• 1.Compress_train.ipynb -> Compress the training snapshots (time steps) using PCA.

• 2.Compress_test.ipynb -> Apply the PCA Compression to the test dataset.

• 3.GAN-training.ipynb -> Train a GAN and save the model

• 4.GAN-Prediction.ipynb -> Predict with the GAN

• 5.GAN-DataAssimilation.ipynb -> Assimilate observed data with the GAN

1.Compress_train.ipynb

2.Compress_test.ipynb

3.GAN-training.ipynb (optional - the trained model and scaler are already on the folder *.h5 / *.pkl)

To execute:

From inside the notebooks

Cell->Run All 

From the command line

$ jupyter nbconvert --to notebook --execute <notebookname>.ipynb

4.GAN-Prediction.ipynb

5.GAN-DataAssimilation.ipynb

To install requirements:

 $ conda env create -f environment.yml 
 $ conda activate py3ml
 $ python -m ipykernel install --user --name=python3 (optional)

Finally, start Jupyter:

 $ jupyter notebook