Easier Method to Add, Update or Delete Features

[atheyers]

It would be nice to be able to loop through the list of subject IDs and make updates to the array of features stored under there, instead of having to create a new datastructure for small changes.

[raamana]

thanks! Looping wouldn't work, as we need to ensure all the samples have same dimensionality, but updating all samples at once can work.

For example, for an MLDataset of tow samples id1 and id2, it can be something like this:

ds.add_feature(name='first', dict(id1=1, id2=0))
ds.add_feature(name='second', dict(id1=2, id2=3))
ds.del_feature(name='first')
ds.update_feature(name='second', dict(id1=5, id2=9))