Version 0.11 Updates¶

Version 0.11.4¶

Enforced the use of headers_only_clone=False when calling GribMetadata.override(). It was a necessary step to fix issues when writing ArrayFieldcontaining GribMetadata to disk. This is considered a temporary change until the issues with headers_only_clone are sorted out (#555).

Fixed issue when to_numpy() could not be called on a FieldList (#551).
Fixed issue when the list-of-dicts source did not raise an exception when the geography was not correctly specified (#549).
Increased the minimum version of multiurl to 0.3.3

Now, dependencies for GeoTIFF support are not installed when earthkit-data is installed with pip install earthkit-data[all]. This step was necessary to make installation work when GDAL is not available. These dependencies need to be installed separately with pip install earthkit-data[geotiff]. See Installation.

Added new Xarray engine called "earthkit". This is the new default when calling to_xarray(). The "cfgrib" engine is still available and can be used by passing engine="cfgrib" to to_xarray(). For details see:
- Xarray engine: overview
- examples_xr_engine (notebook examples)

No array backend is assigned to a Fieldlist any longer. Removed the array_backend property from FieldList, and the array_backend keyword from from_source(). Data accessing methods like to_array() and data() still accept the array_backend option. Now, each Field in a FieldList can have a different array backend reflecting the actual storage type of the values (#471).

You can still create a SimpleFieldList with a single array backend by using the to_fieldlist() method. For example:
```
# Old way
fields = from_source("file", "my.grib", array_backend="pytorch").to_fieldlist()

# New way
ds = from_source("file", "my.grib").to_fieldlist(array_backend="pytorch")
```
Removed ArrayFieldList. Its functionality is covered by SimpleFieldList (#471).
from_array() and to_fieldlist() now return an SimpleFieldList

Added the s3 source to access AWS S3 buckets (#484). See the notebook examples:
- Retrieving data from S3 buckets
Added support for geotiff files (#503). See the notebook examples:
- Using GeoTIFF data
Added stream support for the file source (#500)
Allowed concatenation of stream sources (#500)
Added SimpleFieldList, which can store a list of arbitrary Fields (#471). See the notebook examples:
- GRIB: using array namespaces
Added clone() and copy() to alter field metadata and values (#493, #496, #522). See the notebook examples:
- /how-tos/legacy/grib_modification.ipynb
Reimplemented and documented the list-of-dicts source, which is now generating a SimpleFieldList and is not bound to GRIB specific metadata (#461, #511). See the notebook examples:
- examples_lod
Added serialisation to GRIB fieldlists and Metadata (#463, #474)
Improved in-memory GRIB field implementation (#492)
Enabled to use forcings without providing a source (#495)
Implemented the repr to ArrayField by (#455)
Added remapping option to Field.metadata() (#488)
Added handle property to ArrayField (#464)
Added the Field.to_xarray(), Field.ls() and Field.describe() methods (#513)
Allowed logging control for mars source (#457)
Added support for “lambert_azimuthal_equal_area” metadata (#452)