Initialise a model experiment instance

Download cGENIE output from cluster

The first step starts from getting data from the cluster. If it is in tar.gz format then uncompress it. Otherwise, directly download as a folder.

$ scp -r remote:cgenie_output/experiment_id ~/Downloads

For test, an example output is already attached with the package under directory cgeniepy/src/data/muffin.CBE.worlg4.BASESFeTDTL.SPIN. This is an modern configuration with ecosystem component enabled.

Initialise a model instance

The next step is to initialise a model experiment instance. Two parameters are needed: (1) the directory path (e.g., muffin.CBE.worlg4.BASESFeTDTL.SPIN); (2) the target sub-model (e.g., biogem or ecogem). The default of sub-model is biogem, but if you want to access more data, you can create of gemflag.

from cgeniepy.model import GenieModel

model = GenieModel('/Users/XX/Downloads/path_a', gemflag=['biogem', 'ecogem'])

The model also support multiple experiment directory. The only change is to use a list of directories as input. This list can be create by yourself, or use Python’s Pathlib module.

# Model ensemble
model_dir = ['/path_a/', '/path_b/']

models = GenieModel(model_dir)

## Alternatively
from pathlib import Path
model_dir = Path("XX/model")
models = GenieModel(model_path)