如何从pyspark管道（stringindexer-> onehotencoder-> vectorassembler）中获取原始分类变量值？

Question

我的分类变量有很多不同的层次，我想知道我的模型指向哪个值：

indexer = StringIndexer(inputCols=["a","b"], outputCols=["a_ind", "b_ind"])
df = indexer.setHandleInvalid("keep").fit(df).transform(df)

encoder = OneHotEncoder(inputCols=["a_ind", "b_ind"],
                        outputCols=["a_vec", "b_vec"])
df = encoder.fit(df).transform(df)

df = VectorAssembler(inputCols=["a_vec","b_vec"],
                     outputCol="features").transform(df)

功能列如下：

0: 0
1: 45568
2: 
0: 1
1: 2923
3: 
0: 1
1: 1

当我打印决策树模型时，我看到下面的树

print(dtModel.toDebugString)


DecisionTreeClassificationModel: uid=DecisionTreeClassifier_e9f4ea2ba51e, depth=5, numNodes=13, numClasses=2, numFeatures=45568
  If (feature 0 in {1.0})
   Predict: 0.0
  Else (feature 0 not in {1.0})
   If (feature 1 in {1.0})
    Predict: 0.0
   Else (feature 1 not in {1.0})
    If (feature 3 in {1.0})
     Predict: 0.0
    Else (feature 3 not in {1.0})
     If (feature 5 in {1.0})
      If (feature 3664 in {1.0})
       Predict: 1.0
      Else (feature 3664 not in {1.0})
       Predict: 0.0
     Else (feature 5 not in {1.0})
      If (feature 2933 in {1.0})
       Predict: 1.0
      Else (feature 2933 not in {1.0})
       Predict: 0.0

我希望能够使用这棵树来知道哪些值正在触发1.0的预测（功能3664，功能2993，功能5，功能3等）。如何才能返回到a或b列中的原始字符串？