In this folder we have three subfolders:
/code: the folder that contains the harness function
/data: the folder that contains all external data that will be read in by harness function
/model: the folder that contains pre-trained CatBoost model
How to run the harness function:
conda create -n CDSML poetry
conda activate CDSML
poetry install
cd code
jupyter notebook
Then put the test data file into the data folder.
Open the harness.ipynb file and set the TEST_FILENAME variable. Run the notebook cells.
There should only be one function inside called harness(df).
The output of harness function should be a n*1 vector of predicted default probability, with n=len(df).