Earthdata-varinfo

2023-05-10

This patch update to sds-varinfo makes two changes:

* The CF-Convention attribute names of `add_offset` and `scale_factor` are now
included in the metadata attribute names that will be checked for offset and
scale values, respectively.
* The CF-Convention attribute name of `title` is included in the metadata
attribute names that will be checked when retrieving a UMM-Var Definition.
* The output of `get_all_umm_var` is now a dictionary, with keys corresponding
to the full paths of the variables. These full paths retain their leading
slashes, to allow for easier cross correlation with other dictionary objects
generated by `sds-varinfo`.
* Trailing colons are now stripped from variable references. This is required
for some `grid_mapping` attributes that are of the CF-Convention-compatible
format `crs: dim_1 crs: dim_2` (including some ICESat-2 collections).

2023-05-02

This minor version updates adds functionality to generate UMM-Var records from
the variables contained in a VarInfo instance.

This version also a minor bug fix in the VariableFromNetCDF4 class where
attributes with a value of 1.0 were being returned with a value of True.

UMM-Var attributes currently mapped from metadata attributes:

* Name (required)
* LongName (required)
* Definition (required)
* MetadataSpecification (required) - currently v1.8.2.
* StandardName
* DataType
* Dimensions (see below for more information)
* Units
* FillValues
* Scale
* Offset
* ValidRanges

Dimension types currently supported are `TIME_DIMENSION`, `LATITUDE_DIMENSION`,
`LONGITUDE_TIME` and `OTHER`. Future improvements to sds-varinfo will add
heuristics to identify depth dimensions. In addition, the size of each
dimension can currently only be determined for VariableFromNetCDF4 instances.
Variables parsed from DMR files do not currently contain the required
information for dimension size, and are set to a size of "Varies".

Additional functionality will be required for VariableType and VariableSubType.

In support of UMM-Var JSON output, functionality has also been added to export
either a single record or a list of records to JSON files on-disk. These
functions are `export_umm_var_to_json` and `export_all_umm_var_to_json`,
respectively.

2023-03-28

This major version update does the following:
* Consolidates all variable objects in the VarInfo family of classes into a
single dictionary. This will allow for variables to be identified as more
than simply a metadata variable for a variable with coordinates.
* Converts the class attribute `VarInfoBase.variables_with_coordinates`
into a function that retrieve the same variables.
* Removes the `logger` positional argument to the VarInfo classes, as this
logger has never been used.

2023-03-07

This minor version update adds `VarInfoBase.group_variables_by_dimensions` and
`VarInfoBase.group_variables_by_horizontal_dimensions`, which enable a user to
determine which variables share common dimensions. The
`VarInfoBase.get_all_variables` method has also been added, which will return
all variables, regardless of if they are considered science or metadata.

2024-10-18

Changed:

* CMR native IDs generated from variable names with spaces in them will replace
those space characters with underscores to avoid errors when trying to ingest
such variables.
* The `python-cmr` requirement is updated to v0.12.0, which adds type hints for
that package. A couple of type hints in `earthdata-varinfo` have been updated
accordingly.
* The `numpy` requirement has been relaxed to allow broader compatibility with
client software. Note: This enables Python 3.12 compatibility.
* To ensure compatibility with Python 3.12, the tests are now run using
`pytest`. This allows JUnit style output to be produced, as the previously
used `unittets-xml-runner` package was not compatible with Python 3.12. The
tests themselves are still written using classes and syntax from `unittest`.
The CI/CD for running the tests has been updated to also run the tests under
Python 3.12.

2022-10-05

This major version update changes `sds-varinfo` to use JSON for the
configuration file format in place of YAML. Clients can convert their YAML
configuration files to JSON using the following snippet:

Python
import json
import yaml

with open('path/to/config.yml', 'r', encoding='utf-8') as file_handler:
config_content = yaml.load(file_handler, Loader=yaml.FullLoader)

with open('path/to/config.json', 'w', encoding='utf-8') as file_handler:
json.dump(config_content, file_handler, indent=2)


This version also removes redundant `Applicability` section for GEDI L2
variables from the sample configuration file.