4. [Extra Credit - 10 pts]: Write a program to Implement your own version of the Single Pass method from scratch. Your program should take as input a term-document (or document-term) matrix (this could be in the form of an inverted index structure) and a prespecified similarity threshold (between 0 and 1), and should output the resulting clusters (cluster assignments for each document, as well as the centroid vector for each cluster). In addition, your program should allow the user to specify whether overlapping clusters should be allowed or not (i.e., choose between hard vs. soft clustering). Term weights should be converted to TFXIDF and you should use Cosine similarity measure in the clustering algorithm. Demonstrate how your algorithm works using several input examples