src.constants regroups most of the program's global variables (the remaining are in preprocessing to regroup all ROOT commands in the same file)

src.preprocessing generates .txt saves of the datasets with control over the features to retrieve / compute / remove.

src.trainer trains the specified classifier and stores it in an appropriate folder along with its predictions.

src.plotter takes a trained model, a test set, and generates the associated category content plot.

In progress : SelfThresholdingAdaClassifier provides a wrapper of sklearn's adaboost meta-estimator designed to provide more control over its prediction method (especially optimize some prediction thresholds to minimize a given cost).