Using TfidfVectorizer for document binary classification

Question

I want to extract the ngrams of two sets of documents to make a csv file as follows for a binary classifier using TfidfVectorizer. Where the sets are Pos = {doc1, d