Add entrypoint to the pipeline, and default arguments

ECMWFCode4Earth / ml_drought

Machine learning to better predict and understand drought. Moving github.com/ml-clim

https://ml-clim.github.io/drought-prediction/

90 stars 18 forks source link

Closed gabrieltseng closed 5 years ago

gabrieltseng commented 5 years ago

[x] Add a pipeline_config folder where configurations from different pipelines can be run
[x] Add a default.json file of default configurations, which are overriden by user-provided configurations
[x] Add an entry point to the pipeline
[x] Add tests (some added)

tommylees112 commented 5 years ago

Great work Gabi!!

gabrieltseng commented 5 years ago

Potential to dos, in a future PR:

[ ] Replace arparse with python fire, when more arguments are added
[ ] In addition to the default configuration, add a 'minimal' configuration for running simple models without downloading lots of data
[ ] (Maybe) a more thorough check the input is as we expect it, perhaps by comparing to the defaults