{varName} with query parameters
  lon, lat, time which is used to extract single point data.
  This feature has been added in order to support
  xcube-dev/xcube-viewer#404.

• The xcube server STAC API now publishes all fields available via the
  /datasets endpoint. This includes colormap information for each asset such as
  colorBarName, colorBarNorm, colorBarMin, colorBarMax, tileLevelMin, tileLevelMax.
  (#935, #940)

• xcube server now allows for configuring custom color maps via the configuration file.
  It supports continuous, stepwise and categorical colormaps, which may be
  configured as shown in the section CustomColorMaps of the xcube serve documentation
  (#1055)

Fixes

• Migrated the .github/workflows/xcube_build_docker.yaml and the corresponding
  Dockerfile from setup.py to pyproject.toml. Additionally, updated the relevant
  documentation in doc/source to reflect this change from setup.py to
  pyproject.toml. (related to #992)
• Normalisation with xcube.core.normalize.normalize_dataset fails when chunk encoding
  must be updated (#1033)
• The open_data method of xcube's default xcube.core.store.DataStore implementations
  now supports a keyword argument data_type, which determines the
  data type of the return value. Note that opener_id includes the data_type
  at its first position and will override the data_type argument.
  To preserve backward compatibility, the keyword argument data_type
  has not yet been literally specified as open_data() method argument,
  but may be passed as part of **open_params. (#1030)
• The xcube.core.store.DataDescriptor class now supports specifying time ranges
  using both datetime.date and datetime.datetime objects. Previously,
  only datetime.date objects were supported.
• The xcube server STAC API has been adjusted so that the data store
  parameters and data ID, which are needed to open the data referred to by a STAC item,
  are now included with the item's analytic asset.
  Furthermore, a second assert called analytic_multires will be published
  referring to the multi-resolution data format levels (#1020).
• Improved the way color mapping works in xcube server to support simplified
  color bar management in xcube viewer,
  see xcube-dev/xcube-viewer#390. (#1043)
• The xcube server's dataset configuration extraction methodology has been updated.
  When the data resource ID is provided in the Path field, xcube will attempt to
  access the dataset using the given ID. If wildcard patterns are used, the server
  will crawl through the data store to find matching data IDs. This process may
  result in a long setup time if the data store contains numerous data IDs.
  A UserWarning will be issued for the "stac" data store.
• Corrected extent object of a STAC collection issued by xcube server, following the
  collection STAC specifications
  (#1053)
• When opening a GeoTIFF file using a file system data store, the default return value
  is changed from MultiLevelDataset to xr.Dataset, if no data_type is assigned
  in the open_params of the store.open_data() method. (#1054)
  xcube server has been adapted to always open MultiLevelDatasets from
  a specified data store, if that data type is supported.

Other changes

• Renamed internal color mapping types from "node", "bound", "key"
  into "continuous", "stepwise", "categorical".

1.7.0.dev4

What's Changed

• Bug fix of xcube server config colormaps schema by @konstntokas in #1063
• bundled xcube viewer 1.3.0 by @TejasMorbagal in #1064

Full Changelog: v1.7.0.dev3...v1.7.0.dev4

1.7.0.dev3

What's Changed

• configuring custom color maps via the configuration file by @konstntokas in #1055 #1057 #1059

Full Changelog: v1.7.0.dev2...v1.7.0.dev3