Prediction path explanation should show comparisons in order of node discovery

[parrt]

Currently the output for the prediction path is given in the feature order in the data frame.

[mepland]

Do we want to only display the features used in the prediction? I think the user can print a dataframe row if they want all of the features, but including them is making the text box large, or adding the ellipsis.

[tlapusan]

@mepland I think that @parrt is referring to this viz

[mepland]

@parrt @tlapusan done, see #234

[parrt]

Regarding showing the entire instance of X, probably best to show at all but I can see a case where it's too big. Sometimes knowing what the other variables are even if they are not tested is useful.

[parrt]

Closing per our discussion #234 (comment)