466 add feature importance cli

[msorvoja]

Fixes #466

[nmaarnio]

Couple comments, see below.

Besides those, I looked at the toolkit function and a couple things could be changed there as part of this PR:

• Renamex_test-> X and change description, does not need to be testing data. It can be the same exact data that is fed into any Sklearn model
• Rename y_test -> y and change description, does not need to be testing data
• Change n_repeats default to 10? It is 5 in Sklearn, if we want higher maybe 10 is enough
• Add check that feature_names length matches the dimensions of X (X.shape[1] I think?)
• Maybe we can remove the multiplication by 100 for the importance means – I think the range [0,1] where they reside by default is pretty intuitive and used in other phases of MPM / EIS Toolkit in general