Originally, preprocess downloads raw data from ftp and generates ml data. If we decide to put automatic download into train.py, then it should download the preprocess-generated ml data from ftp. Alternatively, we can add a conditional if statement in train.py that checks whether ml data exists. If it does not, then it calls preprocess.py which generates ml data.
https://github.com/JDACS4C-IMPROVE/Paccmann_MCA/issues/1#issuecomment-1710667737
From @adpartin:
Originally, preprocess downloads raw data from ftp and generates ml data. If we decide to put automatic download into train.py, then it should download the preprocess-generated ml data from ftp. Alternatively, we can add a conditional if statement in train.py that checks whether ml data exists. If it does not, then it calls preprocess.py which generates ml data.