Yeah it will build the dictionary from your poem dataset whicever language it is so it is language agnostic (not sure for chinese and symbolic languages though). Although I have removed the punctuations by accident, you can keep them if you want to capture that grammatical part as well.