entry_points.md

1. python prepare_data.py

• reads training data from TRAIN_RAW_DATA_DIR (specified in SETTINGS.json)
• runs all preprocessing steps
• saves the prepared data in TRAIN_DATA_AUG_DIR (specified in SETTINGS.json)

2. python train.py

• reads training data from TRAIN_RAW_DATA_PATH and TRAIN_DATA_AUG_DIR (specified in SETTINGS.json)
• trains the models
• saves the trained models to MODEL_DIR (specified in SETTINGS.json)

3. python predict.py

• reads test data from TEST_RAW_DATA_DIR and TRAIN_DATA_AUG_DIR, e.g., mean, std, quantiles per cell_type and sm_name (specified in SETTINGS.json)
• loads the trained models from MODEL_DIR (specified in SETTINGS.json)
• uses the trained models to make predictions on new samples
• saves the predictions to SUBMISSION_DIR (specified in SETTINGS.json)