showWaterfall should accept the output of explainPredictions [feature request]

[areeves87]

Each call to showWaterfall runs explainPredictions. For datasets with hundreds or thousands of columns, this is tedious. Runtime can be hours long. To expedite the function call, showWaterfall should also accept an explanations variable that references prediction breakdowns that have already been made, maybe like this:

showWaterfall2 = function(xgb.model, explainer, DMatrix, data.matrix, 
                         idx, type = "binary", threshold = 0.0001, 
                         limits = c(NA, NA), explanations = NULL){
        if(!is.null(explanations)){
                breakdown = explanations[idx,]
        } else{
                breakdown = explainPredictions(xgb.model, 
                                               explainer, 
                                               slice(DMatrix,as.integer(idx)))
        }
etc...