Implementation with Hydra hierarchical configuration

Hydra configuration tool is used to feed multiple input combinations to the Sea ice model and generate a structured output directory based on the given input combination. For example: all combinations of user defined input values for salinity, maximum iterations and time step size are sequentially run using hydra which improves the traceability of simulation along with logging files in their respective output directory.

Check out the main_config.py script for the implementation.

Python psuedo-example for main_config.py script

@hydra.main(version_base=None, config_path="example/conf", config_name="config")
def my_app(cfg: DictConfig) -> None:
    """Runs the model using the provided configuration and output directory."""
    with custom_stdout_logger(log):
        out_hydra_dir = hydra.core.hydra_config.HydraConfig.get().runtime.output_dir
        sys.stdout = SpyiceLogger(log)
        log.debug("Debug level message")
        log.warning("Warning level message")
        main = MainProcess(cfg, out_hydra_dir)
        main.run_model()

You can run for this main_config.py script using the following command in the terminal:

Terminal command

python main_config.py

Hydra configuration for jupter notebook

You can also use hydra configuration for jupyter notebook as well. Check out the main_original.ipynb notebook for the implementation.

Python psuedo-example for main_original.ipynb notebook

import os
from pathlib import Path
from hydra import (
    compose,
    initialize,
)
from omegaconf import OmegaConf

# import the main process class
from spyice.main_process import MainProcess

with initialize(version_base=None, config_path="conf"):
    cfg = compose(
        config_name="config.yaml",
        overrides=["iter_max=iter_max1500", "dt=dt47", "S_IC=S34"],
    )
    out_hydra_dir = Path(output_base_dir, "with_hydra")
    main = MainProcess(cfg, out_hydra_dir)
    main.run_model()