Add graphical outputs for KNN-classifier (and combination scores?)

[aathbt]

In knn_classifier.py, it could be interesting to plot the results of hyperparameters tuning and also to show how the model is performing on the test dataset in terms of accuracy.

• Use several lines plots on the same graph for the 1st part and confusion matrix or classification report for the second?

In check_all_combinations.py, it could be interesting to plot the variation of the number of phenotypes and undefined cells across parameters -ms/--min-samplesxbatch and -mc/--min-cellxsample. This would allow users to adapt the constraints depending on how many phenotypes/undefined cells they want/tolerate.

• Use heat map graph?