Spark MLlib LDA，如何推断一个新的看不见的文件的主题分布？

i am interested in applying LDA topic modelling using Spark MLlib. I have checked the code and the explanations in i am interested in applying LDA topic modelling