Roadmap

[leifdenby]

This issue represents the evolving roadmap of the mllam-data-prep package and is updated monthly based on the mllam developer meetings (4th week of the month).

v0.6.0

• add ability to create top-of-atmopshere radiation field from inputs - WIP: Add selected derived forcings #34

v0.5.0

• Fix for typo in example (unecessary f"" in format string) - Remove f-string from 'name_format' in config examples #35
• Replace global config for dataclasses_wizard with specific config - Remove global config for dataclass_wizard #36
• Add "extra" section which is ignored by mllam-data-prep (used for projection info in neural-lam datastores) or add support for projections in mllam-data-prep, move input coordinates to extra subsection #30

v0.4.0

• add optional output path argument to parser. add output option argument #26
• fix bug by making dependency distributed optional Actually make dask.distributed optional #22 #27
• change config example to call validation split val instead of validation Call validation split 'val' instead of 'validation' in example config #28
• fix typo in install dependency distributed fix typo in pyproject.toml #20
• add missing psutil requirement. add missing dependency #21.

v0.3.0

• add command line flags for running through dask-distributed on the local machine

v0.2.0

• replace existing config-yaml files with python dataclasses that are serialised to yaml for storing the config Move to dataclasses based config #8
• add splitting and calculation of statistics Add splitting and calculation of statistics #10
• add inclusion of source variable units and long_name as additional coordinates as {output_feature}_units {output_feature}_long_name include feature long-name and units in output #11

v0.1.0

---

not currently on roadmap

• add support for projecting/interpolating between different grids