implement Auto-ML for shallow classification

[sammlapp]

AutoGluon provides automatic cross-validation, ensembling, and exploration of various classification methods to create an ensembled optimized classifier, using an allotted time budget:

from autogluon.tabular import TabularPredictor
predictor = TabularPredictor(label="class").fit("train.csv", presets="best")
predictions = predictor.predict("test.csv")

would be good to support this in SongSpace add classifier